If , : ........... ,i.
.'AC. J
^^^^^^^^^^^^^^^^E '
•XHtaHXTVflileM
'' tlH
"^^
sass:^ -
BOUGHT WITH THE INCOME
FROM THE
SAGE ENDOWMENT FUND
THE GIFT OF
lienrg M. Sage
X891
/^>ai$1qA sskji^
Cornell University Library
TN 780.G79
The hydrometallurgy of copper,
3 1924 004 678 755
Cornell University
Library
m.
The original of tiiis book is in
tine Cornell University Library.
There are no known copyright restrictions in
the United States on the use of the text.
http://www.archive.org/details/cu31924004678755
THE HYDROMETALLURGY
OF COPPER
Published by the
McGraw-Hill Book^ Company
Nevvf Yoirlk.
5ucce5sor« to theBookDepartments or the
McGraw F'ublishing Company Hill Publishing Company
Putlishers of Dook^ for
Electrical World The Engineering and Mining Journal
Engineering Record American MacKinist
Electric Railway Journal Coal Age
Metallurgical and Chemical Engineering Rjwer
iriTiriTiTiTiTiTTriTiTiTiTiririririnrimrTnrTiTiTiTiTimn
''I'^
THE H^ DROMETALLl RGY
OF COPPER
BY
WILLIAM E. GREENAWALT, C. E., B. S.
PART I. R()ASTIX(;
PART IT. HYDROMETALLTRUICAL PROCESS
McGRAW-HILL BOOK COMPANY
239 WE.ST 39TH STREET, NEW YORK
6 BOUVERIE STREET, LONDON, E. C.
1912
Copyright, 1912, by the
McGhaw-Hill Book Company
THE. MAPLE. PRESa. YORK. PA
PREFACE
The information available on the hydrometallurgy of copper is
somewhat fragmentary and widely scattered. The wet methods, for
treating copper ores, are diverse; as yet, the industry has not arrived at
any established practice, and it is questionable, on account of the widely
different character of the ores, if at any time one routine practice will
succeed in eliminating other processes entirely. In the discussion of the
various methods it is intended to cover all the most essential phases of the
subject.
Roasting, both oxidizing and chloridizing, has been given a prominent
place in the book, because on many ores, especially the sulphides, hy-
drometallurgical processes are directly or indirectly dependent upon this
step for successful treatment.
The hydrometallurgy of copper differs from the hydrometallurgy of
gold and silver largely on account of the greater percentage of material
recovered. For this reason the discussion of the precipitation plays an
important part. The commercial success of any particular process will
frequently depend on the nature of the precipitant and the cost of
precipitation.
The book is the result of notes, covering a long period of time, from
various sources and from my own experimental work. It is intended, in
the text, to give full credit for the various sources of information.
William E. Greenawalt.
Denver, Colorado,
August, 1912.
CONTENTS
Page
Pkeface V
FART I
CHAPTER I
Prepaiiation of the Ore 1
Relation of copper, gold and silver. Preparation of the ore — Dry crushing
with rolls — Dry crushing with ball rolls.
CHAPTER II
Fuel 6
Wood— Oil— Coal.
CHAPTER III
Oxidizing Roasting 12
Objects of roasting — Chemical combinations of the metals before roasting —
General chemical reactions during roasting — Essential factors in roasting —
Time — Temperature — Valentine's temperature experiments — Air — Rab-
bling— Effect of metallic sulphides if heated with exclusion of air — Sulphur
— Decomposition temperature of the various sulphates — Amount of sulphur
trioxide (SO3) in the sulphur dioxide (SO2) escaping from roasting furnaces —
Sulphur determinations — Tellurium — Copper — Silver — -Gold — Lead — Zinc —
Arsenic — Roasting argentiferous cobalt-nickel arsenides — Antimony —
Bismuth — Nickel — Calcium (lime) — Magnesium — Manganese — Aluminum
— Barium. Alkali metals — Chlorine. Bromine — Loss of weight in roasting.
CAAPTER IV
Chloeidizing Roasting 63
Object of chloridizing roasting — Adaptability of the various ores to chloridiz-
ing roasting — Chemistry of chloridizing roasting — Arsenic and antimony —
Zinc — -Lead — Calcium carbonate — Magnesium — Quartz — Barium sulphate —
Sodium sulphate — Percentage of salt — Time of adding salt — Heap chloridi-
zation — Composition of the roasted ore — Volatilization of the silver — Vola-
tilization of the gold — Chloridization of copper ores — Principal factors in the
loss of silver and gold by volatilization — Temperature — Time — Air or oxy-
gen— Experiments as compared with practice — Relation of sulphur to the
chloridization of silver and gold — Determination of loss by volatilization —
Chloridization determination.
CHiVPTER V
Pyhombtry 79
Color names of temperatures — Pyrometric determinations.
vii
viii CONTENTS
CHAPTER VI
Page
Roasting Furnaces 83
Hand reverberatories — Short reverberatories — Long reverberatories —
Method of operating a long hand reverberatory — Cost of roasting in long
reverberatories — Modified long reverberatories — Mechanical reverberatories
— Cost of mechanical reverberatories — Fuel required in roasting — Hearth
area required in roasting various ores — The Brown furnace — The Pearce
furnace — The Holthoff-Wethey furnace — The Merton furnace — The Edwards
furnace — The McDougal furnace — The Herreshoff furnace — The Wedge fur-
nace— The Greenawalt porous hearth — Bruckner furnace — Howell-White
furnace — Muffle furnaces — Ore coolers — Dust.
CHAPTER VII
Typical Examples of Roasting 146
Roasting of Cripple Creek ores — Roasting arsenical sulphide ore at the
Golden Gate mill, Mercer, Utah — Roasting of Casilas concentrates, Victoria,
Australia — Roasting at Kalgoorlie.
PART II
CHAPTER VIII
Properties and Solubilities of Copper 155
Copper — Influence of impurities on the properties of copper — Cupric car-
bonate— Cupric nitrate — Cupric oxide — Cuprous oxide — Cupric sulphate —
Cupric chloride — Cuprous chloride — Cupric silicate — Cuprous sulphide —
Cupric sulphide — Cupric hydroxide — Copper cyanides — SolubiUty of sulphur
dioxide.
CHAPTER IX
Hydrometallurgical Processes 169
Classification and general consideration — Chemical processes.
CHAPTER X
Chemical Alkali Processes 172
The Mosher-Ludlow ammonia-cyanide process — Sulphite processes — Neill
process — Van Arsdale process — Sulphate processes — ^Acid plants at the
mine — Sulphuric acid leaching of oxidized copper ores at Clifton, Arizona —
Leaching plant at the Snowstorm mine — Copper leaching plant at the Gu-
meshevesky mine, Russia — ^Ferric sulphate — Experiments with ferric sul-
phate at Cananea — Thomas' experiments with ferric sulphate on sulphide ore
— Experiments in Southern Tyrol, Spain — Copper extraction at Kedabeg,
Russia — The Millberg process — The Ehiott process — The Laist process —
Method of extracting copper at Rio Tinto, Spain — Treatment by heap
roasting and leaching — Elimination of arsenic, antimony and bismuth — •
Chloride processes — Hydrochloric acid— Ferric chloride — Doetsch process
— The Froelich process — ^Ferrous chloride process — The ferrous chloride
process as carried out at Ore Knob, Ashe Co., N. C. — Hunt and Douglas
process — The Bradley process — Longmaid-Henderson process for treating
CONTENTS ix
Faob
pyritic cinders — Longmaid-Henderson process at the Helsingbprg copper
works, Sweden — Longmaid-Henderson process at works of the Pennsylvania
Salt Manufacturing Co. — Cost of producing copper by the Longmaid-
Henderson process, having mechanical roasters — Extraction of copper from
atacamite.
CHAPTER XI
Copper Precipitants 270
Iron — Sponge iron — Crucible method of iron ore reduction — Precipitation
with a coke-iron couple — Precipitation with a copper-iron couple — Hydrogen
sulphide — -Lime.
CHAPTER XII
Electrolytic Processes 283
General consideration of electrolytic methods — Definitions — Anode —
Cathodes — Diaphragms — Current density — General laws governing the elec-
trodeposition of copper — ^Faraday's law — Theoretical data for copper deposi-
tion— Loss of energy in electrolytic work: Joule's law — The electrolyte —
Effect of bismuth, arsenic, and antimony in the electrolyte — Iron in the
electrolyte — Purification of the electrolyte — Depolarizers — Rapid depo-
sition of copper — Sulphuric acid process — Electrolytic extraction of copper
from ore at Medzianka, Poland — ^Plant of the Intercolonial Copper Co.,
N. S., Canada — Keith process at Arlington, N. J. — The Siemens-Halske
process — M. De Kay Thompson, Jr.'s experiments on the Siemens-Halske
process^Siemens-Halske process in Spain — Experiments at the Ray mines,
Arizona — Tosizza process — Ramen process — Treatment of ore at the Braden
Copper Go's, mine, Chile — Extraction of copper from matte — Marchese
process — Gunther process — Electrolytic chloride processes — Electrolysis of
oupric chloride — -Electrolysis of cuprous chloride — The Body process — The
Hoepfner process — The Douglas process — ^The Greenawalt process — The
Swinburne-Ashcroft process — Baker-Burwell process.
CHAPTER XIII
Extraction of Precious Metals prom Copper Ores 363
Method proposed by Bertram Hunt — Cyaniding of cupriferous gold ore of
the Bagdad Gold Mining Co. — Extraction of precious metals with chloride
solutions — Electrolytic chlorine, theoretical data — ^Practical data— Treat-
ment of ore at Mt. Morgan mine — Treatment of auriferous copper ores at
Falun, Sweden.
CHAPTER XIV
Treatment op Zinciferous Copper Ores 382
The Hoepfner zinc process.
CHAPTER XV
Treatment of Copper- Nickel Ore and Matte 387
The Hoepfner nickel process — The Browne process — Sjostedt- James process
— Gunther-Franke process — Hybinette process — Cito process.
X CONTENTS
CHAPTER XVI
Page
Precipitation of Copper from Mine Waters
Precipitation of copper from mine waters at Butte — Precipitation of copper
from mine waters at Copper Queen Consolidated Mining Company, Bisbee,
Arizona — Leaching of copper ore in place.
CHAPTER XVII
Refining of Copper Precipitate
Oxidizing stage — Reducing stage — Furnaces used for copper refining.
CHAPTER XVIII
Copper Sulphate, Bluestone ^^"
The Oker process — The Freiberg process — Hofmann's copper sulphate
process.
CHAPTER XIX
Apparatus and Appliances 450
Tanks — Conducting solutions — Regulating the flow of solution — Elevating
solutions — Montejus tanks — Air lifts — Stoneware pumps — Rubber pumps — ■
Pressure tanks — Ore agitation and filtration — Filter presses for filtering acid
copper solutions — Construction of floors.
CHAPTER XX
Power Data 473
Comparison of steam and water-power plants — Cost of water-power in the
Western United States — Comparison of steam and producer plants — Pro-
ducer gas plant using western lignite — Consumption of fuel in a gas
engine — Cost of power with steam-plants — Oil engines.
CHAPTER XXI
Economic Considerations 483
Cost of depositing copper electrolytically — Comparison of electrolytic
methods with the iron precipitation process, using sulphuric acid as the
solvent — Economic relation of current density and voltage — Chemical
treatment of sulphide ores — Comparison of wet methods with smelting —
Chemical and mechanical difficulties — General applicability of hydrometal-
lurgical processes.
THE
HYDROMETALLURGY OF C OPPER
PART 1— ROASTING
CHAPTER I
PREPARATION OF THE ORE
Relation of Copper, Gold, and Silver.— Copper, gold and silver are
chemically, mineralogically, and metallurgically intimately associ-
ated. Chemically, they occur in the same group in the Periodic System;
mineralogically, one of these metals is rarely, if ever, found unaccom-
panied by one or both of the others; and metallurgically, any scheme
which contemplates the profitable recovery of the copper must take into
consideration the profitable recovery of the accompanying precious
metals also.
In the hydrometallurgical treatment of copper ores, or of gold and
silver ores containing copper, it is evident that the extraction of all
three metals, concurrently or consecutively, must be given due consider-
ation. Many operations in the wet treatment of ores of one of these
metals are applicable to the others. In the acid treatment of gold and
silver ores, as in chlorination, the conditions of roasting and extraction
are not essentially different from the treatment of copper ores with acid
solvents.
Many ores may be treated by hydrometallurgical processes without
roasting. Others, especially the sulphides, have to be roasted, and the
sulphides constitute the greater proportion of the available ores of copper.
Some sulphides may be treated without roasting, but such treatment is
the rare exception, and for obvious reasons will probably not find exten-
sive application.
The best conditions of roasting, for the various ores for chemical
treatment by any of the solvent processes, are very much the same,
whether the metals to be extracted are copper, gold or silver, or all
combined in the same ore. The same furnaces are used; the same costs
apply; and the conditions of roasting which give the best extraction of the
precious metals will, in general, also give the best extraction of the copper.
In roasting copper, gold and silver ores, for hydrometallurgical treat-
ment the metals themselves offer no particular difficulty in the operation.
The difficulties encountered in roasting will usually be in the nature of
1
2 HYDROMETALLURGY OF COPPER
the other elements associated with them in the gangue. It is evident
that in considering the roasting of copper ores, or copper ores containing
gold and silver, the foreign elements must be taken into account quite
as seriously as the metals themselves.
Roasting of ores, as a step for their treatment by solvent processes,
is materially different from that required for subsequent smelting.
While the chemical reactions during the roasting are essentially the same
for both methods, a good roast for a solvent process requires vastly
more delicate manipulation and a more thorough elimination of the
sulphur. A roast which would be satisfactory for smelting might be,
and usually is, absolutely worthless for treatment by wet methods; on
the contrary, ore which is satisfactorily roasted for treatment by wet
processes would be satisfactory for smelting also, but the expense of
roasting would be considerably greater.
In ores containing copper, gold and silver, if the precious metals
are not extracted simultaneously with the copper, the roasting of the ore
to make their subsequent extraction satisfactory, either by cyanidation
or chlorination, must be taken into account.
Cupriferous pyritic ores, high in sulphur, are sometimes roasted in
heaps, preparatory to extracting the copper as soluble sulphate, but
this practice is not finding extended application, and at the mine where
it was largely employed its use has been discontinued.
The only roasting which is finding favor for the hydrometallurgical
processes is in suitable furnaces, usually reverberatories, and preparatory
to which the ore is crushed fine enough to be thoroughly roasted in sev-
eral hours.
Preparation of the Ore. — Ore, as roasted in furnaces for hydromet-
allurgical treatment, is usually crushed to a fineness varying from 8 to
40 mesh. Below 8 mesh the particles become too large for efficient
oxidation, and above 40 mesh the dust is likely to give trouble. On
the whole, ore ground to a fineness varying from 10 to 30 mesh will give
the best average results.
When roasting constitutes a step in any metallurgical process, the
ore is crushed dry. Rolls and ball mills are best suited for this work.
Concentrates are usually the product of wet crushing. If the ore is of
too low a grade to admit of direct chemical treatment, concentration
offers a means of increasing the tenor of the material, while at the same
time eliminating the most injurious elements. In this way, lime par-
ticularly may be largely eliminated from the iiltimate product to be
treated by roasting and a chemical process. If concentration forms a
step in the general treatment, there is no need of close work to obtain
a high grade concentrate, and hence there is no need of excessive loss in
the tailings. A concentrate containing 10 per cent, copper would be a
very desirable product for roasting and for treatment by a solvent proc-
PREPARATION OF THE ORE 3
ess, and such a concentrate should be made without excessive loss;
whereas if shipment to a smelter is desired, such a product would not pre-
sent much advantage, and to get a higher grade product would result
also in getting a considerably greater loss.
The moisture contained in the ore before charging into the furnace
should also be considered. Concentrates may be charged in a hand-
rabbled furnace without drying, but in mechanical roasters it is evi-
dently better to remove the moisture sufficiently so that the ore may be
fed uniformily into the furnace by mechanical means. Much moisture
in the ore charged into the furnace has a tendency to cool it unduly.
The moisture in the ores, as well as the moisture in the fuel gases,
has an important bearing on the chemistry of the roasting process.
The principal expense in the preparation of the ore for roasting is in
crushing. This may vary within wide limits, depending on the character
of the ore, the fineness to which it is reduced, and the amount crushed.
Usually it will vary from 25 to 50 cents per ton, with a reasonably large
installation.
Dry Crushing with Rolls. — Rolls are largely used to crush ore to
medium fineness. For grinding finer than 20 mesh they are inferior to
some other type of machines, and it is a question whether, under any
conditions, they are as satisfactory as ball mills.
It costs from 25 to 30 cents per ton to crush Cripple Creek ore to
12 or 14 mesh, on a basis of from 200 to 300 tons per day. In one plant,
having two 48-in. roughing rolls and four 48-in. finishing rolls, 300 tons
of ore are regularly crushed per day. The roughing rolls are run only
during the daytime, but the finishing rolls are run continuously for
three eight-hour shifts, with three men on a shift. One man attends
to the screens.
In another plant, having one 36-in. roll, and three 26-in. rolls, 175
tons are crushed in 24 hours to from 30 to 40 mesh.
A combination of one 36-in. roughing roll; one 26-in. roll doing
medium work, and two 26-in. rolls doing finishing work, will crush from
125 to 175 tons of ore of ordinary hardness to 30 mesh; 200 to 250 tons
to 20 mesh, and 250 tons and more to 16 mesh.
For the best working of a roll crushing plant, it is essential that the
reduction shall be gradual in going from one roll to the next.
Dry Crushing with Ball Mills. — ^For dry crushing ball mills present
certain advantages over rolls in that they are self-contained and the
screening is simplified. Their capacity is also large. A No. 5 Krupp
ball mill will crush 43 tons daily of ordinary sulphide ore, and the No. 8,
100 tons; using from 18 to 23 h. p. for the 5-ton mill, and from 60 to 65
for the 8-ton mill, or 2.1 and 1.6 tons per horse-power respectively.
M. W. Von Bernewitz' has given some valuable information on ball
' Min. and Scientific Press, July 15, 1911.
4 HYDROMETALLURGY OF COPPER
mill practice at Kalgoorlie. The following table gives a summary of the
essential facts.
BALL MILL PRACTICE AT KALGOORLIE (DRY CRUSHING)
Name
of plant
Associated
Associated Northern.
Chaffers
Great Boulder
Kalgurli
Perse verence
South Kalgurli
No. of
mills
4 No. 8
6 No. 5
3 No. 5
3 No. 5
4 No. 8
9 No. 5
8 No. 8
3 No. 8
1 No. 5
Weight
of balls
lb.
4480
2240
2350
2300
4480
2200
4400
4480
2800
Screen
30X30
25X25
27X27
27X27
30X30
26X26
27X27
27X27
r. p. m.
21
25
26
25
24
25
24
24
Power
h. p.
60
23
18
25
60
25
60
65
30
Capac-
ity
daily
tons
92-95
43
40
40
80-90
40
100
95-100
Steel con- [ Life ot
sump- grinding
tion plates
lb. per ton
0.50
0.32
0.74
0.64
0.45
0.65
0.47
days
170
270
180
105
300
118
210
COST OF CRUSHING AT KALGOORLIE, PENCE PER TON
Name
Wages
Power
Other items
9.89
6.25
8.84
6.26
7.49
4.43
9.43
Total
of plant
Pence
-
Dollars
6.45
5.87
2.09
1.81-
13.78
12.08
10.64
9.75
15.23
19.18
20.82
30.11
24.22
21.57
17.82
22.72
0.574
Associated Northern . .
0.463
Chaffers ....
0.413
0.342
0 . 435
3.21
2.05
26.84
0.514
South Kalgurli
32.20
0.619
One man per shift of 8 hours, at lis. 8d. (S3. 22) can look after eight
No. 8 ball mills; but in smaller plants the mill man attends to conveyors,
elevators, dust, pipes, etc. Ball mills should be fed with no larger ore
than will pass a 3-in. ring. With fine ore the balls are likely to bed.
Actual weighing has shown a 3-in. feed is crushed faster and the wear of
steel is less than when 1-in. material is fed with a quantity of fines.
A No. 5 mill, including foundations, may be erected for £600 ($2922.00)
and a No. 8 for £1000 ($4870.00), bin and conveyor inoluded. It
will usually cost about £300 ($1461.00) per year for upkeep of a No. 5.
At the Golden Cycle mill, at Colorado Springs' four No. 66 "Kominu-
ters" have a capacity of 17,000 lb. of Cripple Creek ore per hour for
each mill, when fed with a product that had been reduced by rolls to
pass through a revolving screen made of 1/4 in. steel plate and having
openings 11/2 in. in diameter. The kominuters were equipped with
' Loohiel M. King, Mining and Scientific Press, Jan. 25, 1908.
PREPARATION OF THE ORE 5
a diagonal slotted sciceii, size of opening 5/32 by 1/2 in. No. 8 steel
plate. This opening gave a product varying in size from 1/8-in. cubes
to the finest slimes-. The consumption of power was 50 h. p. at a speed
of 22 r. p. m., the ball consumption being fourteen 5-in. forged steel balls
weighing about 19.5 lb. each, per day of 24 hours. One man can attend
to six mills.
The average results from several types of ball mills show that one
ton of steel balls will crush about 50 tons of ore during 24 hours from a
feed 1 1/2 in. diameter down to a product of from 12 to 20 mesh in one
operation.
CHAPTER II
FUEL
Roasting, as a step in the treatment of ores by the hydrometallurgical
processes, is usually carried out in the immediate vicinity of the mmes.
At industrial centers, the consideration of fuel is a very simple matter,
but not so in copper, gold and silver mining districts where the selection
of a particular kind of fuel is frequently a matter of necessity, based on
local conditions. In vicinities where wood is abundant it will ordinarily
be used in preference to the more expensive coal, which has to be freighted
in. If a mining district has no wood, and is some distance from the
source of fuel supply, the greater calorific power of oil per unit of weight
over that of the coal might make it the cheaper fuel on account of the
difference in cost of freight.
Other things being equal, the relative desirability of fuels for roast-
ing purposes is gas, oil, bituminous coal, wood, lignite. Anthracite is
not often available, but if it is, there is a decided advantage in first
converging it into producer gas.
Most of the expense of roasting, in mechanical furnaces, is in the
fuel. Some of the essential facts pertaining to the various fuels and a
comparison of their relative value is, therefore, pertinent.
Wood. — It frequently happens that in mining camps far removed
from coal supply, wood can be obtained cheaply and in large quantities.
For roasting, if the wood is perfectly dry, it is more desirable than
lignite of the inferior qualities of bituminous. Green wood contains
from 30 per cent, to 40 per cent, moisture. After thorough seasoning,
for about a year, in the open air, the moisture is from 20 to 25 per cent.
The wood of various trees are nearly identical in chemical composi-
tion, which for perfectly dry wood and of ordinary fire wood holding
hygroscopic moisture, is practically as shown in the table on the follow-
ing page.
The ash in most woods varies from 0.5 to 1.5 per cent. Most of the
pines and others of the coniferous family contain hydrocarbons (pitch,
turpentine) which increase their value as fuel.
In steam-boiler tests wood is assumed as 0.4 of the value of the
same weight of coal. It is safe to assume that 2 1/4 lb. of dry- wood is
equal to 1 lb. of average quality bituminous coal, and that the fuel
value of the same weight of different woods is nearly the same. That
is to say, a pound of pine is worth as much for fuel as a pound of hickory,
supposing both to be dry.
6
FUEL
Dessicated wood
Ordinary fire wood
Carbon
50 per cent. | 37 . 5 per cent.
6 per cent. 4.5 percent.
41 per cent. 30.75 per cent.
1 npr ppnt H 7^ npr ppnt.
Hydrogen
Oxvfiren
Nitrogen
Ash
2 per cent.
1 . 5 per cent.
Hygroscopic water
100 per cent.
75 . 0 per cent.
25.0 percent.
100 . 0 per cent.
It is important that the wood be dry, as each 10 per cent, of moisture
in wood will de1>ract 12 per cent, from its value as fuel.
A cord of wood is a pile 4 ft. by 4 ft. by 8 ft. which is equal to 128 cu.
ft. About 56 per cent, is solid wood, and 44 per cent, spaces.
Fire-boxes for burning wood should be built so as to contain a deep
bed of fuel. They should be narrower at the bottom than at the top.
With properly designed fire-boxes, burning thoroughly dry wood, a
very intense heat can be obtained which is quite as effective in roasting
ores as most coals available in copper mining districts.
Where wood is abundant in the Rocky Mountain Region it will
ordinarily cost from $3.00 to $3.50 per cord, cut and piled at the metallur-
gical works ready for use.
Wood burns with a long flame and makes comparatively little smoke,
which are ideal conditions for roasting.
Charcoal gives out much more useful heat than wood, because the
water contained in the wood, or formed by the combustion of its oxygen
and hydrogen, has to be evaporated during its combustion. 100 parts
of wood give only as much heat as 40 parts of charcoal.
Charcoal is made by the dry distillation of wood, at a temperature
of from 460° to 450° C. This may be done in heaps or in closed retorts.
Dry wood in stacks yields about one-fourth its weight in charcoal.
Charcoal develops on burning 8000 heat units, while wood, dried in the
air, does not develop more than 2800 units of heat. Therefore, seven
parts of charcoal gives as much heat as 20 parts of wood, but the 20 parts
of wood are capable of yielding only five parts of charcoal.
If wood has to be transported any considerable distance for roasting,
it might be profitable to convert it into charcoal at the forests and then
burn it in the roasting furnace, after having converted it into producer
gas.
The weight of a bushel of charcoal is usually taken as 20 lb.
HYDROMETALLURGY OF COPPER
RELATIVE HEATING VALUE OF WOOD, COAL, AND OIL
One cord of wood
(128 cu. ft.)
Weight in pounds
per cord
Pounds of coal equiva-
lent to one cord of wood
Pounds of oil equivalent
to one cord of wood
Hickory
4,500
3,850
3,250
2,350
2,000
1,800 to 2,000
1,540 to 1,715
1,300 to 1,450
940 to 1,450
800 to 925
1,000
White Oak
860
Beach
Red Oak I
725
Black Oak
Poplar
Chestnut >
Elm
Pine. . . ,
525
460
It might be said that the approximate heating value of wood, coal
and oil is: 2 cords of average pine = l ton of average bituminous coal =
13/8 tons of lignite = 3 1/2 to 4 barrels of crude oil.
As to the absolute consumption of fuel, in roasting, much depends
on the nature of the ore, the amount of sulphur in the raw ore, and the
extent to which the sulphur is eliminated.
Oil.— Oil, next to gas, is the most desirable fuel for roasting purposes.
It is largely used where it can be obtained cheaply and the supply is
constant. It was for many years the principal fuel used in roasting
Cripple Creek ores. Recently, owing to the uncertainty of the supply,
producer gas has largely displaced" the crude oil and residuum. In
California, where large oil fields have lately been developed, it is dis-
placing wood and coal in the roasting of stamp mill concentrates for
chlorination.
Fuel oil has the following advantages over coal and wood in roasting
ores:
Reduction of weight of fuel by 50 per cent.
Reduction of bulk of fuel by 30 per cent.
Reduction of labor by 50 per cent.
Prompt kindling of fire.
Cleanliness and freedom from ash.
No loss of heat by useless radiation, as in the coal fire-box where
the heat and products of combustion are introduced through the top of
the arch.
Convenience in directing and controlling the heat.
It is possible to get with it either a long rolling flame, or an intensely
hot local flame.
Oil, as it is used in roasting ores, is sprayed with a steam jet directly
into the furnace, either through the sides or through the arch. The
steam is usually kept at a pressure of from 60 to 90 lb. and the oil at
from 30 to 50 lb.
FUEL 9
Most of the oil sold for fuel purposes ranges from 14° to 20° Baunie.
Oil is usually bought by measure and not by weight. The lighter
gravity oils contain more heat units per pound than the heavier oils,
but there are more pounds of fuel in a gallon of heavier oils than in a
gallon of lighter oils. The gravity of the oils, therefore, is not a matter
of much consequence.
A U. S. gallon of oil weighs from 6.5 to 7.2 lb. and 42 gallons are taken
as a barrel. Residuum, that is, the residue of crude oil after the volatile
substances have been driven off by heating, is largely used as fuel for
roasting purposes.
In some of the mills treating Cripple Creek ore both coal and oil are
used in the same furnace. In some of the roasting furnaces coal is used
at the cooler, or feed end, while in others the reverse is the case. The
relative quantity of coal and oil used also varies greatly. The average
consumption might be considered as 100 lb. of coal and 15 gallons of oil
per ton of ore, in roasting 1 1/2 to 3 per cent, sulphur down to about 0.5
per cent. With oil alone, it takes from 0.35 to 0.45 barrels to roast a
ton of ore, in addition to the small amount of fuel necessary to generate
the steam for applying the oil.
In California it takes about half a barrel of oil to roast a ton of stamp
mill concentrates suitable for chlorination, and about 50 lb. of coal
to furnish the steam to pump, heat, and atomize the oil. As a fuel 90
gallons of California oil is equal to 1 ton of coal.
If fuel has to be transported any considerable distance, oil offers
advantages in the cost of freight, since for the same weight it has about
twice the heating value of coal, and about four times that of wood.
Coal. — Coal is the most universally used fuel in roasting. Its quality,
however, varies so much that careful investigation of the different
kinds available is a serious matter. A long-flame bituminous coal, if
direct firing is used, is the best, while lignite, with its short flame and
low heating quality, is the worst. The tendency of short-flame coal is
to give an intense local heat, and such a heat is highly detrimental to
the roasted ore. The best way to distribute the heat is either to gassify
the coal, or if fired direct, get what is known as a semi-producer action
in the fire-box, by the introduction of steam and air under the grate. By
either of these methods, a long rolling flame may be obtained in the
roasting chamber.
Any coal, wheter anthracite, bituminous, or lignite, will give the
most satisfactory result by being first converted into producer gas, and
conducting the gas from the producer mains into the different parts of
the furnace, and there consuming it, so that the atmosphere shall be
highly oxidizing and with as little local heat as possible. This is best
accomplished by introducing the gas in smaller quantities at more points
in the roasting chamber, rather than in larger volumes at fewer places.
10
HYDROMETALLURGY OF COPPER
The advantages of producer gas over direct firing are:
The gas may be produced from inferior coal, and makes more avail-
able heat in the roasting furnace than is possible with any coal burned
in an ordinary grate.
It can be easily introduced into the roasting furnace at any pomt
and in any quantity desired, thereby giving a diffused heat over the
entire bed of the ore.
The construction of the arch of large mechanical furnaces is very
much simplified.
The producers may be centralized, so that the handling of coal and
ashes, by mechanical appliances, may be greatly facilitated.
In all cases, where producer gas is used to roast ore, the air necessary
for its combustion should be pre-heated. This can be done at the least
expense by an air-heating arrangement in the furnace dust chamber.
The relative average value of the several classes of coal may be ap-
proximately determined from the accompanying tables.
CLASSIFICATION OF COALS
(Kent, Min. Ind., 1900)
Anthracite and semi-an-
thracite.
Semi-bituminous
Bituminous — eastern
Bituminous — western
Lignite
Moisture
per cent.
Ash
per cent.
VolatUe
matter
per cent.
1 to3
8 to 12
3 to 12
1 to3
3 to 10
15 to 25
1 to3
3 to 15
25 to 40
4 to 14
5 to 25
35 to 50
12 to 18
5 to 25
over 50
Fixed
carbon,
per cent.
97 to 88
75 to 85
60 to 75
50 to 65
Less than 50
Heating value
b. t. u. per lb.
of combustible
Relative
value of com-
bustible
semi-bit.
= 100
14,700 to 14,900
15,600 to 16,000
14,800 to 15,200
13,600 to 14,800
11,000 to 13,000
94
100
95
90
76
A rough estimate of the relative practical value of the several classes
of coal may be calculated as follows:
Mois-
ture
%
Ash
%
Combus-
tible
%
b. t. u.
per lb.
combus-
tible
Theort-
ical
healing
value
Effi-
ciency
for
boiler
Relative practical
VEilue
b. t. u.
Semi-bit.
= 100
2
2
2
10
15
13
8
8
15
20
85
90
90
75
65
14,800
15,800
15,000
14,200
12,000
12,180
14,220
13,500
10,150
7,800
77
75
70
65
60
9,379
10,665
9,450
6,598
4,680
Semi-bituminous
100
Bituminous — eastern
Bituminous — western
89
62
The relation of the heating value of coal to its ultimate analysis
may be estimated by Dulong's formula, usually within a limit of error
of 2 per cent. This formula with average figures for the constants is :
FUEL
11
Heating value per pound in b. t. u. equals:
1 / 100[14,650C + 62,000 (H - ^) + 4000S]
In which C, H, 0 and S arre respectively the precentages of carbon,
hydrogen, oxygen, and sulphur in the coal.
There is more ash in the smaller size coal than in the larger sizes,
due principally to the greater quantities of dirt and slate, as shown by the
following analyses of different sizes of anthracite.
Size of coal
Fixed carbon
Ash
Egg, 2.5 to 1.7 in
Stove, 1 . 75 to 1 . 25 in
Chestnut, 1 . 25 to 0 . 75 in . . .
Pea, 0 . 75 to 0 . 50 in
Buckwheat, 0 . 50 to 0 . 25 in .
88 . 5 per cent.
83 . 7 per cent.
80.7 per cent.
79 . 0 per cent.
76 . 9 per cent.
5 . 7 per cent.
10.2 per cent.
12.7 per cent.
14 . 7 per cent.
16.6 per cent.
CHAPTER III
OXIDIZING ROASTING
Objects of Roasting. — The object of roasting is to convert the ore
into a condition which will have the least injurious effect on the chem-
icals used, and to simplify their application. Roasting is essentiallj-
oxidation. Many metallic oxides are not as readily attacked by the
solvent in the subsequent chemical treatment as the metals in other
combinations.
The solvents for copper, gold and silver are among the most ener-
getic substances known. Chlorine, for example, combines with those
elements with which oxygen is able to combine, because in many respects
it is equally if not more energetic than oxygen, and replaces it in the
proportion of 2 atoms of chlorine to one of oxygen: Clji 0. Chlorine
cannot displace oxygen from many of its oxide combinations. Iron
is universally associated with copper, gold and silver ores in the form
of oxide or sulphide. Chlorine very rapidly combines with iron in its
sulphide and sulphate combinations, but does not appreciably displace
the oxygen from its oxide combinations. Most of the metals are
less injurious in their oxide than in their sulphide combinations, while
others are not much , improved by the change. If the ore is to be
treated for its copper content by an acid process, the oxide of copper
resulting from the roasting is readily soluble in either hydrochloric
or sulphuric acids, while the sulphide of copper is quite insoluble in
either of these acids. Calcium is acted upon by chlorine and the
acids as readily in its oxide as in its carbonate combinations. Much
of the calcium, however, in roasting, is converted into the sulphate,
which is an improvement, since it is practically neutral and unaffected
by all solvents.
Many injurious metals, such as arsenic, antimony, and bismuth, are
volatile at a high temperature and are expelled during the roasting.
Many oxides are benefited by elevated temperatures. Dehydration
agglomerates the particles and makes a better leaching product. Ores
containing much clay and talc are similarly benefited. And, finally,
roasting makes the ore particles porous, thereby very materially increas-
ing the extraction of the metals. If the ore contains gold and silver
these metals are to a very large extent set free, and are more readily
attacked by the solvent.
12
OXIDIZING ROASTING 13
The objects of roasting, therefore, may be summarized as follows:
1. To oxidize. The common elements oxidized are iron, copper,
lead, zinc, aluminum, calcium and magnesium.
2. To volatilize. The common elements volatilized are sulphur,
arsenic, antimony, bismuth and tellurium.
3. To sulphatize. The common elements sulphatized are calcium,
magnesium and, to some extent, lead and zinc.
4. Dehydration. The object of dehydration is to agglomerate the
ore particles and make them more susceptible to leaching, decantation,
or filtration. Usually ore, which will percolate or filter very slowly
before roasting, will percolate or filter quite rapidly after roasting.
Mill dust, when raw, may be difficult to filter, but after roasting filtra-
tion takes place quite rapidly.
5. To make the ore porous. Oxidation of sulphide and telluride
ores, by the elimination of the sulphur and tellurium, must of necessity
make the ore particles more porous and present a greater surface to the
action of the solvent. Most ores will yield a very much better extraction
after roasting than before, even though they are otherwise equally
susceptible to treatment.
6. To free the gold and silver particles. Neither chlorine nor cyanide
are practical solvents of gold and silver in their telluride combinations,
and in their sulphide combinations they present serious difficulties.
After roasting, the gold and silver are in their metallic state and are
readily soluble if the particles have not been fused.
7. To convert the .desired metals into soluble form. Copper, in its
sulphide combinations, is quite insoluble in either acid or alkaline solu-
tions, while in its oxide combinations it is readily soluble.
Chemical Combinations of the Metals before Roasting. — Copper, gold
and silver ores as they come from the mine, may contain any of the base
elements. The matrix is almost always quartz, but associated with it
will usually be found one or more of the elements enumerated:
Aluminum; usually as a silicate, fluoride or sulphate.
Antimony; usually as a sulphide.
Arsenic; usually as a sulphide.
Barium; usually as a sulphate or carbonate.
Calcium; usually as a carbonate, fluoride, or sulphate.
Cobalt; usually as a sulphide.
Copper; usually as a sulphide, carbonate, oxide or silicate.
Iron; usually as a sulphide, carbonate, or oxide.
Lead; usually as a sulphide or carbonate.
Magnesium; usually as a carbonate.
Manganese; usually as an oxide.
Nickel; usually as a sulphide.
Silver; usually as a sulphide.
14 HYDROMETALLURGY OF COPPER
Zinc; usually as a sulphide or carbonate.
Sulphur and tellurium are usually found in combination with the
metals as sulphides and tellurides.
General Chemical Reactions During Roasting.— The carbonates, on
heating, are readily converted into the oxides of the metals and carbon
dioxide:
MC03 = MO+C02.
When metallic sulphides are heated in the presence of air, metallic
oxides and sulphur dioxide are formed:
MS + 03 = MO + S02.
Most of the sulphur dioxide passes off, but a small portion of it is
converted into sulphur trioxide by contact with the metallic oxides
formed, or with the silica contained in the ore:
S02 + 0 + Si02 = S03 + Si02
Some of the sulphur trioxide will escape, while some will combine
with the metallic oxides to form metallic sulphates :
MO+S03 = MS04.
By heating, the metallic sulphates are dissociated, some giving off
sulphur trioxide, others sulphur dioxide and oxygen:
2FeSO,=FeA + S03 + S02
2CuSO, = 2CuO + 2S02 + 02.
The sulphates of copper, antimony, iron, and nickel, are completely
decomposed at a red heat. A higher temperature decomposes the
sulphates of aluminum, silver, lead, manganese, and zinc. An ordinary
white heat has no action on the sulphates of the alkalies and alkaline
earths, potassium, sodium, barium, calcium, and magnesium, but at the
most intense heat procurable, which is never used in a roasting furnace,
the sulphates of barium and calcium are changed to oxides. At the
same temperature, sodium and potassium sulphates are completely
volatilized.
Essential Factors in Roasting. — The essential factors in roasting are:
Time,
Temperature,
Air, or Oxygen.
These are complementary terms, and each may be carried in excess,
to the neglect of the others. Roasting, as already stated, is substan-
tially oxidation by the application of heat and air. Mineralized veins are
oxidized to great depths by time and the action of atmospheric and
aqueous agencies, without the necessity of any perceptible heat. In
this respect, it differs from roasting. If the temperature is increased the
time of oxidation is diminished. A high temperature, without an
OXIDIZING ROASTING 15
abundance of air, is of little avail. A moderate temperature, with an
abundance of air, is highly efficient, nevertheless time is necessary to
effect complete oxidation, and to get the best roast for subsequent
chemical treatment to extract the metals.
Time. — Time in roasting, and in oxidation, is a most variable factor.
Pyrites may be oxidized almost instantly, in the highly oxidizing atmos-
phere of a shaft furnace, or it may take countless ages, as in the oxidation
of mineralized veins, where the elements of both temperature and air
are lacking.
Some idea of the relation of time, temperature, and air may be
obtained from roasting tests made in Denver, on Gilpin County sulphide
ore. The ore was roasted in a three-compartment shaft furnace at the
rate of 75 tons a day. Each particle was exposed on all sides to a highly
heated oxidizing atmosphere. The ore was thoroughly roasted in 13/4
minutes, which was the total time it remained in the furnace. Similar
ore was roasted in a mechanical reverberatory, with a bed from 4 to 5 in.
deep, and notwithstanding that the ore was rabbled continuously and
remained in the furnace from 4 to 5 hours, the roast was not satisfactory.
The difficulty lay in the inability of the air to penetrate the deep bed
of ore. As soon as an abundance of air was supplied, as for example,
when the ore was discharged from the furnace, innumerable sparks
appeared, showing that the oxidation was taking place more rapidly.
In the Stedefeldt shaft furnace the ore is oxidized almost instantly,
and the time reduced to a minimum, as compared with roasting in the
ordinary reverberatory or revolving furnaces. In the Stedefeldt fur-
nace, which was used only for chloridizing roasting silver ores, the
chlorine acted as an energetic oxidizer, and this materially assisted in
the roasting.
The best results in roasting will usually be obtained when the time
factor is made as great as possible, and the temperature as low as pos-
sible, assuming that the air factor remains the same. Or, again, the
best results will always be obtained by having the time and air factors
as large as possible, speaking, of course, within practical limits. By
increasing the air, the temperature remaining the same, the time will be
greatly diminished without detriment to the roast.
If ore is roasted, as in a shaft furnace of the Stedefeldt type, where
it is showered through a highly heated oxidizing atmosphere, the time
of roasting is reduced to a minimum, but the combustion of the pyrite is
likely to be so intense that the heat developed in the particle itself is
likely to fuse it. Careful panning" of roasted sulphide ore, roasted under
such conditions, will usually disclose some of the grains as fused or even
shotted, which is the worst possible condition for subsequent treatment
by a solvent process, largely because of the inability of the solvent to
penetrate the fused or shotted particle.
16 HYDROMETALLURGY OF COPPER
Temperature.— The regulation of temperature in a furnace to get the
best results in roasting, depends largely on the nature and composition
of the ore. In simple ores, not containing too much sulphur, slow initial
heating is not essential if care is taken not to carry the heat too near the
sintering point. Even on concentrates, and ore high in sulphur, the
initial heat may be reasonably high without deleterious effects, provided
there is an abundance of air and the rabbling in sufficiently frequent.
Copper sulphides are vastly more sensitive to high temperatures than
iron sulphides, and with galena, which fuses at a low temperature, the
utmost care must be taken.
When roasting silicious ore in large furnaces, or even pyritic material
containing only small quantities of copper or lead sulphides, the rear fire
may be pushed quite as hard. as the first one, since the temperature of
the ore must be brought to the ignition- or to the volatilization-pomt of
sulphur, before roasting can begin. This fact was demonstrated by
interesting experiments made in roasting Cripple Creek gold ore, in four
100-ton furnaces. Three of these furnaces were of the ordinary me-
chanical reverberatory type; the fourth had a revolving hearth, with a
gas producer in the center, and which was so fired that the temperature
throughout the entire hearth was practically the same. The raw ore
entering the furnace was subjected to almost the same heat as the
roasted ore being discharged. A comparison of several thousand tons
of tailings from the different furnaces showed no material difference in
the extraction.
The ordinary roasting starts with a low initial heat, and finishes at
the highest temperature the ore will stand without sintering. This is
particularly true of all revolving furnaces, and to a large extent in rever-
beratories also. While this may be the best for some ores, careful com-
parative tests in large furnaces would indicate that it is better to bring
the ore as quickly as possible to a dull red (or even cherry red) heat, and
that the finishing temperature should not be too high. In roasting ore
containing from 2 per cent, to 4 per cent, sulphur, in a furnace having
say four fire-boxes and roasting 100 tons per day, the best results will be
obtained by firing the finishing fire box at a lower temperature than the
one preceding it. The dark magnetic oxide is, to a large extent, con-
verted into the ferric oxide by the prolonged roasting at a moderately
low temperature. Similarly the cuprous oxide, which is with some
difficulty soluble in ^cids, may be reduced to the cupric oxide, which is
quite readily soluble.
If the ore is overheated, as is frequently the case when the finishing
heat is high, it will have a dark appearance; whereas, if finished at a
lower temperature, the ore will have the red appearance of ferric oxide.
Overheating, or lack of air, will convert the ferric oxide into the magnetic
OXIDIZlNd ROASTING 17
o^ide, which, at a lower temperature and with an abundance of air, may
t»e reconverted into ferric oxide.
In order to determine the effect of temperature on the extraction,
the following laboratory experiments were made on Cripple Creek ore.
Chlorine was used as the solvent. Head assay of raw ore, gold 5.32 oz.;
head assay of roasted ore, gold 5.56 oz.; sulphur in raw ore, 4.02 per cent.
Test No. 1. — The ore was given what appeared to be an ordinary
roast in an assay muffle. The finishing heat was that ordinarily given
in large furnaces. The ore did not show any sintering. Sulphur in
roasted ore, soluble, 1.16 per cent.; insoluble, 0.26 per cent.; total, 1.42
per cent. Average tailing from 10 bottle tests, 0.50 oz. Extraction,
91 per cent.
Test No. 2. — The ore was given a prolonged, roast to reduce the sul-
phur content. The finishing heat was quite high (about 1575°F.) but
the ore was not sintered. Sulphur in roasted ore, soluble, 0.90 per cent. ;
insoluble, 0.10 per cent.; total, 1.00 per cent.. Average tailing from 10
bottle tests, 0.33 oz. Extraction, 94 per cent.
Test No. 3. — The ore was roasted at a high heat, and finished at a
very high temperature (about 1650 to 1700° F.) ; it had a dark appearance
and was slightly fused. Sulphur in roasted ore, soluble, 0.34 per cent.;
insoluble, 0.09 per cent.; total, 0.43 per cent. Average tailing from 10
bottle tests, 1.23 oz. Extraction, 7S per cent. Two bottles were
recharged and treated 12 hours; recharge tailings ran 0.79 oz. Extrac-
tion, 86 per cent.
Test No. 4. — The ore was roasted 10 hours in the mufHe. It was
brought quickly to a red heat, and finished at a moderately high tem-
perature. The ore had a dark appearance, and the finer particles were
slightly fused. Sulphur in roasted ore, soluble, 0.75 per cent.; insolu-
ble, 0.11 per cent.; total, 0.86 per cent. Average tailing from 10
bottle tests, 0.73 oz. Extraction, 88 per cent.
From Tests 3 and 4 it is evident that if the ore is fused or sintered,
a close extraction is impossible.
Test No. 5. — ^The ore was roasted 5 hours at a very low temperature
(scarcely visible red) ; it was taken out of the muffle and divided into
two parts. One half was returned to the muffle and roasted 8 hours
longer at a low dull red heat. The finishing heat was a dull red. The
entire roasting was performed at a prolonged low temperature. Sulphur
in roasted ore, soluble, 0.74 per cent.; insoluble, 0.25 per cent.; total, 0.99
per cent. Average tailing from 10 bottle tests, 0.15 oz. Extraction,
97.3 per cent.
Test No. 6. — The other half of the ore taken from No. 5, after roast-
ing 5 hours, was then put into the muffle and roasted 5 hours more and
finished at a higher temperature. The ore was not fused or sintered.
Sulphur in roasted ore, soluble, 0.69 per cent.; insoluble, 0.24 per cent.;
18 HYDROMETALLURGY OF COPPER
total, 0.93 per cent. Average tailing from 10 bottle tests, 0.48 oz.
Extraction, 91 per cent.
In the bottle tests for the different roasts, the conditions were kept
the same. The chemicals corresponded to 15 and 20 lb. of bleach, and
30 and 40 lb. of sulphuric acid, per ten of ore. The time of treatment,
on account of the high grade of the ore, was 5 hours. The ore was ground
to 16 mesh.
It will be noticed that the best results were obtained from No. 5,
where the finishing heat was quite low. By increasing the heat, the
extraction was not improved, as shown by No. 6. In No. 1, the ore was
not roasted sufficiently, as indicated by the sulphur analysis of 1.42
per cent. In No. 3 the ore was roasted at too high a temperature; the
sulphur content is low, 0.43 per cent. The sulphur in No. 2 represents
more nearly the mill roast. For ore having 4 per cent, sulphur, the
soluble and insoluble sulphur in the roasted ore, as represented by No.
2, might be considered normal for Cripple Creek ore. A prolonged low
heat, as represented by No. 5, will give the best average extraction.
The conditions there represented, however, could probably not be fully
realized in practice, on account of the reduced capacity of the furnaces.
These test were repeated in large furnaces roasting 100 tons of ore
daily, and each test represents a day's run, or 200 tons. The extraction
is based on the mill tailings. Furnace No. 1 was fired with low initial
heat and a higher finishing heat. Furnace No. 2 was fired with a higher
initial heat and a lower finishing heat. Care was taken to get the best
possible roast under both conditions.
Furnace Test No. 1
Furnace No. 1 Furnace No. 2
Sulphur, raw ore, 2 . 75 per cent. 2 . 50 per cent.
f soluble, 0 . 75 per cent. 0 . 79 per cent.
Sulphur in roasted ore, I insoluble, 0 . 09 per cent. 0 . 14 per cent.
[ total, 0 . 84 per cent. 0 . 93 per cent.
Assay of roasted ore, gold, 1 .05 oz. 1 .85 oz.
Assay of chlorination tailings, 0.08 oz. 0. 11 oz.
Extraction, 92 . 4 per cent. 94 . 6 per cent.
Furnace Test No. 2
Sulphur, raw ore, 2 . 66 per cent. 2 . 66 per cent.
f soluble, 0.71 per cent. 0.77 per cent.
Sulphur, roasted ore, < insoluble, 0 . 07 per cent. 0 . 08 per cent.
[ total, 0 . 78 per cent. 0 . 85 per cent.
Assay of roasted ore, gold, 1 . 23 oz. 1.11 oz.
Assay of chlorination tailings, 0.12 oz. 0.07 oz.
Extraction, 90.0 percent. 94.0 percent.
These comparative tests in actual mill practice, in furnaces roasting
100 tons of ore per day, are characteristic of many others made along
OXIDIZING ROASTING
19
the same linea. The results clearly indicate that the lower finishing
heat gives the best average results; that ore roasted at a low temperature
gives up its values better than ore roasted at a high temperature, and
that sintering or overheating is highly injurious. The ore in these tests
was ground to 10 mesh. Time of chlorination, 3 hours. The chemicals
used were 15 lb. of bleach and 30 lb. of sulphuric acid, per ton of ore.
Valentine's Temperature Experiments. — Valentine' made interesting
experiments on the effect of temperatures on iron pyrite, with and with-
out free access of air. The results of his experiments undertaken to
ascertain the effect of heat on FeSj, when air is freely given access, are
given as follows :
Approximate
Duration of
1
S. in residue, j
Loss,
Per cent, of
temperature, deg. F.
heat
per cent.
j
per cent.
S. expelled
Original pyrite
1250
53.43 '
1 hour
4.27
49.15
92.05
1250
2 3/4 hours
0.70
52.73
98.68
1600
20 minutes
0.78
52.65
! 98.54
1600
45 minutes
0.08
53.35
99.85
1800
20 minutes
0.13
53.29
99.75
1800
1 hour
0.65
52.78
98.79
2200
15 minutes
3.23
50.19
93.95
2200
20 minutes
5.92
47.51
88.93
2200
35 minutes
1.56
51.87
: 97.10
2200
2 hours
1.18
52.24
97.78
It will be noticed that a larger amount of sulphur remains in the
residues when higher temperatures have been applied.
Valentine, from his experiments, draws the following conclusions
in roasting pyritic ores:
1. Heat alone without access of air, can remove at best only one
half of the sulphur present.
2. Atmospheric oxygen is absolutely necessary for a proper desul-
phurization.
3. Even at a low heat, ore is properly desulphurized if air can gain
access freely to the FeSj in it.
4. Sulphate of iron can be decomposed equally well with or with-
out air.
.5. In order that the residuum sulphur in roasted ore may consist as
far as possible of sulphates, the roasting must be done under free access
of air.
' Trans. A. I. M. E., Vol. XVIII.
20 HYDROMETALLURGY OF COPPER
6. Fusion or sintering of ore is likely to retard further desulphur-
ization.
7. Sintering does not allow much of the remaining sulphur to be in
the form of sulphate.
Air. — Much oxygen is consumed in roasting sulphides, and a highly
oxidizing atmosphere is essential to good results. With the sulphur
fumes, and the products of combustion from the fire-boxes passing over
the partly roasted ore in the rear of the furnace, the atmosphere, while
it may not be strongly reducing, is certainly not highly oxidizing. A
comparatively small amount of sulphur dioxide in the furnace gases
will greatly retard oxidation; and if, in addition to sulphur dioxide, the
atmosphere is charged with carbon dioxide from the combustion of the
fuel, effective roasting is impossible. The only advantage to be gained
in passing these deleterious gases over the fresh ore as it is introduced
into the roasting furnace is to heat it so that oxidation can proceed more
rapidly when it reaches a more highly oxidizing atmosphere.
It is desirable to bring the sulphides to the ignition temperature as
soon as possible after the ore has been introduced into the furnace.
The value of the vitiated hot gasses passing over a long stretch of cold
ore to heat it and thereby save fuel is largely overestimated. The loss
will exceed the gain.
With a deep bed of ore in the furnace, say from 3 1/2 to .5 in., even
in a highly oxidizing atmosphere, only the ore on the surface is under
thorough oxidizing conditions, while that below the surface is not so
advantageously placed. If overheated, therefore, fusion or matting is
likely to occur in the early stages of the roasting, and when the ore is
heated too quickly to get a correspondingly quick oxidation. If fusion
occurs, the particles assume a dark, and sometimes glazed, appearance.
In this condition it is more difficult to sufficiently eliminate the remaining
sulphur. The metals, too, are difficult to extract in the chemical process,
owing to the inability of the solvent to penetrate the pores of the ore
particles.
Speaking within practical limits, it is not so much the high tempera-
ture as the lack of air that is fatal to rapid and thorough roastino- in
reverberatory furnaces. Fusion of the sulphide particles invariably
occurs when the ore is brought too suddenly against a high temperature
with insufficient air. The tendency is to convert the sulphide into
matte. Much of the sulphur in the deeper portions of the bed vola-
tilizes as such, and when it reaches the surface it burns to sulphur dioxide.
Heat, without access of air, can remove only about 50 per cent, of the
sulphur originally in the ore.
In order to determine the effects of time and air on roasting and
extraction, comparative tests were made with furnaces roasting 100 tons
of Cripple Creek ore daily, containing about 2.75 per cent, sulphur.
OXIDIZING ROASTING
21
The ore in furnace No. 1 was roasted under normal conditions; the bed
of ore was about 2 1/2 in. deep; the angle of the rabble blades was 22 1/2
degrees, and the ore was about 2 1/2 hours in passing through. In
furnace No. 2 the angle of the rabble blades was changed to 12 degrees,
which resulted in having the bed about 4 1/2 in. deep; the ore remained
in the furnace about 5 hours to get the same capacity. All the other
conditions remained the same, so far as they could be kept the same.
Test No. 1
Sulphur, raw ore,
I soluble,
insoluble,
total,
Assay, roasted ore, gold.
Assay, chlorination tailings,
Extraction,
Sulphur, raw ore,
r soluble,
Sulphur, roasted ore < nsoluble
[ tota ,
As.say, roasted ore, gold.
Assay, chlorination tailings.
Extraction,
Sulphur, raw ore,
{soluble,
insoluble,
total.
Assay, roasted ore, gold,
Assay, chlorination tailings.
Extraction,
Furnace No. 1
90 tons per day
2 . 63 per cent.
0 . 73 per cent.
0.17 per cent.
0 . 90 per cent.
1.07 oz.
0 . 09 oz.
91.5 percent.
Test No. 2
Furnace No. 1
100 ons per day
2. 80 per cent.
0.79 per cent.
0. 14 per cent,
0.93 per cent.
1.80 oz.
0.12 oz.
93 . 3 per cent.
Test No. 3
Furnace No. 1
100 tons per day
2 . 62 per cent.
0.76 per cent.
0 . 14 per cent.
0 . 90 per cent.
1 . 90 oz.
0.12 oz.
94 . 0 per cent.
Furnace No. 2
90 tons per day
2.53 per cent.
0.94 per cent.
0 . 26 per cent.
1 . 20 per cent.
2.22 oz.
0.28 oz.
87 . 4 per cent.
Furnace No. 2
70 tons per day
2 . 75 per cent.
0 . 72 per cent.
0.25 per cent.
0.97 per cent.
1.19 oz.
0.19 oz.
85 . 3 per cent.
Furnace No. 2
70 tons per day
2 . 50 per cent.
0. 59 per cent.
0 . 23 per cent.
0 . 82 per cent.
1.08 oz.
0.29 oz.
73 . 0 per cent.
In f.urnace No. 2 it was soon found that a capacity of 100 tons per
day was out of the question. The capacity was at once reduced to 70
tons to give, what at least appeared to be, a fair roast. The bed of ore,
which with a capacity of 100 tons, was about 5 in. deep, with 70 tons,
was reduced to 4 in. The high tailings in No. 2 may have been due, in a
measure, to the higher temperature frequently necessary to eliminate
the sparks from the roasted ore.
22 HYDROMETALLURGY OF COPPER
It will also be seen from the sulphur analyses that whilein test No. 1,
the sulphur is higher in furnace No. 2 than in furnace No. 1, in tests 2
and 3 , it is lower ; nevertheless the, extraction was not improved. It will be
noticed, however, that the insoluble sulphur in furnace No. 2 is abnormally-
high as compared with the insoluble sulphur in furnace No. 1. This is
evidently due to lack of air, and perhaps higher temperature, in furnace
No. 2 to get approximately the same total sulphur elimination as in furnace
No. 1.
By increasing the bed from 2 or 2 1/2 in. to 4 or 5 in. which makes
the penetration of the air more difficult, the capacity was reduced from
100 to 70 tons per day, and the quality of the roast was very inferior,
notwithstanding that the time of roasting was practically doubled.
The ore in these tests was crushed to 12 mesh, and chlorinated 3
hours with a chemical charge of 15 lb. of bleach and 20 lb. of acid per
ton of ore.
The time of roasting, of 5 hours, in reverberatory furnace No. 2,
may be compared to the time the ore is subjected to roasting in a shaft
furnace, which may be considered about half a minute ; or 1 / 600 of the time.
The temperature in both cases may be considered the same; the difference
in the results, therefore, is due to the difference in air supply.
Interesting experiments were made in Denver to determine the effect
of an abundance of air supply in roasting charges of 2000 lb. of ore in
a hand-rabbled reverberatory furnace. In these experiments, some of
the charges were roasted in the ordinary way, while in others arrange-
ment was made to pass air through the incandescent roasting ore, both
by up-draft and down-draft; other conditions remained the same. The
experiments proved that the capacity of the furnace, due to the extra
air supply, was trebled in roasting a heavy sulphide ore; an appreciable
saving of fuel was effected, and the sulphur elimination was more perfect.
The amount of air required in practice in roasting is enormously in
excess of that required to combine with the sulphur and other elements.
Theoretically, at least, the air in all parts of the furnace should be kept
as pure as possible; on the other hand, the cost of heating a large volume
of excess air is considerable. In practice, therefore, the best results will
be obtained by carefully balancing these two opposing factors.
The amount of sulphur dioxide in the flue gases for effective roast-
ing should not exceed 2 per cent. When the sulphur dioxide in the
furnace atmosphere reaches 4 per cent., roasting becomes slow; when
it reaches 8 per cent., it becomes very slow; and when it reaches 9 per
cent, and over, the reactions practically cease.
Rabbling. — Rabbling is an important operation in roasting. Its
object is essentially to expose fresh particles of the ore to the direct action
of the air and heat, and to facilitate bringing the entire mass of ore to
incandescence, and thus assuring a uniform roast.
OXIDIZING ROASTING 23
It is evident that the roasting is facilitated by frequent rabbling,
but the frequency of the rabbling is limited by the rabbling mechanism
in mechanical furnaces, and by the fatigue of the roasterman, in hand-
rabbled furnaces. Theoretically, the more the ore is rabbled, the better
will be the roast, and this theoretical condition should be approached as
closely as possible. It is for this reason, more than all others, that
mechanical furnaces give a much better roasted product than hand-
rabbled furnaces. No hand rabbling, on a large scale, can approach
the frequency and uniformity of mechanical rabbling.
Effect' of Metallic Sulphides if Heated with Exclusion of Air.— Gold
and platinum can be completely desulphurized. The sulphide of silver
(AgS) remains undecomposed. The sulphides of arsenic, antimony,
and mercury, volatilize unchanged. Iron pyrites (FeSj) gives up 23
per cent, of its sulphur, and is reduced to magnetic pyrites (FegSg), which
by a strong heat may be reduced to ferrous sulphide (FeS). The ferrous
sulphide is not further reducible. Of the copper minerals, chalcocite (CujS)
is not decomposed, but the chalcopyrite (CuTeS,) loses only one part of
the sulphur which is combined with the iron. Galena (PbS) is reduced
to a lower stage with separation of metallic lead.
Sulphur. — Sulphur usually occurs combined with the base metals as
sulphide, but not infrequently the ore is highly charged with sulphates,
due to partial decomposition by atmospheric and aqueous agencies.
Sulphur, combined with some of the metals as sulphide or sulphate, is
highly injurious in the hydrometallurgical process; and if occurring in
large quantities, it is fatal. Many of the sulphates are acted upon by
acids; in any cases the soluble sulphates affect the leaching solution
injuriously. In the cyanide process some of the sulphides, as for example
pyrite, are not particularly injurious, while most of the sulphates offer
difficulties. In the chlorination process, the sulphur in combination
with some of the metals is displaced by the chlorine, which itself unites
with the metal or acts as an oxidizer. In either case, the chemicals are
consumed by reacting with the base elements, and are not available for
action on the desired metals. Roasting, in any event, largely overcomes
these difficulties, and in many cases practically eliminates them entirely.
Sulphur is rarely, if ever, entirely eliminated during the roasting.
Frequently that which remains is not injurious to the process. The
sulphur, as sulphide or sulphate, may be encased in quartz particles, or
it may be in the ore as sulphates of the alkali metals or of the alkaline
earths. So combined, it is not replacable by any of the chemicals ordi-
narily used. The sulphates of sodium, potassium, barium, calcium, and
magnesium appear to be unaffected by either hydrochloric and sulphuric
acids, chlorine, bromine, cyanide, or sodium hyposulphite. Their pres-
ence, in the solution, may however have some effect on the solubilities
of the various solvents.
24 HYDROMETALLURGY OF COPPER
Most of the sulphur in either copper, gold or silver ores is usually
combined with iron, as iron pyrite (FeS^) . One of these atoms of sulphur
may be distilled, or be burned to sulphur dioxide at a low temperature.
In the roasting of sulphides, sulphur dioxide is exclusively formed. In
the presence of air, by catalytic action with indifferent substances such
as silica or iron oxide, there is always formed a small amount of sulphur
trioxide, which with the moisture of the air and that contained in the ore
and fuel gases, gives sulphuric acid.
The elimination of sulphur from concentrates or heavy sulphide
ore is accompanied by the evolution of considerable heat. Concentrates
containing from 25 to 35 per cent, sulphur are frequently roasted down to
5 per cent, by the heat generated from their own oxidation. Thirty-two
parts 0^ sulphur, in combining with 32 parts of oxygen (that is, forming
SO2), evolves 69,260 heat units; and if the oxidation proceeds to SO3,
91 ,900 heat units are evolved. These figures may be compared with those
which correspond to the passage of carbon into carbon monoxide (CO)
and carbon dioxide (COJ when 29,160 and 97,200 units of heat, respec-
tively, are evolved. The evolution of heat by the rapid oxidation of
sulphur is practically demonstrated in the various sinter-roasting
processes in which copper and lead sulphide ore and fines are fused into
a coherent mass by the heat from the sulphur alone.
The elimination of sulphur, in roasting, varies greatly with different
ores. Some forms of pyrite are more difficult to roast than others.
Unoxidized ores from the deeper levels of a mine are more difficult to
roast than the partially oxidized ores nearer the surface, even though
the sulphur content of both is approximately the same. The chemical
composition of the ore, aside from its sulphur content, has much to do
with the roasting. Ferrous sulphate, for example, is much more easily
broken up than zinc sulphate, or than the sulphates of the alkalies or
alkaline earths. Ore containing much lime is likely to be high in sulphur
after roasting.
In some of the Cripple Creek ores having 2.75 per cent, sulphur, the
best extraction is frequently made, and without undue consumption of
chemicals, when the roasted ore contains from 0.60 to 0.80 per cent,
sulphur. When the total sulphur in the roasted ore is less than 0.40 per
cent, the tailing are usually high. Other Cripple Creek ores, which are
partially oxidized, give the best extraction when the sulphur is from 0.30
to 0.50 per cent., the insolubles usually going from 0.03 to 0.08 per cent.,
and the solubles from 0.35 to 0.40 per cent.
It is customary in many of the mills to make frequent sulphur deter-
minations. Sometimes they are made for such shift for every furnace;
sometimes once a day. These sulphur determinations are made both
for soluble and insoluble sulphur. The insoluble sulphur is more par-
ticularly relied upon to indicate the roast. The soluble sulphur is that
OXIDIZING ROASTING 25
which is soluble in boiling water; usually a little sodium carbonate is
added before boiling.
The progress in the elinaination of the sulphur, when treating 100
tons of Cripple Creek ore daily in large mechanical furnaces, is shown
by the following samples taken at various points in the furnace during the
roasting. The ore remained in the furnace about 2 3/4 hours, so that
the distance, in feet, from the feed will also closely represent the time
in minutes for the ore in the furnace, when the respective samples were
taken. The ore was rabbled every 17 seconds. The results are averages
of a large number of sets of samples taken from three different furnaces.
The samples were taken so as to fairly represent the total cross section of
the ore. No. 3 was taken from a type of furnace totally different from
the others. The ore before roasting was crushed to 12 mesh.
No. 3 furnace had a revolving hearth and was fired at a lower tem-
perature than the others. Notwithstanding the high sulphur content
in the roasted ore from No. 3, the extraction by chlorination was about
the same as for the others. The tailings from No. 2 were somewhat
higher than from No. 1; this was doubtless due to the fact that in No. 2,
with three fire-boxes, the ore had to be roasted at a higher temperature
than in No. 1, which had four fire-boxes.
PROGRESSIVE SULPHUR DETERMINATIONS MADE IN ROASTING
CRIPPLE CREEK ORE, IN FURNACES ROASTING 100 TONS
OF ORE DAILY
Furnace No. 1 (Four Fire-boxes)
Sample taken, feet (also
approximate time in
min.) from feed
Raw ore
45 ft. Minutes. .
65 ft. Minutes. .
100 ft. Minutes. ,
160 ft. Minutes. .
Cooler
Distance
of fire-box
from feed
in feet
55
72
120
140
Insol.
per cent.
Sulphur
Soluble
per cent.
2.60
0.86
0.41
0.23
0.14
0.10
0.00
0.34
0.49
0.60
0.63
0.66
Total
per cent.
Difference
and differ-
ence in per
cent.
Per cent
of S com-
pared to
raw ore
Per cent
of S com-
pared to S
eliminated
2.60
1.20
1.40
76
54
76
0.90
0.30
16
65
92
0.83
0.07
4
68
96
0.77
0.06
4
70
100
0.76
0.01
Furnace No. 2 (Three Fire-boxes)
2.58
1.54
0.74
0.35
0.16
0.14
0.00
0.19
0.40
0.51
0.54
0.62
2.58
1.73
1.14
0.86
0.70
0.76
1
0.85 ;45
0.59 ;31
0.29 ; 15
0.16 ; 8
33
56
65
73
45
65 ft. Minutes
100 ft. Minutes
160 ft. Minutes
72
120
140
86
91
100
26
HYDROMETALLURGY OF COPPER
Furnace No. 3 (Revolving Hearth)
2.54
1.12
0.30
0.17
0.10
0.00
0.29
0.80
0.93
0.90
2.54
1.41
1.11
1.10
1.00
1.13
0.30
0.01
.0.10
73
20
1
6
44
56
56.4
60.4
73
93
94
165 min. from teed (roasted ore)
100
In these results it will be noted that much of the sulphur was driven
off early in the operation, and before passing the first fire-box. This
is a practical demonstration of the instability of the first atom of sulphur
in iron pyrite. The difficulty of eliminating the last 25 per cent, is
apparent; the difficulty of eliminating the third 25 per cent, is considerable.
Only a small fraction of the total sulphur is expelled during the last 100
minutes of the 160 minutes of roasting, while a large portion (about one-
fourth) still remained in the ore. Roasting a low-sulphur ore down to a
trace is evidently as difficult as to extract all but a trace of the met-
als. Roasting sulphide ores down to a "trace of sulphur" and extract-
ing all but a "trace of the metals" are operations frequently spoken of
but rarely truthfully realized.
It will be noticed that only from 60 to 70 per cent, of the sulphur
was eliminated in these roasts. It is safe to say that fully two-thirds
of the total fuel was consumed in expelling only a small fraction of a
per cent, during the latter half of the operation. In the ordinary
reverberatory furnace there does not appear to be any adequate gain,
in the latter part of the roasting, for the fuel expended.
The tables of the progress of roasting give a fair idea as to the rate of
decrease of the insoluble sulphur and increase of the soluble sulphur.
After the first 65 or 70 minutes, the principal result accomplished by
roasting, is the changing of the remaining insoluble sulphur to the
soluble. In furnace No. 3, for example, there is only a difference of 0.11
per cent, in the total sulpur, between the roasted ore and the first 70
minutes of roasting, but for the remaining 85 minutes, 0.30 per cent,
insoluble sulphur was changed to 0.10 per cent., and this represents the
difference between a go.od and a poor roast.
It sometimes happens that the sulphur in the roasted ore from the
cooler is higher than the discharge from the furnace; this may be ac-
counted for by the fact that the rabbles frequently push more or less
partly roasted ore ahead of them in the grooves made by the preceding
rabble, or the rabbles themselves may carry partly roasted ore through
the furnace and discharge it on the cooler.
Within certain limits, the sulphur content of the roasted ore does not
appear to affect the extraction; beyond these limits the effect is marked.
Nothing would be gained in extraction by roasting the ore represented
in the tables to say 0.05 per cent, insoluble and 0.20 per cent, to 0.40 per
OXIDIZING ROASTING 27
cent, soluble sulphur, while the extra cost of roasting to such a low sul-
phur content would be enormously increased. There would also be
high tailings from overheating.
There is quite as much danger of over-roasting as under-roasting.
Ore roasted too much will give high tailings, nor can these tailings be
materially reduced by repeated charges of chemicals. If ore is under-
roasted, repeated charges of chemicals may be necessary to get the
desired extraction; but the tailings will be reduced each time, and ulti-
mately the limit of extraction may be obtained.
An experiment was made with a 100-ton furnace to determine the
effect of ultimate roasting on the extraction. The furnace was fired
under normal conditions; but instead of treating the ore after its first
passage through the furnace, it was returned again and again for 12 hours.
It is needless to say that the ore was roasted "dead"; nevertheless four
charges of chemicals on this ore failed to give even the average extraction.
One of the essential features of roasting is to find the sulphur deter-
minations which will give the best results, and to find the point where a
lower sulphur content will not appreciably increase the extraction.
The sulphur in the roasted ore, from one mine or from one district, which
has proved to give the best results, might be fatal to the treatriient of
ore from another district. This is largely due to the way in which the
sulphur is combined. Soluble sulphur, as sodium or potassium sulphate,
is unaffected by the chemical solvents; if the same amount of sulphur
were combined with iron, as ferrous sulphate, the roast would be abso-
lutely worthless. Again, the insoluble sulphur, as barium or calcium
sulphate, is not particularly detrimental to the subsequent treatment;
but if the same amount of sulphur is combined as sulphide, it is almost sure
to be fatal. Insoluble sulphur does not necessarily imply that the
sulphur is in the form of sulphide.
In the roasts, as shown in the accompanying tables, the iron sul-
phate was practically eliminated at 100 ft. from the feed, and totally
eliminated ait 120 ft. In some instances it was totally eliminated at
100 ft., as shown by the ferricyanide test, which did not produce the
usual delicate reaction for iron. The ferricyanide test, except for making
a rough determination, is absolutely worthless, since it shows only the
sulphur combined with the iron as soluble sulphate. It frequently
happens that the ore is far from being roasted, when the sulphate of iron
is all decomposed.
The barium chloride test, except as a rough indication, is also worth-
less, since it precipitates sulphur that might be considered as perfectly
harmless. If the ore, as shown in the tables, had been roasted so
thoroughly that no sulphur had been precipitated with barium chloride,
it is safe to say that only a comparatively few tons of ore could be roasted
in a day, and the tailings would be quite sure to be high. A considerable
28 HYDROMETALLURGY OF COPPER
temperature would be required to break up the sulphates of the alkaline
earths, and such a temperature would be detrimental to the extraction of
the desired metals.
A direct sulphur determination seems to be the only way of indicating
the roast with any degree of accuracy. It is not so much a matter of
absolute refinement in the sulphur determination as long as the results
are uniform. Nevertheless, care and accuracy are essential to uni-
formity. The essential idea of the sulphur determination is to indicate
the quality of the roast as compared with the extraction. It is not very
material whether the sulphur is relatively high or low. If a certain
sulphur determination indicates a good or a bad roast one day, theoretic-
ally at least, it should indicate the same at any other time. It is evident
that these ideal conditions cannot always be realized, because the work-
ing conditions of the furnace change from time to time. For example,
ore which is over-roasted for four hours on a shift and under-roasted
the other four, if averaged, might give the same sulphur determination
as ore which had been evenly and uniformly roasted for the total eight
hours; but the tailings would be entirely in favor of the latter roast.
Unless it is known that the conditions of operating the furnaces have
changed, it is fair to assume that they have remained the same, and this
is usually the case in well conducted plants.
In order to get uniformity in the roasting operation and in the
roasted product, it is essential that the raw ore fed into the furnace
should average about the same in sulphur. If the ore comes from
different mines, or from different levels of the same mine, it should be
mixed. Extreme care in mixing is neither necessary nor profitable.
It is more economical to build the furnace of ample capacitj^, so that
small variations in the ore will naturally be taken care of, under uniform
conditions of operation.
The quantity of sulphur which roasted ore may contain without
particular detriment is variable, depending largely upon the lime, mag-
nesia, and lead.
Careful tests have demonstrated that an abundance of air is not con-
ducive to the formation of sulphates. Air does not appear to be neces-
sary to decompose sulphates; nevertheless, when it is supplied in abun-
dance the decomposition of the sulphates is greatly facilitated. -Other
substances, by catalytic action, may also aid in their decomposition.
Decomposition Temperature of the Various Sulphates. — With the ex-
ception of lead sulphate, all the common metallic sulphates are com-
pletely decomposed upon heating, into metallic oxide, sulphur trioxide,
sulphur dioxide, and oxygen. Some give up their trioxide readily at
low temperatures, others require considerable heat and much time, to
be completely freed from sulphur. Kerl, in 1881, arranged the prin-
ciple metallic sulphates, as they are decomposed by a rising temper-
OXWlZINd ROASTING
29
ature, in the following order: silver, iron, copper, zinc, nickel, cobalt,
manganese, and lead. Lead sulphate is decomposed only at a white
heat. Bradford' found that ferrous sulphate is decomposed at 590° C,
cupric sulphate at 653° C, and silver sulphate at 1095° C. H. 0.
Hofman- found that zinc sulphate is decomposed at 739° C.
In the presence of air and other gases, and various other substances
in the ore, the temperature of the decomposition of the various metallic
sulphates may be vitally affected. The decomposition of silver sulphate
takes place at from 860 to 870° C. in the presence of cupric oxide, silica,
and ferric oxide. In the presence of reducing gases, silver sulphate
is decomposed at a very moderate heat, resulting in the formation of
rnetallic silver.
Prof. H. O. Hofman and W. Wanjukow determined the decomposition
temperature of some metallic sulphates in a current of air, as follows:
DECOMPOSITION TEMPERATURE OF VARIOUS SULPHATES
Sulphate
FeSO,
Fe,(S0j3
Bi,(S0j3
CuSO.
MnSO,
2CUO.SO3
NiSO,
CoSO,
ZnSO,
CdSOj
5Bi,0,.4(S03)3
SCdOSOj
CaSO.
BaSO,
Temperature
Product
Deg. C.
Deg. F.
150 302
FeA(S03)3
530 j 986
Fc.O,
540 1 1004
581203.4(803)3
653 ' 1207
2CUOSO3
680
1256
MnjO,,
704
1299
CuO
708
1306
NiO
718 I 1324
CoO
739 1362
ZnO
830 1526
5CdO.S03
850 1562
Bi,03
875 1607
CdO
1200
2192
CaO
1500
2732
BaO
(New York meeting of the American Institute of Mining Engineers, Feb., 1912.)
Amount of -Sulphur Trioxide, (SO3) in the Sulphur Dioxide (SOj)
Escaping from Roasting Furnaces. — In the roasting of Spanish py-
rites for the manufacture of sulphuric acid it was found that from 2 to 3
per cent, of all the sulphur dioxide was converted into sulphur trioxide.
Lunge'' found in two experiments with burning Spanish cupriferous
' Trans. A. I. M. E., 1903, Vol, XXXTIL
2 Trans. A. I. M. E., 1905.
' "Sulphuric Acid and Alkali Manufacture."
30 HYDROMETALLURGY OF COPPER
pyrites, containing 48.62 per cent, sulphur, in a glass tube, in a current
of air :
1 2
Sulphur obtained as SO2, 88 . 02 per cent. 88 . 78 per cent.
Sulphur obtained as SO3, 5 . 80 per cent. 6 . 05 per cent.
Sulphur in residue, 3. 42 per cent. 1 ^ I7 per cent.
Sulphur lost, 2 . 75 per cent. /
Of the sulphur of the burner gas there were present:
1 2
As SO2, 93 . 83 per cent. 93 . 63 per cent.
As SO3, 6 . 1 7 per cent. 6 . 37 per cent.
Two experiments were made in this way; in the glass tube 50 grm. of
cinders, from the same pyrites, in pieces of the size of a pea, were com-
pletely freed from sulphur by ignition, and fresh pyrites burned as before,
the gas passing through the cinders. There were found:
3 4
Sulphur as SO^, 79 . 25 per cent. 76 . 90 per cent.
Sulphur as SO3, 16 . 02 per cent. 16 . 84 per cent.
Sulphur in residue and loss, 4 . 73 per cent. 6 . 26 per cent.
Of the sulphur in the burner gas itself there were present :
3 4
As SO2, 83 . 18 per cent. 82 . 00 per cent.
As SO3, 16. 82 per cent. 18.00 per cent.
Experiments made by Scheurer-Kestner with gases from fur-
naces roasting pyrites for the manufacture of sulphuric acid show that the
sulphur trioxide is quite variable. One set of determinations were
made from two samples taken at various times from a lump kiln burner,
and the other from a Maletra fine ore burner.
The average results of these experiments were :
Vol. Sulphur converted
per cent. into SO3; per cent,
of SO2 of total sulphur
Lump burner, 7 . 5 per cent. 3 . 1 per cent.
Fine ore burner, 8 . 9 per cent. 3 . 5 per cent.
The presence of sulphuric acid in the sulphur fumes, especially
those from muffle furnaces, is interesting as showing the formation of
sulphuric acid, probably mostly by catalytic action. This is clearly
shown in Lunge's experiments Nos. 3 and 4, where the sulphur dioxide
was passed through a column of roasted ore, mostly ferric oxide, and
the amount of sulphur trioxide increased from 6.17 per cent, and 6.37
per cent, as shown in experiments 1 and 2, to 16.82 per cent, and 18.00
per cent., as shown in experiments 3 and 4. The amount of sulphuric
acid in the fumes is also interesting from the fact that the acid gases
absorbed in water has been used as the solvent in leaching copper ores,
mostly, however, after chloridizing roasting.
OXIDIZING ROASTING 31
Sulphur Determinations. — The following method of making sulphur
determinations is used in Cripple Creek mills, where from 30 to 40
analyses are frequently made daily. Instead of determining the soluble
and insoluble sulphur from one weighing, two weighings are usually
made; one for the total and one for the soluble sulphur. A complete set
of samples for one shift consists of an average sample of the raw and roasted
ore from each furnace. In this way a check is kept on the work done
by the different shifts.
The frequency with which sulphur determinations are made depends
upon the uniformity or changeableness of the ore. When ore is roasted
for a hydrometallurgical process it is the most important and one of the
most delicate steps in the entire treatment, and any indication of the
work done by the different shifts is desirable.
Total Sulphur. — Weigh 1.373 grm. of the finely powdered ore into a
No. 4 casserole. Add 10 c.c. of a saturated solution of potassium chlorate
in nitric acid. Cover with watch glass and boil to dryness. Add 10 c.c.
(15 c.c. for concentrates) hydrochloric acid. Evaporate down to about
one-half. Add 100 c.c. hot water and boil slightly. Add ammonia until
a precipitate of ferric hydrate forms, and then add 10 c.c. of a saturated
solution of ammonium carbonate. The ammonium carbonate is to convert
any lead sulphate to carbonate and thus render the combined sulphur
trioxide soluble as ammonium sulphate. Heat to boiling, remove from
the heat, let settle, filter, and wash thoroughly five or six times. Acidu-
late the filtrate with hydrochloric kcid, and then add 5 c.c. in excess.
Boil, and while boiling add a hot solution of barium chloride in slight
excess. Boil a few minutes longer and let settle. It is best, before
filtering, to let the mixture remain hot or boiling slightly, as long as pos-
sible, which greatly facilitates the filtering. Filter through a 9 cm. fil-
ter, and wash at least six times with boiling water. Ignite, and weigh
the barium sulphate.
Since 1.373 grm. were taken, the percentage of sulphur in the ore may
be read directly from the scales, 100 milligrm. of barium sulphate being
equal to 1 per cent, of sulphur in the ore. In weighing out the 1.373 grm.
of ore, instead of making the weighings with the usual gram and milligram
weights, a lead button or disc is carefully made so as to weigh 1.373 grm.,
and this lead button is then always used as the standard weight in making
sulphur determinations.
If 1/2 grm. of ore is taken, as may be desirable with ores high in
sulphur, the weight of the barium sulphate must be multiplied by
0.1373 to obtain the weight of the sulphur. To ignite the barium sul-
phate, the filter, with the precipitate, is placed in an annealing cup and
heated to redness in the muffle. The ignited barium sulphate should be
perfectly white.
32 HYDROMETALLURGY OF COPPER
In some of the Cripple Creek mills the step of adding ammonia and
ammonium carbonate is omitted.
Soluble Sulphur.— Weigh 1.373 grm. of the finely pulverized ore
into a No. 3 casserole. Add about 1/2 grm. sodium carbonate and
20 c.c. of water. Boil 5 minutes. Remove and filter; wash thoroughly
four or five times with boiling water. Add 10 c.c. hydrochloric acid
to the filtrate, boil, and while boiling add a hot solution of barium
chloride in slight excess. Boil a few minutes longer and let settle.
Filter and wash at least six times with boiling water. Ignite by placing
the filter and precipitate in an annealing cup and burn in the muffle till
white. Weigh as for the total sulphur.
Insoluble Sulphur. — The insoluble sulphur is determined by sub-
tracting the soluble sulphur from the total sulphur, by taking the two
weighings from the same carefully mixed sample, and making a soluble
determination on one, and a total determination on the other.
If it is desired to make the two determinations from one weighing
it is first treated for the soluble sulphur, and then for the insoluble, by
treating it in the same way for the total sulphur.
Tellurium. — Tellurium, the analogue of sulphur, is a common asso-
ciate of copper, gold, and silver ores. In recent years it has been found
that this clement is associated with gold in almost all of the great mining
districts of the world, even where not long ago its presence was unsus-
pected. It is very common in Cripple Creek, Colorado, in Goldfield,
Nevada, and in the Kalgoorlie mines of Australia. It occurs, though
less conspicuously, in the black hills of South Dakota, in the Mount
Morgan mine in Australia, and in the San Juan mines of Colorado. The
gold in many of the richest mines in the world is associated with tellu-
rium and sulphur; the ore is then known as a sulpho-telluride, although
the tellurium is rarely, if ever, found in gold, silver, or copper mining,
unaccompanied by sulphur. Tellurium also occurs quite universally
associated with copper, but in quantities so minute as to be of no special
metallurgical importance. Many of the copper ores of Arizona, Butte,
Montana, and of Australia contain small quantities of tellurium — rarely
exceeding two or three hundredth per cent. The matte from the Copper
Queen, Arizona, contains 0.00088 per cent, tellurium, while that from
Butte contains from 0.001 to 0.01 per cent. The anode slimes from
electrolytic copper refining, at Butte, contains from 2 to 3 per cent,
tellurium and selenium, and in some electrolytic refineries the slimes
contain as high as 5 per cent, of these elements.
In the roasting of telluride ores, the tellurium of itself is not of any
great metallurgical importance. At the most, the quantity of it is usually
exceedingly small as compared with the sulphur and other constituents.
Cripple Creek is widely known as a telluride camp, and yet mill samples
rarely show more than a trace of tellurium, and frequently not that
OXIDIZING ROASTING 33
Kalgoorlie is probably the richest tellurium district j'et discovered, and
yet typical ore analyses show only from 0.03 to 0.10 per cent, of tellurium.
Nevertheless, tellurium is often a source of anxiety to the metallur-
gist. In sulphide ores the metals, principally gold, are usually fairly
evenly distributed through the rock, and the mineral contained in the
rock; but not so in telluride ores. By far the greater gold values are
intimately associated with the tellurium, so that where a speck of tellur-
ium occurs, there is likely to be -associated with it an appreciable quantity
of gold also. Tellurium, unlike sulphur, is not usually evenly distributed
through the rock. It is ordinarily concentrated within small areas and
cleavage planes. The greater portion of the gold in a ton of telluride
ore is frequently concentrated in a few rich places. It is this character-
istic of telluride ores, as in ores containing free gold in particles of
appreciable size, that makes their treatment difficult by a chemical
process.
Tellurium fuses at 500° C. (930° F.) and volatilizes at a higher tem-
perature (from 550 to 575° C). When roasting in an oxidizing atmos-
phere, it burns with a blue flame edged with green. Sylvanite melts
easily tinging the flame greenish-blue. The tellurium combines with
oxygen to form tellurium dioxide (TeOj) which corresponds with the
sulphur dioxide (SOj) formed in roasting sulphides.
The gold compounds of tellurium, usually sylvanite, petzite, and
calaverite, all fuse at a low heat, forming at first a globular mass, which,
when the tellurium is all volatilized, leaves behind a speck of gold of
definite proportions — frequently like a pin head. If the telluride par-
ticle is roasted at a low temperature, this speck of gold will be very porous,
and present a large surface for attack by the chemical solvent. If
roasted at a sudden very high temperature, it is likely to be round and
smooth; in this condition the solvents have no appreciable effect on it in
the time ordinarily given by a chemical process. When, however, the
tellurium is largely associated with sulphur, the sulpho-telluride particle
will not fuse, but the sulphur and tellurium will be driven off, as in the
case of sulphides, leaving behind the gold disseminated through the
ferric oxide particle. If the gold, in roasting, issues from its telluride
combination in a shotted form, the best recourse, after the chemical
treatment, is concentration. This has proved quite effective. The
gold, although usually having a bright yellow appearance, does not amal-
gamate well. Amalgamation has been tried repeatedly, but has not
proved the success that was anticipated; nevertheless, there seems to be
no logical reason why it cannot be successfully accomplished.
The difference in the roasting of a grain of sulphide and a grain of
telluride, both of which contain the same amount of gold, is likely to be
this; in the sulphide the sulphur is gradually expelled leaving the result-
ing grain of ferric oxide extremely porous and with the gold scattered
34 HYDROMETALLURGY OF COPPER
through it in perhaps microscopic particles; in the telluride, on the
contrary, unless the utmost care is taken, the grain is likely to fuse into a
plastic mass, from which all the gold contained in it will finally emerge
concentrated into one particle. The solvent, as subsequently applied
to the roasted sulphide particle, will extract a very high percentage of
the gold in a very short time; while the gold resulting from the roasting
of the telluride particle would scarcely be affected. Some telluride ores
are easily treated and show but little coarse gold, but the illustration
given shows why it is usual to find coarse gold in the tailings of telluride
ores, even when the chemical treatment has been apparently satis-
factory. It is highly probable also that some of the tellurium, by partial
fusion, may be converted into a compound which is insoluble, and which
resists further oxidation at higher temperatures. This is frequently the
case in the corresponding sulphur combinations, which are insoluble, and
from which it is difficult to drive off more sulphur by increasing the heat.
Tellurium is insoluble in water and in dilute sulphuric and hydroch-
loric acids. It is practically unaffected by chlorine, bromine, and pot-
assium cyanide. Gold and tellurium probably form true chemicals
compounds; if this is so, it is evident that the gold cannot be closely
extracted unless, in a measure, the tellurium is decomposed. Gold tel-
lurides are very compact and do not permit of much penetration by a gold
solvent. It is largely due to these facts that roasting of tellurides is
desirable to get low tailings by a solvent process.
There is no appreciable loss of gold by volatilization in the roasting
of telluride ores, although at very high temperatures some gold is doubtless
volatilized with the tellurium. Ordinarily, however, the conditions in
roasting are such that the tellurium is volatilized before the temperature
is sufficiently high to volatilize any of the gold with it. On comparing
the assays of the raw ore with the roasted ore, in mills treating sulpho-
telluride ores, no loss by volatilization is apparent with an oxidizing
roast. Neither does the dust in the dust chambers show a much higher
value in gold than the ore from which the dust was obtained. The slight
difference in value can be accounted for by the fact that the gold val-
ues in the ore are largely confined to the sulphides and tellurides, and
owing to their friability a larger proportion of the dust will result from
these than from the other constituents. If the gold were appreciably
volatile with the tellurium, the fumes on cooling in the dust chamber
would condense and appear, to some extent, in the furnace dust, and
there manifest itself in the assays.
Iron. — Iron is inseparably associated with copper, gold and silver ores.
While it is of great importance mineralogically and metallurgically, it
presents no serious problems. In raw ore it is frequently troubelsome
for wet methods, but all difficulties are practically eliminated by careful
roastings.
OXIDIZING ROASTING 35
Ferric oxide (Fefi,), which should be the ultimate condition of the
iron in all roasted ore, is insoluble and practically unaffected by all chem-
ical solvents of copper, gold, and silver. It is immaterial whether the
solvents are acid or alkaline, or whether they are dilute or somewhat
concentrated. If iron gives any serious trouble in roasted ore, it is
entirely due to imperfect roasting.
Many raw ores contain considerable quantities of iron as soluble
sulphate. This is very pronounced in some mines, and particularly,
in the zone of partial oxidation. When this occurs in appreciable
quantities, the treatment of the raw ore by any of the chemical processes
becomes difficult, and frequently impossible, since all of the chemicals
used in solvent processes are quickly affected by it. Washing the ore,
either with water or dilute acid or alkaline solutions, is not always effective.
Roasting effectively removes the difficulty by converting the ferrous
sulphate into the insoluble ferric oxide.
When copper, gold, and silver ores contain simply iron in any of its
various combinations with the usual quartz matrix, and without appre-
ciable quantities of any other foreign matter, they can be roasted quickly,
thoroughly, and cheaply, no matter what the sulphur content may be,
and the values easily recovered with a high percentage of extraction.
Iron, associated with copper, gold, and silver ores usually occurs in
the form of:
The Oxide, Hematite (Fe203), with or without the water of hydration.
The Sulphide, Pyrite (FeS^).
The Carbonate, Siderite (FeCO,) , which, while common, is not general.
When siderite is roasted it is decomposed according to the reaction:
FeC03=FeO + C02
and the molecule of FeO is subsequently converted to FeaOg by taking
oxygen from the air. The roasted carbonate may be strongly mag-
netic. The temperature must be carefully regulated, to avoid sintering
the ore, which because of the fusibility of ferrous oxide and silica, may
easily happen. According to Le Chatelier, the decomposition of fer-
rous carbonate takes place at 800° C. (1472° F.).
The oxidized iron, in copper, gold, and silver veins is usually the
result of the natural decomposition of the pyrite by aqueous and atmos-
pheric agencies. With depth in mines, below the influence of these
agencies, pyritic ore will be encountered, while at the surface the ore may
by perfectly oxidized.
The principle objects to be gained in roasting thoroughly oxidixed
ores are dehydration and agglomeration, which much facilitates the sub-
sequent chemical treatment. Some of the base elements are invariably
expelled or are put in better condition to resist the action of the chemicals.
Unless the iron is thoroughly oxidized to the dehyrated ferric oxide,
much of it is likely to go into solution with an acid solvent.
36 HYDROMETALLARGY OF COPPER
Most of the iron in oxidized ores is in the form of ferric hydrate,
known mineralogically as limonite (FezOg+Fe^ (OH)e). Ores which
consist largely of other substances, such as quartz and clay, usually
have the characteristic yellow appearance of ferric hydrate. By roasting,
the water of hydration is driven off, which converts the iron into the
ferric oxide (FejOj). The color at the same time changes from yellow
or brown to the familiar red of well-roasted ore. In some ore the red is
very intense.
The ferric hydrate gives off part of its water at a teniperature between
80 and 100° C. (176 and 212° F.) and all of it at a red heat. Intense
heat, in roasting oxidized ores, is not usually necessary, since sulphates
in appreciable quantities are ordinarily absent. It is the sulphates
which usually require a higher temperature for their decomposition.
Oxidized ores can be quickly and cheaply roasted; about all that is neces-
sary is to bring the ore to a good red heat.
The magnetic oxide, magnetite (FcjOJ, sometimes, though not
frequently, occurs associated with copper, gold and silver ores. When
roasted at a moderate temperature, with an abundance of air, it may
be converted into the ferric oxide. Roasting with salt appears to be
much more effective in bringing about this change than a simple oxidizing
roast. Both magnetic oxide and the ferric hydrate dissolve to some extent
in acids, and in a smaller degree are converted by chlorine into the
chloride.
If iron is contained in the raw ore as pyrite (FeSj), the first
action of the roasting is to expell one atom of sulphur. This should be
accomplished at a moderately low temperature and with an abundance
of air. The temperature in the early stages of the roasting should not
exceed a dull red. As long, however, as the ore does not show any tendency
to adhere and form into small lumps there is not much danger of over-
heating. It is well known that one of the atoms of sulphur in pyrite is
quite tenaciously combined with the iron, while the other is held only
by a feeble bond. In an oxidizing atmosphere at a temperature of
about 315° C. (600° F.) the molecule of pyrite beings to be decomposed.
The sulphur in pyrites, exposed to the direct action of the highly
heated atmosphere in the furnace, is converted at once into dioxide. In
the deeper portions of the bed, where it is difficult for the air to pene-
trate, the sulphur may be first volatilized, and on appearing at the sur-
face, also burns to sulphur dioxide:
FeS2 + 02=FeS + S02
FeS2-Fheat=FeS-fS
S+02 = S02
Some of the sulphur dioxide in the presence of large quantities of
incandescent oxides or quartz is converted, by catalytic action into
OXIDIZING ROASTING 37
sulphur trioxide, and the sulphur trioxide combining with the moisture
of the air and fuel is converted into sulphuric acid.
The ferrous sulphide, by combining with the oxygen of the air, is
converted into ferrous oxide:
FeS + 03=FeO + S03.
The ferrous oxide, by combining with more oxygen, may be converted
mto the magnetic oxide, or by combining with the sulphur trioxide,
may be converted into ferrous sulphate:
3FeO + 0=Fe304
FeO+S03=FeSO,.
The magnetic oxide, by proper heating or by combining with sulphur
trioxide may be converted into ferric oxide :
2Fe30, + 0 = 3Fe203
2Fe30i + S03 = 3Fe203 + S02
The ferrous sulphate, at a red heat, is decomposed into sulphur
dioxide, ferric oxide, and ferric sulphate, which on further heating is
ultimately decomposed into ferric oxide and sulphur trioxide:
6FeSO,=Fe2(SOj3 + 2Fe,03+3S02
Fe2(SO,),=FeA+3S03.
The ultimate result of all the reactions is :
4FeS2 + 1102 = 2Fe203 + 8S02.
It is desirable to convert all of the iron into the ferric oxide. If the
heat is properly adjusted, and the ore remains in a highly oxidizing
atmosphere for a prolonged time, the final result will be the ferric oxide.
If the heat has been too high or if there has been insufficient air in the fur-
nace, large quantities of magnetic oxide will be formed and remain un-
changed. The presence of magnetic oxide in the roasted ore in considerable
quantities indicates an inferior roast. Whether it is due to the presence
of the magnetic oxide itself, or the condition which produced it, or both
is difficult to determine. It has been said that the injurious effects of
magnetic oxide is due to its inability to resist the action of the chemicals as
well as the ferric oxide; however, in the chlorination and acid processes the
consumption of acid or chlorine is less when there is considerable magnetic
oxide in the well-roasted ore, nevertheless the tailings are invariably
high. The presence of magnetic oxide usually indicates a high tempera-
ture roast, and ore roasted at a high temperature certainly resists the
action of chemicals better than when roasted at a lower temperature.
The unsatisfactory extraction of high temperature roasts is probably
due to the formation of silicates with iron. If the ore has an unusually
38 HYDROMETALLURGY OF COPPER
dark appearace, high tailings may be expected. The extraction is always
the best when the roasted ore has the red appearance of ferric oxide.
The reaction:
2Fe304 + 0 = 3Fe203
by which the magnetic oxide is converted into the ferric oxide, is revers-
able:
3re203 + heat = 2Fe304 + 0
so that if the heat is too high (about 1700 to 1800° F.) , one atom of oxygen
of the ferric oxide is driven off, and the ferric oxide is converted into the
magnetic oxide. This is more likely to happen if the atmosphere in the
furnace is not highly oxidizing, or if the bed of ore is too deep for the air
to penetrate, or if the ore is insufficiently rabbled.
Magnetic oxide is probably formed in considerable quantities in the
earlier stages of roasting. The color of many ores, especially those
having considerable sulphur, is quite dark while the greater portion of
the sulphur is being eliminated. It is well known that sulphur dioxide
is a highly reducing agent. The heated top layer of the incandescent
ore, as it is turned over by the rabble, is ploughed under, so that the
ferric oxide particles are surrounded by the highly heated reducing atmos-
phere of sulphur dioxide, which results in reducing the ferric to the
magnetic oxide. The reaction is probably represented by the equations:
FeS, + 0,=FeS+SO,
FeS + 10Fe2O3 = 7Fe3O4 + SO
so that this action, and the corresponding reversible reaction by which
the magnetic oxide is reconverted back to the ferric oxide, is likely to
continue until the sulphur is pretty well eliminated.
The greater part of the iron sulphide becomes sulphate for only a
very brief period, especially in the early stages of the roasting. It is
probable also that sulphate is not formed to any very great extent in the
ordinary roast. Sulphur determinations taken every hour in roasting
sulphide ores show very little soluble sulphur in the early stages of the
roasting, and at any time only small quantities of ferrous sulphate.
Ferrous sulphate is decomposed at about 950° F. (510° C.)
Ferrous sulphate in roasted ore is highly injurious, and shows a very
defective roast. It is not difficult to eliminate, but it does not follow
that when the ferous sulphate has all been decomposed that the ore is
sufficiently roasted.
There are two substances which offer simple chemical tests for iron
in solution — potassium ferricyanide (FeK3CeN8) and potassium thiocya-
nate (KCNS) . The ferricyanide gives with /errows salts a blue precipitate
which imparts a blue color to the solution, but with ferric salts shows no
reaction, but only a brown color.
OXIDIZING ROASTING 39
The test may be made by filtering a sample of the ore with
water, then taking a drop of the liquid on paper or on a porcelain plate,
and adding a drop or two of the ferricyanide. A blue color indicates the
presence of ferrous iron. If the iron is likely to be in the ferric condition, as
in the solutions issuing from the ore by the acid or chlorination processes,
it should be reduced from the ferric to the ferrous condition before
applying the test. Ferric salts are easily reduced to ferrous salts by
applying such reducing agents as zinc, stanous chloride, sulphur dioxide,
or hydrogen sulphide.
The theocyanate does not give any marked coloration with ferrous
iron, but with ferric iron in the most dilute state it forms a bright red
soluble compound. The test is made as with the ferricyanide. If the
iron is likely to be in the ferrous condition, it should be tested with the
ferricyanide, or the ferrous salt converted to the ferric salt before apply-
ing the test with thiocyanate.
If no color appears, either with the ferricyanide or thiocyanate, it
indicates a thorough roast only in so far as the soluble sulphur com-
pounds of iron are concerned, but not as to the other constituents of the
ore. As a final indication of the roast, except perhaps in pure pyritic
concentrates, or iron accompanied only by silica, these tests are worthless.
Leaching, or boiling, a little of the roasted ore with water and pre-
cipitating with ammonia will usually indicate the soluble iron.
If the ore is so poorly roasted as to show undecomposed sulphides,
the roast is worthless. The best way to ascertain if there is any unde-
composed sulphides in the roasted ore is to pan it.
Copper. — The mineralogical combinations of copper are quite varied,
frequently in the same mine. It may occur as the oxide, carbonate,
silicate, or sulphide. The sulphide is by far the most common.
In a typical copper mine, the limonite gossan, usually stained more
or less with copper, appears at the surface; below the gossan, in the oxi-
dized zone, are the oxides, carbonates, and silicates; then comes the
zone of secondary sulphides consisting of chalcocite, bornite, and chalco-
pyrite; and below this the primary zone, consisting largely of pyrite
interspersed with chalcopyrite.
As the oxide, copper occurs as:
Cuprite, CujO
Tenorite, CuO.
As the carbonate, it usually occurs as:
Malachite, 2CuO, CO^, H^O
Azurite, 2CuO, 200^, H^O.
As the silicate, it usually occurs as:
Chrysocolla, CuSiOj, 2H2O.
40 HYDROMETALLURGY OF COPPER
As the sulphide, it usually occurs as:
Chalcocite, CujS
Chalcopyrite, CuFeSj
Bornite, CujFeSg
Enargite, CugAsSi
Tetrahedrite,
in which the copper occurs as CujS, associated with antimony and arsenic
sulphides, and frequently with iron, lead, zinc, and silver sulphides.
The carbonates, malachite and azurite, when roasted at a low heat
are converted into cupric oxide (black oxide), while the carbon dioxide
and water of hydration are driven off:
CuCOj, Cu (OH) 2 + heat = 2CuO + CO^ + H^O.
The silicate is also converted at a low red temperature to the oxide,
the color changing from the characteristic greenish-blue of the silicate
to black.
Of the sulphides of copper only the cuprous sulphide (CujS) is of any
metallurgical importance. Cupric sulphide (CuS) is not stable at high
temperatures, but is decomposed on heating into cuprous sulphide and
sulphur dioxide. When cuprous sulphide is roasted, the copper is first
converted into cuprous oxide and sulphur dioxide:
Cu,S + 03 = Cu20+S02
and by contact action is further oxidized into cupric oxide and sulphur
trioxide :
Cu20 + S02 + 02 = 2CuO + S03.
Some of the sulphur trioxide, combining with cupric oxide, forms cupric
sulphate:
CuO+S03 = CuSO,.
At a higher temperature, 653° C, the cupric sulphate undergoes decom-
position, sulphur trioxide being more or less expelled, so that ultimately
the sulphate will be reconverted into the oxide.
If chalcopyrite, with a quartz matrix, is roasted at a low heat, the
following reactions take place:
3CuFeS2-l-Heat-hOi8 = Cu2S+3reSO, + CuSO, + S02.
At 590° C. the ferrous sulphate decomposes, and acting upon the
cuprous sulphide remaining, converts it into the cupric sulphate:
Cu,S+FeS0, + 0e = 2CuS0,+FeA-
At 650° C. the copper sulphate is decomposed into basic sulphate
and sulphur trioxide, and at 700° C, into cupric oxide and sulphur
trioxide as follows:
2CuSO,= CuO,CuSO,-i-S03
CuO, CuSO, =2CuO + S03.
OXIDIZING ROASTING 41
The ultimate condition of the roasted product, therefore, when carried
to above 700° C. are ferric oxide, and cupric oxide. If the temperature
is not carried above 700°, sulphate of copper may remain, while if the
temperature is not carried above the decomposition point of ferric
sulphate, both copper and iron sulphates will remain in the roasted ore.
For a sulphatizing roast, the temperature should not exceed 650° C.
Cuprous sulphide fuses readily, and if contained in the ore in any
considerable quantity, must be heated carefully to avoid fusing.
As long as sulphur dioxide is being produced by the oxidation of the
sulphur, cupric oxide (CuO) cannot be formed. As soon as all the sul-
phide is converted into a mixture of cuprous oxide and sulphate, the
cuprous oxide begins to be converted into the cupric oxide; and if the
roasting is continued long enough, all the copper in the ore will be con-
verted into the cupric oxide, with the probable formation of silicates
also, if there is silica present.
With a low temperature, the copper may be contained in the roasted
ore as cupric oxide (CuO), cuprous oxide (CU2O), and cupric sulphate
(CuSOJ, and this is usually the best condition for copper extraction,
provided there are no deleterious sulphates in the roasted ore, or undecom-
posed sulphides. If the ore is poorly roasted, some of the sulphides may
remain undecomposed, and this would unfit it for a solvent process. If
undecomposed sulphides are suspected, it is best to pan some of the
roasted material, when the sulphides will be made apparent.
If the amount of copper in the ore is small, as for example in cup-
riferous gold and silver ores, all the copper sulphide will be converted
into the cupric oxide when roasting sufficiently low in sulphur to make
the ore suitable for a solvent process, provided the finishing heat has
not been excessive. Whatever the condition of the copper in the raw
ore, in the roasted ore it will appear as cupric oxide if the ore has been
properly roasted.
If the ore is to be treated principally for its copper content, it will
not ordinarily be necessary nor desirable to roast to such a complete
state of oxidation as in ore treated principally for the precious metals.
For the acid processes, sulphate of copper is riot harmful, and is usually
highly beneficial, but no sulphides should be in evidence. If the same
ore is to be treated for gold or silver by the chlorination process the
preliminary acid treatment for the extraction of the copper puts it in
the best possible condition for the extraction of the gold and silver. If
cyanidation is to follow the acid treatment, a thorough alkaline wash is
necessary.
The greatest danger in roasting copper ores is in the early stages of
the process. Cuprous sulphide, as previously stated, is readily fusible.
Cuprous oxide melts at a red heat, but the cupric oxide is quite infusible.
Cupric oxide is more readily soluble than cuprous oxide. If the tempera-
42 HYDROMETALLURGY OF COPPER
ture is excessive during the roasting, ferrites (CuOjFejOj) and silicates
are likely to form, and the copper in these combinations is soluble only
with the greatest difficulty. It will usually be more satisfactory to
slightly under-roast copper ore than to take chances in getting the best
possible roast by overheating. Copper ores are particularly sensitive
to high temperatures, that is to say, temperatures above a dull red (650°
C. or 1202° F.) and if sintering or fusion occurs it is practically impos-
sible to get a satisfactory extraction. Cupric oxide is reduced to cu-
prous oxide at 1050° C.
The best way to determine the best conditions of roasting, prin-
cipally as to the temperature, is by direct experimenting, and leaching
the roasted ore with dilute hydrochloric or sulphuric acid. If the
roasting has been properly done, there should be no difficulty in extract-
ing at least 90 per cent, of the copper. If, however, the temperature
has been excessive, a very poor extraction of the copper may be expected.
In making preliminary tests, it is well to roast at least one lot of ore at a
temperature no higher than scarcely a visible red, and then increase the
temperature on successive charges. The precentage of copper extracted
in a certain reasonable time from the different roasted samples will,
by comparison, give the highest temperature the ore will stand without
detriment, and that is the temperature at which the ore should be
roasted.
Whether or not there is any copper sulphate in the roasted ore can
easily be ascertained by placing a small portion in a funnel, leaching it
with hot water, and then adding ammonia to the filtrate, when, if there
is any soluble copper, the familiar blue will appear. If no blue appears,
it may be assumed that the copper in the roasted ore is all in the condition
of oxide.
In gold and silver ores, where the copper does not occur in sufficient
quantities to attempt to recover it at a profit, the principal injurious
effect of the copper is in the consumption of chemicals, and in its pre-
cipitation with the precious metals. It is also undesirable in the gold
and silver bullion, unless there is sufficient copper to make electrolytic
refining possible. The extent to which copper in the ores of the precious
metals is fatal will depend largely on the price of the chemicals, on the
consumption of chemicals, and, in a measure, on the roast. The best
ultimate condition of the copper in roasted ore of the precious metals is
in the form of oxide, and, fortunately, this is the way it usually occurs.
If the ore is roasted with salt, much of the copper will be converted
into cupric chloride (CuClj), but since at a red heat cupric chloride gives
up half its chlorine, some of the copper will be in the form of cuprous
chloride (CuCl) . If the temperature is very high most of .the cupric
chloride will be converted into the cuprous chloride and oxychloride
of copper, both of which are wholly or partly insoluble in water. The
OXIDIZING ROASTING 43
chloridizing roasting of copper ores is taken up more fully under "Chlo-
ridizing Roasting." — "Longmaid-Henderson Process."
The ideal roast for copper ores is one in which all the copper is in the
form of sulphate and oxide, and all the iron in the ferric condition. This
represents a roast in which practically all the copper is soluble in water or
dilute acids, and all the iron insoluble. Since ferrous sulphate is decom-
posed at 590° C, and curpic sulphate at 650° C, the best roast for the cop-
per, theoretically at least, should be obtained by maintaining the tempera-
ture between 590 and 650° C. This also represents the best conditions
in practice.
The temperature at which ferrites and silicates of copper are formed
has not been definitely determined, but if the ore is heated much above
700° C, there is great danger, and the ferrites and silicates once formed,
the satisfactory extraction of the copper presents a problem of some
magnitude. The presence of ferrites and silicates is usually indicated by
the dark appearance of the ore, instead of the red color of ferric oxide.
The sulphur elimination in roasting copper ores depends largely
on the amount of copper in the ore, as well as on the amount of sulphur.
If much of the copper in the roasted ore is soluble, it is soluble as the
sulphate, and the soluble sulphur will be quite large. In this respect
roasting of copper ores differs somewhat from roasting gold and silver
ores, where the sulphur content of the roasted ore is necessarily quite
low. It is not desirable to eliminate the same amount of sulphur from
copper ores as from gold and silver ores, so that the roasting of copper
ores suitable for hydrometallurgical extraction will usually be somewhat
cheaper than if the same ore were roasted for the extraction of the
precious metals. On the other hand, copper ores usually require a
more delicate roast than gold and silver ores.
For high grade cupriferous concentrates or sulphide ore, the roasted
product may contain from 3 to 5 per cent, sulphur, and be well roasted
for chemical treatment.
It is possible to make most of the copper soluble by prolonged roasting
at a low temperature and with insufficient air, but such a roast will also
leave much of the iron and other constituents soluble. Whether this
is desirable or not depends almost entirely on the method of precipitating
the copper, and on the chemical composition of the soluble iron.
In cases where gold and silver ores contain copper, the roasting
must be effected to get the best extraction of those metals, but if a little
care is used in the temperature of the roasting, a high extraction of the
copper may also be made, even if the ore is roasted with a view of getting
the best extraction of the gold and silver.
It is difficult to understand why some sulphide ores with careful roast-
ing will give up their values while with others of apparently the same
or similar composition it is difficult to get a reasonable extraction.
44 HYDROMETALLURGY OF COPPER
In order to determine the effect of roasting on chalcopyrite, per se,
several pounds of the pure mineral, containing as a matrix pure quartz
and some galena, was carefully roasted in an assay muffle. The heat
was never above a dull red. The copper content of the raw ore was 29.3
per cent., and the sulphur 28.4 per cent. The sulphur in the roasted ore
was 5.5 per cent., of which 3.5 per cent., was soluble. The high insoluble
was probably largely due to the galena, forming lead sulphate.
The extraction, by agitating with a 5 per cent, sulution of sulphuric
acid, was as follows:
Raw ore, 29 . 3 per cent. Cu.
Roasted ore, 1 hour's treatment, 3 . 2 per cent. Cu. Extraction, 89 . 1 per cent
2 hours' treatment, 2.2 per cent. Cu. Extraction, 92.5 per cent
4 hours' treatment, 1 . 2 per cent. Cu. Extraction, 95 . 9 per cent
On the other hand, in just as carefully roasting a sulphide concen-
trate from Mexico, containing 6 per cent, copper, it was difficult to get
an extraction of 80 per cent. The mineralogical combination of the
copper in these concentrates was not determined, but presumably it was
in the form of chalcopyrite or bornite.
Silver. — Silver is universally associated with copper and gold. It
may be said that gold in ores is never found unaccompanied by silver,
while copper is a common associate of both gold and silver.
Silver is an important factor in the treatment of copper and gold
ores. The silver is not readily recovered, and if contained in the tailings
in any considerable quantity, may act as an obstacle to the close extrac-
tion of the gold, the fact that silver is not readily soluble in any of the
ordinary solvents of copper and gold adds somewhat to the difficulty of
its extraction by the wet processes. Roasting, in the metallurgy of
silver, is a very important factor.
If there is any free gold in the ore, the silver will be more or less
alloyed with the gold. In thoroughly oxidized ores it probably occurs
in this way. In unoxidized ores it will almost always be found as the
sulphide, associated more or less with arsenic, antimony, and copper, as
will be noticed from the common minerals of silver:
Argentite (silver glance), AgjS.
Pyrargyrite (ruby silver), 3Ag2S,Sb2S3
Proustite, 2Ag2S,As2S3
Stephanite, 5Ag2S,Sb2S3
Stromeyerite, Ag2S,Cu2S
Polybasite, _ 9(Ag2S,Cu2S),Sb2S3,As2S3
Cerargyrite (horn silver), AgCl
Hessite, Ag2Te
Petzite, (Ag,Au)2Te.
OXIDIZING ROASTING 45
While silver is more or less associated with tellurium, it is a strange
fact that only a comparatively small quantity of silver is found associated
with gold in telluride or sulpho-telluride ores, and is not of any great
consequence either in the oxidizing roasting or subsequent chemical
treatment. When it does occur combined with gold and tellurium, the
tellurium is volatilized in roasting, leaving behind an alloy of gold and
silver.
In the oxidizing roasting of sulphide copper and gold ores containing
silver, the silver sijlphide is first converted into the sulphate at an early
stage of the operation. The silver sulphide reacting with the sulphur
trioxide, formed principally from iron and copper sulphides by catalytic
action with hot silica and metal oxides, forms silver sulphate, and reduces
the sulphur trioxide to the dioxide:
2Ag2S + 4SO3 = Ag,SO, + 4S0,.
If there is any free gold in the sulphide ore, as there frequently is in
small quantities, some silver will also be free but alloyed with the gold.
This silver in the first stages of roasting is likely to be converted into the
sulphate :
2Ag + 2S03 = Ag,SO, + S02.
As the ore is rabbled against a higher temperature and a more highly
oxidizing atmosphere, the silver sulphate is gradually converted into
metallic silver. The silver sulphate is partly reduced by the direct
action of heat alone:
Ag2S0,+ Heat =2Ag + S03 + 0
but the temperature required for this reaction — 1095° C. — is rarely if
ever attained in a roasting furnace. In the presence of reducing gases,
silver sulphate is decomposed at a very moderate heat, metallic silver
being deposited. In the presence of copper oxides, silica, and iron oxides,
silver sulphate is decomposed at temperatures from 860 to 870° C. :
Ag2SO, + 4Fe304 = 2Ag + 6Fe203+S02
Ag2S04 + Cu20 =2Ag + CuS04 + CuO.
In the roasting of copper, gold, and silver ores suitable for treatment
by chemical processes, in a highly oxidizing atmosphere, the silver in the
ore will always be found in the metallic condition, alloyed with the gold.
It is for this reason that silver plays so important a part in the hydro-
metallurgical treatment, if alloyed in appreciable quantities with gold.
Unless the extraction of the silver is quite thorough there may enough
remain in the tailings to protect, ia a measure, the gold from the action
of the solvent.
If the sulphide ore is improperly roasted, some of the silver may
46 HYDROMETALLURGY OF COPPER
remain as sulphate. As such it is readily soluble in water. One hundred
parts of water dissolves 0.58 parts of silver sulphate.
It has been conclusively proved that silver is volatilized in oxidizing
roasting. It is possible that in many instances the volatilization of
silver in oxidizing roasting is due to the presence of small quantities of
cerargyrite, or natural silver chloride. The chloride of silver seems to be
quite generally distributed in the various silver ores. The chloride
volatilizes at a strong red heat, so that if an excessive loss of silver is
discovered in oxidizing roasting it is well to examine the ore to ascertain
the presence of chlorine.
Since metallic silver is with difficulty soluble by any of the commercial
processes for the recovery of copper, gold and silver, chloridizing roasting
is frequently resorted to, in order to convert the silver into the more
soluble silver chloride. This subject is taken up in detail under the head
of "Chloridizing Roasting."
Any chemical process having for its primary object the recovery of
copper or gold from its ores must take cognizance of the silver usually
associated with them. If the quantity of silver is small there is no
difficulty in recovering a fair percentage by either the cyanide or chlorina-
tion processes. If the quantity is large, the best average results will be
obtained by chloridizing roasting, when the silver may be extracted quite
closely by either the hyposulphite, cyanide, or chlorination processes.
Gold. — Gold, of itself, is of no metallurgical importance in the process
of oxidizing roasting. It always occurs native or mixed or combined
with sulphur or tellurium; but whether mixed or combined, on roasting it
emerges as metallic gold, which at all stages of the roasting is unaf-
fected by any temperature or condition of the furnace. If the gold is
free and microscopically fine, coarse, flaky, solid, or porous, it will, of
itself, remain so. It will appear in the finally roasted ore as it appeared
in the raw ore, or after being liberated in the early stages of the
operation.
Much has been said about the loss of gold in roasting. Careful inves-
tigation has shown that in oxidizing roasting thei'e is none but a mechan-
ical loss, which is subject to the same conditions as the handling of dry
ore under any circumstances. If the ore contains silver chloride, which
usually has associated with it some gold, a loss of gold may be expected
with that of the silver, in oxidizing roasting, but this loss will ordinarily
be very small if the temperature is properly regulated, and in no case
will it ever be serious.
Lead. — Lead usually occurs as the
Sulphide (PbS), Galena,
Carbonate (PbCOg), Cerussite,
Sulphate (PbSOJ, Anglesite.
OXIDIZING ROASTING 47
Lead is more or less associated with copper and gold ores, but not
usually in large quantities. Silver is more commonly associated with
it. The presence of lead in ores in small amounts is not particularly
harmful, either in roasting or in the subsequent treatment by the hydro-
metallurgical processes. When lead occurs in ores in large quantities, it
is so desirable as a smelting material that its treatment by solvent process
is quite remote.
According to Plattner, if the sulphide, galena, is roasted at a low
temperature, to prevent fusion, it will at first be converted into the
oxide and sulphur dioxide:
PbS + 03=PbO+S02.
A part of the sulphur dioxide on coming in contact with the heated
silica combines with the oxygen of the air to form sulphur trioxide, and
this combines with the lead oxide to form lead sulphate :
PbO + S03=PbSO,.
The lead in the roasted ore will usually be in the form of oxide and
sulphate. Silicate will not occur unless the ore has been fused, in which
case, in any event, the ore would probably be unfit for subsequent
chemical treatment. The proportion of the oxide to the sulphate will
depend upon the presence of other sulphides, the method of roasting,
and to a large extent on the proportion of the galena to the other con-
stituents in the ore.
It is probable that much of the sulphide may be converted directly
into the oxide and sulphate by the slow roasting and at the low temper-
ature usually employed in roasting copper, silver, and gold ores for
subsequent chemical treatment:
2PbS + O7 = PbSO 4 + PbO -h SO 2.
"When the ore contains considerable lime, as sometimes happens,
some metallic lead may possibly be formed:
4PbS+4CaO=3CaS + CaS04-t-4Pb.
CaS-I-0, = CaSO,.
The transformation of the calcium oxide to the calcium sulphate
is a desirable change for the subsequent chemical treatment. Quartz,
clay, and silicates remain inert to lead sulphide.
Galena is difficult to roast. It fuses at a low temperature, and if
excessively heated, is likely to agglomerate into a mixture of lead oxide
and lead sulphate from which it is difficult to expel the sulphur trioxide,
even when heated so high as to melt, and from which it is absolutely
impossible to satisfactorily extract the copper and precious metals.
Lead sulphate, with heat alone, is decomposed only at a white heat.
48 HYDROMETALLURGY OF COPPER
Lead carbonate (PbCOa) is readily decomposed at a low temperature
(200° C; 392° F.) into lead oxide and carbon dioxide:
PbC03=PbO+C02.
By prolonged roasting, lead carbonate, or the monoxide, at a temper-
ature of not exceeding 450° C. (842° F.) may be converted into the higher
oxide, minium (PbaOJ. At a still higher temperature the red lead
or minium again gives up its oxygen and is reconverted into the monoxide
or litharge.
The lead in roasted ore, on cooling, is likely to be in the form of mon-
oxide (PbO); rarely perhaps as the red oxide (PbgOJ; some sulphate
(PbSOJ, and, if the ore is fused, as silicate.
Ore containing as high as 10 per cent, lead, can with care, be satis-
factorily roasted for subsequent chemical treatment.
Zinc. — Zinc frequently occurs associated with copper, gold, and
silver ores as:
Sulphide (ZnS), Sphalerite.
Oxide (ZnO), Zincite.
Carbonate (ZnCOj), Smithsonite.
Silicate (Zn^SiO^), Willemite.
In the oxidized ores the zinc usually occurs as the oxide or carbonate;
in the sulphide ores it is always found as sphalerite.
If the zinc is in the form of carbonate, roasting readily drives off the
carbon dioxide, leaving the oxide of zinc;
ZnC03 = ZnO+C02.
If the zinc is in the form of sulphide, oxidation at the temperature at
which copper, gold, and silver ores are usually roasted, takes place
slowly, and yields a mixture of oxide and sulphate. The amount of
sulphate, however, is small as compared with the oxide.
Zinc sulphide begins to oxidize at a dull red heat. As the temperature
is increased the oxidation takes place more rapidly, with the formation
of zinc oxide and sulphur trioxide:
ZnS + 03 = Zn0 + S02.
2ZnS + 07 = ZnO + ZnS04 + SOj.
At a prolonged high temperature the sulphate is converted into the
oxide:
ZnS04 + heat = ZnO + S03.
Some of the sulphur dioxide released in roasting by catalytic action
with the glowing ore is converted into the trioxide, which may then com-
bine with some of the zinc oxide to form the neutral sulphate, ZnS04.
OXIDIZING ROASTING
49
In the decomposition of the neutral sulphate, basic sulphates may be
formed which require a high and prolonged temperature to ultimately
resolve them into the oxide. Zinc sulphate, by heat alone, is decomposed
at 739° C.
Zinc sulphide does not oxidize as readily as iron or copper sulphides,
and a higher temperature is required to start oxidation. It is infusible
at any temperature attained in the roasting furnace, in roasting copper,
gold and silver ores. Its presence, so far as roasting is concerned, is not
harmful.
The relative proportion of zinc oxide and zinc sulphate formed in
roasting will depend upon the temperature, the oxidizing qualities of the
atmosphere, and, a certain extent, the relative quantity of zinc in the ore.
The higher the temperature in a highly oxidizing atmosphere, the more
zinc oxide, and the less zinc sulphate, will be formed.
Zinc, though quite generally distributed, is not frequently found in
copper, gold, and silver ores in sufficient quantity to interfere with the
metallurgical treatment by wet methods. If contained only in small
amounts, and the ore is thoroughly roasted, practically all the zinc will
be in the form of oxide. The oxide is readily soluble in acids, but is
not so readily acted upon by chlorine, cyanide or sodium hyposulphite,
although it affects these solvents injuriously.
Zinc oxide is not volatile at the highest temperature used in roasting
ores— from 900 to 970° C. (1652 to 1778° F.). Zinc in its metallic con-
dition is quite volatile, even at a moderately low temperature. Reducing
gases, such as carbon monoxide from the fuel, have a tendency to re-
duce the oxide to the metallic zinc and thus volatilize it:
ZnO+CO = Zn-t-CO,.
RESULTS OF HEATING ZINC SULPHATE WITH AND WITHOUT
FREE ACCESS OF AIR
(H. O. Hofman, Trans. A. I. M. E., 1905)
Heated ia
Air
Air
Carbon dioxide.
Carbond dioxide
Temp,
deg. C.
578
58S
599
602
Total S.
eliminated
1 . 63 per cent
1 . 79 per cent.
0 . 50 per cent.
0 . 50 per cent.
S. eliminated
as SO,
0.14 per cent.
0.14 per cent.
0.18 per cent.
0.16 per cent.
S. eliminated
as so.
1 . 49 per cent.
1 . 65 per cent.
0.32 per cent.
0 . 34 per cent.
Ratio .
S as SO2
S as SO 3
9.06
8.26
55.2.3
47.88
In these experiments, the temperature of the furnace was brought
slowly to the point at which the first acid was given off, then raised about
10° C. and maintained constant for several hours. While only a very
small proportion of the total sulphur was driven off in the tests, they
4
50 HYDROMETALLURGY OF COPPER
show that in heavy zinc sulphate the tendency of the salt to split into
ZnO, SO2, and O is greater when oxygen is absent than when it has free
access of air. .
Arsenic. — Arsenic usually occurs associated with copper, gold, and
silver ores as the sulphide:
Arsenopyrite (FeAsS), Mispickel.
' Realgar (AsjSj)
Orpiment (AsjSg)
Arsenic is almost universally associated with sulphide copper ores.
If occurring in small quantities it is of no special importance in the
roasting except that its elimination should be as complete as possible
under the conditions of the roast. Its presence in the precipitated copper
is very harmful, and hence effort should be made to keep the solutions as
free from it as possible.
The sulphides of arsenic fuse readily so that care must be exercised
in the first stages of the roasting. Arsenic volatilizes at a comparatively
low temperature, in the condition of arsenous oxide. In the presence of
an excess of oxygen there is a tendency to form arsenates of iron and other
metals, and these arsenates are decomposed only at an exceedingly high
temperature. After the arsenic is driven off in the first stages of the
roasting, no harm can result in elevating the temperature to that re-
quired for the other constituents of the ore. Arsenates are undesirable
in the roasted ore, as they interfere with the close extraction of the copper
and precious metals. Arsenic in the ore, as a rule, is not particularly
detrimental to any of the solvent processes, if the roasting is properly
done.
Mr. R. R. Rothwell, in speaking of roasting arsenical pyrites at
Deloros, Canada' says, "It was asserted by some metallurgists that the
roasting of arsenical pyrites presents many difficulties. I can affirm, on
the contrary, that they roast with much greater facility and in about
two-thirds of the time necessary to roast simple sulphides. They stand
almost any amount of heat without fusing, and the arsenic, wljich forms
about 49 per cent, of the mispickel, volatilizes at comparatively low
temperature, seems to leave the mass porous, thus facilitating the oxida-
tion of the sulphur." An extraction of 95 per cent, of the gold was made
on these arsenical pyrites by the chlorination process.
At Murcur, Utah, where the ore has been treated successfully for
many years by the cyanide process, roasting has been found to eliminate
any injurious effects from the arsenic, which occurs in considerable
quantities in the raw ore.
Arsenic, when heated in air, easily oxidizes into white arsenous
oxide, AS2O5, and is easily volatilized. When arsenical ores are roasted,
' Trans. A. I. M. E., 82-83.
OXIDIZING ROASTING 51
the sulphur and arsenic are converted into arsenous oxide and sulphur
dioxide. The former is a solid at ordinary temperatures, and the latter
gaseous, and therefore the arsenous oxide is deposited as a sublimate in
the cooler portions of the flues and dust chambers, through which the
fumes escape from the furnace.
In roasting cupriferous pyritic ores there will usually be no difficulty
in eliminating from 75 to 80 per cent, of the arsenic.
Roasting Argentiferous Cobalt -nickel Arsenides \ — The ore used in this
investigation was chiefly smalltite, containing 689 oz. silver per ton, and
56 per cent, arsenic. The object of the investigation was to ascertain
(1) the temperature at which the arsenic is most rapidly expelled; (2)
the thoroughness with which it is expelled by prolonged roasting at this
temperature, and (3) the effect of adding charcoal near the end or at the
beginning of the roast.
It was found that 15 per cent, of arsenic per 100 of ore, that is, 27 per
cent, of the total arsenic, is expelled below 700° C, but that the rest of the
arsenic is not expelled until the temperature reaches about 840° C. when
rapid expulsion sets in. By rabbling at temperatures above 840° C, the
percentage of arsenic can be further reduced by about 34 per cent., that
is down to 17 per cent, in the ore, from the original 56 per cent; in this
range of temperature the arsenic is removed much faster than at lower
temperatures. Raising the temperature quite suddenly to 800° C. does no
harm as the ore remains porous. The addition of charcoal either at the
beginning or toward the end of the roast failed to increase the expulsion
of arsenic. Finer grinding of the ore, after it had been roasted once,
and re-roasting at about 880° C. showed no further expulsion of arsenic,
due to fine grinding.
Antimony. — Antimony, like arsenic, usually occurs as the sulphide.
It is almost universally in the form of Stibnite (SbjSa) and is frequently
associated with silver and gold, and quite commonly with copper. In the
small quantities in which it usually occurs in the ores of these metals, it
does not present any special difficulties in roasting. Care must be used
in the early stages of the operation. If an attempt were made to remove
the antimony by rapid oxidation, there would be danger of converting
it into the insoluble antimonates of the metals in the ore. This would be
undesirable for some of the chemical processes, while for others it might
be somewhat serious. In the early stages of the roasting it is therefore
necessary to employ a very low heat. The presence of steam, largely
supplied by the burning fuel and the water of hydration in the ore, is
found to be useful as a source of hydrogen, which removes sulphur as
hydrogen sulphide:
SbjSg -I-3H2 -l-heat = 3H2S -l-Sb.
' Bi-Monthhj Bulletin of the A. I. M. E., .Jan. 1907.
52 HYDROMETALLURGY OF COPPER
The antimony then combines with oxygen and escapes as a volatile
oxide.
When the temperature of the roasted ore is brought to about 350° C.
(662° F.) the atmospheric oxygen converts the antimony trisulphide into
antimonous oxide and sulphur dioxide. Antimonic acid is formed in the
presence of the oxides of the other metals, and combines with them to
form antimonates. Sulphates of antimony are not formed during the
roasting. If the ore contains large quantities of foreign sulphides,
which on being roasted would form sulphates, antimonates of the foreign
metals are formed instead of the sulphates.
Carbon, such as coal or charcoal, finely ground or mixed with the ore,
has been used to break up the antimonates and arsenates, and expel the
arsenic and antimony, but it has not been found of utility enough to find
a permanent place in practice.
All ores of copper, gold, and silver, containing appreciable quantities
of arsenic and antimony, are difficult, if not impracticable, to treat raw,
by any of the solvent processes. By roasting, and consequent volatili-
zation of these elements, .they are largely eliminated from further consid-
eration, except, perhaps in the case of silver ores to be treated by sodium
hyposulphite, when the ores may contain -considerable quantities of arsen-
ates and antimonates, after roasting.
Antimonous oxide (SbjOg or Sb^OJ resulting from the roasting of
antimony sulphide (SbjSg) is insoluble in water, but is soluble in hydro-
chloric acid and in alkalies.
Bismuth. — Bismuth is one of the most injurious alloys of copper. It
may be present in copper ores in the metallic state, or in sulphides,
arsenides, and antimonides. The metal and sulphides are volatile at the
roasting temperature but much less readily than in the cases of arsenic
and antimony. The minerals containing bismuth are readily oxidized
to fixed compounds.
In the incomplete roasting of copper ores, arsenic, antimony, and
bismuth may remain in the roasted product in the same combinations
in which they occurred in the ores, accompanying the fixed compounds
that are formed during the roasting operation.
The degree of elimination of these impurities in roasting varies
necessarily with the minerals in which they occur, as well as the copper
ore, and the conditions under which the roasting is carried on. The
following analytical data by Allan Gibb' shows the elimination from fairly
typical ores, when roasted in heaps and in reverberatory furnaces for
smelting, which does not represent as complete a roast as that required
for the wet methods.
' Trans. A. I. M. E., Vol. XXXIII.
OXIDIZING ROASTING
53
ROASTING COPPER ORES; ELIMINATION OF ARSENIC, ANTIMONY AND
BISMUTH
Raw
ore
Roasted ore
Per cent,
actual
Per cent,
relative
Cu = 100
Per cent,
actual 1
Elimination
per 100
of Cu
Per cent,
relative
Cu = 100
Total per-
centage of
elimination
No
. 1
Copper
5.55
1.18
100.00
21.36
7.68
0.407
100.00
5.29
Arsenic
15.97
75.0
Antimony
0.035
0.63
0.035
0.47
0.160
25.4
Bismuth
0.011
0.198
No
0.011
2
0.143
0.055
27.8
Copper
12.15
0.967
100.00
7.96
14.68 '
0.454
100.00
3.09
Arsenic
4.87
61.2
Antimony
0.46
0.378
0.045 '
0.307
0.071
18.8
Bismuth
0.014
0.115
0.015 1
0.013 1
0.013
11.3
No. 1 was a cupriferous iron pyrites which was roasted in heaps and
subsequently smelted in blast furnaces.
No. 2 was a dressed ore containing the copper mostly in the form of
copper pyrites, with a small proportion of bornite and copper glance.
It was roasted in a reverberatory furnace.
Nickel. — Nickel is quite frequently associated with copper ores, and
when it so occurs in paying quantities its recovery is advisable.
Nickel usually occurs as the
Sulphide, Millerite, NiS,
Arsenide, Niccolite, NiAs,
Silicate, Garnierite, H (NiMg)Si04, H^O.
When the sulphide is roasted, the nickel is oxidized and the sulphur
passes off, mostly as the dioxide, and some as the trioxide. The tri-
oxide produces sulphuric acid and forms some nickel sulphate. When
the sulphate is strongly heated the nickel is converted into nickelous
oxide and sulphur trioxide is driven off. By prolonged roasting, at the
proper temperature, nickelous oxide, NiO, alone may be obtained. If
imperfectly roasted, there will be a mixture of oxide, ■ sulphate, and
unaltered sulphide.
If a mixture of nickel and iron sulphides is carefully roasted, a mix-
ture of nickelous oxide and ferric oxide is obtained. As sulphate of
nickel is a very stable compound, the roasting may be so conducted,
that the greater part of the nickel is obtained as sulphate, while the iron
will be in the condition of ferric oxide.
54 HYDROMETALLURGY OF COPPER
By roasting nickel and copper sulphides in the same way, it is possible
to get nickelous oxide and cupric oxide, or a mixture of oxides and sul-
phates. As nickel sulphate is stable at a higher temperature than
copper sulphate, the nickel may be roasted to the sulphate and the copper
to the oxide.
If nickel, iron, and copper sulphides are all roasted together, the
nickel may be in the condition of sulphate and the other two metals as
ferric and cupric oxides.
If nickel arsenide is roasted, the arsenic forms arsenous oxide, and
the nickel sesquioxide. Part of the arsenous oxide escapes unaltered,
part is further oxidized to arsenic oxide, and this combines with the
nickelous oxide to form an arsenate. Nickel arsenate is not decomposed
when heated alone, so the result is basic nickel arsenate.
Copper ores containing nickel usually contain also magnetic and iron
pyrites, and often arsenic and antimony compounds as well as silicates,
quartz, and earthy matter. In the roasting, the arsenic and antimony are
mostly driven off, the sulphur partly escapes as dioxide, and is partly
converted into trioxide by contact with the red-hot masses of ore and
furnace walls. Iron, copper, and nickel oxides combine with this trioxide
to form sulphates. As the roasting proceeds, and the temperature is
raised, the sulphates are again decomposed into oxides and sulphur
trioxide, or sulphur dioxide and oxygen. Iron sulphate is first decom-
posed, next the copper, and lastly the nickel compound. If the roasting
were continued at the proper temperature, the product would be a
mixture of ferric oxide, cupric and cuprous oxides, and nickelous oxide.
Nickel sulphate is readily and abundantly soluble in water. The
oxide is soluble in mineral acids, especially dilute hydrochloric acid,
when warmed. The chloride is soluble in water, but not as readily as
the sulphate. The mineral garnierite is soluble in sulphuric and hydro-
chloric acids, but with some difficulty.
Calcium (Lime). — The compounds of calcium, on account of their
prevalence and positive action on almost all of the chemical solvents used
in the hydrometallurgical processes, are among the most important to be
considered. In the alkali process, calcium compounds are not particu-
larly harmful, and frequently, as in the case where calcium is combined with
oxygen to form lime, it imparts a desired alkalinity before applying the
solvent. In the acid processes, like chlorination or the treatment of copper
ores with dilute acids, the amount of calcium in the ore and the way it is
combined will usually be the most important factor in determining the
applicability of the process and, to a large extent, indicate its success or
failure.
There would be no difficulty in treating most of the copper, gold and
silver ores successfully, by the acid processes, if it were not for a few
interfering elements, and of all the interfering elements, the presence of
OXIDIZING ROASTING 55
lime in large quantities presents the most common and the most difficult
problem. Fortunately, most of the ores of copper and the precious
metals do not contain enough lime to seriously interfere with the treat-
ment. The vast majority of all metalliferous deposits have quartz as
the matrix, and usually the lime is not found in quantities sufficient to
make an acid treatment prohibitive, if the ore is otherwise suited to the
process, particularly if the ore is amenable to preliminary concentration.
Calcium usually occurs associated with copper, gold and silver ores
in the form of
The Carbonate (CaCOg), Calcite (Limestone).
The Fluoride (CaF^), Fluorite.
The Sulphate (CaSOJ, Gypsum.
The carbonate is not readily attacked by chlorine, but is immediately
decomposed by acids. Roasting converts the carbonate into the oxide
and carbon dioxide:
CaCOj + Heat = CaO + CO^.
When cold, the oxide (lime) does not absorb chlorine, but at a red heat,
in the presence of chlorine, it forms calcium chloride with the evolution
of oxygen:
CaO + 2Cl = CaCl2 + 0.
If the ore contains considerable sulphur, the sulphur trioxide released
during the roasting combines to a greater or less extent with the lime to
form sulphate:
CaO + S03 = CaSO,
which is practically unaffected by all the acids, only very slightly acted
upon by chlorine, and remains neutral to cyanide or sodium hyposulphite.
It is almost insoluble in water; one part of calcium sulphate requires 432
parts of water for its solution. Its solubility is increased by the presence
of alkaline chlorides and free hydrochloric acid.
It is desirable, therefore, that ores containing considerable lime should
be mixed with ores containing considerable sulphur before roasting.
Some of the calcium, however, will unavoidably remain as oxide after
roasting, which, when coming in contact with water in the subsequent
chemical treatment, is converted into the hydroxide (slacked lime) in
which form it is desirable in the alkali processes, but is readily attacked
by chlorine and the acids.
The lime, when coming in contact with sulphuric acid, as in the sul-
phuric acid copper processes, and the barrel chlorination process where
chlorine is generated from bleach and acid, is converted into calcium
sulphate:
Ca(OH)2 + H2S04 = CaSO, + 2H20,
56 HYDROMETALLURGY OF COPPER
which accounts for much of the excess of acid sometimes used in treating
ores containing considerable lime.
. In order to economize acid, it is desirable to convert as much as
possible of the lime into the sulphate, by judicious roasting, if the ore is
a sulphide.
If the chlorinating is done by the "Plattner" or by the "Percolation"
processes, in which acids are not ordinarily used, then the lime, instead
of combining with the excess of acid, will combine with the chlorine:
2Ca(OH)2 + 4Cl = CaCl2 + Ca(C10)2 + 2H20,
forming the chloride and hypochlorite, as in the manufacture of bleaching
powder. Since chlorine acts more readily on lime than on copper, gold,
and silver, in the ore, sufficient chlorine must be provided to chlorinate '
the lime and have an excess after all other base elements have been
satisfied. To avoid the large consumption of chlorine when it is applied
directly as gas in ores containing much lime, the ore is frequently roasted
with salt; in this way the lime is converted into chloride in the furnace and
is no longer harmful.
Calcium, in the form of fluorite, is a common associate of copper, gold
and silver ores. It occurs abundantly in Cripple Creek, intimately associ-
ated with calcite. Fluorite is peculiarly a constituent of metalliferous
veins. In minute quantities it is widely diffused.
Fluorite is unaffected by chlorine, cyanide, or dilute acids. Hot
concentrated sulphuric acid decomposes it. By roasting, the fluorite is
converted into the oxide, as in the case of carbonate:
CaF2 + H20 = CaO+2HF.
" The fluorine probably combines with the moisture of the air, and
water combined in the ore as hydrate, to form hydrofluoric acid.
Mixed with silica and sulphur, as the fluorite usually is in metallifer-
ous ores, the sulphuric acid formed in the roasting converts some of
the calcium into the sulphate:
CaF2-FH2SO, = CaSO, + 2HF.
Car2-t-2H2SO,-fSi02 = CaS04-h2H20-FSiF2.
The principal point of interest, so far as roasting for the subsequent
chemical treatment is concerned, is, like the carbonate, the fluoride is
converted into lime, and that in the presence of sulphur it is converted
into the sulphate.
A specimen sample of Cripple Creek ore, composed largely of fluorite,
after roasting had a white appearance, and analysis showed 39.25 per
cent. CaO.
Calcium sulphate is largely associated with copper, gold, and silver
ores, in the form of gypsum or anhydrite. It also occurs largely as the
OXIDIZING ROASTING 57
result of the decomposition of pyritic ores acting on the calcium carbonate.
Pyritic ore is oxidized by the action of water and air, forming ferrous
sulphate and sulphuric acid.
FeS2 + H20 + 0,=FeSO, + H2SO,.
The sulphuric acid then acts on the carbonate, forming calcium sulphate
and water:
CaC03 + H2SO, = CaSO, + H20.
Calcium, therefore, in oxidized ores is largely in the form of sulphate,
and is not particularly injurious in any of the chemical processes, or in
the roasting operation. The sulphate, once formed, can only be converted
into the oxide by the most intense heat procurable. Such a heat is never
realized in a roasting furnace.
Gypsum gives off its water of hydration at 200 to 250° C. (392 to
482° F.) . The dehydrated gypsum melts at a red heat mthout decomposi-
tion. On coming in contact with water, the dehydrated calcium sulphate
again takes up its water of hydration, just as in the case of ordinary
plaster of Paris. In doing this, if the ore contains considerable sulphate,
it sometimes happens that the ore during leaching sets so hard that picks
have to be used to remove it from the vats.
An analysis made on unoxidized ore from Cripple Creek, showed:
Calcium Sulphate (gypsum, CaS04-|-2H20), 0.83 per cent.
Calcium Fluoride (fluorite, CaFj), 0.78 per cent.
The amount of lime in an ore which may be fatal to chlorination or
to an acid treatment depends largely on other conditions. Ordinarily
from 5 to 6 per cent, is the limit. In Cripple Creek ores the lime varies
from 1.5 to 2.5 per cent., although in some mines it is much higher. The
Potsdam ores of the Black Hills, which have been successfully chlorin-
ated, contain as much as 8 per cent. CaO.
That only a small portion of the lime in roasted ore combines with
chlorine or the acids is evident from the treatment of 800 to 1000 tons
daily of Cripple Creek ores by the barrel chlorination process, where it
may be assumed that the ore averages 2per cent, lime, or 40 lb. per ton.
The average chemical charge may be assumed to be 15 lb. of bleach and
30 lb. of sulphuric .acid. Theoretically, it takes 6 parts of acid to com-
bine with 7 parts of bleach, but in practice, owing to the impurities of
the bleach and acid, equal parts of each are required. Of the 30 lb. of acid,
therefore, per ton of ore charged into the barrel, 15 lb. are consumed in
reacting with the bleach to generate chlorine. The solutions issuing from
the barrels after treatment are always strongly acid, so that much
acid remains unconsumed, and some is also consumed in reacting with
other base elements. It is safe to say, therefore, that only from 5 to 10 lb.
58 HYDROMETALLURGY OF COPPER
of the acid actually combines with the calcium or lime in the ore. But
if the calcium in the ore were all present as lime, that is 40 lb. CaO, it
would take at least 70 lb. of acid to neutralize this lime, instead of only
5 or 10 lb. actually required in practice. Some of the Cripple Creek ores
are chlorinated with only 10 lb. of bleach and 15 lb. of acid, which makes
the acid consumed considerably less. From this it will be seen that the
injurious effect of the lime in ore in an acid process depends largely on its
chemical combination, and that much of the lime in sulphide ores may
be converted into a comparatively harmless condition by roasting.
Magnesium. — Copper, gold and silver ores frequently contain small
quantities of magnesium, but usually not in sufficient quantity to seri-
ously interfere with any operation in the hydrometallurgical processes.
For all practical purposes of hydrometallurgy it may be considered as
equivalent to its analogous element, calcium. Magnesium usually occurs
combined with calcium as:
The Carbonate, (CaMg)C03, Dolomite,
The Sulphate, MgSO^.H^O, Kieserite,
The Silicate, H^MgjSiOg, Serpentine,
The Silicate, H2Mg3(Si03)4, Talc.
In oxidized ores the magnesium is largely in the form of carbonate and
silicate. It may also be present as sulphate, formed by the decomposi-
tion of pyrites, as the corresponding calcium sulphate. The magnesium
sulphate, kieserite, is very slowly soluble in water — about like gypsum.
The hydrous sulphate epsomite (MgSO^jXHjO) is readily soluble. In
roasting, this water of hydration is driven off. Much of the magnesium
sulphate formed in the oxidation of pyrites in mineralized veins is carried
away in solution.
In roasting sulphide ores, the magnesium carbonate is partly con-
verted into the oxide and partly into the sulphate. The oxide, like the
corresponding calcium oxide, is practically insoluble in water. It
reacts readily with chlorine, bromine, hydrochloric and sulphuric acids, to
form the chloride, bromide, and sulphate. Magnesium chloride is very
soluble in water — 100 parts of water will dissolve about 52 parts of
magnesium chloride at ordinary temperatures.
Magnesium sulphate is practically unaffected by any of the chem-
ical solvents. All the harmful hydrous sulphates may be converted
into the harmless anhydrous Sulphate by roasting. As in the case of
calcium, therefore, ore containing magnesium should be roasted with a
view of converting as much of it as possible into the form of sulphate.
Talc is insoluble in acids both before and after ignition. Roasting
greatly improves the talc for subsequent treatment by the wet methods,
especially in the leaching or filtering qualities of the ores containing it.
OXIDIZING ROASTING 59
Manganese. — Manganese is one of the most deleterious substances in
the extraction of metals by wet methods. It affects injuriously the acids,
the halogens, and cyanide. Fortunately it does not frequently occur
in ores of copper and the precious metals in quantities so great as to be
fatal.
Manganese almost universally occurs as the oxide; sometimes as the
sulphide and silicate. After roasting it is always in the form of oxide,
and roasting does not materially lessen its injurious effects on the solvent.
Manganese is readily soluble in acids and difficult to eliminate from the
solvent. Its principal injurious effect is in the consumption of chemicals.
Aluminum. — Aluminum, as it occurs in copper, gold, and silver ores,
affects the chemical processes somewhat injuriously. Its mineralogical
combinations are numerous and varied. It may occur as the oxide,
hydroxide, sulphate, or silicate. It usually occurs as the silicate, more
or less intimately associated with calcium, magnesium, iron, and the
alkali metals.
Roasting converts some of the aluminum compounds into aluminum
oxide (AI2O3), which is infusible at all temperatures ever attained in a
roasting furnace. It is not decomposed by heat alone. It is not decom-
posed by chlorine at any temperature. Anhydrous aluminum oxide is
perfectly insoluble in water. After strong ignition, it is likewise insoluble
in most acids. The lower the temperature at which aluminum oxide is
heated, the more soluble it is in the acids and alkalis.
All the silicates of aluminum are insoluble in water, with the exception
of the alkali salts, and these are soluble only when the ratio of the base to
the acid is above a certain limit. Many of the silicates are decomposed
by dilute sulphuric and hydrochloric acids. Chlorine, bromine, and
potassium cyanide react very slowly.
Aluminum sulphate, AljCSOJg, when heated to redness, is converted
into the oxide. The sulphate is very soluble in water. Chlorine reacts
very slowly with it. The basic sulphate, AljOg, SO3, lOHjO, is insoluble
in water, but soluble in sulphuric and hydrochloric acids.
When ores containing considerable aluminum are properly roasted,
and a sample filtered with water, it will be found on testing that there is
no soluble aluminum in the ore. If the sample is then filtered with di-
lute sulphuric or hydrochloric acid, some aluminum will be dissolved.
If the sample is treated with chlorine, bromine, or potassium cyanide,
only traces will be found in the solution.
The compounds of aluminum are so numerous, varied and compli-
cated that it is difficult, if not impossible, to determine their exact
composition either in the raw or roasted ore. The only alternative seems
to be to resort to direct tests with the chemical solvents. If acids are
used in the chemical treatment of the ore, some of the consumption of
the acid is due to combining with aluminum. Beyond the slightly
60 HYDROMETALLURGY OF COPPER
increased cost of treatment, due largely to increased consumption of acid,
no great inj ury to its presence in the ore is apparent. Its presence, even
in large quantities, is not fatal, or even serious, to any chemical process.
Cripple Creek ores, which are very successfully treated after roasting, by
cyanidation, and by chlorination with or without the us& of acid, fre-
quently contain as high as 20 per cent, alumina (AI2O3), and the average
is about 18 per cent. Copper ores at Clefton, Arizona, containing 16 per
cent, alumina have been successfully leached for many years with sul-
phuric acid.
Usually copper, gold, and silver ores do not contain more than several
per cent, alumina; frequently it is less than 1 per cent. Whatever the
condition of the aluminum in the raw ore, where it may be injurious, the
tendency in roasting is to convert it into the harmless aluminum oxide.
The higher the temperature at which the ore is roasted, the less difficulty
will result due to the presence of aluminum, but the ultimate temperature
of roasting ores containing much aluminum will depend on the other,
more or less fusible, constituents.
The hydrate of aluminum occurs mineralogically as Gibbsite; it is
easily dissolved by acids. The monohydrate occurs native as diaspora;
it gives up its water of hydration at 360° C. (680° F.).
Clay. — This is the term applied to hydrous silicates of aluminum,
produced for the most part by the decomposition of feldspar rocks, and
generally mixed with other substances, chiefly lime, magnesia, and oxide
of iron. Clay is frequently a constituent of ores, usually occurring as
"Gouge" matter in the vein.
As a rule clays contain from 45 to 60 per cent, silica; from 20 to 30
per cent, alumina; from 0.5 to 3 per cent, lime; from 0.5 to 3 per cent,
magnesia, small quantities of iron, and about 19 per cent, water. Clays
always contain a hydrous compound of alumina and silica, which is able
to give up the alumina contained by it as a base to sulphuric acid.
Clays are very much improved by roasting, both as to filtration and
chemical consumption.
Barium frequently occurs associated with copper, gold, and silver
ores in small quantities. It is usually in the form of sulphate, Barite
(heavy spar, BaSOJ. Sometimes it occurs as the carbonate, Witherite
(BaCOa).
If the carbonate is heated in an atmosphere free from sulphur, the
barium oxide, BaO, will be produced, which reacts with the halogens and
the acids. The temperature required for the decomposition of the
carbonate by heat alone is very high. In the presence of sulphur, the
carbonate is converted into the sulphate.
Barium sulphate is practically unaffected by any operation of the
chemical processes. Any heat obtainable in a roasting furnace does not
decompose it. It is insoluble in water and in acids.
OXIDIZING ROASTING 61
Alkali Metals. — The alkali metals, sodium, and potassium, are fre-
quently found in considerable quantities associated with ores. They
usually occur as the feldspars or hornblende, and as such are unaffected
by roasting or any of the chemicals used in the solvent processes.
Chlorine, Bromine. — Chlorine and bromine are sometimes found in
copper, gold, and silver ores, and when they do so occur are of considerable
metallurgical importance in roasting. The compound which is most
common is the silver chloride, cerargyrite (AgCl) . The minerals embo-
lite, Ag(ClBr), and bromyrite, AgBr, occur occasionally, and in roasting
may be considered the same as cyrargyrite. Chlorine also occurs quite
frequently in combination with lead.
The surface ores of Tonapah, Nevada, show much of the silver
combined with chlorine — frequently as much as 20 per cent. As depth is
attained, the silver chloride gradually merges into the sulphide, although
the chlorides appear never to be entirely absent.
The principal point of importance in connection with the roasting of
silver chloride, is the danger of volatilization, even with an oxidizing
roast. In making exhaustive tests in Denver, on a working scale, on
some of the Tonapah ores, it was found that the volatilization, with an
oxidizing roast, was about the same as with a chloridizing roast, but in no
case was the volatilization serious. If volatilization is known to take
place in an oxidizing roast, chlorides in the ore may be suspected.
To ascertain the amount of silver chloride in the ore, leach, or treat
a sample with sodium hyposulphite (sodium thiosulphate) and compare
the hypo tails with that of the original ore. Also test for chlorine with
silver nitrate.
Loss of Weight in Roasting. — There is always some loss of weight in
ore due to roasting. The loss is usually largest in pyritic ores and in
pyritic concentrates, but it may also be considerable in ores which are
oxidized and highly silicious. In sulphide ore the loss is due mostly to
the expulsion of the sulphur; in oxidized ores it is mostly due to driving
off the water of hydration. The water so combined, in many ores may
be quite large. Iron in oxidized unroasted ores is almost always in the
form of ferric hydrate (limonite), 2re03, SHsO, which contains 14.4 per
cent, water, all of which is driven off in roasting. Similarly othei' sub-
stances give up their water of hydration, and some of the elements are
eliminated by volatilization.
The loss of weight in sulphide ores is represented by the substitution
of oxygen for sulphur. From the equation
4FeS2 + 1102 = 2Fe203+8S02
the loss of weight of pyrites can readily be calculated that 3 parts of
FeS2 = 2 parts Fefis, but the matter is usually not so simple as this, owing
to other constituents in the ore and the manner in which the remaining
62 HYDROMETALLURGY OF COPPER
sulphur is combined. If, for example, there is galena (PbS) in the ore
and is oxidized to sulphate (PbSO^), there has been an actual gain of
weight of 4 atoms of oxygen, or 27 per cent.
The loss of weight can readily be ascertained from the difference in
weight between the raw and roasted, as it is charged and withdrawn
from the furnace. This method is expensive and not quite accurate
because the dust loss cannot usually be taken into consideration. The
loss of weight is best and most conveniently obtained by direct experi-
ment. This is done by weighing a small average sample of the ore, then
thoroughly drying it; weighing it again, and then roasting it in a roasting
dish, in a muffle, to the same extent as the ore is roasted in the mUl. A
sulphur determination will show this. From the differences in weight
between the raw ore, the dried ore, and the roasted ore, the loss due to
moisture and the loss due to roasting can easUy and accurately be
ascertained.
A ton of roasted pyritic concentrates will occupy about 24 1/2 cu. ft.
This is derived from 2800 lb. of raw ore, which will occupy about 23 2/3
cu. ft. per ton. A ton of the concentrates after roasting will weigh from
1450 to 1700 lb., and will occupy about 17 1/2 cu. ft. The loss of weight
in Cripple Creek ores, due to roasting, is usually from 5 to 7 per cent.,
based on the control samples. Of this loss, about 2 per cent, is for
moisture, and from 3 to 4 per cent, dust and volatilization loss. Of the
volatilization loss about 1 per cent, is accounted for by the elimination
of the greater portion of the sulphur. The accountable dust loss is
about 2 per cent., and the unaccountable loss amounts to about 1 per cent.
Much of this unaccountable loss is due to unsettled dust going out of the
furnace stacks, and some also due to unrecovered dust in crushing and
roasting other than flue dust.
At Butte, in roasting copper concentrates, containing 35 per cent,
sulphur down to 7 per cent, sulphur, the loss of weight, including flue dust,
is about 20 per cent.
In roasting Black Hills ore, containing 11 per cent, sulphur, down to
0.08 per cent., there was a loss in weight of 21 per cent., even though the
ore was apparently thoroughly dry. This ore was very talcy, and the
great loss was evidently due principaUy to the water of hydration.
CHAPTER IV
CHLORIDIZING ROASTING
Object of Chloridizing Roasting.— Most of the chlorides, at elevated
temperatures and in the presence of sulphides or sulphates, have the
power of converting copper and silver into their respective chlorides, and,
to some extent, the gold also. To roast in the presence of chlorides,
usually sodium chloride (common salt), is known as "Chloridizing
Roasting." The term "chloridizing" is limited to the production of
chlorides by the interchanging of chlorine from its chloride combinations,
usually at elevated temperatures; while the term "chlorinating" is
limited to the production of chlorides, usually in the wet way, by the
application of free chlorine.
The objects of chloridizing roasting are:
1. In copper ores, or in gold and silver ores containing copper, to
convert the copper into chlorides, which will not react with chlorine or
the acids, but which are directly soluble in water or in chloride
solutions.
2. In silver ores, or in gold and copper ores containing silver, to con-
vert the insoluble metallic silver or its insoluble compounds, into the more
soluble silver chloride.
3. In any ore, to convert the harmful elements into less harmful
compounds.
4. To assist in a more efficient oxidizing action than is possible under
the same conditions, in ordinary oxidizing roasting.
Metallic silver is not readily soluble in any of the commercial chemical
solvents. The silver chloride is readily soluble, either in chloride solu-
tions, sodium or calcium hyposulphite, or in potassium or sodium
cyanide. If, therefore, a high percentage of the silver can be converted
into the chloride, a quick and correspondingly high percentage of the
silver can be extracted.
If the ore contains copper, or if a copper ore is treated by a chloride
process, it is frequently desirable to convert the copper in the ore into
the soluble cupric chloride, so as to save acid, if an acid process is used.
It may be cheaper to convert the copper into chlorides at the expense of
a cheap material, such as salt, than to let the oxides react with the more
expensive acids. Chloridizing roasting is largely used in the extraction
of copper from its ores.
63
64 HYDROMETALLURGY OF COPPER
Most of the chlorides are soluble in water; if desired, many of the
objectionable elements in the ore may be removed by a preliminary
washing, after roasting, and before applying the chemical solvent.
Of the metallic sulphides usually associated with copper, gold and
silver ores, those of iron, copper, lead, and zinc are the most common.
Of these, only the iron and copper sulphides, are available to react with
the salt; while those of lead and zinc remain quite indifferent.
Adaptability of the Various Ores to Chloridizing Roasting. — Ottokar
Hofmann' aptly classifies the adaptability of the various ores to chloridiz-
ing roasting, as follows:
1. Those like iron and copper pyrites, gray copper ore, and silver
copper glance, which in roasting form sulphates, and decompose salt,
liberating chlorine.
2. Those like galena and zinc blende, which form sulphates remaining
indifferent to salt.
3. Antimonial and arsenical silver minerals which form antimonates
and arsenates of silver.
The gangue remains indifferent, like quartz or porphyry, or it takes
an active part, like limestone, talc, spar, manganese, and minerals con-
taining magnesia.
If ore consists of minerals of the first group together with an indifferent
gangue, chloridizing roasting offers no difficulty and a high chloridization
can be obtained without much loss of silver by volatilization and no
special skill is required in the roasting; neither does it matter if the salt
is added to the charge before entering the furnace or after it has been
subjected to partial oxidizing roasting.
The process of chloridizing roasting becomes more difficult if one or
both of the minerals of the second class are present in large quantities,
even if associated with an indifferent gangue. With such ores the time
of adding the salt becomes very important. If added before the charge
enters the furnace a very inferior chloridization is obtained, as is also the
case if the salt is added before the oxidizing period has sufficiently
advanced. 'Moreover, the temperature and air supply require much
attention.
The roasting is still more difficult if all the classes of ore are repre-
sented in connection with a gangue like limestone which takes an active
and injurious part in the operation.
Chemistry of Chloridizing Roasting. — The sulphides in the ore,
mostly relied upon for chloridization, are those of iron and copper. The
sulphates of these metals, formed during the roasting, react with the
salt to form sodium sulphate and the chlorides of the metals. Some
hydrochloric acid and chlorine are formed at the same time, largely due
' Mineral Industry, 1896.
CHLORIDIZING ROASTING 65
to the action of the sulphur trioxide and sulphuric acid. The following
reactions, represent in a general way the chloridizing action:
2NaCl+ FeSO, = Na2S04+FeCl2.
2NaCl+ CuS0, = Na,S0, + CuCl2.
2NaCl+ 2S03 = Na3SOi + 2Cl + S02.
2NaCl+ H2SO, = Na2SO, + 2HCl.
The chlorine and chlorides thus formed react with the silver and silver
sulphate to form the silver chloride:
FeCl2+ AgS0, = 2AgCl+FeS0,.
CuCl,+ AgS0, = 2AgCl + CuS0,.
2NaCl + AgSO, = 2AgCl + Na2SO,.
2HC1 +2Ag + 0=2AgCl + H20.
CI + Ag = AgCl.
Any or all of these reactions may take place at the same time. The
salt, reacting with the sulphates of iron and copper, converts those metals
into their higher chlorides, while the chlorine and hydrochloric acid are
formed at the same time. Both chlorine and hydrochloric acid, at the
temperature of the roasting furnace, react readily with metallic silver or
its sulphate, to form the silver chloride, while the chlorides of iron and
copper, in chloridizing the silver, may pass repeatedly from the ferric
and cupric condition to that of the ferrous and cuprous:
2FeCl2 + 2Cl =2FeCl3.
2FeCl3 + 2Ag = 2FeCl2 + 2 AgCl.
2CuCl +2C1 =2CuCl2.
2CUCI2 + 2Ag = Cu^Cl, + 2 AgCl.
The ferric chloride, FeClj, is volatile and at a red heat, chloridizes
the silver with great avidity. The ferrous chloride at the same time is
resolved into ferric oxide and ferric chloride:
3FeCl2 + 03 = Fe203 + 2FeCl3.
In contact with aqueous vapor, and the fuel gases, at a red heat, the
ferrous chloride may be converted into the magnetic oxide :
3FeCl2 + 4H20=Fe30, + 6HCl + 2H.
The magnetic oxide may be again reconverted into the ferric oxide,
in the presence of salt and at a lower temperature, as was shown con-
clusively by Stetefeldt' who succeeded in converting an ore containing
67.2 per cent, magnetite to 1.4 per cent, after 4 1/2 hours' roasting
with 5 per cent. salt.
Cupric chloride (CuClj) is easily decomposed at a red heat into cuprous
"Trans. A. I. M. E., 1885-1886.
5
66 HYDROMETALLURGY OF COPPER
chloride (CujClj) and free chlorine, which gives free chlorine available
for the chloridization of the silver.
Arsenic and antimony form chlorides, which are easily volatile
and which may be decomposed into arsenous and antimonous acids
and chlorine and hydrochloric acid, by means of the oxygen, and the
vapor from the burning fuel. These chlorides, however, will mostly
escape withput decomposition. If the temperature is low and the salt
has not been added until the arsenic and antimony have been largely
driven off, the soluble arsenates and antimonates, in the roasted ore,
will not usually be present in sufficient quantities to seriously interfere
with the extraction. If the raw ore contains arsenic and antimony in
large amounts, much of the silver may be converted in the early stages
of the roasting, into arsenate and antimonate. Ottokar Hofmann found
in roasting arsenical ore that 53.8 per cent, of the silver was soluble in
sodium hyposulphite, probably as arsenate of silver, before the salt was
added to the ore.
Arsenous oxide volatilizes at 218° C. Chlorine, with the aid of heat,
decomposes the sulphide of antimony completely, forming the trichloride
of antimony and sulphur dioxide. The trichloride of antimony melts at
70° C.
Zinc. — Zinc blende, in oxidizing roasting, is converted into zinc
oxide and zinc sulphate, while sulphur dioxide escapes. In the presence
of salt, zinc blende remains indifferent and does not decompose salt, at
least at the temperature used in chloridizing roasting. Salt does not
decompose zinc sulphate. Zinc oxide may be converted into the chloride
at a red heat. By the action of chlorine and hydrochloric acid zinc
chloride is formed, which is very volatile. In ore which has been given
a chloridizing roast, the zinc is usually found as the oxide, sulphate, and
chloride. Zinc oxide, like the calcium and magnesium oxides, is com-
pletely soluble in acids, so that when an acid process is employed to
extract copper or silver, zinc must be regarded more or less as equivalent
to calcium and magnesium. In the chlorination of gold ores, when the
chlorine is applied direct without the use of acids, considerable quanti-
ties of zinc will not seriously interfere with the treatment. Both zinc
oxide and zinc sulphate react very slowly with chlorine.
Lead. — If galena is subjected to chloridizing roasting, especially in the
presence of sufficient air, most of the lead is converted into sulphate,
which does not react on the salt, and oxide, which may be converted
into the chloride. Both lead oxide and chloride are volatile, while the
sulphate remains indifferent. In the roasted ore, the lead will be in the
form of sulphate and chloride, but the sulphate will predominate.
Calcium Carbonate. — Carbonate of lime, when roasted with metallic
sulphides, will change partly into calcium sulphate and partly into the
oxide (lime). The calcium sulphate does not act on salt, but the oxide
CHLORIDIZING ROASTING
67
decomposes the metal sulphates and chlorides, and also, to some extent,
the silver chloride. Calcium oxide, or carbonate, does not absorb
chlorine when cold, but at a red heat combines with it to form calcium
chloride, with the evolution of oxygen. If the ore contains calcium
carbonate in large excess, only a small quantity of iron and copper sul-
phates will be formed, to decompose the sodium chloride. Most of the
iron and copper sulphides in the ore will be converted directly into the
oxides. Since the sulphates of iron or copper are necessary to release the
chlorine in the salt, and these sulphates are not formed or are immediately
appropriated by the lime, the salt will not react to release chlorine or hy-
drochloric acid, which are the most active elements in chloridizing roast-
ing of copper, gold, and silver ores. The lime itself is quite indifferent to
silver chloride at low temperatures, but decomposes it energetically
when the temperature reaches red heat. If there are more sulphides in
the ore than are necessary to convert the lime into sulphate or chloride,
usually a good chloridization of the silver and copper may be obtained.
The practical effect of lime, in the formation of silver chloride, in
chloridizing roasting, is clearly shown by a well conceived experiment by
Ottokar Hofmann' on concentrates containing large quantities of sul-
phur, arsenic, iron, considerable zinc, some lead and aluminum. The
object of the experiment was to ascertain the effect of varying quantities
of calcium carbonate, in the formation of silver chloride, all other condi-
tions remaining the same. The ore was roasted one-half hour with 7
per cent. salt.
Sample
No.
Per cent, of con-
centrates in
mixture
100
75
62
50
25
Per cent, of barren
gangue in mixture
mostly CaCOg
25.0
37.5
50.0
76.0
Value of mixture
per ton 02. silver
96.0
72.0
60.0
48.0
24.0
Value of leach
tails per ton oz.
silver
Chloridiza-
tion per cent.
2.91
5.38
4.72
5.38
5.47
97.0
92.6
92.2
88.8
77.2
The deleterious effect of the lime is very evident from these results.
Magnesium usually occurs as the carbonate, and in chloridizing
roasting, as in oxidizing roasting, has about the same effect as calcium.
Magnesium carbonate is decomposed at 170° C. (338° F.) into magnesium
oxide. If there are sulphides in the ore, much of the magnesium will be
converted into the sulphate. The sulphate is quite infusible, melting
only at about 1 100° C. In chloridizing roasting the magnesium combines
with the chlorine to form magnesium chloride (MgClj), with the liberation
of oxygen. Magnesium chloride fuses at a red heat, 708° C. (1300° ¥.).
Magnesium chloride is more positive in its action than sodium chloride.
' Min. Ind., 1896.
68 HYDROMETALLURGY OF COPPER
Quartz.— Quartz is the most desirable gangue in chloridizing, as
it is in oxidizing roasting. Silica is indifferent to any action in chloridiz- ■
ing roasting, unless perhaps, it promotes the formation of chlorides and
oxides by catalytic action.
Barium sulphate, which occurs quite frequently associated with
silver ores, remains inert during chloridizing roasting.
Alumina is not fused by heat alone, nor is it decomposed by chlorine
at any temperature.
Sodium sulphate, so abundantly formed during chloridizing roasting,
may be considered a neutral substance in any of the hydrometallurgical
processes. It is usually filtered off before the solvent is applied.
Silver sulphate is completely decomposed by sodium chloride, at the
temperature of the roasting furnace.
Percentage of Salt. — The percentage of salt used at various mills
differs greatly, depending largely on the character of the ore, principally
the gangue. More salt has sometimes been used than was really needed.
Aaron when roasting a pyritic ore with 4. per cent, salt, found an enor-
mous loss by volatilization; later he reduced the amount of salt to 3 lb.
per ton of ore, and got satisfactory results. Ordinarily, the amount of
salt for silver ores will vary between 1 and 5 per cent., although much
greater percentages than these have been used. Only 3 per cent, was
used at Panimint, California, and gave a chloridiz ation of 95 per
cent. The amount of salt is largely proportional to the amount of
lime or magnesia in the ore. An excess of salt does not improve the
chloridization.
If copper, instead of silver ores are to be chloridized, the amount of
salt required will be larger. If the copper contained in the ore is consider-
able, the amount of salt will be roughly proportional to the copper. From
5 to 10 per cent, might be considered fair averages for ores containing
only several per cent, of copper.
The minimum amount of salt that may be used for any ore is best
determined by direct experimenting. First determine the conditions of
time, temperature, and fineness of the ore, which will give the highest
satisfactory chloridization with an abundance of salt, and then reduce it
in successive roasts until a minimum is obtained which will show no ap-
preciable difference as compared with the highest chloridization possible,
with an abundance of salt.
Time of Adding Salt. — The time of adding salt is governed almost
entirely by the composition of the ore. If the ore is low in sulphur, the
salt may be added before the ore is charged into the furnace, preferably
before it is crushed, so as to get an intimate mixture of ore and salt.
If the ore contains considerable sulphur, combined with iron or copper,
the ore may be given practically a full oxidizing roast before adding the
salt, and still have enough sulphur in the ore to chloridize the silver. If
CHLORIDIZING ROASTIM; 69
copper is to be chloiidized, the ore should contain at least as much
sulphur as copper before the salt is added. 'If the ore contains consider-
able zinc and lead sulphides, the sulphur combined with the zinc and lead
may be disregarded for the purpose of chloridization, and the ore given
an oxidizing roast previous to adding salt, if the sulphur combined with
the iron and copper is large; if the sulphur so combined is small, the
salt is best added at once to the raw ore.
If the ore is thoroughly oxidized, and does not contain sufficient sul-
phur, either as raw ore, or after a thorough oxidizing roast, the salt and
pyrites, both finely ground, may be added to the ore. Ferrous sulphate
may be used instead of pyritesj but is much more expensive.
If the salt is added while there is a large excess of sulphur in the ore,
it will largely be volatilized without doing any good. If the salt is added
at the proper time, the chloridization takes place very rapidly.
Heap Chloridization. — Imperfectly roasted ore, after being drawn from
the furnace and placed in a mass on the cooling floor, or in a pit, will gain
in chloridization, largely in proportion to the imperfectness of the roast.
On very poorly roasted ores it may gain as much as 50 and 75 per cent.
The reactions which take place in heap chloridization are essentially the
same as those which take place in the furnace. In any well regulated
mill, the ore is probably never so poorly roasted but that all the iron and
copper sulphides are decomposed.
From the reactions given for chloridizing roasting, it is evident that
air is not essential to the chloridization after the sulphides have been
converted into the sulphates. Small quantities of air, however, permeate
the mass and promote the reactions.
Ores which are well roasted in the furnace, and which is the only safe
way to roast, do not show any increase in chloridization in the heaps, pit,
or cooling floor.
If the ore does not contain lime in considerable quantity, moistening
the hot ore adds to the chloridization of the silver and this is especially
the case if the ore contains copper, or is moistened with a solution of
cupric chloride. If the ore contains appreciable quantities of lime, then
instead of adding to the chloridization, there is likely to be a diminution.
A loss of chloridization of 10 per cent, has been known to occur in this
way.
Composition of the Roasted Ore. — Ores which have been subjected to
chloridizing roasting contain a great number of soluble salts. Of these,
sodium sulphate, resulting from the decomposition of the salt, and the
undecomposed sodium chloride, predominate. Besides these there may
be the sulphates of manganese, zinc, copper, iron, aluminum, and mag-
nesium; the chlorides of the same metals and of calcium and barium.
The barium chloride will be immediately decomposed on solution, and
precipitated as insoluble barium sulphate. Sodium arsenate is also
70
HYDROMETALLURGY OF COPPER
present if the ore contains arsenic. Salts not easily soluble in water, are
cuprous chloride, lead chloride, calcium sulphate, sodium antimonate,
and calcium oxide. Lead chloride, on solution, will be precipitated as
lead sulphate. Silver chloride, lead sulphate, and antimonate* are al-
most insoluble in water but are soluble in solutions of other chlorides.
Cuprous chloride, calcium sulphate and calcium oxide are more soluble
in a chloride solution than in water. If the ore contains large quantities
of lime, the soluble metals may be precipitated as hydroxides.
The composition of Ontario raw ore, and of the ore roasted with 13
per cent, salt in a Stetefeldt furnace, from analyses made by Stetefeldt,
is given by Kustel' as follows:
Per
cent.
Zinc
9.45
Lead
6.07
Iron
2.77
Copper
1.41
Manganese .
0.45
Silver
0.60
Sulphur. . . .
7.68
Antimony. .
1.20
Arsenic ....
0.20
Silica
5.5.21
Alumina. . .
13. U
Potassium
1.00
sodium.
Bismuth. . .
trace
Cadmium.. .
trace
Lime
trace
Magnesia. . .
trace
Roasted ore; shaft
Per
cent.
Roasted ore; Hue
Per
cent.
Zinc chloride
Copper chloride
Aluminum chloride
Sodium chloride
Traces of chlorides of other
metals.
Aluminum sulphate
Lead sulphate
Sodium sulphate
Traces of sulphates of other
metals.
Rest, metallic oxides and
gangue.
Sulphur in undecomposed sul-
phides.
1.38
0.25
1.51
3.68
0.56
3.26
4.62
0.18
Aluminum chloride
Sodium chloride
Traces of chlorides of other metals.
Aluminum sulphate
Lead sulphate
Sodium sulphate
Copper sulphate
Zinc sulphate
Traces of sulphates of other
metals.
Rest, metallic oxides and gangue
Sulphur in undecomposed sul-
phides.
1.07
3.08
2.88
5.18
10.01
0.74
1.47
0.°64
These results are interesting as showing the condition of the various
constituents of the ore, after chloridizing roasting.
Of the silver contained in the ore, 81.32 per cent, was chloridized.
Volatilization of the Silver. — The volatilization of the silver, in chlorid-
izing roasting, is largely due to the presence of other chlorides which
are more volatile than the silver chloride. The volatilization of the
silver is roughly proportional to the volatilization of the base metal
chlorides, or to the loss in weight the ore sustains. Manganese seems to
be particularly active in causing loss by volatilization. Cupric and
cuprous chlorides, both of which volatilize at a low heat, are likely to
cause a heavy loss of silver. Arsenic and antimony are also effective in
assisting the volatilization of the silver chloride. A high temperature
indirectly causes a high loss of silver by the expulsion of the volatile
chlorides.
Much of the loss due to volatilization, is chargeable to the manipula-
' Kustel, "Roasting of Gold and Silver Ores".
CHLORIDIZING ROASTING 71
tion of the ore in the furnace. Any condition which wUl produce the
chloridization of the silver, if carried to excess, will also cause its volatil-
ization. Silver chloride, under the conditions of roasting, is formed at a
comparatively low temperature by the chemical reaction between the
salt and sulphates. A scarcely visible red heat is quite sufficient for
these reactions, and if this temperature is not exceeded, only a small loss
by volatilization will occur. If, however, the temperature is elevated to,
say, a bright red, a high loss of silver is sure to take place. A safe rule to
follow, is to keep the ore at the lowest possible temperature at which it
will give off visible fumes. It is best to maintain a deep layer of ore, and
plenty of air. A small charge of ore spread thinly over a large hearth
area, will show a greater loss by volatilization, than a large charge with
a deep bed spread over the same area.
The stirring of the ore should not be too frequent, but this is not an
essential if the temperature is not too high. These conditions for good
chloridizing roasting are contrary to those desirable in the best oxidizing
roasting, where the ore should be in a thin layer and be stirred as fre-
quently as possible. Chloridizing roasting can be done with a very small
loss by volatilization — frequently only an inappreciable loss — and it is
very probable that the great losses recorded are due entirely to improper
manipulation.
In the chloridizing roasting of any ores, at an exceedingly low temper-
ature, a difficulty may arise, in the subsequent chemical treatment. If
roasted at too low a temperature, some of the injurious elements may not
be decomposed sufficiently, so that trouble may arise in the consumption
of chemicals when the solvent is applied to the ore. This, however, is a
matter for adjustment for each particular ore, and will usually, in such
cases, resolve itself down to roasting at the highest temperature the ore
will stand without serious loss by volatilization.
Almost any ore, likely to be treated by a solvent process, can be
effectively chloridized, but the essential of such roasting is that the loss
during the process should not be serious, or if serious, its recovery should
be carefully considered. With care, many silver ores can be given a
chloridizing roasting with not much greater loss of silver than in oxidizing
roasting.
That there is sometimes a considerable loss of silver in oxidizing roast-
ing is pretty well established. Plattner in his " Metallurgische Rost-
prozesse" goes very minutely into the loss of gold and silver in oxidizing
roasting. By a series of mufile roasts on a small scale, he comes to the
conclusion that while there is no loss of gold, the loss of silver is unavoid-
able. From numerous tests, varying from 3/4 to 1 1/2 hours he records
a loss of from 0.5 to 18 per cent, of the silver. He concludes that the
percentage loss of silver increases with the temperature, the porosity of
the charge which facilitates the supply of air throughout the ore mass, the
72
HYDROMETALLURGY OF COPPER
freedom of the silver from combination with other substances, and with
the time of the roasting.
In order to verify the work done by Plattner, Christy' cites some
experiments in . oxidizing roasting made by himself and others. The
material used in the experiments were concentrates from Nevada City,
California, which consisted chiefly of pyrite, with small amounts of
chalcopyrite (0.05 to 1.5 per cent. Cu), a little galena, a small amount
of quartz, traces of arsenic and antimony, but no tellurium. The ore was
given an oxidizing roast of from 1 1/2 to 8 1/2 hours; in the early stages
at incipient dull red, and finished at dull red to full red. The results
are tabulated as follows:
Time of
Raw ore, ounces per ton
Roasted ore, ounces per ton
Percentage loss per ton
roasting, hours
Gold
Silver
Gold Silver
Gold 1 Silver
1 1/2
2 1/2
8 1/2
4.. 58
4.58
4.50
27.50
27.65
28.39
4.58
4.58
4.50
26.44
27.07
27.39
0.00
0.00
0.00
3.85
2.09
3.52
These results verify the conclusion of Plattner and others, that while no
loss of gold occurs in oxidizing roasting, by volatilization, the loss of
silver may be considerable.
Butters^ found in roasting a hard white quartz, intimately mixed
with about 7 per cent, calcite and a very little pyrite, assaying 5.55 oz.
silver and 0.65 oz. of gold, per ton, that there was a loss by volatilization
in oxidizing roasting, of 2 to 9 per cent, of the silver, but none of the
gold.
It is possible that losses of silver, which have been attributed to
chloridizing roasting may have been partly due to the loss in oxidizing
roasting, and especially if some of the silver in the ore is in the form of
chloride.
Volatilization of the Gold. — It is pretty well established both by care-
fully conducted experiments and by the experience of practical metal-
lurgists, that no loss of gold takes place either in oxide or sulphide ores,
in oxidizing roasting. There seems to be some doubt in the case of tellur-
ides, but the experience with Cripple Creek ores, containing tellurium, of
which hundreds of tons are roasted daily, is, that no appreciable loss, if
any at all, occurs by volatilization. Kustel records a loss of 20 per cent,
of the gold during the oxidizing roasting of certain telluride ores of gold
and silver, and states that this is not a mechanical but a volatilization
loss. There can be no doubt about the gold, combined with tellurium,
> Trans. A. I. M. E., 88-89.
2 Trans. A. I. M. E., 88-89.
CHLORIDIZTNG ROASTING 73
volatilizing at elevated temperatures, but whether any volatilization
takes place at the low temperatures and under the practical conditions
of roasting, seems very doubtful.
Tellurides, even in small quantities, are extremely sensitive to chlorine
at almost any temperature, at which salt is decomposed. Experience
with Cripple Creek ores, in large 100-ton furnaces, showed appreciable
loss of gold when only a very small amount of salt — from 1/2 to 2 per
cent. — was added during the roasting. A loss was shown even when the
salt was added to the hot ore dropping on the cooling hearth.
Many ores are known to contain chlorine, frequently as chloride of
silver or chloride of lead. That ore containing a part of its silver as
chloride, if given an oxidizing roast, will volatilize small amounts of both
gold and silver, was proved conclusively by the author in exhaustive
tests on Tonapah ore. It is probable when gold losses occur in any ore
in oxidizing roasting, and especially in the tellurides, it may be due to
small quantities of chlorine.
Prof. Christy' made some interesting experiments on the volatiliza-
tion of gold in the chloridizing roasting of pyritic ores. As the result of
a large number of experiments he comes to the following conclusions:
At 100° C. (212° F.) the volatility of the gold in an atmosphere of
chlorine, is almost zero; that the loss begins, above this temperature, to
rapidly increase to a maximum at a temperature of about 250° C. (482°
F.); that it rapidly diminishes to a temperature somewhere below red
heat; that it again increases, but more slowly, to another maximum, at
a temperature above a melting heat, and that this increase is apparently
continuous between a red heat and a white heat. The ratio of losses at
various temperatures is also instructive; at incipient redness the standard
loss is already 0.05 per cent.; at a cherry red it is five to seven times as
great as at incipient redness; at incipient yellow it is more than eight
times what it is at incipient red; while at melting heat it is nearly thirty
times as great.
Crosly^ found with a certain California pyritic ore, assaying about
$110.00 in gold, and $40.00 in silver, that' an oxidizing roast showed no
appreciable loss, but when the salt was added, losses appeared rapidly.
Thus, according to his tests, with 3 per cent, salt the gold loss was 30
per cent, and the silver loss 50 per cent, of the assay value. He at-
tributed the loss to the presence of tellurides, which he supposed were
present.
Aaron^ found a large loss in roasting a simple pyrite in a 3-hearth
reverberatory furnace, with 1 to 2 per cent, of salt, which was added
on account of the silver. He then made two tests on a small scale;
' Trans. A. I. M. E., 1885.
' Trans. A. I. M. E., 1888.
= "Leaching Gold and Silver Ores," 1881.
74 HYDROMETALLURGY OF COPPER
one with 4 per cent, salt, the other without any salt, and purposely
pushed the roasting to an extreme as to time and temperature, and
found on assaying that the salted ore contained less than half as
much gold as the unsalted one. He also found that the ore, in
roasting, sustained a loss of 18 per cent, in weight, and consequently
should have assayed 18 per cent, more than the raw ore, which was not
the case. By modifying the roasting, so as not to add the salt until the
dead roasting of the ore was finished, not only did the roasted ore assay
20 per cent, more than the raw ore, but the yield overran the guarantee,
while the tailings, nevertheless, contained considerably more gold than
before. He afterward found that a very small quantity of salt — not more
than 3 lb. per ton of ore — might be mixed with the raw ore without
detriment to the gold and with decided advantage to the extraction of
the silver.
The principal object of roasting gold ores, containing silver or copper,
with salt, is to chloridize the small amounts of silver and copper, and in
some cases to neutralize substances in the ore, which might be injurious
to the solvent. By a partial chloridizing roast, or even with an oxidizing
roast, it is practicable to get a high extraction of both the gold and silver
by either the cyanide or chlorination processes. If it is simply a matter
of neutralizing injurious substances in the ore, this can be done in
chloridizing roasting by not pushing the operation to the limit, and if not
carried beyond the point required to satisfy the base elements, no
appreciable volatilization of either gold or silver will occur.
Chloridization of Copper Ores. — According to Von Kothny^ by roast-
ing copper sulphide mixed with iron oxide and sodium chloride prac-
tically all the sulphur goes into sulphate and about half the copper
is transformed into chloride. Anhydrous cupric chloride mixed with
sodium chloride and heated in a current of air to 250° C. gives off
chlorine. The decomposition of copper sulphate by sodium chloride
begins at 280° C. Ferric chloride converts copper oxide into chloride
very rapidly at temperatures from 500 to 600° C. The formation
of copper sulphate by roasting with copper oxide in the presence of
sodium chloride plays no part. At temperatures of 300 to 600° C. ferric
sulphate converts copper oxide slowly into sulphate. Chlorine is with-
out direct action on cuprous sulphide. The reactions involved in the
Hargreaves process by which hydrochloric acid is formed plays no part
in converting cupric oxide into chloride. Von Kothny concludes that
the mechanism of the chloridizing of pyrite cinder containing a small
amount of copper and sulphur, is as follows : The copper is present largely
as sulphide, which by an oxidizing roasting is converted into sulphate and
oxide. Sodium chloride acts directly on the sulphate and ferric chloride
on the oxide. To insure chloridizing of such material it must be finely
• Metallurgie,Ju\j 8, 1911.
CHLORIDIZING ROASTING 75
ground; a large amount of air must be admitted in the oxidizing roasting
period and stirring must be resorted to to insure contact with oxygen;
sufficient pyrite must be present to furnish the required amount of ferric
chloride and is best added in a weathered form; for 4 per cent, copper
content at least 7.5 per cent, salt must be added; the process should be
carried^ out at temperatures between 500 and 600° C. For a full discussion
of chloridizing roasting of copper ores see " Longmaid-Henderson process, "
Part II, page 246.
Principal Factors in the Loss of Silver, Gold, and Copper by Volatiliza-
tion.— -The principal factors, controling the loss of silver, gold and copper
by volatilization, in chloridizing roasting, have been well established both
by practice and careful experiments. These, in the order of their im-
portance, are:
1. Temperature.
2. Time.
3. Amount of air, or surface exposed.
The amount of salt has some influence on the volatilization, but it is
supposed that the amount of salt used is the least that will give satisfac-
tory results, and once determined, becomes constant.
The presence of volatile substances, such as arsenic, antimony,
selenium, tellurium, and the chlorides of copper and iron, also affect the
volatilization. Gold is particularly sensitive to tellurium in chloridizing
roasting. But as these are constituents of the ore, they cannot be con-
sidered as variable, or controllable factors, except in so far as preliminary
oxidizing roasting may eliminate them.
Temperature is the all important factor in chloridizing roasting. Any
ore chloridized at an excessive heat will volatilize much of the metals,
irrespective of any considerable time, or in any atmosphere attainable in
a roasting furnace. If the temperature is kept at the lowest possible
point at which the metals can be chloridized, then the time of roasting
and the amount of oxygen in the furnace atmosphere i;i immaterial. By
merely changing the temperature, from 10 to 80 per cent, of the metals
may be volatilized in a short time; or only a few per cent, may be vola-
tilized after several hours roasting, all other conditions remaining the
same.
Russel, experimenting with Ontario ores, found a volatilization of 8.3
per cent, of the silver at a dark red heat, and of 17.6 per cent, at a cherry
red. Ottokar Hofmann' found in roasting calcareous ores containing
large quantities of zinc and arsenic, that the ore lost 3.5 per cent, of
its weight and 1.8 per cent, of its silver was volatilized when roasted
at a low temperature; the same ore roasted at a high temperature
with insufficient air, lost 7 to 13 per cent, of its weight, and 15 to 25
' Min. Ind., 1896.
76 HYDROMETALLURGY OF COPPER
per . cent, or more of the silver. He also found' on an ore consisting
essentially of 25 per cent, zinc, 12 per cent, lead, 21 per cent, sulphur,
7 per cent, iron, and 10 per cent, calcium carbonate, that the loss by
volatilization varied from 1.7 to 15 per cent. The least increase of tem-
perature above a dull red, caused a heavy loss, even if the increase
lasted for only a short time. The average of 31 days roasting at a high
(almost white) heat was:
Chloridization of the silver, 72.7 per cent.
Loss by volatilization, 17.9 per cent.
Roasted at a low heat (not above a dull red) :
Chloridization of the silver, 81.5 per cent.
Loss by volatilization, 1.2 per cent.
The chloridization in favor of the lower heat was 8.8 per cent, and a
decrease of loss by volatilization of 16.7 per cent.
Time. — The volatilization of the silver, gold, and copper, in chloridiz-
ing roasting, is approximately proportional to the time of roasting, other
conditions remaining the same. If in chloridizing roasting, an ore will
lose, say 1 per cent in the first hour after the salt is added, it will lose
approximately 5 per cent, after five hours roasting, if the conditions
remain the same.
Air or Oxygen. — Time and temperature remaining the same, the vol-
atilization will be approximately proportional to the amount of air
supplied to the ore. If a ton of ore is roasted on a hearth area of 100 sq. ft. ,
shows a volatilization of say, 1 per cent, per hour, it is likely to show 2
per cent, per hour if spread over a hearth area of 200 sq. ft.
Experiments as Compared with Practice. — Almost all the chloridizing
roasts made in preliminary tests, in a muffle, will show a higher loss by
volatilization than will subsequently be found in practice. There is no
appreciable loss in heap chloridization, when improperly roasted ore is
withdrawn from the furnace and the chloridizing allowed to proceed on
the cooling floor. Neither does any appreciable loss occur when the
damper of the furnace is closed, so that the furnace has no draft and no
fresh air supply.
Relation of Sulphur to the Chloridization of Silver and Gold. — Chlorid-
ization of the silver, in chloridizing roasting, may take place very rapidly
under proper conditions. If the^ ore contains an excess of sulphur,
chloridization will not take place to any appreciable extent, until some
of the sulphur has been eliminated, even if there are sulphates present.
This may be due to the reducing action of sulphur dioxide, or other reduc-
ing gases, which are likely to occur in abundance in the early stages of
the roasting.
' Engineering and Mining Journal, 1888-89.
CHLORIDIZING ROASTING
77
In order to determine the relation of the chloridization of the silver
and gold to the sulphur, and the progress of chloridization during the
roasting, the following interesting results were obtained by the author
on Tonapah concentrates, which consisted largely of silica and iron
pyrites, about 3 per cent, lead, and small quantities of zinc, copper,
manganese, and antimony. The raw ore had 16.35 per cent, sulphur, and
assayed 615.0 oz. silver, and 6.50 oz. gold per ton. The roasting was done
in a furnace having a hearth area of 100 sq. ft. The concentrates were
first given an oxidizing roast for two hours, after which 10 per cent,
salt was added and samples taken every hour.
Time,
chloridizing
roasting
Sulphur, per cent.
Roasted ore, ^ Hypo tails,
value ounces ! value ounces
Chloridization,
per cent.
Total I Soluble 'insoluble' Silver Gold ; Silver Gold Silver Gold
0 hours 1 16.35
1 hour 9.45
2 hours i 7.10
3 hours j 6.65
4 hours I 6.45
5 hours 6.25
8 hours 6.15
3.00
6.45
4.85
2.45
5.32
1.33
5.60
0.85
5.70
0.55
5.80
0.35
615.0
610.0
612.0
610.0
680.0
580.0
575.0
6.50
6.45
6.60
6.35
6.29
6.13
6.00
689.2
700.0
420.0
140.0
78.0
29.8
7.30
6.80
6.00
5.80
3.40
3.00
31.7
77.2
87.7
95.5
7.1
10.5
47.4
53.6
It will be noticed that no chloridization took place the first 2 hours of
chloridizing roasting, notwithstanding that there was from 3.0 to 4.65
per cent, soluble sulphur in the ore at that time. The same results in
chloridization would doubtless have been obtained if the salt had been
added three hours later than it was, or after the ore had been given an
oxidizing roast for 5 hours.
The high value of the hypo tails after 1 and 2 hours chloridizing
roasting, is due to the fact that there was no silver chloride formed, and
in leaching with the hypo the soluble matter was removed, thereby
somewhat concentrating the value of the ore.
The high soluble sulphur in the roasted ore is mostly due to the
sodium sulphate formed by the roasting. Some of the insoluble sulphur
may have been in the form of lead or calcium sulphate.
Determination of Loss by Volatilization. — In order to roast skillfully
it is of great importance to frequently ascertain the loss by volatilization,
but to do this it is necessary to know the loss of weight the ore sustains.
In practical handling of the ore this is difficult and inconvenient. Hof-
mann* gives the following method, which can be performed in an assay
office in a few hours :
"Ten grams of the raw pulp, containing the same percentage of salt
as the ore in the furnace, is placed in a roasting dish and roasted in the
' "Hydrometallurgy of Silver," Page 22.
78 HYDROMETALLURGY OF COPPER
muffle for half an hour or an hour; then the sample is removed from the
muffle, allowed to cool, weighed, returned to the muffle, roasted again for
half an hour, and then weighed again. This is repeated until two weigh-
ings are alike, or until in the last half hour the ore does not lose more than
2 or 3 mg., then the difference between the original weight and that of
the last weighing, expressed in percentage, gives the highest possible
loss the raw ore can suffer.
Ten grams of a sample of roasted ore, corresponding with the sample
of raw pulp, is placed in a roasting dish, and also roasted in the muffle
until two weighings agree, or the difference between two consecutive
weighings is not more than 2 or 3 mg. The difference between the first
weighing (10 grm.) and the last, expressed in percentage, gives the weight
which the roasted ore is still capable of losing if subjected to prolonged
roasting. If we deduct, therefore, the capable loss from the highest
possible loss, we obtain in percentage the loss in weight the ore has
suffered during roasting in the furnace by volatilization."
Chloridization Determination. — To determine the amount of silver
and gold chloridized, it will usually be sufficient for practical purposes to
take several ounces of an average sample of the ore, and treat it thor-
oughly with a solution of sodium hyposulphite. It will be found most
satisfactory to put the ore and hypo in a beaker for several hours at least,
stirring it occasionally, and then thoroughly filter and wash the ore in a
funnel. It is then dried, bucked, and assayed. If the ore contains large
quantities of soluble salts, the sample should be weighed before and after
the hypo treatment, and the difference allowed for in the results of the
Chloridization determinations are sometimes made by taking an
assay ton, or less, of the ore treating it with hypo, and assaying the
residue. While this rectifies any error of soluble salts, it introduces a
more or less uncertain element in the assaying.
CHAPTER V
PYROMETRY
Color Names of Temperatures. — The temperatures corresponding to
different colors have been determined quite accurately by White and
Taylor, by Howe, by Janivier, and by Pouillet. The difficulty in deter-
mining a certain temperature, by its corresponding color, lies in the
personal equation of the observer and the time and conditions of observa-
tion. Much depends on the susceptibility of the retina of the observer
to light as well as the degree of illumination under which the observation
is made. A furnace looks very much hotter at night than at day, and
hotter in a dark room than in a bright one. The most experienced
roasterman is unable to compensate fully for these factors, nevertheless,
the information given by these color temperatures is often convenient.
White aad Taylor
Name of color
Dark red, blood red, low red.
Dark cherry rod
Full cherry red
Light cherry, bright cherry,
bright red.
Orange
Light orange
Yellow
Light yellow
White
Temperature
C.
566
635
746
843
899
941
996
1,079
1,205
F.
1,050
1,176 J
1,375
1,550
1,650
1,725
1,825
1,975
2,200
Howe
Name of color
Temperature
C. F.
Lowest red -visible in the dark.
Lowest red visible in daylight.
Dull red
Full cherry .
Light red. . .
470
475
r550
[625
700
850
FuU yellow. . .
Light yellow.
White
950
1,050
1,150
878
887
1,022
1,157
1,296
1,562
1,742
1,922
2,102
(£. and M. J., Jan. 20, 1900.)
79
80
HYDROMETALLURGY OF COPPER
Pouillet
Name of color
Temperature
F.
Janivier
Name of color
Temperature
I
Incipient red
Dull red
Incipient cherry red
Cherry red
Clear cherry rod. . . .
Deep orange
Clear orange
White
Bright white
Dazzling white
525
700
800
900
1000
1100
1200
1200
1400
1500
to
1600
977
1292
1472
1652
1832
2012
2192
2372
2552
2732
to
2912
Very dull red ....
Dull red
Bright red. ...'...
Cherry red
Bright cherry red
Very deep orange
Deep orange red .
Orange red
Whitish
Brilliant white. . .
Dazzling white. . .
Blue white
525
700
800
900
1000
1050
1100
1200
1300
1400
1500
1600
977
1292
1472
1652
1832
1922
2012
2192
2372
2552
2732
2912
(E. and M. J., July 20, 1905.)
Pyrometric Determinations. — The only way of accurately determining
the temperatures in various parts of a roasting furnace, under all condi-
tions, is by the use of reliable pyrometers, and every roasting plant should
be equipped with at least one of these instruments.
Attempts have frequently been made to get uniformity in the quality
of ore, roasted for treatment by the chemical processes, by establishing a
system of absolute temperatures in certain parts of the furnace, and so
firing as to keep those temperatures constant. On theoretical grounds
this appears quite feasible. The difficulty lies in assuming that the ore
fed into the furnace is of uniform quality, and that the other essential
factors, such as air supply, always remain the same. Ore which is well
bedded, and containing about 2.5 per cent, sulphur, may vary as much as
0.35 per cent, to 0.50 per cent., in 24 hours. The conditions which
would be ideal for ore having 2.25 per cent, sulphur would be far from
ideal for ore having 2.50 or 2.75 per cent.
Much also depends on the physical and chemical composition of the
ore. The condition of temperature which would give the best results for
partly oxidized ore would not give satisfactory results with ore containing
the same amount of sulphur, from the deeper workings of the mine, in
which no oxidation had taken place, assuming of course, that the amount
of ore roasted remains the same. When ore contains an excessive amount
of dust, it cannot be roasted at the same rate and at the same temperature
as ore which contains only the normal quantity, and the dust is likely to
vary, especially when the supply bins get low.
The temperature of a roasting furnace appears to be very much
hotter at night than during the day. Inexperienced roastermen are
PYROMETRY ' 81
frequently misled by this, and even experienced men cannot judge
accurately within the desired limits. The tendency in the daytime,
especially in well lighted buildings, is to get the temperature too high,
and at night to get it too low. In such cases pyrometers are of
great service in establishing temperatures. They are also of great
service in determining the temperature beyond which it is unsafe to roast.
Experience and skill in the appearance of the ore as it progresses through
the furnace, and its appearance after roasting, however, are the best gen-
eral guides to obtaining uniform results. It is questionable whether,
even with a perfect system of pyrometry, the experience and skill of the
operator will not always remain the dominant factor.
It is a curious fact that when furnaces are overheated, the amount
of sulphur in the roasted ore is abnormally high. One of the dangers in
employing new roastermen is, that in their anxiety to get a good roast,
they invariably fire at too high a temperature, with the result, that the
roasted ore contains an unusually large amount of sulphur, partly fused,
and is in the worst possible condition for treatment by a solvent process.
Overheated partially roasted ore is also likely to run more or less like a
liquid and in this way emerge insufficiently roasted. Pyrometers, in
such cases, are invaluable as a warning to the roasterman when the safe
limit of temperature is being exceeded. Frequently pyrometric deter-
minations are essential to intelligent work, but they must he supple-
mented by experience and skill, and not dominate them.
Of the pyrometers in general use, those of the Le Chatelier type will
be found most satisfactory in roasting work. If the thermo-electric
couple is protected, it may be inserted into the furnace and kept there a
very long time — in fact more or less permanently — without appreciable
injury. The limit of temperature at which it is safe to use these pyrom-
eters is a little below the melting point of platinum, which is about 3250°
F., although readings above 3000° F. cannot be relied upon as perfectly
accurate. In roasting work these temperatures are never approached.
It is rarely that 1700° F. (927° C.) is exceeded. From 1400° F. to 1600°
F. (760 to 871° C.) is the usual range in the hottest part of the furnace,
for the various ores. Frequently ores are encountered which give the
best results at as low a temperature as 1000° F. (528° C). It will be
seen, therefore, that the only danger to the thermo-electric couple of the
pyrometer is from the furnace gases. Even this danger is remote in
any case, and is entirely obviated if the thermo-electric couple is
protected.
Several pyrometers inserted at various points of a large roasting fur-
nace will give invaluable information as to the limits of temperature,
which for any particular ore, will give the best extraction. Once these
extreme limits have been determined, it is an easy matter to fire the
furnace so that they shall not be exceeded.
82 HYDROMETALLURGY OF COPPER
While the determination of the absolute temperatures in roasting is
not essential, nevertheless it is highly desirable. One of the gratifying
features of the Le Chatelier type of pyrometer is, that absolute tempera-
tures may be determined with greater facility and accuracy than the
relative temperatures may be determined by other means, and absolute
temperatures are always reliable for comparison.
The elements of a Le Chatelier pyrometer consist
1. Of a thermo-electric couple, which generates, when heated, a
slight electric current, which is proportional to the heat applied.
2. A galvanometer so arranged that the deflection of the needle, due
to the current, indicates the temperature on the scale of the galvanometer.
3. Flexible wires connecting the thermo-electric couple with the
galvanometer.
In using the pyrometer, the thermo-electric couple may be inserted
directly into the furnace at the points where the temperature is desired,
and the reading taken. Such a proceeding is awkward and troublesome.
When only one galvanometer is used, the most satisfactory arrangement is
to permanently insert the thermo-electric couples at the various points
of the furnace, as desired, and by small switches and the wires connecting
the couples with the galvanometer, the temperatures of the furnace at
the different po'nts may be quickly determined. In the same way, the
temperatures of different furnaces may be readily ascertained. The
galvanometer should be located at a convenient point, away from the
dust and fumes of the furnace room. The thermo-electric couples may be
inserted into the furnace through the arch, but care must be taken not to
project them down far enough to be injured by the rabbles, although it
is desirable to get them as close to the ore as possible. The thermo-
couple, where it is intended to remain permanently in the furnace, should
be protected by porcelain tubes.
If a continuous record of the temperature at any one point is desired,
it is best to use a recording pyrometer. This consists essentially of:
1. A recorder, which is composed of a galvanometer and a clock
arrangement, so that a pencil indicates the temperature and time on a
moving chart.
2. The thermo-electric couple, the fire end of which is inserted into
the space, where the temperature is to be measured.
3. Flexible wires connecting the recorder with the thermo-couple.
With this apparatus, a continuous, automatic, and permanent record
of temperature and time may be made, which will give an accurate idea
of the firing of the roasterman during the entire shift. A comparison of
charts, will quickly establish the best temperature at which it is desirable
to roast any ore, and locate the responsibility of any defects in the ore due
to the temperature in roasting.
CHAPTER VI
ROASTING FURNACES
Roasting furnace design as applied to roasting ores preparatory to
treatment by the hydrometallurgical processes, is rapidly resolving
itself down to the various types of mechanical reverberatories. Hand
reverberatories are still in use in small reduction works, but even for
small output they are rapidly being displaced by the more efficient me-
chanical roasters. Labor, in hand roasting, has been the most impor-
tant factor in the cost of operation, especially in mining districts, where
labor is from $2.50 to $4.00 a day. Besides, the quality of labor is an
exceedingly variable factor, and as the quality of a roast depends much
on the efficiency, conscientiousness, and skill of the workmen, it is a
disturbing element in any metallurgical plant where hand roasting
is used.
The tendency, therefore, has been to eliminate the labor item, and the
personal factor, by the substitution of mechanical for. hand furnaces. A
mechanical furnace, once set in motion under the conditions which liy
experiment have been found to give the best results, will alwaj-s give
practically the same results if the same conditions are maintained, and
it is possible to keep the conditions practically constant. The personal
factor of the workmen, in mechanical furnaces, is largely though not
entirely eliminated. Maintaining proper conditions in a roasting furnace
requires skill, but as the extremely hard labor of hand rabbling the ore is
eliminated in mechanical furnaces, there is not the same temptation to
slight the work.
The output per shift is very much larger in a mechanical than in a
hand furnace, so that it makes it possible to pay the men better and get a
superior quality of labor. In mechanical furnaces, the principal factor
under the control of the roasterman, is the temperature of the furnace,
and even this may be made largely automatic by the judicious use of
pyrometers. The principal function of the roasterman, in large mechan-
ical roasters, is that of an overseer, and to regulate the temperatures as
indicated by the pyrometers.
Wonderful strides have been made in roasting and in mechanical
roasting furnaces in recent years. Roasting can no longer be considered
either difficult or expensive if fuel is available at a reasonable price.
Roasting costs of ten or twenty years ago, in hand furnaces, or even in
mechanical furnaces, are now obsolete, and the future will see still further
reductions.
83
84 HYDROMETALLURGY OF COPPER
The mechanical difficulties of roasting furnaces, prevailing some
years ago, which are peculiar to mechanism working under high tem-
peratures and in the presence of dust, have been practically overcome
in all of the successful mechanical furnaces now in use. It is not unusual
for a large 100-ton mechanical roaster to run from three to six months
without a single shut-down except, perhaps, the stopping of the mechan-
ism for some minutes to change the rabbles; but this does not interfere
with the daily output of roasted ore. It is not unusual for a furnace to
run from six months to a year without cooling for repairs.
Roasting, in mechanical furnaces, is better and cheaper than in any
type of hand furnace, whether the amount roasted is from 5 to 10 tons
a day, or from 100 to 200 tons, for a single furnace. The capacity of
200 tons has not yet been realized, nevertheless, there are no mechanical
or chemical difficulties to its realization. Roasting furnaces of that
capacity, in large works and on low grade sulphur ores, will soon be an
established fact, and will considerably reduce present costs of roasting.
It is not now unusual, on Cripple Creek ores containing from 1 to 3 per
cent, sulphur, to roast 125 tons a day in furnaces designed to roast, nor-
mally, 100 tons. Since it usually takes one man on a shift to attend to a
mechanical roaster, whether the capacity is 25, 50, or 100 tons a day, the
saving in the larger units is manifest, as well as the saving in fuel and other
items. If the furnaces are fired with oil, or with well designed centralized
gas producer plant, one man on a shift can attend to several furnaces,
irrespective of their size, with an extra man occasionally to assist in
attending the producers and changing the rabbles.
In small installations, the first cost of a mechanical furnace, over a
long reverberatory, is not usually a serious item. The difference need not
exceed $3000 to $3500 'and the cost of roasting can usually be reduced
from 50 to 75 per cent. A good mechanical furnace to roast, say, 10
tons of pyritic concentrates a day, or 25 to 30 tons of low sulphur silicious
ore, can be erected for about $6000. It will take three of the ordinary
long hand reverberatories to do the same Avork. Three such furnaces
would cost more than the mechanical furnace, and would require one man
on a shift for each furnace, making nine men in all; whereas the mechan-
ical furnace would require only one man on a shift, or three men in all.
As the amount of ore roasted per day becomes larger, the difference in
cost of roasting, between the mechanical and hand furnaces, becomes
more pronounced.
The variety of roasting furnaces, evolved and suggested, have been
numerous. The practice has all been toward greater simplicity. Of
the hand reverberatories, the multiple hearth, for roasting ore for sub-
sequent treatment by chemical processes, has become obsolete. The
"Long Reverberatory," the "Fortschauflungsofen" of the Germans,
1^0 A STING FURNACES Hr,
has supplanted all other types of hand furnaces, and proved itself the
survival of the fittest.
The furnaces to consider most seriously in the treatment of copper,
gold, and silver ores by the solvent processes, are:
Hand reverberatories,
Mechanical reverberatories.
Revolving cylindrical furnaces.
Muffle furnaces.
All of these types are in actual and successful use. Shaft furnaces,
like those of the Stetefeldt type, have long since become obsolete. There
are none now in operation, and it is questionable whether a roasting
furnace, based on that principle, can ever be devised which will success-
fully compete with the various types of mechanical reverberatories. The
chemical conditions of roasting in a shaft furnace are all that could be
desired, but the physical and mechanical difficulties are well-nigh
insurmountable.
Hand Reverberatories. — A hand reverberatory roasting furnace,
fired with solid fuel, consists essentially of a hearth, a fire-box, a bridge
separating the hearth from the fire-box, a reverberatory arch over the
hearth and which reverberates the heat and flame toward the hearth, a
flue, and means, such as exhauster or chimney, of acquiring a draft
through the furnace. If gas or oil are used as fuel, the fire-box and
bridge may be dispensed with, and the gas or oil injected through the
side walls or through the arch.
A hand reverberatory is one in which the ore is stirred and advanced
by hand labor; in a mechanical reverberatory the ore is stirred and
advanced by mechanical means.
The hand reverberatories may be subdivided into two general classes,
based essentially on the method of operation. These are:
1. Tlie short reverberatory, in which the ore is all charged, roasted,
and withdrawn, in successive complete operations.
2. The long reverberatory, in which the ore is charged at one end,
and then advanced by stages, while at each stage another charge is intro-
duced and one withdrawn. Several charges are, therefore, in the furnace
at the same time and each going through its cycle of treatment, inde-
pendent of the others.
Short Reverberatory. — This type of reverberatory is used only in
works where small quantities of ore are treated. These furnaces, while
cheap to construct, are expensive to operate. They labor under the
disadvantage that the conditions of the furnace itself change as the ore
progresses in the roasting operation; while in the long reverberatory the
conditions of the furnace remain practically the same all the time, but as
86
HYDROMETALLURGY OF COPPER
the ore progresses in the roasting it is advanced against the purer air and
more highly oxidizing atmosphere.
Figs. 1 and 2 show section and plan respectively of a short reverbera-
tory, which is the usual form and construction for furnaces of that size.
It has a hearth area of approximately 120 sq. ft., and is capable of roast-
ng a ton of ore at a charge. The number of charges that can be roasted
in_24 hours depends upon the ore; if the ore is silicious and low in sulphur,
three charges a day can be roasted; if the ore consists of pyritic concen-
trates, from one to two charges a day is about all that could be put
through. If the hearth is made longer than 12 ft., it is better to throw
the arch longitudinally over the hearth, instead of transversely, as shown.
In Figs. 1 and 2, A represents the hearth, B the reverberatory arch,
C the fire-box, D the bridge wall, E the flue holes leading from the rever-
beratory chamber to a small fine chamber before entering the stack,
^T^;--^/'
Fig. 1. — Short hand reverberatory. Longitudinal section.
F the charging hole through the arch from the top of the furnace, and
H a hole through the hearth for discharging the roasted ore through a
spout on to the cooling floor, or into a car or wheel-barrow, to be taken
to the cooling floor. The flues, E, are arranged so that the flame from
the fire can be equally distributed over the entire body of ore on the
hearth. The holes, K, in the outer wall, easily admit of the regulation of
the flue holes by means of bricks placed in the flues. The stack, also,
should be provided with a suitable damper.
It is not necessary to build the entire interior of the furnace of fire
brick. The fire-box, bridge, and arch immediately over them must be
built of fire brick; the rest may be built of any good common red brick,
preferably a pressed brick of the cheaper quality. Common pressed
brick for the hearth is very desirable; it is hard enough to withstand the
wear, and smooth and even enough to make the rabbling easier than it
ROASTING FURNACES
87
would be otherwise. The hearth brick should be set on edge, 4 in. thick,
and laid without mortar. The joints may afterward be filled in with
fine sand or tailings.
On account of the bridge being exposed to injury by the high tempera-
ture on one side, and by rabbling on the other, it is desirable that the top
course should be made of fire clay tile, say, 12 in. by 24 in. by 2 in. thick.
The grates are 12 to 16 in. below the top of the bridge, and the top of the
bridge is 8 to 10 in. above the hearth. Through the middle of the arch
is an opening of cast iron, F, with a well-fitted cast-iron cover, through
which the ore is charged into the furnace. The walls of the furnace
should be reasonably thick, in order to retain the heat as much as possible.
Fig. 2. — Short hand reverberatory. Plan.
and to prevent the furnace-room from getting uncomfortably hot. The
arch should be at least 8 in. thick, with brick set on edge. Another 4-in.
arch may be placed on top of this, if desired. Most of the heat radi-
ated from a furnace passes through the arch, so that the cost of the extra
thickness of arch is money well invested. For the arch, fire clay should be
used for the joints, and the joints are best made by dipping the brick
into the thin clay before setting them in place. If a 4-in. arch is placed
over the 8-in. arch, it may be made of brick or brickbats, laid in mortar.
The rabble doors are usually about 8 in. high and 14 to 16 in.^ wide. In
front of the door is an iron bar laid across it from projections on the
■casting, to facilitate the rabbling. The furnace is bound together by
1-in. rods, attached either to cast iron, or railroad iron buck staves.
Long Reverberatory. — The typical hand roasting furnace is the long
reverberatory, of which Figs. 3, 4, 5, 6, and 7 show a typical example.
It is the type of hand reverberatory almost universally used, and is
HYDROMETALLURGY OF COPPER
r ' I n // '
#x
a
^08*
m
or;,
r
J
>.'f)
\lr.
hU'
:^OfJi
ROASTING FURNACES
s<)
essentially the same as those used in smelting works. The length of the
furnace is largely governed by the character of the ore to be roasted. If
the ore is highly silicious, and contains only a small amount of sulphur,
there is no advantage in having the length more than 40 to 50 ft. If the
ore is highly pyritic, or consists of ordinary sulphide concentrates, a
length of 60 to 75 ft. will be found the most satisfactory. With furnaces
longer than 60 to 75 ft., it will be found difficult to ignite the ore at the
rear, and unless it is ignited, except for drying and heating, there is no
advantage in having the furnace much longer than the zone of ignition.
This zone, under any conditions, is considerably shorter in furnaces
roasting for the chemical processes than for smelting, because the initial
heat at the fire-box, and hearth adjoining, must always be kept below
the sintering point of the ore, whereas if the ore is roasted for smelting,
.K-16->K-
mtijjj uujj I u uj u ujjj u ti ijjjjmujjjiiiJJiJiJJiD ' jjjjj ouJiUUJij i ijiii': u
— 16->
I'lG. 5. — Long hand reverberatory. Transverse section.
the sintering is of no particular consequence. It will usually be found
advisable, therefore, if increased capacity is desired, to build two furnaces,
rather than to attempt to get it by increasing the length of the hearth,
or perhaps by the addition of a second fire-box, toward the rear.
The practical width of the furnace is controlled by the convenience of
working from both sides. From 14 to 16 ft. has been found by experi-
ence to give the best general results. When the width exceeds 16 ft., or a
reach of rabbling of about 8 ft., the labor of stirring the ore and advancing
it becomes tiresome for the workmen, and hence the quality of the roast
is likely to be defective.
The hearth is sometimes made continuous, in one plane, and some-
times with slight breaks of several inches, corresponding in length of
hearth to the amount of ore charged into the furnace at one time. The
object of the stepped hearth is to enable the roastermen to clearly dis-
tinguish the different charges, and keep them separate as they progress
through the furnace. As one charge is sufficiently roasted and withdrawn,
the next charge is moved forward to the position occupied by the previous
90
HYDROMETALLURGY OF COPPER
one, and a new charge introduced in the section at the flue end. It is
very desirable that mixing of the different charges should not occur. If
mixing occurs to any considerable extent, the insufficiently roasted ore
will contaminate that which is well roasted and serious difficulty in the
chemical treatment will be the result. If the hearth is not stepped, the
roastermen work to imaginary lines in the furnace to keep the charges
separate. Sometimes, in order to enable the workmen to more easily
advance the ore, the hearth is built with a gentle slope from rear to
front. The hearth should be built 4 in. thick with brick laid on edge
and without mortar. The ordinary quality of pressed brick make an
excellent hearth, and these are satisfactory for the arch also.
Fig. 6.
Fig. 7.
Figs. 6 and 7. — Long hand reverberatoiy. Details of construction.
The height of the arch above the hearth is dependent largely upon
the nature of the ore to be roasted. As a matter of fuel economy, the
lower the arch, the more flame and fuel gases will come directly in con-
tact with the ore, but the limit in this direction is governed by the condi-
tions of the furnace atmosphere. When roasting highly silicious ores,
the furnace gases will be highly oxidizing under almost any conditions,
and the height of the arch above the hearth, in such a case, is therefore
limited by other considerations, such as convenience in rabbling. If,
however, the ore is high in sulphur, as in roasting pyritic concentrates,
the fumes of sulphur dioxide from the ore and carbon dioxide from the
ROASTING FURNACES 91
fuel, would tend to make the furnace atmosphere reducing, instead of
highly oxidizing, and the object of roasting would be largely defeated.
With such ores, the reverberatory arch should be high so as to permit of
large volumes of air passing over the ore in order to keep the atmosphere
within the furnace as highly oxidizing as possible. From 20 to 36 in.
will usually be found to be within the practical limits of the height of the
crown of the arch above the hearth. The rise of the arch should not be
less than 3/4 to 1 in. for every horizontal foot of width. That is to say,
the least rise that an arch 16 ft. wide, should have to be safe, is from 12
to 16 in. Theoretically, a flat arch is the best, but there are practical
difficulties in constructing and maintaining a flat arch under the strenuous
conditions to which it would be subjected in furnace work. So far as the
roasting is concerned, the best results are obtained by having the arch as
nearly flat as possible, and the limit in this direction is governed by con-
structional difficulties. As to the best practical rise for the reverberatory
arch, much will depend on the quality of the brick; but the rule given
above will be found best under average conditions and conform with the
best practice for all types of reverberatories for roasting ores.
The method of constructing the arch is largely a matter of choice.
The brick should be set on edge, the 8-in. way. The arch may be built
of independent successive rings, as shown in Fig. 8, or be bonded so as to
make a continuous whole, as shown in Fig. 9. The method of independ-
ent rings has much to commend it. With this construction, every brick
Fig. 8. Fig. 9.
Figs. 8 and 9. — Reverberatory arch construction.
is under full compression, whereas in the bonded arch, if the brick or
joints are not of equal size, it is not possible to have them all under the
same compression, so that there is danger of the thinner ones dropping
out.
The reverberatory arch is usually built of ordinary straight brick,
with the difference in the thickness of the joint between the intrados and
extrados made up with clay. It is better, however, to build the arch of
straight brick and occasionally insert a row of wedge brick. In this way
the joints can all be kept of even thickness and the arch will have greater
stability. It will not be necessary to carry the fire brick in the arch
more than 10 or 20 ft. beyond the fire-box, in the roasting chamber.
92 HYDROMETALLURGY OF COPPER
The remainder of the arch, as also the side walls, may be built of common
brick.
The side and end walls should not be less than 16 in. thick, or two
courses of brick laid the 8-in. way. A little extra cost in thick walls and
reverberatory arch, will be more than compensated for in the saving of
fuel and comfort to the workmen.
The foundations are best and cheapest made of concrete. By putting
up the necessary side boards, and leveling the top edges, the foundation
can be quickly and cheaply laid in the best possible condition for the
superstructure. The concrete may be brought up to within a foot of the
hearth level, and this also will be found cheaper than brickwork and be
just as good. Brickwork, in mining districts, and especially in isolated
camps, is expensive, since all the material has to be hauled on the ground,
and brick-layers command high wages. With concrete, only the cement,
which is a small proportion of the whole, has to be supplied from without,
since rock and sand are usually available and common labor all that is
required.
The lower tie-rods should not be built in solid, but ducts should be
provided so that they can be removed or inserted at will, should the rods
at any time become disabled. Worn out iron piping, or common earthen-
ware pipes, are the best for this purpose, although ducts made of brick
will answer about as well.
The space between the walls, below the hearth, may be filled with rock
and earth or loam. It should be well tamped, so that there will be no
danger of the hearth settling when the furnace is in operation.
The rabble door frames should be set flush with the hearth and also
with the exterior walls. The buckstaves will then lap the j oint between
the exterior brickwork and the iron castings, and catch the face of the
channel which takes the thrust of the arch. The rabble door castings
are in this way securely fastened without bolts. Instead of the channel
as shown in the drawings, cast iron beams in the form of T's may be used
to support the arch, and which alternate with the rabble door castings.
The buckstaves may be made of old railroad rails, I-beams, cast T-irons
or two small channels secured together, back to back, with a separator to
permit the tie-rods to go between them. It is well to have the tie-rods
abundantly heavy to take the enormous horizontal thrust of the arch.
For the ordinary span of from 12 to 16 ft., 1 1/4-in. rods have been found
satisfactory. For the longitudinal rods 3/4-in. diameter will be large
enough.
The number of working doors should be sufficiently numerous to
permit of easy rabbling. A distance of about 6 ft., from center to center,
is satisfactory. When the doors are too far apart, rabbling of the ore in
the intermediate spaces becomes difficult, and may be neglected. Figs.
6 and 7 show the details of the working doors and the method of setting
ROASTING FURNACES 93
them. The details may vary somewhat according to the material most
convenient, and upon local conditions. Each door casting has attached
to it lugs which receive an iron bar about an inch square upon which the
rabble may slide while the ore is being rabbled. These bars are remov-
able, and sometimes it is more convenient to work without them. The
door castings have flanges on the sides so that they can be secured in
place by the buckstaves. Each door casting weighs about 150 lb.
When odd shaped bricks are used, it will be found cheaper and
better to have them especially made for the purpose if they cannot
be obtained in standard shapes. The cutting, and consequent breaking,
of a large number of brick will be more expensive than the extra cost of
special forms, and the work will not be as good. Usually the manufac-
turers of fire brick have special shapes enough to fill any want.
All brickwork about a furnace should be "shoved" and well grouted,
especially that part of the work which forms the skewback of the rever-
beratory arch. The spandrels of the arch may be filled in with brickbats
and mortar. The filling in of the spandrels will make the arch stronger
and also enable the furnace to better retain the heat.
After the furnace is finished, the buckstaves and tie-rods should be
put in place. The rods should be made reasonably tight, so that they
will vibrate when struck. It is best to take up any looseness in the arch
l)efore its weight is taken off the centers. The centers may then be lo-
moved, either before firing, or they may be burned out afterward, which-
ever is the most convenient and economical. The furnace should be
allowed to stand as long as possible before firing. When the fire is
started the furnace should be heated quite gradually for at least 24 hours,
after which there will be no harm in bringing it up to heat. It should be
fired long enough to get the hearth quite hot before charging the ore,
since a cold hearth greatly retards the roasting.
The furnace should be provided with several rabbles, 4 in. by 8 in.,
and 12 to 14 ft. long, and several paddles 8 in. by 12 in., 14 ft. long. The
handles are best made of strong wrought iron pipe to which the rabble
' and paddle blades are fastened, by welding.
The stack of the furnace may be built of brick, iron, or re-enforced con-
crete. For single furnaces an iron stack will usually be found the most
economical. The stack for a furnace as shown in Figs. 3, 4, and 5, should
be about 30 in. in diameter and from 60 to 75 ft. high. The stack, or
the flue leading to the stack, should be provided with a damper. The
position of the stack in reference to the furnace, is largely a matter of
convenience and local conditions. If the works have several furnaces,
they may all connect with a common stack. If there is only one furnace,
and no dust chamber is desired, the arrangement of the stack as shown
in Figs. 3 and 4 will be as satisfactory as any. Provision should be made
for a car track to bring the ore to the hopper over the rear of the furnace.
94 HYDROMETALLURGY OF COPPER
Hand reverberatories should not make more than 3/4 to 1 1/2 per
cent. dust. Whether this dust is worth recovering by building a large
dust chamber will depend largely on the value of the ore. Small dust
chambers are not very effective. It will ordinarily be found that the
recovery of the dust is a matter worth careful consideration.
The cost of building a long reverberatory, as described and shown,
will be between $3000 and 14000. It will take about 50,000 common
brick, 8000 fire brick, and 20,000 lb. of iron. If concrete is largely used,
a great saving may be effected in the number of brick, and a saving also
in the cost of the furnace.
Method of Operating a Long Hand Reverberatory. — When the furnace
is hot, the ore, which may be assumed to be pyritic concentrates, is
charged from a car into the hopper, and into the furnace, in the section
nearest the flue, or at the rear. It is then spread out evenly over this
section of the hearth. The weight of the charge, for the best work, should
not exceed 10 to 15 lb. per square foot of hearth area; 10 lb. is better than
15 lb. if the ore contains very much sulphur. In any event, there is no
advantage in using a deep bed of ore, for what is gained in the amount of
ore charged, is lost by a correspondingly increased time of roasting. The
depth of the charge will usually be from 2 to 3 in.
The working doors are all closed until the sulphur is well ignited. The
moisture is first driven off, after which the ore will soon become slightly
incandescent and the sulphur begin to burn with a blue flame. This is
one of the most delicate stages of the roasting, and should be done at the
lowest permissible temperature and in the presence of a maximum
. amount of air. The charge should be rabbled energetically, and with the
intervals between the rabbling as short as possible. At this stage the ore
will be very unstable and is likely to run somewhat like a liquid. If the
rabbling is neglected, or if the temperature of the furnace is too high,
partial fusing or matting is likely to occur, which forms lumps that are
difficult to eliminate; and if not eliminated, will result in improperly and
insufficiently roasted ore. Such ore will be highly detrimental in the
subsequent chemical treatment. The excess of air required during this
stage of the roasting may be obtained by keeping the working doors open,
and by free admission of air through the bridge wall. The air should be
introduced as much as possible through the bridge, since this will tend to
keep the charge nearest the fire from becoming too hot, by interposing a
layer of cooler air between the ore and the flame, and give a highly oxidiz-
ing atmosphere at the surface of the ore where it is most desired.
When the sulphur flame has abated, which will be in about 8 hours
after charging, the ore is moved forward into section No. 2, and a new
charge introduced into section No. 1. The ore on section No. 2 is spread
out over a large area to give it as much surface as possible. There is
still much sulphur in the ore, and most of the oxidation takes place in the
ROASTING FURNACES 95
middle section. The ore being brought closer to the fire, is brought to a
dull red heat. During this stage the ore swells somewhat, and becomes
more or less inert. As the sulphur is eliminated, the ore has no power of
generating heat within itself and hence the fire is urged, to keep the ore
at the desired temperature. The rabbling in this section need not be as
frequent as in section No. 1; a thorough stirring every 15 or 20 minutes
will suffice. The ore should be uniform throughout, and as it is turned
over, the newly exposed incandescent surface should quickly turn dark
and not show any live sparks of burning sulphur.
The charge, after reaching this stage, which will usually be about 16
hours after it has been put into the furnace, is transferred to section No. 3;
the ore in No. 1 is advanced to section No. 2, and a new charge introduced
into section No. 1. There is now no danger of lumps forming in section
No. 3, and the temperature may be raised somewhat, but must never
approach the sintering point. The temperature permissible in this
section is the controlling factor in firing the furnace. The firing should
always be done with a view of throwing as much heat and flame as possible
to the rear of the furnace without danger of sintering the ore on the
finishing section. In this section, the roasting will largely consist of
decomposing the soluble sulphates, and while this is going on the odor of
sulphur dioxide can be detected when a sample of the incandescent ore
is removed for inspection. As the ore becomes more nearly completely
roasted, it becomes more coherent, and remains as placed by the rabble.
"Sweet" or "Dead" roasting are more or less indefinite terms, used
to denote the condition of the ore when all the sulphur has been elimi-
nated. But as the elimination of all the sulphur is practically impossible,
and as the approximation thereto is a very indefinite matter varying with
the different ores, it may be taken to mean ore sufficiently roasted to give
the best results in the subsequent chemical treatment.
After the ore has remained on the finishing section for about 7 hours,
the roasting is completed. It is then withdrawn through the holes in the
hearth, near the last door, into a pit, or into a car and taken to the cooling
floor. The charge in section No. 2 is then moved forward to No. 3, and
the charge in section No. 1 moved forward to No. 2, while a new charge is
introduced into section No. 1. In this way a charge is withdrawn and a
new one added every day, so that there are always three separate
charges in the furnace, and each charge remains in the furnace almost
24 hours.
The fuel used in roasting should be either wood or long flame coal.
Oil gives better results than either wood or coal, but is not usually avail-
able. It is best, in order to get a long flame, with almost any coal, to
fire the furnace fire-box more or less as a gas producer. This is easily
arranged by keeping a deep bed of ash and fuel on the grates, and intro-
ducing steam and air through the closed ash pit. Much of the air needed
96 HYDROMETALLURGY OF COPPER
to completely consume the fuel gases may be introduced through the
bridge and some through the working doors. The draft is regulated by
the damper in the flue or stack, and by opening and closing the working
doors. The time of roasting depends largely on the ore, but somewhat
also on the amount of rabbling. Roasting, as already stated,. is essen-
tially an oxidizing process, so that any operation, such as continuous and
energetic rabbling, which will expose the greatest amount of ore to the
highly oxidizing furnace atmosphere, will materially reduce the time of
operation.
It is customary when roasting pyritic gold concentrates, or ore con-
taining lime, or silver and copper in appreciable quantities, to add* a small
amount of salt, usually just before drawing the charge. The amount of
salt may vary from 0.5 to 5 per cent. If the ore is not sensitive to vola-
tilization, the salt may be added to the ore as it is advanced from the
middle to the finishing section and thus become thoroughly incorporated
with it. Usually, however, the salt is added about 30 minutes before the
charge is withdrawn, and thoroughly mixed with the ore. Chloridization,
under proper conditions, takes place rather quickly, and as explained
under " Chloridizing Roasting" air is not essential to the chloridization.
The salt may therefore be added a short time before discharging the ore,
and by permitting the ore to cool slowly after it is discharged, the neces-
sary degree of chloridization can well be realized without any appreciable
loss by volatilization. This gives the ore a thorough oxidizing roast
before chloridization, and the sulphur is never so thoroughly eliminated
but that there are always enough sulphates left to sufficiently chloridize
the silver and small amounts of copper.
After the salt is added, the ore begins to fume, increase in bulk, and
has a "woolly" appearance. After the salt is added the temperature
should be kept low — not over a dull red heat. Much of the gold is
chloridized as well as the silver; careful tests have shown it to be from 10
to 20 per cent, of the gold contained in the ore.
If the ore contains galena, great care must be exercised in the first
stages of the roasting to keep the charge at the lowest practicable tem-
perature, as the lead sulphide fuses at a very low heat, and agglomeration
in the early stages of the roasting will make the subsequent work more
difficult.
Furnaces having considerable of a drop between the different sections
of the hearth have been recomended and built, but have not come into
general use. The cause of this is evident. While theoretically the
showering of ore through this drop, as in a shaft furnace, appears good, it
is evident that the draft in the furnace will whip the dust along with it
and cause excessive loss in that way. The great dust loss is not com-
pensated for by the small gain in the time of roasting.
Most furnaces, like the one illustrated in Figs. 3, 4, and 5, are de-
ROASTING FURNACES 97
signed to take a charge of from 3 to 3 1/2 tons on a section. It is usual,
however, instead of charging and withdrawing this amount of ore all at
once, to still furthur subdivide it so that each shift of eight hours will
charge and withdraw one third of this amount or from 2000 to 2400 lb.
The advantage of this is that the quality and quantity of the roasted
material can be checked up for the different shifts, and the furnace can
be worked with greater regularity, than when so much ore is charged and
withdrawn at the same time.
One man on a shift, working three shifts, will roast from 3 to 3 1/2
tons of pyritic concentrates a day. Each shift will draw a charge of
from 2000 to 2400 lb. and introduce one. The amount of fuel used is 1/2
cord of wood per ton of concentrates.
The following tabulated statement gives the essential facts of roast-
ing pyritic concentrates in California:'
ANALYSES OF CONCENTRATES
Eureka and Idaho
mines, Graas Valley
0.85
0.78
0.02743
0.0068
0.00
40.63
trace
32.80
12.64
0.10
3.50
8.65
Washington mine,
Mariposa County
0.00
1.50
0.00914
0.0035
1.34
30.85
0.00
31.33
33.30
0.00
0.00
1.67
Black Bear mine,
Klamath County
0.00
Lead
0.00
Gold
Silver
0.0137
0.003
0.00
42.05
31.25
25.10
10.35
0.85
0.85
Oxygen and loss by difference . .
0.38
These analyses give a very good idea of the composition of California
pyritic concentrates, which have been treated for many years by roasting
and chlorination. The size of some of the furnaces, the time of working
on a shift, and the quantity of ore treated, is given in the following table:
Name of works
Size of furnace
Time of working (shift) Quantity of ore in 24 hours
Heywood's
Zeile
Amador. . .
Plymoth...
Maltman's.
Merri field's
60X12
75X12
80X12
50X12
60X12
70X10
' Eggleston,
7
'Met. Silver, Gold, Mer.
8 hours
8 hours
8 hours
8 hours
8 hours
12 hours
3 tons
3 tons
4 1/2 tons
3 tons
2 1/2 tons
3 tons
98
HYDBOMETALLURGY OF COPPER
>iX:
Ph
>fO;
ROASTING FURNACES
99
The hearth of these furnaces lasted from four to six years.
Cost of Roasting in Long Reverberatories.— The cost of roasting in
long reverberatories is quite large. For pyritic concentrates, which is
about the only material roasted in these furnaces, it is about $4.25 per
ton, distributed as follows:
Roasterman, 1 shift, roasting one ton, $2 . 50
Fuel, 1/2 cord of wood, for one shift, 1 . 50
Other expenses, 0 25
$4.25
The cost of roasting silicious ore, low in sulphur, is very much less.
Such ore can be roasted about as rapidly as it can be worked through the
furnace.
In California' at one of the mills where oil was substituted for wood,
it was found that the capacity of the furnace was increased from 4 tons
to 6 tons per day. The furnace was 14 ft. wide and 75 ft. long. Bakers-
field crude oil, of 14 to 16° gravity, was used. In roasting 2647 1/4 tons
of pyritic concentrates, 1290 barrels of Bakersfield crude oil was used
which cost, delivered, $1917.63 or 48/100 barrel per ton of ore, cost-
ing 72 cents. There was also used 66.76 tons of coal to generate steam
for pumping, heating, and atomizing the oil, which cost delivered,
1867.84, or 35 cents per ton of ore treated; making a total cost of fuel for
roasting, of $2785.47 or $1.05 per ton.
Fig. 12. — Modified long hand reverberatory. Transverse section.
Modified Long Reverberatories. — An important modification of the
long reverberatory, especially for copper roasting, is shown in Figs. 10, 11,
and 12. Fig. 10 shows a longitudinal section; Fig. 11 the plan, and
Fig. 12 a transverse section through the front of the furnace at the pro-
tecting arch and discharge openings in the hearth.
One of the difiiculties in hand reverberatories in roasting ore for the
hydrometallurgical processes is the liability to fuse the ore near the fire
' E. C. Vorhies, Scientific and Mining Press, March 26, 1904.
100
HYDROMETALLURGY OF COPPER
end, when the fire is urged sufficiently to ignite the charge at the rear.
In chloridizing roasting a similar difficulty presents itself. In order to
eliminate the sulphur sufficiently in hearth section No. 2, by oxidizing
roasting, a temperature nearly as high as the ore will stand without fusing
is desirable to expedite the process as much as possible. To accomplish
this the ore in section No. 3 may become unduly heated, even if it does
Fig. 13. — Modified hand reverberatory. Longitudinal section.
not approach the sintering point. The temperature for chloridizing
roasting should be the lowest at which the reactions take place, and this
condition conflicts with that required for efficient roasting in the middle
and rear end of the furnace.
To overcome these difficulties, a protecting arch is thrown over
section No. 3, as shown in Figs. 10 and 12, which shields the ore from the
Fig. 14. — Modified hand reverberatory. Plan.
direct action of the heat and flame. In this way the ore in the middle
section may be kept the hottest, and that in the- rear section may readily
be ignited. The flame and heat from the fire-box pass through the
space between the protecting arch and the reverberatory arch, so that
the ore in the finishing section is heated only indirectly, as in a muffle.
The protecting arch should be as thin as possible, consistent with good
construction.
COASTING FURNACES
101
In the chloridizing roasting of copper ores, this modified furnace has
been used in preference to the ordinary long reverberatory. It can
readily be seen, that for chloridizing work especially, this modification of
the long reverberatory offers many advantages. The furnace may be
still further modified by returning the flues under the hearth, and thus
heat the ore from below. In a long furnace this is a doubtful utility.
Much will depend on the temperature of the furnace gases.
Figs. 13, 14, 15, and 16 show a modification of the hand reverber-
atory, at one time used in Europe for roasting copper ores.
V///////,
Fig. 15. Fig. 16.
Figs. 15 and 16. — Modified hand reverberatory. Transverse' sections through A-B
and C-D (Fig. 13).
Mechanical Reverberatories. — ^Practically all the roasting at the
present time, in preparing ores for treatment by wet methods, is done in
mechanically operated reverberatories. With the exception of the
rabbling mechanism, these furnaces are not essentially different from the
hand reverberatories.
The mechanical reverberatories differ mostly from each other in the
rabbling mechanism. They may have one long roasting hearth, or
several hearths superimposed above one another.
The single hearth furnaces have the advantage of making less dust
than the multiple hearth roasters; while, on the other hand, the multiple
hearth furnaces are more compact and better conserve the heat.
In roasting ores for chemical treatment, where a more or less com-
plete roast is necessary, single hearth furnaces have usually been pre-
ferred, largely on account of the low dust loss. Multiple hearth furnaces
have also been largely used, and their use is becoming more general.
In roasting pyritic concentrates or heavy sulphide ore down to 6 or
8 per cent, sulphur, which may be done without fuel, the multiple hearth
roasters have found more favor than the single hearth furnaces.
Recently the multiple hearth furnaces have been modified by the
102 HYDROMETALLURGY OF COPPER
addition of fire-boxes for the different hearths, so that the sulphur can
be eliminated quite as completely as in single hearth furnaces.
The advantages of the multiple hearth furnace is that the heat from
the ore and fire-boxes in the lower hearths, heats the bottom of the upper
hearths, and this heat is very effectively applied. The disadvantage is
that when ores are roasted with external fuel, the volume of air necessary
for the combustion of both fuel and the sulphur in the ore is very large,
and the ore dropping from one hearth to another against a strong current
of ascending gases, makes considerable dust, much of which is lost
unless efficient means is provided for its recovery. This difficulty may
to some extent be obviated by providing separate flues for the ore and
gases between the hearths, or by exhausting the gases from one or
more of the intermediate hearths as well as from the upper hearth.
In roasting heavy sulphide material down to 6 or 8 per cent, sulphur,
fuel is not ordinarily necessary, hence the volume of air and gases going
through the furnace is comparatively small and its rate of passage com-
paratively slow, hence the dust loss in dropping the ore from one hearth
to the next need, not be serious or excessive. In roasting ores for treat-
ment by wet methods, the furnace operates under much higher tem-
peratures than in roasting for smelting, where the sulphur most difficult
to eliminate, remains in the ore. Usually furnaces, for sweet roasting,
work under a temperature of from 1200 to 1700° F. The ore itself may
not be at these temperatures, but the reverberatory chamber in which
the rabbling mechanism operates, is. The temperature of the rever-
beratory chamber, at the fire-boxes may even exceed 1700° F. in regular
roasting work, and for that reason, if anything goes wrong with the
rabbling mechanism, it is necessary to at once cool the furnace or the
top layer of ore would take the same temperature as the reverberatory
gases, become sintered, and be unfit for extraction by hydrometallur-
gical processes.
The problem, therefore, in mechanically roasting ores, has been to
provide a rabbling mechanism which will work in grit and dust and at a
temperature varying from 1200 to 1700° F., without serious delays or
excessive repairs. The different mechanical roasting furnaces are
based fundamentally, on the method of overcoming these difficulties.
O'Harra dragged the rabbles through a straight line reverberatory
by means of a chain and track within the reverberatory chamber. Brown
conceived the idea of placing a supplemental chamber on both sides of
the reverberatory hearth, containing the tracks on which run the rabble
trucks, shielded more or less from the direct heat and dust of the rever-
beratory chamber. Holthoff-Wethey placed the trucks entirely outside
of the furnace and devised a slot arrangement which opened and closed
automatically as the rabble moved along. Pearce supported the rab-
bling mechanism by a column in the central open space of two concen-
ROASTING FURNACES 103
trie walls forming the reverberatory chamber, the inside wall of which
is slotted and the reverberatory arch supported from above. Edwards
and Merton have a longitudinal series of rabbles of a diameter correspond-
ing to the width of the hearth, projecting through the reverberator^'
arch, and as the ore is advanced by one of these rabbles, it is delivered
to the next, and so on until it issues from the furnace. The multiple
hearth furnaces are mostly based on the support and protection of the
central column carrying the rabbles for the different hearths and the
arrangement for cooling and replacing either the rabble arms or the
blades.
Any single hearth reverberatory may be constructed with multiple
hearths, but when so modified, it is questionable whether they are as
efficient and present the same advantages as the regular circular multiple
hearth furnace of the McDougall type.
For furnaces where the rabbles are alternately within and without the
roasting chamber, as in the O'Harra, Brown, and Holthoff-Wethey, no
special cooling provision is necessary, since the rabbles never get danger-
ously hot, and never, except in case of accident, do the rabbles have the
same temperature as the reverberatory chamber. When the rabbling
mechanism remains in the furnace indefinitely, as in the Pearce, Edwards,
Merton, McDougall, Herreshoff, and Wedge, water cooling is necessary or
desirable. Air cooling, for sweet roasting, has not given satisfactory
results.
There is no great difference in the cost of installation of the various
mechanical reverberatories, and on the same ore, for a thorough roast for
wet processes, there is no great difference in the cost of operation.
In roasting Cripple Creek ore, for example, there are three 100-ton
Pearce roasters in one reduction works; six 100-ton Holthoff-Wethey at
another, and eight 100-ton Edwards at still another, all in successful and
satisfactory operation, roasting in all about 1500 tons of ore daily and
having a precious metal content of approximately $30,000.
Cost of Mechanical Reverberatories. — The average cost of a good
mechanical reverberatory is about 115 per square foot of hearth area for
the smaller sizes, and about $12 per square foot of hearth area for the
larger sizes, installed, ready to operate, but not housed.
Fuel Required in Roasting. — The fuel consumption in wood or coal
will usually be from 10 to 15 per cent, of the weight of the ore roasted,
and will be more or less independent of the original sulphur content. For
ore low in sulphur, considerable fuel is required to bring it to the roasting
temperature and to maintain it at that temperature. For ore high in sul-
phur, the sulphur itself develops considerable heat, so that extreme firing
is not necessary except to remove the last few per cent, of sulphur. On
sulphide concentrates or heavy sulphide ore, the sulphur content in the
roasted material may be reduced to 6 or 8 per cent, without any fuel, and
104 HYDROMETALLURGY OF COPPER
to 4 and 6 per cent, with only one fire-bo-x at the finishing end of the
furnace. It is in the elimination of the last few per cent, of the sulphur
that most of the fuel is consumed. But if the fuel consumption is more
or less independent of the sulphur content of the ore, the capacity of the
furnace is more or less proportion to the contained sulphur.
Hearth Area Required in Roasting Various Ores.— The capacity of a
roasting furnace is dependent on the amount of sulphur in the raw ore,
and on the amount of sulphur to be eliminated. For roasting ores
suitable for the hydrometallurgical processes, the hearth areas required
are approximately as follows:
For silicious ore containing from 1.5 to 3.5 per cent, sulphur will
require from 10 to 15 sq. ft. of hearth area per ton of ore roasted per
day; ores containing from 10 to 15 per cent, will require a hearth area
of from 25 to 30 sq. ft., and pyritic concentrates which carry from 35 to
45 per cent, sulphur will require from 40 to 50 sq. ft.
Pyritic concentrates and heavy sulphide ores, carrying 28 per cent,
or more of sulphur are self roasting down to 6 or 8 per cent. After that
fuel has to be used to complete the roast, to make the ore suitable for
a solvent process.
The Brown Furnace. — The latest and most approved type of Brown
furnace is shown in Figs. 17 and 18, which is a straight line reverberatory,
in which the ore is stirred and advanced by rabbles mounted on trucks
attached to endless chains, one on each side of the furnace, moving in
a supplemental chamber. Brown was the first to work out a successful
method of protecting the rabbling mechanism in a straight line furnace
from the heat and dust of the reverberatory chamber.
The Brown furnace has had various modifications, but in its most ap-
proved form it is a single hearth reverberatory, with supplemental chambers
on each side of the roasting chamber, to protect the rabbling mechanism.
The supplemental chambers are formed as shown in Fig. 18 by a tile
projecting up from the hearth, above the level of the ore, and a corres-
ponding tile, forming part of the reverberatory arch, projecting down
from above. This construction forms a supplemental chamber and leaves
a slot between the roasting and supplemental chambers just large enough
for the rabble arm to pass through. The rabble arm, at the slot, is usu-
ally made rather wide and about an inch thick so that the slot may be
as narrow as possible.
The rabbles, extending from one side of the furnace to the other,
are mounted on trucks on each side, and these trucks run on tracks in
the supplemental chambers. The trucks are attached to endless
chains which move about sprocket wheels at each end of the furnace.
One pair of these sprockets is driven by means of spurs and gears, which
in turn are actuated by a counter shaft driven by belt and pulley.
The rabbles in passing through the furnace stir and advance the
ROASTING FURNACES
105
I. .'-i^
^ g
Ph
pq
^
106 HYDROMETALLURGY OF COPPER
ore, and issue quite hot; then elevated by the sprocket wheels at the other
end of the furnace to a track above, and returned to the feed end where
they are again lowered by the driving sprockets and enter the reverbera-
tory chamber to again complete the circuit. While returning, outside
of the reverberatory chamber, the rabbles are cooled enough so that no
special cooling device is necessary. As each rabble enters the furnace
it takes with it from the feed bin the proportionate amount of ore
required to make the daily output. This may be regulated by the feeding
device or by the depth of raw ore in the path of the rabble as it enters
the furnace.
Counterweighted sheet iron doors at both ends of the furnace,
hinged at the top, keep out the cold air; they remain closed except when
lifted by the rabbles in passing in and out. When the doors are opened,
even momentarily, a strong inward draft is likely to set in; to avoid this,
two doors are sometimes inserted at each end a short distance apart so
as to form a neutral air chamber; one door being always closed while
the other is opened by the moving rabble.
The skewbacks of the arch are steel channels supported by the
buckstaves and short cast iron columns, the spaces between the columns
being 3 ft. 6 in. long by 12 in. high. These openings extend the entire
length of the hearth; they are closed by sheet-iron doors, lined with
asbestos. This construction permits of ready access to the hearth at any
point for repairs and observation.
From the ground to the hearth the furnace may conveniently be
built of concrete or uncut stone and the remainder of the furnace con-
structed of brick. The furnace is bound together by steel I-beams which
also carry the tracks on the top of the furnace.
The furnace is regularly made with a roasting hearth 10 ft. wide, and
in lengths varying from 60 to 200 ft., or even longer. The roasted ore
may be elevated to the top of the furnace to a sheet iron hearth supported
by the steel I-beams, and there cooled and advanced by the returning
rabbles, or the roasting hearth may be somewhat extended and used as
a cooler, unless it is preferred to cool the ore independently of the furnace
mechanism.
The frequency of stirring the ore may be regulated both by the speed
of travel and the number of rabbles on the moving chains.
The size of the sprockets is governed by the vertical distance between
the roasting and cooling hearths. In driving the chain mechanism both
spockets at the drive end are made tight to the shaft, while at the other
end one of the sprockets is tight to the shaft and the other loose, so that
any uneven strain in the chain is self adjusting and prevents the chain
from riding the sprockets and being thrown off.
The material for a standard straight line Brown furnace 10 ft. wide
by 140 ft. long, actually erected, was as follows:
ROASTING FURNACES 107
Weight of ironwork, 63,000 lb.
Weight of tiles, 35,000 lb.
The Brick, etc., required for this furnace are as follows:
For Hearth
46,200 common red brick,
420 partition tiles,
14 cu. yd. of sand,
16,800 lb. of ground fire clay,
11 barrels of cement,
126 cu. yd. of stone for concrete work.
Mortar for Hearth; 5 cu. yd. of sand and 42 bu. of lime.
For Arch
1260 common red brick,
1344 skewback brick.
For Five Fire-Boxes (one double and three single fire-boxes)
15,000 common red brick,
10,000 common fire brick,
509 arch tiles 12 in. long,
115 arch tiles 6 in. long,
10 fire-door tiles.
For Mortar; 2 1/2 cu.yd. of sand and 25 bu. of lime.
Power. — The Power required to drive the furnace mechanism, etc.
was supplied by a 9 in. by 12 in. slide valve engine, which is rated at
about 25 h. p. but the engine was never taxed to its capacity at any time.
Grate Area. — The total area of the grates in the five fire-boxes is
about 63 sq. ft.
In supplying the above material 5 per cent, extra was allowed for
tile and skewback brick, and 10 per cent, on the remainder of the items.
H. 0. Hof man' gives the following data on roasting with Brown furnaces
with a hearth 8 ft. wide and 135 ft. long :
Silicious ore with pyrite, crushed to 30 mesh and containing 2.5 per
cent, sulphur was roasted at the rate of 95 tons per 24 hours with 6 cords
of wood; sulphur in roasted ore, 0.3 per cent.; roasted ore per square
foot of hearth area, 131 lb.; ratio of hearth to grate area 32 to 1; 15.83
tons of ore were roasted per cord of wood.
Silicious ore with pyrite crushed to 20 mesh, containing 2.3 per cent,
sulphur, was roasted at the rate of 65 tons in 24 hours, with 6 cords of
wood; sulphur in roasted ore, 0.5 per cent.; ore roasted per square foot
of hearth area, 90 lb.; ratio of hearth to grate area, 32 to 1; 10.85 tons of
ore were roasted per cord of wood.
' "The MetaUurgy of Lead," p. 185.
108
HYDROMETALLURGY OF COPPER
r Air Passage at sides of
I-BeumB U> Cooler Hcaitb
Coal II upper
-H^i<^
Figs. 19 and 20. — Plan and section, Pearce furnace.
With combined roasting and cooling hearth.
ROASTING FURNACES 109
Silicious ore with pyrite crushed to 16 mesh and containing 2.5 per
cent, sulphur was roasted at the rate of 50 tons per 24 hours with 4.5
tons of refuse slack coal; sulphur in roasted ore 0.5 per cent.; ore roasted
per square foot of hearth area, 77.0 lb.; ratio of hearth to grate area
28.66 to 1; per cent, of fuel on ore, 8.5.
Copper matte, crushed to 40 mesh and containing 20 per cent, sulphur
and 40 per cent, copper, and 15 per cent, lead, was roasted at the rate
of 20 tons per day; sulphur in roasted ore, 6.0 per cent, the aim of the
roast being to peroxidize the iron and convert as much as possible of
the copper into soluble sulphate; ore roasted per square foot of hearth
area, 38 lb.; ratio of hearth grate area, 35 to 1; coal consumed 3.25 tons;
per cent, of fuel on ore, 16.
The Pearce Furnace. — ^The Pearce furnace. Figs. 19 and 20, is of the
circular hearth type. The hearth is formed by two concentric walls,
usually about 10 ft. apart. The operating mechanism is at the center,
from which the rabble arms radiate.
The ore is fed into the furnace from a hopper, to a tapering screw
located beneath the hearth, which distributes and raises the ore across its
width, so that each rabble blade takes its proportionate share as it comes
along. The rabble blades are attached to the horizontal pipe rabble arms,
which in turn, are attached to a rigid iron framework radiating from a
hollow hub, at the center. The hub revolves on ball bearings, around
a stationary cast iron column.
The rabble arm is made of two concentric pipes, the smaller one being
fitted into the larger, and having holes at the end. These rabbles are
continuously water cooled by a gravity system. The water from the
main is run into a trough located above the level of the rabbles and re-
volving with the rabbling mechanism. It then flows by gravity into
the inner concentric pipe of the rabble arm to the further end, where it
is delivered to the rabble pipe, which is exposed to the heat of the fur-
nace, and flows back again to the other end of the rabble arm, from
whence it is exhausted into a stationary circular trough about the hub,
■near the floor. A complete and continuous circulation is kept up in
this way. Air was at first used in the rabble arms, but for oxidizing roast-
ing, satisfactory for wet methods, it had to be abandoned.
The inner circular wall of the furnace has a continuous slot for the
passage of the rabble arms. It is made reasonably tight by travelling
. steel shields, counterweighted so as to press gently against the walls
forming the slot. The upper part of the wall, above the slot, rests on a
casting suspended from I-beams and cross-beams, and braced by radial
struts and angle irons. The I-beams are supported by the other wall
and central column, and the cross-beams by the I-beams. The outer
wall is 18 in. thick.
The rabble blades, which are made of 3/8-in. sheet steel, are graduated
no HYDROMETALLUHGY OF COPPER
in length and direction from the inner to the outer circle, so that the ore
on the outer circle, which has to travel a greater distance, may be at
the same height and remain in line with that near the inner circle. In
other words, the rabbles are so proportioned that all the ore, through a
cross-section of the furnace, travels uniformly through the furnace and
is discharged at the same time, notwithstanding that the ore in the outer
diameter has considerably further to travel.
In a furnace having an outside diameter of 60 ft., the smallest blade
is 3/4X6X3/8 in., and the largest blade is 8X6X3/8 in., and the inter-
mediate blades are proportioned to these dimensions. There are six
rabble arms, and each rabble arm has 20 blades, for a hearth 10 ft. wide.
The rabble arms are fastened to the radial struts by a clamp which
permits the raising or lowering of the rabble arm, so that the bed of ore
may be kept level by raising or lowering the further end of the rabble.
If the rabble is not properly adjusted, the ore may pile up on one side
of the furnace, instead of having a uniform thickness.
The depth of the ore varies from 2 1/2 to 3 1/2 in., depending on the
amount being roasted; 2 1/2 in. is the depth when roasting from 80 to
90 tons per day, and 3 1/2 in. when roasting 100 tons or more in a
furnace having an external diameter of 60 ft. Such a furnace has six
rabble arms, which make three complete revolutions in 5 minutes, or
one revolution in 13/4 minutes. The ore is therefore rabbled every
17 seconds. The life of the rabble arm, which is a heavy 4-in. pipe, is
one and one-half years. The life of the 3/8-in. sheet steel blades is three
months in roasting low sulphur silicious ore, and from five to six weeks
in roasting 30 to 40 per cent, sulphide ore. The blades are changed
without cooling the furnace, and without interfering with the daily
output of roasted ore. The rabble arms may be changed by cooling the
furnace somewhat.
Two to three per cent, sulphur ore remains in the furnace from
2 1/2 to 3 hours, and on the cooling hearth, 1 hour.
The height from the top of the ore to the spring of the arch is 16 in.,
and from the spring to the crown of the arch, 12 in.
The Pearce furnaces, for sweet oxidizing roasting, are usually built
with one roasting hearth above and a cooling hearth below.
The following gives a summary of data on the largest size Pearce
roasters Figs. 19 and 20, when roasting ores low in sulphur.
Outside diameter, 60 ft.
Width of hearth, 10 ft.
Average length of hearth, 168 ft.
Mean diameter of hearth, 53 1/2 ft.
Total hearth area, 1689 sq. ft. Available for roasting 1500 sq. ft.
Number of fireplaces, 4. Or 3 fireplaces and one oil burner.
ROASTING FURNACES 111
Grate area of each fire place, 4 ft. by 4 ft. 8 in.
Numlier of rabble arms, 6.
Number of blades on each rabble arm, 20.
Angle of blades 22 1/2 degrees.
Depth of ore, from 2 to 3 1/2 in.
Rabbles make one revolution in 100 seconds.
Ore stirred every 17 seconds.
Capacity per 24 hours, roasting 2.5 to 3.0 per cent, sulphur ore
down to 0.5 or 0.75 per cent., 100 tons.
Power required to drive furnace, 6 h. p.
Fuel required, 10 tons good bituminous coal.
The Holthoff-Wethey Furnace.— The Holthoff-Wethey furnace,
shown in Figs. 21 and 22, is regularly constructed with a roasting hearth
above and a cooling hearth below. The roasting chamber is supported
on structural steel, arranged so that there is a space between the side
walls of the furnace and the supporting posts. In this space, on both
sides, and attached to the posts are the tracks for the rabble trucks,
running longitudinally with the furnace. The rabble trucks are mounted
on chains, which are moved, raised, and lowered by sprockets at both
ends of the furnace.
The power for driving the mechanism is applied to the shaft at one
end of the roaster on which one pair of sprockets are mounted. The
driving mechanism, rabble trucks, chains, and tracks are all located
entirely outside of the roasting chamber and at all times exposed to the
air. Theslot through which the rabble arms pass are opened and closed
automatically by tripping doors, or flexible steel sheathing, as the rabble
progresses through the furnace. One half of the number of rabble blades
of each rabble are set at one angle and half at an opposite angle, thus
removing all end thrust.
The reverberatory arch is firmly held in place between two h«avy
I-beams, suspended from above and from beams resting on the channel
iron posts. Opposite posts are securely joined together and take the
end thrust of the arch.
The ore is fed evenly into the furnace at the drive end, and after
travelling the full length of the roasting chamber, is dropped to the cool-
ing hearth and again carried back to the charging end, thus allowing the
ore the same time to cool that was required to roast it. The cooling
hearth may be provided with water pipes or water jackets to help cool
the ore, but this, on account of the length of the cooling hearth, will
usually not be necessary.
The furnaces are usually built from 10 to 12 ft. wide and from 100 to
130 ft. long. The ordinary dimensions are 12 X 120 ft. Such a furnace,
for roasting ores low in sulphur, is equipped with eight rabbles, which
HYDROMETALLURGY OF COPPER
Ph
o
"o
a
ROASTING FURNACES
113
make a complete revolution in 4 1/2 minutes, thus stirring the ore every
33.5 seconds. The upper sprocket shaft, at the drive end, is counter-
weighted so as to keep the rabble chains taut.
The iron work for a 10 X 100 ft. Holthoff-Wethey furnace, with roast-
ing and cooling hearths, weighs approximately 130,000 lb. Such a
furnace usually has four fire-boxes and a grate area of 60 sq. ft.
In the erection of this size furnace there were required:
68,000 common brick,
4,000 fire brick.
It was driven by a 1-5 h. p. motor.
r'^
<3 ■'
r.V^iWi.iViit.tmiivyfn
>■ . jj
/I. .nil i Jiin.n nm nm ^mmMMmmi ^h
" 'T^"'|' — "1 r" T'"':~rv~r' f— r ~ r~'f^'^.~ i 1-1 -ft i—--^ 2: nrMT""
^^rv*«-.v
~|— — r
Fig. 22. — Holthoff-Welhey furnace. Transverse .section.
This size furnace erected at Dalonega, Georgia, roasted concentrates
containing over 15 per cent, sulphur down to 0.15 or 0.20, all but a trace
of which was as soluble sulphates. The capacity was 35 to 40 tons per
day of 24 hours. The ore contained 10 to 12 per cent, moisture when
charged into the furnace. The fuel used was 12 cords of wood, which
for 35 tons per day would equal 0.34 cord per ton. At Colorado City
six Holthoff-Wethey furnaces, 12 ft. wide and 120 ft. long have been in
operation for many years roasting Cripple Creek ores for chlorination,
with a capacity of 100 per furnace per day.
The Merton Furnace.— The Merton furnace. Figs. 23 and 24, is a
rectangular multiple hearth furnace in which the rabbling is done by
vertical shafts arranged in a line, and passing through the respective
hearths. To these shafts are attached rabble arms of a radius equal to
114
HYDBOMETALLURGY OF COPPER
ROASTING FURNACES 11.5
one-half the width of the furnace. The vertical shafts are a little
further apart than the radius of the rabble arms, so that their areas of
revolution intersect, and the ore is delivered from one rabble to the
next, through the different hearths.
Each shaft is set in motion by a worm gear or by spur and pinion,
and arranged to revolve at the desired speed — usually from one and
one-half to three times per minute. A rabble arm is attached to each
shaft for each hearth. In the adjacent hearths the rabbles on each shaft
are placed at right angles to one another. Three of the rabble arms in the
same hearth are made to point in the same direction, while the fourth,
or end rabble, is at right angles to them.
The finishing hearth has a special rabble to itself which is larger than
those working on the other hearths; it may be water cooled, but for ores
not requiring a high finishing temperature the water cooling is often
dispensed with.
The entire furnace is enclosed in 1/4-in. plates, well braced with
buckstaves and rods.
For a standard type furnace the total length is 32 ft. 6 in. The main
body of the furnace measures 23 ft. 9 in. long inside the plates, 8 ft. wide,
and the height, with three hearths, is 6 ft.
The hearths are horizontal, and are not given any inclination what-
ever. The height from the hearth to the crown of the arch, inside
measurements, is 16 1/2 in., and 9 in. at the sides. The thickness of the
crown of the arch is 4 1/2 in. Each of the hearths has a door opposite
each shaft. At the end of each hearth is a slot connecting with the
hearth below.
In roasting, the ore is fed in at one end of the upper hearth, and
passed from one rabble to the next until it reaches the other end, where,
discharging through the slot, the ore is delivered into the next hearth
below, and is carried by another series of rabbles to the opposite end,
again it is discharged to the third hearth and travels along it to the
finishing hearth. Near the end of the third hearth is another slot in which
is a gate permitting the control of the discharge to what is termed the
finishing hearth. In this hearth the roast is completed.
The weight of the iron work of this furnace is about 12 tons. About
10,000 ordinary brick and 3000 fire bricks are required in its construction.
About 2 h.p. is required to drive the mechanism.
The Edwards Furnace. — ^In the Edwards furnace. Figs. 25, 26 and
27, the ore is advanced by a series of revolving rabbles of relatively
small diameter and of intersecting areas, so that the ore from the first
rabble is delivered to the next, and so on through the series until it
issues as the roasted product.
The furnace is regularly made in two general types; one having a
single row of rabbles, known as the "Simplex," and the other having a
116
HYDROMETALLURGY OF COPPER
double row of rabbles, known as the "Duplex." These two types are
built either "Tilting" or with "Fixed Hearth." In the tilting con-
struction there is a double continuous girder, balanced on a center sup-
port. The girder supports the iron work of the rabbling mechanism
and the drive therefor. The furnace is adjusted vertically at the end
Fig. 25. — Edwards furnace. Perspective view.
s'l'/J-J* — 3'iV,6- ->^
Fig. 26. — Edwards furnace. Transverse section.
to give the hearth any desired slope. The iron construction is encased
in brick. The fixed hearth construction is the same as in standard
reverberatory practice.
The tilting furnace is a straight single-hearth reverberatory, 63 ft.
long by 9 ft. wide over all, and 58 ft. by 6 ft. 5 in. in the clear, which is
ROASTING' FURNACES
117
118 HYDROMETALLURGY OF COPPER
designed foi' sweet roasting of- concentrates or sulphide ores. The shell
of the furnace is a rectangular chamber of plate-iron stiffened with angle
iron; this is lined inside with brick. Good common red brick are found
suitable for ordinary temperatures, although at the fire end fire bricks
may be used. 10,000 bricks are required for the lining of this furnace.
On top of the furnace is mounted gearing for driving the rabbles,
the spindles of which extend down through the cast-iron boxes, built
into the brickwork of the reverberatory arch. The whole furnace rests
on two pivots, one for each side. These pivots are arranged near
the middle of the length of the furnace, so that by tilting the fur-
nace from the horizontal to various angles, the discharge of the ore may
be regulated according to the rate at which it is being roasted. The
mechanism for tilting the furnace is simple, and can be worked without
altering the usual speed of the rabbles. The weight of the whole iron
work including the shell, bracing, shafting, gearing, rabbles, etc., is 19
tons. No single part of the furnace exceeds 2500 lb. in weight, while
most of the parts do not exceed 225 lb.
The bottom of the furnace is supported on No. 14 corrugated steel;
directly on this is placed a thin layer of non-conducting material, and on
this is laid the brick floor.
There are 15 rabbles placed side by side along the length of the furnace,
the blades of which nearly touch the hearth. Each rabble has a circular
path, the circumference of which almost touches the brickwork on either
side, and as the distance between the rabbles is a little greater than the
radius of the circle described, each rabble works ore almost up to the heel
of its neighbor, and as each rabble rotates in the opposite direction to the
one next to it — ^they are alternately right- and left-hand rabbles — ^the
ore is not passed along too rapidly from one end of the furnace to the
other, but gets a thorough stirring and exposure to the air as it
proceeds on its course.
The speed of the first 13 rabbles from the feed end is one revolution
per minute, while the fourteenth has two, and the fifteenth, or discharge
rabble, four revolutions per minute. All the rabbles being driven from
the same shafting, the change in speed is arranged by alternating the
ratio of the bevel wheels. The speed of the fourteenth and fifteenth
rabbles is so arranged that while they stir into the areas of the adjacent
rabbles, they do not come in contact with them. The object of this
increased speed is to give the ore a brisker stirring in the final stages of
the roast.
In order to protect the rabbles in the hottest part of the furnace from
the destructive action of the high temperature, water is circulated
through the last three to keep them cool; by this means they are found
to last for years.
There are two different kinds of rabbles used; one is solid and flat-
ROASTING FURNACES 119
footed, the front edge of which is beveled; the other is hollow; on the
arm of it cast-iron shoes are fitted. The latter rabble is the one used
at the fire end of the furnace. The cast-iron shoes can be slid on or off
the rabl)lc arm without lowering the heat of the furnace or removing the
rabble from it. When roasting concentrates or ores high in sulphur,
the first 10 rabbles, counting from the feed end, are generally of the
solid, flat-footed type; these pass through the ore close to the hearth
and effectively stir and expose the particles of ore so long as they carry
a fair percentage of sulphur. The last five rabbles are provided with
cast-iron shoes, as by the time the ore reaches this part of the furnace it
has lost most of its sulphur and is less lively. These shoes pass through
and under the ore.
When roasting ores that do not contain much sulphur, shoes are
used on all the rabbles, although water need only be circulated through
those subjected to the greatest heat. The ordinary flat-footed rabble
arm is fastened to the spindle by placing the end in a socket and passing
a pin through both. The water rabbles have a 3 1/8 in. cast-iron hollow
spindle with a flange at the bottom, which is bolted on to a correspond-
ing flange on the upper part of the arm.
The ore is conveyed by an automatic feeder from the hopper into the
hearth at the upper end of the furnace; after traveling to the lower end,
near the fire, the ore is discharged down a pipe, located near one of the
sides; the bottom of this pipe passes through and works in a case leading
to the conveyor, to prevent any escape of dust. This conveyor pushes
the ore into the pit.
During the roasting air is admitted through the holes, situated above
the fire bars. The fumes pass into the main fine through a short flue,
which is attached to the furnace. To allow for the movement of this
short flue when the angle of the furnace is altered, the hole in the main
flue through which it passes is made larger than actually required for a
nice fit; in order to cover the space left between the short and main flues,
and prevent cold air from passing into the latter, a sliding cover-plate
moves in a frame, which is bolted onto the brickwork of the main flue.
The power required to work the furnace is from 2 to 3 h. p.
The Duplex furnace is a stationary structure, designed for large
capacity. The principle of rabbling is similar to that adopted in the
tilting furnace, but consists of two lines of rabbles driven from two
horizontal line shafts, and the walls can be built of brick or concrete.
The concrete can be brought to within two layers of brick to the fire
zone. The brackets carrying all the mechanical superstructure are
firmly fastened to anchor bolts in the furnace walls, and angle stays
and cross bars are so arranged as to make the entire superstructure firm
and substantial. Buckstays and tie rods hold the whole of the arch and
120 HYDROMETALLURGY OF COPPER
brickwork together with straps on the face of the walls to give solidity
to the skewbacks to protect the arch.
This furnace when complete and ready for operating consists of
100,000 lb. of iron and steel and 70,000 bricks. Fire bricks are only
used when in close proximity to the fire-box.
The hearth area of the standard Duplex type, 112 ft. long by 13 ft.
wide gives 1456 sq. ft. of working area, and the outside measurement
overall is 116 ft. by 16 ft.
The usual fall of the hearth in this furnace is 1/2 in. to 1 ft., and the
mechanism driving the furnace can be so arranged to drive the rotating
rabbles at various speeds so that the roasting material can be under the
control of the roasterman.
The furnace requires from 6 to 9 h. p. to operate, according to the
speed at which the rabbles are driven.
The water required to cool the rabbles is 400 gallons per minute.
The McDougal Furnace. — The McDougal Furnace,^ has long been
used to roast copper ore and concentrates for smelting, and has been
modified by Herreshoff, Evans, Klepetko, Wedge, and others to adapt
it to modern requirements. This furnace, so successful in imperfectly
roasting copper ores for smelting, has recently been improved to adapt
it to the more thorough roasting required for the hydrometallurgical
processes.
The McDougal furnace. Figs. 28 and 29, is essentially a multiple hearth
upright cylinder with central shaft carrying the radial rabble arms. The
hearths are horizontal arches having discharge openings alternately at the
center of one hearth and the periphery of the next. The central revolving
shaft is provided with radial rabble arms for the different hearths, and
the rabble blades are so arranged at an angle with each arm that for the
odd numbered hearths they push the ore toward the center, and on the
even numbered hearths toward the periphery. In doing so the ore is
turned over and over by the rabbles and describes a spiral coiirse around
the shaft. The ore, in its descent from hearth to hearth, describes a
zig-zag course through the furnace from top to bottom, passing alter-
nately through the holes at the center and at the periphery. !>;
The size of the McDougal furnaces, as largely used at Butte, is 16 ft.
in diameter, and 18 ft. 3 1/2 in. high. It is sheathed with 3/8-in. boiler
iron and lined with a full course of red brick. It has six arched hearths
with a 9-in. spring and 3 ft. apart; each hearth has two rabble arms
making one revolution per minute. Each furnace has two exhaust
flues 24 in. in diameter and 12 ft. apart, passing out of the roof, and
flues from three furnaces lead to one main 6 ft. in diameter, having
openings along the top and bottom for removing the flue dust.
'H. O. Hofman, Trans. Am. Inst. Mng. Eng., Vol. XXXIV; L. S. Austin, Tras.,
Vol. XXXVII; Peters, "Practice of Copper Smelting."
HO A S TING F UR XA( 'I'JS
121
The cciitial shaft of tlic fui'nace is driven from below. The cooling
water for the rabbles is forced down to near the bottom of the revolving
hollow shaft, which is 9 in. in diameter, through a 3-in. pipe and out to
the ends of the horizontal rabble arms through 1-in. horizontal pipes.
In its upward passage between shaft and pipe it takes up the return
Fig. 28. — AUis-Chalmers McDougal furnace (fire-box type).
section.
Perspective view and
water from the rabble arms and discharges at the top through two
spouts into a secondary launder. Shafts and arms are made up of
flanged sections to permit of easy exchange. Running the overflow
water at 80° F. 20 gallons of cooling water are required per minute to
cool the rabbles.
122
HYDROMETALLURGY OF COPPER
The two rabble arms of a hearth have seven and eight cast-iron
blades; these are 8 in. long and 6 in. wide, and 5/8 in. thick; the lower
12 in. of the blade which comes in contact with the ore are chilled. The
blades on the top hearth last from 25 to 35 days; those on the sixth, from
6 to 8 months.
The six circular flat arches which form the roof of one hearth and the
floor of the next above, require care in construction. There must be a
large central opening for the main vertical shaft which carries the
rabbles, and the brick at the periphery must be well anchored. There
must also be drop holes from each hearth to the next, and these are
K 7—^iA: M
FiQ. 29. — Elevation and section. McDougall roaster.
arranged alternately at the central opening and at the extreme periphery
of the hearths and protected by iron castings. The first, third and
fifth hearths have one drop hole at the center; the second and fourth
have six and the sixth has two drop holes near the periphery.
The hearths with peripheral openings are provided with a central cast-
iron ring, cut in halves. This ring circles the shaft, leaving an annular
clearance space of 3 in. The brick of the hearth butts against the ex-
terior of this ring in its entire circumference, and is keyed into a groove
in the ring. The center drop hole is formed by stopping the brickwork of
the hearth so as to leave an annular space of 16 in. encircling the shaft.
The peripheral drop holes are 14 in. wide on the first and fifth hearths
and 18 in. wide on the third hearth, where there is a strong evolution of
sulphur gas.
In roasting at the Washoe smelter at Butte, the moist concentrates
ROASTING FURNACES 123
are dumped into the feed hopper on top of the furnace, which holds 33
tons. The composition of the material is:
Moisture, 8.1 percent.
Cy, 7.42 per cent.
SiOj, 21.2 percent.
1^6, 26 . 1 per cent.
S, 33.2 percent.
"■'2O3) 2.7 percent.
CaO, 0.3 percent.
98 . 94 per cent.
To this is added 5 per cent, of limestone, of which the diameter of
the largest piece does not exceed 1 in.
The ore is fed contjnously into the furnace and is spread on top of
the hearth to the thickness of 3 in. by the rabbles. The ore dropping
down through the holes in the hearths showers through the ascending
air, which actively roasts it, but at the same time this air current
carries away the finer particles as dust, which amounts from 4 to 5
per cent, of the ore charged. The gases escaping from the upper
hearth have a temperature of 315° C; by the time they reach the flue
chamber they have cooled to 117° C.
The appearance of the roasting at the different hearths is as follows:
On the first hearth the ore is dropped at the circumference and, containing
6 to 10 per cent, moisture, is drying out, but attains no visible heat.
Entering the second hearth it still looks dark, but shows a blue flame
by the time it reaches the borders of the hearth, where it is 600° C. On
the third hearth some sparks show where the rabble passes, together
with blue flame, and with a flame temperature of 900° C. On the fourth
hearth the sparking has ceased, the ore having attained an orange-red
heat. In falling upon this hearth from the one above, the ore as it showers
down burns freely, hastening the roasting by this momentary, but thorough
exposure to the ascending tir. On the fifth hearth the sulphur is elimi-
nated sufficiently so that the discharge temperature is less than the enter-
ing temperature; that is, the ore is brighter near the periphery than at
the center. On this hearth, the maximum temperature of 960° C. is
attained. On the sixth and final hearth the heat has become uniform,
but is lowered to 860° C. As the ore leaves the hearth it seems brighter,
but speedily cools to 660° C. as it falls, smoking freely, into the hopper.
Fig. 30 shows diagramatically the progress of reactions and temperatures.
The composition of the escaping gases is:
By weight By volume
per cent. per cent.
SOj, 4.95 2.25
SO,, 1.46 0.53
0, 19.60 18.45
N, 75.00 78.77
124
HYDROMETALLURGY OF COPPER
Thirty-two pounds of air is needed per pound of sulphur and since
13.3 lb. of the latter is burned off per minute, there is needed in that time
6384 cu. ft., which passes up the central hearth openings at the rate of
8.8 ft. per second. A screen analysis of the flue dust shows:
On 10 mesh screen, 9 . 7 per cent.
Between 10 and 30, 25 . 3 per cent.
Between 30 and 80, 30 . 7 per cent.
Passing 80, 33.4 per cent.
99 . 1 per cent.
The ore takes about 2 hours and 15 minutes to pass through the
furnace.
In starting up a furnace, a small fire of dry, soft, long flame wood is
started from the three side doors of the third and fifth hearths. A new
1st Hearth
Snd Hearth
8rd Hearth
4tb Hearth
5th Hearth
0^
Percentage ol Contained Sulphur
20^ 40^ 60»
IMfi
1
1
1
f
1
1
1
1
1
1
1
1
'•'•«
''i^;
1
'^^
/
f
/
in
as
•
\
s
'5
1 ®
\
\
1
/
/
1
1
1
/
1
!
6th Hearth .
« c 201) 40U eoo eoa looo c
Fig. 30. — Progress of reactions and flame-temperature in the McDougall roaster.
furnace is brought to a darlc red heat in 3 to 4 days, an old
furnace requires only 2 days. Concentrates are then fed. After charg-
ing for 5 or 6 hours, it sometimes happens that the furnace cools down
too much, and this makes it necessary to start on the third and fifth
floors a new fire for 1.5 to 2 hours; occasionallyfeeding of the ore is stopped
and half a ton coal charged. Under normal conditions a furnace does its
best work when the flue shows a depression in water of 0.3 in. If it is
ROASTING FURNACES 125
less the furnace gets cool. The temperature may be regulated by the
admission of air; closing the bottom doors drives up the heat, opening
the doors draws it down; opening doors higher up checks the draught.
The rate of feed once settled upon is usually not altered, and the number
of revolutions the rabbles make per hour remains the same.
A section of six furnaces is tended to in 8-hour shifts by one-third
foreman, one furnaceman, one helper, one-sixth oiler, and one-ninth
repair man and one trimmer.
The dust which collects in the flues connecting the furnaces forms
4 to 5 per cent, on the ore, is raked out every day. The loss in weight,
including the flue dust, is about 20 per cent.
A furnace treats under normal conditions, 40 tons of sulphide ore,
with 35 per cent, sulphur, and 10 per cent, copper, or 0.042 tons per
square foot of hearth area, reducing the sulphur to 7 per cent. Roasted
ore with 14 per cent, copper treated in the same manner, retains about
10 per cent, sulphur. The product can of course be varied with the speed
of travel of the rabbles, and the sulphur more thoroughly eliminated by
the addition of fireplaces.
The following partial analysis of roasted ore represents two determina-
tions from the average day and night samples taken during an experi-
mental run of 15 days. SiOj, 26.9 per cent.; Cu, 18. 3 per cent, of which
9.9 per cent, was present as CuO; Fe, 30.0 per cent, of which 17.9 was
present as FeO; S, 9.3 per cent, of which 0.81 was present as SO3.
At Butte, in the regular roasting of concentrates the results are:
Amount roasted in 24 hours, 40 tons.
Sulphur in raw concentrates, 35 per cent.
Sulphur in calcines, 7 per cent.
Hearth area, 952 sq. ft.
Concentrates roasted per square foot of hearth area, 84 pounds.
Coal, none
Cost of roasting, 35 cents.
The large sizes of the standard McDougal roasters have an outside
diameter of 18 ft. 5 in., containing 6 hearths, with an enclosed fire-box
under the sixth hearth. The weight of the entire iron work for such
a furnace is about 90,000 lb. There are required for its construction
about 37,000 red brick and 500 fire brick.
At the Washoe smelter of the Anaconda Copper Mining Co., 64
McDougal furnaces are in operation; these have a height of 18 ft. 3 in.,
and an outside diameter of 16 ft. They are enclosed in boiler iron shell
3/8 in. thick and are lined with a full course of red brick, leaving an
inside hearth diameter of 14 ft. 6 in.
At Garfield, Utah, where there are 24 McDougals in operation* the
ore and concentrate mixtures were generally such that it was not neces-
sary to roast below 10 or 11 per cent, sulphur. At the same time 30
' R. R. Moore, E. and M. J., May 14, 1910.
126 HYDROMETALLURGY OF COPPER
per cent, of the total charge was added on the fifth hearth, which
gave an average of 55 tons per day at a cost of 22 cents per ton.
Other averages of over 50 tons per furnace day were maintained for
six months at a cost of less than 25 cents per ton. The concentrates
added on the fifth hearth of the McDougals were high in copper.
They were added there on account of the fineness and tendency to
produce excessive amounts of flue dust. At Garfield the McDougals
produce about 6 per cent, flue dust. This flue dust carries more silica
and sulphur and less copper than the charge. Notwithstanding the
elaborate system of flues constructed at the Garfield plant there was
a stack loss of about 500 lb. of copper per day from these roasters.
The McDougal furnaces are regularly built in the "Self -roasting"
and "Enclosed Fire-box" type. If a more thorough roast is desired
in the self-roasting type for hydrometallurgical work than can be obtained
without fuel, satisfactory results are obtained by firing with oil, in which
case the oil is injected into the various hearths, as desired. If solid
fuel is used, it is desirable to use the enclosed fire-box type. This type
has two grates at the bottom, each having an area of 29 sq. ft. or a total
area of 58 sq. ft. of grate surface to each furnace. In one furnace
of the enclosed fire-box type, partly muffled, for roasting pyritic ore con-
taining 45 per cent, sulphur and reducing it to an average of 2 per cent, in
the calcines, there is obtained a capacity of 14.4 tons per day of 24 hours.
The Herreshoff Furnace. — In this furnace. Fig. 31, when used in
roasting pyrites for sulphuric acid manufacture, or for the preliminary
roasting of sulphide ores for metallurgical treatment, the rabbles are cooled
with air, through the central column. This column in a double ver-
tical hollow shaft. Attached to this shaft are one or more arms at each
hearth, and the replacable rabbles are slipped on these arms. The air
for cooling the rabbles is forced into the bottom of the column and then
delivered through the central part of the shaft, from which it passes in
multiple at once to all the arms. After cooling the arms it returns to the
annular space between the inner and outer shaft, and finally discharges
at the top of the outer shaft.
The temperature of the iron in the shaft and arms is kept above the
condensing point of acid to prevent corrosion, and at a point where the
strength of the metal is the greatest. The rabbles are made in sections.
There are five sections on each arm of from one to five blades per section,
depending on their position on the arm. The sections can be slipped
on or off the arms and any blade can be taken out of the section and
replaced without disturbing the rest. The hearths are made of special
moulded arch fire brick.
The shaft is driven from the bottom by means of a cast iron gear and
pinion, and makes one revolution in from 70 to 150 seconds, depending
on the kind of roast. The six-hearth furnace, 15 ft. 9 3/4 in. diameter,
ROASTING FURNACES
127
requires about 1 h. p. Speed reductions are made by gear, reducing
worm gear, or sprocket, as desired. A shear pin is provided in the driving
mechanism which acts as a safty device in case of undue strain. The
following table gives dimensions and data for some of the standard
size furnaces.
Fig. 31. — Herreshoff furnace.
HERRESHOFF FURNACE
Outside diameter
Number of
hearths
Hearth area
square feet
Weight, metal
parts
Weight, special
fire brick
Pounds, sulphur
per 24 hours
11 ft. 7 1/2 in
lift. 7 1/2 in....
115 ft. 9 3/4 in...
20 ft
5
7
6
5
7
381
547
912
1,308
1,810
16,000 lb.
25,000 lb.
43,000 lb.
68,000 lb.
82,000 lb.
16,000 lb.
32,000 lb.
79,000 lb.
132,000 lb.
168,000 lb.
3,000 to 6,000
4,500 to 12,000
8,000 to 21,000
12,000 to 30,000
20 ft
16,000 to 42,000
128
HYDROMETALLURGY OF COPPER
in the above tables the capacities given in pounds of sulphur per
24 hours must be used to form a general idea only, as the chemical com-
position and physical character of each ore, together with the kind of
roast required, will have to be determined for each particular case.
The Wedge Furnace.— The Wedge furnace, Figs. 32 and 33, has for
many years been successfully used in the east for chloridizing roasting
"Jlf^'^'^-^B^v.^'oflifflu)!!' 1
Ki,_,>i»....,
JH^: iiininmniHUp |
j
': ^.Vi"
^«J|^R
^^H^B
<
1
- IP***'
//
j
'-
nam
u
i, ,SP
JS^^^
ii
Fig. 32. — Wedge furnace.
copper ores, and its use is rapidly being extended into the field of oxidizing
roasting. The Wedge furnace is of the McDougal type and built in var-
ious diameters, with one, three, five, or seven hearths, as may be required.
The top of the furnace is used as a dryer, and a bottom hearth, below the
roasting hearths, may be used as a cooler.
The ore or concentrate is fed to the top of the furnace at the peri-
phery, and is mechanically fed across the top, entering the center of the
ROASTING FURNACES 129
furnace dry and hot. The feed entering the furnace is so arranged that
the material forms a lute, making the furnace gas-tight at this point.
One of the most distinctive features of the furnace is the central
shaft whicli is hollow, open at the top and bottom, 4 ft. in diameter,
and is covered with tile which are attached to and revolve with the
shaft. The advantage of the large hollow shaft is that an arm can be
changed easily and without losing heat in the furnace. The arms are
held in position by breech blocks placed upon the inside of the central
shaft; this makes it possible for workmen to enter the central shaft while
the heat is in the furnace, and remove the breech block, when workmen
1^^
^^^^^^^^^^P^^^- ^^^^^1
S^S^^^F
■|F;::,^;/'.
*wS
mt
E^
^fei^-.--
f.. . . _
liMJ
■
Fig. 33. — Rabble details, Wedge furnace.
on the exterior of the furnace will withdraw the wornout arm through
one of the doors and insert a new arm, when the workman on the inside
of the shaft will replace the breech block. All parts in connection with
the rabbling mechanism are interchangeable.
The arms are built for either water cooling or air cooling. Each arm
has its own supply pipe and discharge pipe, so that it is possible for the
furnace operator to know at all times that each arm is receiving its
proper supply of water.
The hearths are all level. This is made possible by building them
of specially shaped fire brick. These are pressed brick and the arches
are laid up dry. The result is that falling arches are eliminated. In
furnaces which have been in operation for eight years, the arches are still
in good condition without the expenditure of one dollar for repairs.
The weight of the central shaft and arms, including the arms on the
dryer hearth, is carried on roller bearings. This reduces the power
required to operate the furnace to a minimum. The indicated power
on a large 21 ft. 6 in. diameter furnace, with seven hearths, is less than
2h. p.
The furnace is built with either a full steel shell or skeleton construc-
tion, as may be desired. Common red brick may be used in the side walls.
The furnaces may be fired with oil, gas, or solid fuel. When solid fuel
9
130
HYDROMETALLURGY OF COPPER
is used the fire-boxes are placed at the sides of the furnace, and the heat
or gases to the various hearths regulated by suitable dampers.
The single hearth chloridizing furnace, made for direct firing with
oil, is 32 ft. in diameter and has a capacity of 100 tons of pyritic cinder
per day of 24 hours. The fuel used is 14 gallons of oil or 210 lb. of coal,
per ton of ore roasted, or approximate 10 per cent, of coal on the weight
of ore.
The multiple hearth mufHe fired furnace has been successfully used
for sulphating roast, in which, on some ores, 88 per cent, of the copper
was soluble in water. By the use of a weak acid, which can be made at
low cost from the escaping gases, the extraction has been increased to
98 per cent.
The rabbles and rabble arm, Fig. 33, can easily and quickly be removed ,
and replaced.
The furnaces are built both with open hearth or muffle, and of vary-
ing diameters and number of hearths. The following table gives the
essential figures for some of the standard sizes.
No.
Diameter outside
Number of
hearths, sq. ft.
Hearth area,
sq. ft.
Weight of '' Capacity in 24 hours,
metal parts, lb. : tons of 2000 lb.
1
2
3
9 ft. 9 in
9 ft. 9 in
12 ft
5 217
7 ] 304
5 1 373
7 522
5 ! 75!.';
18,075
22,075
22,200
26,600
76,500
91,800
97,000
113,000
101,200
118,000
76,300
2.5 10.8
3.5 15.0
4.3 18.6
4
12 ft
6.0 26 1
5
16 ft
8 3 36 0
6
16 ft
7
1015
11.6 51.0
7
20 ft
5
7
1245
14 3 62 0
8
20 ft
20 0 87 0
9
10
11
21ft. 6 in
21 ft. 6 in
22 ft. 6 in
5 i 1470
7 2058
3 i 978
16.9 73.5
23.6 103.0
11.2 48.9
12
32 ft
1
787
154,400
100
No. 12 furnace is designed more especially for chloridizing purposes,
and the capacity shown above has been demonstrated in this service.
In the capacity column the figures at the left are based on roasting
pyrites containing 50 per cent, sulphur, and reducing the sulphur to
2 per cent, in the roasted ore.
The figures at the right in the capacity column are based on smelter
practice where concentrate containing 35 to 38 per cent, sulphur is
roasted, the sulphur being reduced to from 7 to 9 per cent.
When oil is used as fuel it can be introduced through port holes
anywhere at the sides of the furnace, as in Fig. 32; if coal is used, regular
fire-boxes are necessary, as shown in Fig. 41 .
The Greenawalt Porous Hearth. — ^The porous hearth. Fig. 34, invented
by John E. Greenawalt, and patented in 1906, is applicable to any
ROASTING FURNACES
131
furnace. Some rather remarkable results have been obtained in roasting
ores by the use of this device, both as to saving of fuel and capacity per
square foot of hearth area.
The essential principle involved is the method of supplying sufficient
air for the ready oxidation of the incandescent sulphide particles not
directly exposed to the oxidizing atmosphere of the roasting chamber.
One of the greatest objections to reverberatory furnaces is that the heat
and air cannot be most effectively applied. The top layer of the ore is
^■w^^^
Fig. 34. — Greenawalt porous hearth furnace.
exposed to the highly oxidizing atmosphere, but that below the surface
is in an atmosphere which, if not reducing, is certainly not highly
oxidizing.
To obviate these difficulties Greenawalt conceived the idea of placing
the roasting ore on a porous bed, or filter, and percolate the air, either up
or down, through the ore mass and porous bed. In carrying out the
first experiments in a hand reverberatory furnace, certain interesting
results were obtained. The air, in roasting heavy sulphide ores, was
not found to be of much benefit in the early stages of the roasting, but
in the later stages it proved of the greatest advantage. Another difficulty
in roasting heavy sulphide ores was that the oxidizing effects were so
violent that the heat evolved sintered the charge so that the sintered
portion had to be screened from the roasted ore before it was learned
how to properly regulate the draft or suction. The sintering, or agglom-
eration, was entirely due to the air supply, so that if the ore was to be
132
HYDROMETALLURGY OF COPPER
roasted without agglomeration, the air was percolated through the ore
and porous bed with moderation, while if agglomeration or sintering
was desired the air was used with considerable suction or pressure,
depending upon whether down-draft or up-draft was employed. Roast-
ing, with or without sintering, was found to be purely a matter of air
supply. The down-draft gave more uniform results than the up-draft.
The amount of air that can be passed through a bed of incandescent
ore, on a porous hearth, is also surprising; as much as 14,000 cu. ft. of
air per hour have been passed through a hearth 10 ft. square, continu-
ously, without disturbing the ore particles.
The following results were obtained on tests made by the New Jersey
Zinc Co. with the demonstration furnace at Denver. The furnace has
a hearth area 100 sq. ft. It is arranged for hand rabbling and fired
with coal. The tests were made under the supervision of W. C. Wetheril,
consulting engineer and metallurgist of the Empire Zinc Co., and Wm.
H. Faul, assistant engineer. The chemical determinations were made
by the company's chemist, E. M. Johnson.
WITHOUT AIR (24 HOURS)
Sulphur
11.00 a. m. Charged raw ore
1.00 p. m. — second hour
3.00 p. m. — fourth hour
5.00 p. m. — sixth hour
7.00 p. m. — eighth hour
9.00 p. m. — tenth hour
11.00 p. m. — twelfth hour
1.00 a. m. — fourteenth hour
3.00 a. m. — sixteenth hour
5.00 a. m. — eighteenth hour
7.00 a. m. — twentieth hour
9.00 a. m. — twenty-second hour
11.00 a. m. — twenty-fourth hour
Roasted ore, final average of entire charge.
Crude concentrates. . . .
Roasted concentrates
Weight of charge, raw, 2000 lb.
AVeight of charge, roasted, 1690 lb.
Shrinkage, 310 lb.
Coal consumed (slack), 3052 lb.
Total S I Sol. H2O
Sulphates
NazCOi
Insoluble S
34.3
0.044
0.11
34.15
19.5
0.760
1.17
17.57
17.6
0.150
0.37
17.08
16.6
0.150
0.37
16.11
12.8
0.130
0.39
12.28
11.4
0.120
0.29
10.99
5.4
0.190
0.62
4.59
2.9
0.110
0.56
2.23
3.2
0.180
0.65
2.37
2.3
0.25
0.78
1.27
1.1
0.15
0.81
0.14
1.4
0.23
0.87
0.30
2.5
0.72
0.19
0.59
3.1
0.72
1.17
1.21
Si02
Fe
Zn
Pb
1.6%
10.2%
46.8%
5.7%
1.6
11.6
53.4
5.0
ROASTING FURNACES
WITH AIR (12 HOURS)
133
9.45 a. m. Charged raw ore
10.45 a. in. — first hour
11.45 a. m. — second hour
12.45 p. m. — third hour
1.45 p. m. — fourth hour
2.45 p. m. — fifth hour
3.45 p. m. — eighth hour
4.45 p. m. — seventh hour
5.45 p. m. — eighth hour
6.45 p. m. — ninth hour
7.45 p. m. — tenth hour
8.45 p. m. — eleventh hour
9.45 p. m. — twelfth hour
Roasted ore, final average of entire charge.
Total
34.3
26.18
21.03
19.70
17.90
15.04
11.61
6.75
3.17
1.23
1.12
0.85
0.68
0.68
SiOz
Sulphur
Soluble
0.11
0.21
0.25
0.32
0.34
0.40
0.42
0.48
0.56
0.60
0.63
0.66
0.50
0.48
Insoluble
Fo
Zn
34.5
25.87
20.87
19.38
17.56
14.64
11.19
6.27
2.61
0.63
0.-49
0.19
0.18
0.19
Pb
Crude concentrates. . .
Roasted concentrates.
1.6%
1.6
Furnace charge raw ore, weight 1800 lb.
Furnace charge roasted ore, weight 1551 lb.
Shrinkage, 249 lb.
Coal (alack) consumed, 1160 lb.
10.0"
11.8
-16.9"
.-,r,.(i
In these comparative tests, alternate charges of ore were roasted
without and with air passing through the hearth, other conditions
remaining the same.
From these tests it was concluded that, with the porous hearth the
capacity of the furnace was increased from two and one-half to three
times; that the amount of fuel required was only 35 per cent, of the
amount required without the air, and that the total cost of roasting was
reduced by 65 per cent.
Two mechanical furnaces, 10 ft. wide and 100 ft. long were then
erected to roast this and similar material, but the lead and other ingre-
dients of the ore formed a smooth crust due to the friction of the rabbles
with the stationary ore on the hearth, which soon became impervious
to the air. To what extent such a crust would form in roasting other
ores is somewhat questionable. In silicious ores, like those of Cripple
Creek, no crust of any kind is discernible between the moving and
stationary ore of the hearth, even after furnaces have been in operation
for years. The stationary ore is as permeable as the moving ore. Under
such conditions no crust difficulties should arise, and the capacity of
the furnace should be enormously increased, and the cost of roasting
materially diminished.
134
HYDROMETALLURGY OF COPPER
ROASTING FURNACES
135
REVOLVING FURNACES
Bruckner Furnace. — ^The Bruckner furnace was introduced in Colo-
rado in 1867, and since that time has been more or less generally used
bgth in oxidizing and chloridizing roasting.
The furnace, as shown in Figs. 35 and 36, consists of a cylindrical
shell of steel plate, with circular openings at each end. Two circular
tracks are fastened around the cylinder, at equal distances from the ends.
With these tracks the cylinder rests on four riding wheels or rollers,
which are mounted on strong cast iron frames. Two of the rollers have
double flanges to keep the cylinder in proper alignment.
A revolving motion is given the cylinder by means of cogs and
pinions, so adjusted that the proper reduction in speed is made between
the driving pulley and the revolving cylinder. A large peripheral cog
is fastened to the cylinder, which meshes into a pinion of suitable size.
Fig. 36 shows the mechanical details of the cylinder and driving mech-
anism for a furnace 8 1/2 ft. diam. by 18 1/2 ft. long, and having a
capacity of 8 to 11 tons of ore at a charge.
A fire-box is placed at one end of the cylinder, the throat of which
corresponds with the central end opening; while a similar opening at the
other end communicates with the stationary flue or dust chamber. The
fire-box for the furnace is built of brick, and in some instances it is
mounted upon wheels so that it may be removed to facilitate the relin-
ing of the shell. When a complete roast is desired, a stationary fire-box
is to be preferred over a removable one.
There are four doors in the cylinder; two of these are placed diamet-
rically opposite the other two. The doors serve for charging and discharg-
ing the ore. A hopper is placed above the furnace large enough to hold
a charge of the material to be roasted; this hopper has two outlet spouts,
each provided with a slide, which corresponds with the furnace doors.
The shell, as well as the ends, are lined with brick. Provision is
made in the driving mechanism to regulate the speed from one revolution
per minute to one revolution in 3 minutes.
The ends of the cylinder are sometimes contracted to facilitate
discharging of the ore, but this is not necessary, except for large furnaces.
These furnaces are usually built of various standard sizes; the di-
mensions, weights, and capacities, are approximately as follows:
BRUCKNER ROASTING FURNACES
Size of furnace
6 X12
7 X18
8 1/2X18 1/2
8 1/2X28
Weight of iron work
17,800 lb.
30,000 lb.
52,000 lb.
69,000 lb.
Number
fire brick
Number
GommoQ brick
Capacity, per
charge, tons
1,300
18,000
4
1,700
20,000
6 to 8
2,800
25,000
8 to 11
3,300
27,000
15 to 25
136
HYDROMETALLURGY OF COPPER
,,..^;;' .' ^
■'. "^' "^ "
\ \ \ \
J0^^^
i^^^H^^^
M^W^'^'^^tttSi^B^^^^^^
Km ^i ^^?v^j{|^^^9BH^^^^^^^|
H^^B'\ ^ \ '
H^^^i.\\\\
■i^iP***^!*: ;
^BP^vfMi^^
^^^^^^■^^^b''<
jlur ^^^Hi^^H
HiH^B'V\
^w '^^^S^^l^^^^l
HIShU'^ \ '
^M^r^. ;■■,■;,':". ■;;i;;
:'<""' IH
mg^ga^kJA \
(^*^-.^,
>Jfl^^^HES^^^^Vv
■^^^HV$\' '
^^^^^^V^v\
^F^ "v^^hI
1^ ^^^H
^^^L -' ' .ii^^^l
^^'
ROASTING FURNACES 137
In the operation of the furnace two of the doors in the cylinder are
opened and brought directly under the hopper spouts. The slides are
then withdrawn and the charge allowed to run in. The furnace is then
given a slight turn to bring it in a convenient position to close the doors,
which are then fastened as tight as possible.
The cylinder is then revolved and a strong fire maintained, until
the sulphur is well ignited. If the charge being roasted is heavy sul-
phide ore or pyritic concentrates, the firing is discontinued after the
sulphur is thoroughly ignited. The heat developed by the burning
sulphur is considerable, and if augumented by external firing, caking
and balling of the ore would take place.
Usually two charges of ore are roasted in 24 hours. If the ore is a
heavy sulphide, the furnace can be run from three to four hours with-
out firing. If the ore is being chloridized, salt is then added, and the
roasting continued with a moderate fire. The temperature may be
observed, and scoop samples taken, from a hole back of the flue. When
the roasting has fairly well progressed, the charge in the furnace loses its
tendency to run like a liquid, and assumes an inclined position, up to
45 degrees. This unequally distributed weight acts against the direction
of the motion, and if the clutch is thrown out in order to stop the furnace,
this weight will pull the furnace back nearly a quarter of a turn. If the
ore is given a chloridizing roast, the salt is introduced through two of the
doors and well scattered over the charge.
When the roasting is finished two cars are pushed under the furnace,
one for each door. All four doors are then opened and the furnace
again revolved. The receiving cars are made narrow so that no ore
is dropped beyond the confines of the car, because the ore discharges
through the doors over a large arc.
The revolving motion of the Bruckner furnace should be slow. The
ore, in any case, is continuously changing its position and exposing new
surfaces. There is no advantage in moving the ore more rapidly than
the ore at the surface can be oxidized. One revolution in two or three
minutes should be ample.
In the larger sized Bruckner furnaces, the ore is exposed to rather
an uneven heat. That near the fire-box is at a considerably higher tem-
perature than the ore at the further end, from 18 to 28 ft. distant. To
overcome this defect, O. Hofmann modified the Bruckner furnace by
placing a fire-box at both ends, and arranged so that the fire-boxes could
be used intermittently, by damper arrangements connecting with the
flue, or dust chamber.
The cylinders are made to revolve slowly; the smaller ones by apply-
ing power to a shaft carrying the friction rollers, the larger ones by a
pinion which engages a spur gear surrounding the cylinder.
Howell-White Furnace.— The Howell- White furnace. Figs. 37 and 38,
138
HYDROMETALLURGY OF COPPER
consists essentially of a long telescope-shaped cast-iron cylinder, made
in sections with cast flanges, carefully fitted, and bolted together. The
cylinder is supported on a system of rings resting on friction wheels, and
guided in a central position by rollers in upright frames, and revolved by
friction of the wheels operated by gears and pulleys. The flame passes
through the revolving cylinder from a fire-box at one end to the flue at the
other.
That one-third portion of the cylinder nearest the fire has a larger
external diameter than the part next the flue, but it is lined with fire
brick to make its internal diameter the same as that of the smaller part,
which, although unlined, stands the heat very well. In some instances
the entire cylinder is lined, and this is probably the best when the ore is
given a thorough oxidizing roast.
Fig. 37. — Howel-White furnace.
Projecting fire brick, arranged spirally in the brick-lined portion,
assist in oxidation by raising and showering the ore through the flame,
which passes directly through the cylinder. When the feed end of the
furnace is not lined with brick cast-iron projections are provided for the
same purpose. These projections make large quantities of dust, some-
times from 30 to 50 per cent, of the ore charged, and for that reason are
frequently omitted.
The cylinder is inclined slightly toward the discharge end in order
to advance the ore gradually against a constantly increasing temperature.
The furnace is fed at the upper end with dry ore by means of a screw
feeder, and then makes its way automatically toward the lower end of the
furnace, where it passes out, dropping between the end of the cylinder
and the fire-box, into the vault.
On account of the excessive dust from the furnace, an auxiliary
fire-box is usually placed at the flue end of the furnace for roasting the
flue dust as it passes, suspended in the air, into the dust chamber. If
the spiral projections are omitted from the cylinder, the auxiliary fire
to roast the dust may also be omitted. The ore as fed into the furnace
has its sulphur fairly well eliminated before it comes against the more or
less direct flame from the fire.
ROASTING FURNACES
139
The furnace requires very little repair, and very little power to run
it, while its capacity is quite large. The revolving speed of the furnace
is adjustable; and may vary from three to eight revolutions per minute,
the larger furnaces revolving more slowly than the smaller ones.
Fig. 37 shows a longitudinal elevation of the furnace and Fig. 38, a
section through the discharge end.
-13-10^^ 1 ; >
Truck to 5 * ^
WoodYarJ / Hrlck Pftvios
■siasW-l
J'W'^Si'/'^W':^.
Wln-'tliiig Flour
Fig. 38. — Howel- White furnace. Discharge end and ore vault.
The following table gives the standard sizes of the Howell-White
furnace, with weights, material, capacities, etc.
HOWELL-WHITE ROASTERS
Diameter
inches
Length,
feet
23
27
34
36
Capacity in
tons
15 to 20
30 to 50
50 to 60
60 to 70
Weight of iron
work, lb.
25,000
43,500
54,000
59,000
Fire brick
required
Common brick
required
31X41
52X62
52X62
52X62
1,900
2,700
2,700
3,000
22,000
28,000
28,000
31,000
140
HYDROMETALLURGY OF COPPER
ROASTING FURNACES 141
This furnace has been largely used for chloridizing roasting; when
so used the salt must be fed in with the ore. A portion of the salt may
be added as the ore drops into the vault, after passing through the
cylinder.
Muffle Furnaces.— All of the well-known designs of standard rever-
beratories may be modified to mufHe furnaces by placing flues under the
hearth, and by having two concentric arches over the reverberatory
chamber.
A modification of the hand reverberatory for muffle roasting, in
chloridizing pyritic cinders, is shown in Fig. 49, page 263.
Section D-D
Fig. 40. — Edwards muffle furnace. Transverse section.
A modification, of the Brown furnace for muffle roasting is shown in
Fig. 39; for the Edwards, in Fig. 40; for the Wedge, in Fig. 41, and for
the Allis-Chalmers McDougal, in Fig. 42. These illustrations fairly
represent the modifications required for the straight line chain-driven
roasters; the straight line circular rabbled roasters, and the multiple
hearth furnaces with central revolving shaft.
Muffle furnaces are not as economical as reverberatories, but for
certain work present recognized advantages. The heat can be regulated
more uniformly and the amount of dust is minimized. In chloridizing
142
HYDROMETALLURGY OF COPPER
roasting, muffle furnaces present the decided advantage that the volatile
metals in the fumes can be more readily condensed and recovered. In
reverberatories, or other direct-fired furnaces, the gases from the fires are
mixed with those from the ore, so that the volume of gases passing
through the condensers for the recovery of the volatilized values is quite
large, and by far the largest amount of air in such cases is due to the
combustion of the fuel. In chloridizing roasting only that part of the
furnace containing the ore after the salt is added need be muffled.
A combination of reverberatory and muffle can frequently be used to
advantage to avoid excessive dust loss, and where it is intended to make
Fig. 41. — Wedge muffle furnace.
sulphuric acid to leach the copper from the roasted ore. Fig. 42 shows
such a furnace, in which the three lower hearths are muffled with special
tile to keep the products of combustion from the fire-boxes separate
from the sulphur dioxide gases; the furnace is equipped with uptakes for
conveying the gases, and a separate stack is arranged with connections
to the muffles for taking away the gases of combustion. If sulphuric
acid is to be made, furnaces may be muffled so as to get the maximum
sulphur dioxide content in the furnace gases for the manufacture of the
acid, and the remaining sulphur in the ore can then be eliminated by
a continuation of the same hearth into a reverberatory chamber in
which the fumes from the ore and from the fuel commingle and pass
out together, and separately from the sulphur gases from the muffle.
no A H TI A'G F UR NA CES
143
^s^^iO.ll -'. «\^-\^-.^'I»^;5 ■ .^-.^ ; .„ isS^^^^Sl^
Fig. -12. — AUis-Clialmcrs McDougal muffle furnace.
144 HYDROMETALLURGY OF COPPER
Fig. 41 shows a muffle furnace now largely used in the chloridizing roast-
ing of cupriferous pyritic cinders for the extraction of the copper. This
type of furnace is replacing the hand muffle furnaces used for that
purpose.
Ore Coolers. — After roasting, it is necessary to cool the ore before
charging it into vats or agitators for chemical treatment. The cooling
in almost all of the furnaces is done in combination with the roasting by
a mechanical device similar to the rabbling mechanism of the furnace.
Usually, as in the case of the Pearce, Holthoff-Wethey, and Wedge
furnaces, by dropping the ore to a lower open hearth and continuing the
rabbling as in the roasting. Frequently the cooling hearth is made of
water j ackets, or pipes, through which cold water is continually circulated.
When the design does not conveniently lend itself to the use of a lower
hearth, the ore is cooled by a rabbling device similar to the roasting
mechanism, and either connected with it, or as a separate apparatus.
A push conveyor, with water jacketed bottom, arranged to deliver the
cooled ore to the leaching plant, makes a good cooler. In Figs. 20 and
22 the lower hearth of the furnace is used to cool the ore.
Dust. — Dust presents one of the most serious problems in roasting. '
If the ore fed into a furnace contains an excessive amount of dust, the
capacity is reduced. In one of the Cripple Creek mills, when the ore in
the supply bins became low and contained an excessive accumulation of
dust, the capacity of the furnaces was reduced from 100 to 90 tons per
day, but this decreased capacity was of short duration, as the bins soon
became empty and a fresh supply of ore had to be provided. Ore
crushed exceedingly fine cannot be roasted at the same rate as ore
crushed from 8 to 20 mesh.
All furnaces make more or less dust in roasting, which if not recovered,
represents a serious loss; and if recovered, an additional expense. A
dust chamber is usually a necessary adjunct to a roasting furnace, and
must be taken into account in a roasting installation.
For straight line reverberatory furnaces the dust recovered in suitable
dust chambers varies from 1.5 to 3.5 per cent, of the ore roasted. Much
depends on the nature of the ore and the speed of travel of the rever-
beratory gases. In multiple hearth furnaces the recovery in the dust
chambers will usually be from 3.5 to 5 per cent, in careful work. In re-
volving furnaces, the dust recovery in the flues or chamber is quite large.
M. W. von Bernewitz gives as an example' a battery of six roasting
furnaces with a capacity of 20 tons each of Kalgoorlie ore daily, that
has been crushed to 25-mesh screen. These furnaces are connected with
a flue 100X7X7 ft. and a stack 100X6 ft. with a 1/2-in. draft. Twenty
tons of flue dust are collected monthly, which amounts to about 3.5 per
cent, of the ore roasted.
ȣ. and M. J., Feb. 26, 1910.
ROASTING FURNACES 145
As a rule the dust collected from the dust chambers is not well
roasted. This is largely due to the fact that the partially heated ore dusts
more than after it has been heated to incandescence, but more largely
due to the rapid travel of the fuel and furnace gases over the partially
heated ore at the rear of the furnace. Roasting furnaces, as usually
built, exhaust all the gases from one opening at the rear of the furnace,
so that as each fire-box discharges its gases into the roasting chamber,
the draft, or speed of travel of the gases against the roasting ore must
be proportionately increased with each addition. At the rear of the
furnace all the fire-boxes have poured their gases into the roasting
chamber, and as the ore is rabbled, the rapid movement of the gases
over it, whip a large proportion of the finer unroasted material into the
dust chamber.
There is no reliable data available on the dust loss through the stacks.
It is probably more than usually supposed. In one of the Cripple Creek
mills there was at one time an unaccountable loss of values amounting
to 1 per cent. It was thought that much of this was due to dust losses
in the furnace stack.
10
CHAPTER VII
TYPICAL EXAMPLES OF ROASTING
Roasting of Cripple Creek Ores. — Most of the ores in Cripple Creek
are roasted before chemical treatment. The amount roasted per day-
is about 1500 tons. The ore after roasting is treated both by the chlorina-
tion and cyanide processes. All the furnaces now in use are designed to
roast 100 tons of ore daily. Of the furnaces in regular operation, there
are' three Pearce furnaces; six Holthoff-Wethey; eight Edwards, and
one^Holthoff revolving hearth. All of these furnaces are giving satis-
factory results.
The average chemical composition of Cripple Creek ores which are
roasted, is about as follows:
SiOj, 60 to 70 per cent. S, 1 . 5 to 3 . 0 per cent.
AI2O3, 18 to 22 per cent. Pb, trace to 0. 10 per cent.
CaO, 1 . 5 to 2.5 per cent. Zn, trace to 0.10 per cent.
MgO, 0 . 1 to 0 . 25 per cent. Mn, trace to 0.15 per cent.
BaO, trace to 0 . 25 per cent. K and Na, 5 . 0 to 10.0 per cent.
Fe, 3 . 0 to 7.0 per cent. Te, trace.
Cu, trace.
The ore as it comes from the mines is crushed through rock breakers
and rolls to pass through 8- to 16-mesh screens. The oxidized ore, as it
comes from the mines, contains as low as 0.25 per cent, sulphur; the un-
oxidized, from the deeper levels, contains as much as 5 per cent.; the
average for the different mills ranges between 1.5 and 2.75, depending
on the condition of the mines from which most of the ore comes.
Ore low in sulphur — from 1.0 to 2.0 per cent. — is roasted until it
contains between 0.25 and 0.60 per cent.; that carrying from 2.25 to 3.25
per cent, to between 0.50 and 0.85 per cent. All of the mills are at pres-
ent equipped with furnaces which regularly roast 100 tons of ore con-
taining from 1.5 to 2.5 per cent, sulphur; when the sulphur exceeds 2.5
per cent, the capacity is likely to be reduced to 90 tons, while on the
contrary, if the sulphur is about 1.5 per cent, or less, the capacity may
exceed 120 tons per day of 24 hours. The hearth area of these furnaces
is between 1200 and 1500 sq. ft.; the grate area about 75 sq. ft.
Many of the furnaces are fired with producer gas. The ore coolers
are usually located directly under the roasting hearths and form part of
the furnace, so that the ore on issuing from the roasting hearth drops to
the cooling hearth, to be turned over, cooled, and advanced by the cool-
ing hearth rabbles.
146
TYPICAL EXAMPLES OF ROASTING 147
The consumption offuel varies fronil0tol5tonsofWestern bituminous
coal per 100 tons of ore roasted. Lignite is used to some extent, but is
not as effective as the longer flame bituminous. Oil and residuum are
also used, but in connection with coal. If they are used together it
takes 5 tons of coal and 150 gallons of oil per 100 tons of ore.
The ore, in roasting, is given an initial temperature at the first fire-box
of 1200 to 1300° F. and at the last fire-box from 1400 to 1500° F. The
temperature of the fuel gases entering the reverberatory chamber is
approximately 1800° F. The ore never attains this temperature, because
it is being continually stirred and advanced, so that only the top layer
is momentarily exposed to the higher temperatures.
The charge remains in the roasting furnaces from 2.5 to 3.0 hours,
and on the cooling hearth from 1.5 to 2.5 hours. The bed is from 2.5 to
3.5 in. thick. In some of the furnaces the ore is rabbled every 17 seconds;
in others every 35 seconds.
One man attends a furnace; he usually has a helper to wheel the coal
and ashes. One such helper attends several furnaces.
The cost of roasting ranges from 45 to 55 cents per ton, estimated
approximately as follows, per 100 tons:
Coal, 12 tons, at .12.00 per ton,
Three f urnacemen, 8-hour shifts, at $2 . 50,
Coal and ash trammer,
General repairs, oil, water, etc.,
Power,
Interest on investment, at 6 per cent..
Cost per 100 tons.
Cost per ton,
To this must be added the proportional share of administration, etc.,
which will bring the cost per ton about 50 cents. These estimates in-
clude cooling and conveying.
The furnaces do not give much trouble. It is not unusual to have a
100-ton furnace in continuous operation for months without a serious
shut-down and no appreciable repairs; with occasional changing of the
rabbles the roasting proceeds indefinitely.
The accumulation of the dust in the dust chambers is usually from
1.5 to 2.5 per cent.; in some mills it is re-treated with the ore, while in
others it is shipped to the smelters, after briquetting. The dust is not
well roasted, and contains much undecomposed pyrites, and is high in
soluble sulphates. When the dust is treated in the mills, it is automatic-
ally again fed, with the ore, into the furnaces, and in this way worked
up with regular charges.
The average value of the dust is higher than that of the ore; this is
doubtless due to the fact that in crushing, the sulphides and tellurides
$24
.00
7
.50
2
.25
5
.00
5
.00
3
.50
$47
.25
0
.4725
148 HYDROMETALLURGY OF COPPER
are pulverized more than the quartz. There has not been any appreci-
able loss found by volatilization in oxidizing roasting.
Roasting Arsenical Sulphide Ore at the Golden Gate Mill, Mercur,
Utah. — The ore contains about 5 per cent, sulphur, and from 4 to 6 per
cent, arsenic, and small quantities of lead and copper. It is crushed to
10 mesh, when it is delivered to the roasters. The gold in the ore is
found in minute cleavage planes and crevices, and is easily attacked by
the cyanide solution; for this reason, crushing to 10 mesh is required
only for the sake of quick and complete roasting, and not for the purpose
of facilitating the leaching.
The plant consists of four coal-fired roasters, 125 ft. long by 12 1/2 ft.
wide. Each furnace consists of a single roasting hearth, whose roof
serves as a cooling hearth. The ore is fed into the lower level, and is
moved continuously in one direction by travelling rabbles. Escaping
from the end, it is elevated to the upper hearth, where it again travels the
length of the furnace, exposed to the air.
The ore occupies about four hours in roasting, and the same length
of time in cooling. Its volatile components are reduced from 5 per cent,
sulphur, and 4 to 6 per cent, arsenic, in the raw state, to 0.6 per cent, sul-
phur and 0.8 per cent, arsenic in the roasted ore.
During the year ending June 30, 1906, four roasters were in operation
for 46 days, five roasters for 237 days, and six roasters for 82 days.
The total number of roaster days was, therefore, 1861. As 126,358 tons
of sulphide ore were roasted, the average work was 68 tons per roaster
per day. The operation costs, including maintenance and repairs, for
the year 1906 were as follows:
Coal,
Labor,
Power and other items,
$126,039 $0,998
Slack coal is burned which costs 15.25 delivered. The labor consists
largely of firemen who receive $2.75 for eight hours. ^
Roasting of Casilas Concentrates, Victoria, Australia.^ — The following
is an average analysis of the concentrate: Lead, 4.48 per cent.,; zinc,
5.26 per cent.; iron, 31.65 per cent.; arsenic, 15.16 per cent.; sulphur
31.63 per cent.; unestimated, 11.91 per cent.; the copper rarely exceeded
1 or 2 per cent. It was usual to make from 15 to 20 per cent, concen-
trate, running from 3 to 4 oz. of gold, and containing over 50 per cent,
of the total gold value of the ore crushed. The tailings from the stamp
battery and concentration tables are cyanided, and the concentrate
chlorinated.
^E. and M. J., Nov. 10, 1906.
^Francis B. Stephens, E. and M. J., .July 29, 1905.
Total
Per ton
$73,759
$0,583
33,736
0.267
18,544
0.146
TYPICAL EXAMPLES OF ROASTIXG 149
It wiis found that short hand-rabbled reverberatories were unsuited
for roasting the concentrate on account of the charge fusing too easily.
As the galena could not all be separated, some form of mechanical
furnace was necessary, and two Edwards' mechanical furnaces with
60-ft. hearths were installed. One man per shift of 8 hours, attended
to all the work of the two furnaces, with a weekly capacity of 30 to 35 tons
each. The furnace man charged the hoppers, stoked the two furnaces,
and looked after the engine and dynamo for lighting the works. The
fume was lead into a brick flue 300 ft. long and 5 ft. by 4. ft. inside, with
a 40-ft. iron stack 2 1/2 ft. in diameter.
The height of the top of the stack above the hearth of the furnaces
was 70 ft.; small dust chambers were built between the ends of the fur-
nace and the flue. The iron stack did not suffer at all, and acted as an
excellent arsenic condenser. It was necessary to clean the whole length
of the flue every three months; about'30 tons of deposit being obtained.
The flue dust for the first 100 ft. consisted partially of roasted con-
centrate and arsenic soot, assaying about 3 oz. in gold per ton, or about
the same value as the concentrate roasted. The last 200 ft. of the flue
contained arsenic soot comparatively free from concentrate, and assayed
7 1/2 dwt. The arsenic at the base of the stack had to be cleared out
weekly.
Just before the finish of the roast, 1 to 1 1/2 per cent, salt was added
to the concentrate in order to obtain a sweet roast. No evidence of loss
of gold by volatilization could be obtained.
The roasted ore discharged into a push conveyor which carried the
ore to a steel bucket elevator; this took it to a cooling bin over the treat-
ment vats. Dry wood was used as fuel; any green sticks getting in
generally had the effect of throwing back the charge to magnetic oxide.
Badly roasted ore set hard in the vats, while roasted ore did not. The
ore in the hand-rabbled furnaces always roasted black; while in the
mechanical furnaces, it roasted chocolate color, but never bright red.
The brighter the color obtained in roasting, the better the roast; although
the magnet failed to show any difference. Practically no zinc was sent
to the flue, the heat not being high enough; it was in the roasted ore
mostly as sulphate or basic sulphate.
During the chlorination treatment, the zinc was almost all leached
out by the sulphuric acid, but the amount of zinc seemed to have no
influence on the extraction, other than to prolong the period, through
packing of the charge in the vats as the zinc leached out. The chlorine
solution used had a strength of 0.09 to 1.2 per cent, of chlorine, and from
0.5 to 1 per cent, sulphuric acid, over and above the amount required to
combine with the beaching powder. The copper gave very little trouble.
An extraction' of 85 per cent, was obtained on well-roasted ore, the
loss amounting to 1 to 2 dwt. per ton crushed. Numerous experiments
150
HYDROMETALLURGY OF COPPER
were carried out to try and better the extraction, but with no success,
although they led to a steady decrease of the cost of chemicals. Fine
grinding after roasting gave no better results. The cost of roasting the
concentrate at the rate of 54 tons per week, was as follows :
Wages,
Fuel,
Repairs,
General charges,
Cost per ton roasted,
$0,946
1.066
0.040
0.446
$2,498
Notwithstanding the extremely refractory nature of the concentrate,
over 50 per cent, was saved by roasting and chlorinating, over the cost
of shipping to the smelters. The concentrate sometimes carried as high
as 20 per cent. zinc.
Roasting at Kalgoorlie.' — The ore is crushed dry in Krupp ball or Griffin
mills and roasted in Edwards or Merton furnaces. At the Kalgurli mine
the nine No. 5 Krupp mills, eight of which are in continuous use, are capable
of putting through between 1 0,000 and 1 1 ,000 tons per month. These mills
maintain a duty of 45 tons per 24 hours when crushing through a 37-
mesh screen. The load of balls weighs about 2300 lb. ; one 18-lb. ball is
added every day to compensate for wear and tear. The mills run at
25 r. p. m. and require, including their share of counter-shaft friction
and dust fans, 25 h. p. The cost for crushing through 37-mesh screen
at the South Kalgurli is 79 cents per ton; the cost at the Associated
and Associated Northern, crushing through 27-mesh screen in ball mills,
is 59 cents.
The following table gives the essential facts for roasting at the Kal-
goorlie mines:
ROASTING FURKACES AT KALGOORLIE MINES
Mine
Type of
furnace
Associated Northern
Kalgurli
Kalgurli
Perse verence
Great Boulder
Great Boulder
South KalgurU
Associated
Merton.. .
Edwards..
Edwards..
Edwards..
Edwards..
Merton.. .
Merton.. .
Merton.. .
Area of
hearth
630 sq. ft.
70X9 ft. 6 in.
63X9 ft. 0 in.
121X13 ft. 6 in.
64X6 ft. 6 in.
422 sq. ft.
617 sq. ft.
445 sq. ft.
Average
tonnage
(24 hours)
23
23
22
93
25
25
36
20
Sulphur
in ore
per cent.
Fuel represents about 50 per cent, of the total roasting costs.
Gerard W. WilHams, E. and M. J. Feb. 15, 1908.
5.0
4.2
4.2
4.0
3.5
3.5
3.8
6.1
Fuel con-
sumption
per cent,
(wood)
12.2
13.0
13.0
11.0
9.2
9.2
10.6
10.0
Cost
per
ton
61.0*
68.4*
65.8*
59.2*
59.2*
64.0*
63.0*
TYPICAL EXAMPLES OF ROASTINd 151
Recently ' at the Associated, the smaller Merton furnaces have been re-
placed by the larger sized Edwards furnaces. The sulphur in the ore
averages about 5.5 per cent. The ore gets a dull red heat about the
fifth rabble from the end. The furnaces average about 95 tons per day
each. They are motor driven, and use 6 amperes at 550 volts each.
(4.4 h. p.). The end fire-boxes are not used much, just two or three
logs are kept burning to warm the air passing through the fire-bars. The
middle fire-boxes are fired heavily, and the sulphur continues to burn
till the fourth rabble from the discharge, namely 22 ft. from the end,
and then discharges quite cool. Fuel consumption averages 11 per cent,
of the roasted ore. The flue temperature is 700° F. One man attends
two furnaces. The roasting cost is 60 cents per ton. About 81,000
tons of ore and concentrate are roasted at Kalgoorlie monthly.
' M. W. von Bernewitz, Mining and Scientific Press, May 13, 1911.
PART II
HYDROMETALLURGICAL PROCESSES
CHAPTER VIII
PROPERTIES AND SOLUBILITIES OF COPPER
Copper. — Atomic weight, 63.6; specific gravity, 8.94. Weight per
cubic foot: cast copper, 542 lb.; rolled copper, 555 lb. Weight per
cu. in., 0.32 lb. Copper occurs chemically as cuprous compounds,
formula CuA', or cupric compounds, formula CuA", where AMs a
univalent or monad acid radical, and A" a bivalent or dyad acid radical.
As a monad atom, copper has a chemical equivalent of 63.6, as a dyad
element 31.8. The amounts of copper dissolved into or deposited from
a cupric or cuprous salt are proportional to the chemical equivalent of
copper in these two states and to the amperes flowing. Assuming
that one ampere liberates electrolytically 0.00001036 grm. of hydrogen
per second, the amount of copper deposited by the passage of one ampere
will be as follows:
Cuprous compounds
Cupric compounds
One ampere per second .
One ampere per m nute .
One ampere per hour . . .
One ampere per day ....
One ampere per year. . .
0.0006589 grm.
0.03953 grm.
2.372 grm.
56 . 93 grm.
20 . 78 kilogrm.
0.0003295 grm.
0.01977 grm.
1 . 186 grm.
28 . 46 grm.
10 . 39 kilogrm.
The melting-point of copper is 1080° C. It is a red metal, but thin
sheets transmit a greenish-blue light, and it also shows the same greenish-
blue tint when in a molten condition. Of the metals in ordinary use,
only gold and silver exceed it in malleability. In ductility it is inferior
to iron and cannot be so readily drawn into exceedingly fine wire. Al-
though it ranks next to iron in tenacity, its wire bears only about half
the weight which an iron wire of the same size would support. As a
conductor of heat it is surpassed only by gold. Next to silver it is the
best conductor of electricity.
Dry air has no action upon it; in moist air it becomes coated with a
film of oxide which protects it from further action of air or of water. It
forms a number of very important alloys with other metals; with tin
it forms bronze; with i;inc and sometimes with small amounts of lead and
tin, it forms brass; and with nickel and zinc it forms German silver.
155
156 HYDROMETALLURGY OF COPPER
Copper which has become hardened by mechanical work may be
again made malleable by heating. The boiling-point of copper is about
2000° C. Molten copper has a great tendency to dissolve hydrogen,
carbonic oxide, and sulphur dioxide, which it evolves again on solidi-
fying. Aluminum, cobalt, nickel, zinc, cadmium, tungsten, molyb-
denum and iron, are more or less readily dissolved by it, as also are cup-
rous oxide, sulphide, and phosphide, and the arsenides, arsenates, anti-
monides and antimonates.
On heating copper to a low red heat, far below its melting point, it
becomes covered with a film or scale which consists of a mixture of the
cuprous and cupric oxides.
Copper exhibits a greater affinity for sulphur than do any of the other
metals. It also unites directly with the metalloids, excepting hydrogen,
nitrogen, and carbon.
The best solvents for copper are nitric acid, concentrated sulphuric
acid, and aqua regia. Hydrochloric acid and dilute sulphuric acid only
dissolve the metal when air or some other oxidizing substance is present;
under these conditions it is more readily soluble in dilute hydrochloric
acid than in dilute sulphuric acid.
Cupric chloride acts on metallic copper to produce cuprous chloride:
Cu + CuCl2 = 2CuCl.
Cuprous oxide has the property of mixing with molten copper in all
proportions. Small amounts of cuprous oxide have no injurious effect
upon it, but large quantities make it cold-short, and when a certain
limit is exceeded, also red-short. Copper containing about 2 per cent,
cuprous oxide is still as fit for ordinary use as ordinary cast-refined
copper.
Cathode copper is exceedingly pure, usually about 99.93 per cent,
copper, with hydrogen as the chief impurity. Objectionable cathode
impurities are of two classes — those which depress the electrical conduc-
tivity and those which make the metal brittle. Arsenic and antimony
represent the first class; tellurium and lead the second. Good cathode
copper should show but a few thousandths of a per cent, of arsenic and
antimony. Experiments have indicated that it takes but 0.0013 per
cent, of arsenic or 0.007JL per cent, of antimony to lower the conduc-
tivity 1 per cent. Any conductivity troubles in electroljrtic copper
can almost invariably be traced to the presence of undue amounts of one
or both of these elements. Impurities of the brittle-making class are
rarely met with, and if present are due to mechanical contamination
of the cathode, either in the bath or in the subsequent furnace treatment.
Influence of Impurities on the Properties of Copper.
Arsenic and Antimony. — Hampe, in 1892, found that 0.5 per cent,
arsenic produces no bad results and that even when the percentage was
PROPERTIES AND SOLUBILITIES OF COPPER 157
increased to 1 per cent, only a slight degree of red-shortness, but no cold-
shortness could be noticed. He found that copper with 0.8 per cent,
arsenic could be drawn into the finest wire. Stahl, in 1886, stated that
a small percentage of arsenic prevents copper from becoming porous.
Hiorns, in 1906, showed that copper with arsenic up to 0.4 per cent, was
very malleable when cold; that with 0.2 per cent, each of arsenic and anti-
mony, the same is true; and that arsenic in the presence of antimony
makes the copper more malleable than it is with antimony alone, though
antimony when not above 0.2 per cent, only slightly impairs the mal-
leability of copper. He adds that arsenic in copper is highly beneficial
because it deoxidizes cuprous oxide, which tends to destroy the mallea-
bility of copper. Johnson, in 1906, stated that cast copper with 0.5 per
100
60
60
I
6 40
1
i
i
1
Condu
jtlvlty
w
Cu-As
Alloys
\
Cu-Sb
ire Cop
Alloys
per=l(
10
\ Te
nperat
are 13.4
"c.
%
"^»-^^
- —
-•
1 2
Percentage of Arsenic
Pjq 43 Conductivity diagram of pure electrolytic copper with arsenic and antimony
as alloys.
cent, arsenic has a tensile strength of 10 long tons per square inch, and
a 24 per cent, elongation. After forging, the tensile strength was raised
to 12.75 tons, and the elongation to 35 per cent. Upon rolling, the
tensile strength became 14 tons and the elongation 48 per cent.; and
finally upon being highly wrought and cold drawn, the tensile strength
of the same cast copper was raised to 15.9 tons, and the elongation varied
158 HYDROMETALLURGY OF COPPER
from 24 to 50 per cent, while the specific gravity was increased from
8.83 for copper in the cast state to 8.866 when the metal was wrought.
H. S. Hiorns and S. Lamb' prepared alloys consisting of pure elec-
trolytic copper and arsenic and antimony in quantities va,rying from
0.05 to 3.5 per cent. These alloys were drawn into wires 0.0325 in. in
diameter, and were tested for conductivity, shown by Fig. 43.
A. H. Hiorns found that with between 0.5 and 1 per cent, of arsenic
the malleability seemed to diminish, but with over 1 per cent, and up to"
2 and 3 per cent, arsenic, the copper rolled perfectly and was harder than
pure copper. With less than 0.5 per cent., arsenic the copper should
be less malleable when cooled slowly than when cooled quickly. With
a certain small quantity of arsenic introduced into copper the first small
portions appear to act by reducing the cuprous oxide, the remainder
retaining the metallic form and toughening the copper.
Arsenic invariably improves the forging properties where added to
impure copper. Arsenical copper is largely specified for in materials,
such as locomotive and boiler tubes, which are required to withstand high
temperatures, since mechanically hardened arsenical copper is not
softened at so low an annealing temperature as pure electrolytic copper,
which has undergone the same treatment.
Arsenic appears to improve the hot-working properties of copper
vitiated by traces of bismuth.
Bismuth. — Bismuth is the most injurious impurity in copper, as
very small quantities render the copper unworkable. According to
Hampe, copper containing as little as 0.02 per cent, of bismuth is red-
short, and 0.05 per cent., cold-short. With 0.1 per cent, the copper
crumbles under the hammer at a red heat. The presence of certain
proportions of arsenic and antimony somewhat counteracts the tend-
ency of bismuth to produce cold-shortness.
Lead. — ^Lead can be melted with copper in all proportions, but the
greater part of it can be liquated out of the alloy by a gentle heat. Lead
is to be found in all ordinary commercial copper, but it is not desirable
in any proportion over 0.10 per cent., and the lower the proportion
under 0.10 per cent, the better. According to Hampe's experiments,
0.15 per cent, of lead does not affect the malleability of copper in any
way; with 0.3 per cent, of lead it becomes slightly red-short, and with
0.4 per cent, slightly cold-short. With 1 per cent, it is unworkable.
The lead reduces the strength, ductility and toughness of copper.
The solvent action of copper for lead is very small. The addition of
lead to copper has the effect of lowering the affinity of copper for
reducing gases.
Iron. — Iron forms no true alloy with copper; small admixtures of
iron such as are contained in many varieties of copper have no injurious
•Min. Ind., Vol. XVIII, 1909.
PROPERTIES AND SOLUBILITIES OF COPPER 159
effects upon it. It occurs in refined copper in the merest traces,, which are
quite harmless. When all the sulphur has gone from a charge of copper,
in refining, a sample taken from the furnace shows an unblistered surface,
and is said to be "set-copper." At this point all the iron has been elimi-
nated except the merest trace.
Iron acts as a deoxidizer when added to copper. Copper containing
only 1 per cent, iron is rendered feebly magnetic, will forge well at a red
heat, is quite malleable, tough and strong, even in the presence of arsenic.
It lacks the fluidity of pure copper when poured at the same temperature.
Nickel. — Nickel alloys with copper in all proportions. Traces
of nickel in copper are beneficial, imparting strength and toughness.
Nickel must be kept low if arsenic is present, 0.10 per cent, being quite
sufficient to harden arsenical copper which has to withstand severe
working. Below 0.05 per cent., even in the presence of arsenic, its
effect upon the physical properties, excepting electrical conductivity,
may be considered insignificant.
Cobalt.— Little is known about the influence of this metal on copper.
According to F. Johnson, it toughens and strengthens copper when
present up to at least 1 per cent, without impairing its hot-working
qualities. Probably it acts very similar to nickel, conferring greater
durability at high temperatures, while toughening, hardening, and
strengthening the copper in the cold. Cobalt, in the low percentages in
which it is found in copper, is, if anything, beneficial, and, moreover, it
does not disagree with arsenic to so great an extent as nickel does.
Tin. — ^Tin hardens copper, more than any other element. It occurs
very rarely in commercial copper being readily eliminated during the
process of reducing the copper. Low percentages of tin improve the
tensile strength, ductility and resistance to corrosion, and maintain
these improvements at high temperatures, but the natural softness of
copper and its red color are both materially removed. Its malleability
is also decreased.
Tellurium. — ^A few thousandths of 1 per cent, of tellurium renders
copper appreciably red-short; but very little is known of the effect of
tellurium in commercial copper.
Sulphur rarely occurs in more than harmless traces in commercial
copper, yet it may occur to the extent of 0.03 per cent. (SOj) in electro-
lytic copper having a conductivity of 102.2 per cent. It derives its
origin from the incomplete removal of sulphur from the sulphate liquor
in which the copper cathodes were deposited, and from the sulphurous
gases from the fuel of the reverberatory refining furnace where it is
partially dissolved by the molten metal, as they pass over it on their way
to the flue. In the first case the sulphate would probably be reduced to
sulphide by the reducing action of poling, and in the second case, sulphur
dioxide would be absorbed and retained as such. Cuprous sulphide is a
160 HYDROMETALLURGY OF COPPER
highly undesirable constituent of copper, and its presence in any alloy
would be detrimental. Hampe finds that copper with 0.25 per cent,
sulphur is still moderately malleable, but with 0.5 per cent, it becomes
very cold-short, although not red-short.
Carbon is not at all absorbed by copper.
Pure copper or copper of more than usual purity, assaying, say, 99.8
per cent., is inferior to impure copper in mechanical properties, dura-
bility, and resistance to corrosion.
Copper, when cast in moulds, has the property of rising and becoming
porous. Sound castings can only be obtained by means of special
precautions, such as pouring at the lowest possible temperature, or
pouring in an atmosphere of carbon dioxide.
Cupric Carbonate. — The normal carbonate has not been obtained.
The two most important basic carbonates are:
(1) CuC03,Cu(OH)2, which occurs native as malachite.
(2) 2CuC03,Cu(OH)j, which occurs native as azurite.
The first is obtained when sodium carbonate is added to a solution
of copper sulphate. When these carbonates are slowly heated to 220° C.
the carbonate is slowly converted into black cupric oxide.
The carbonates of copper are readily soluble in dilute sulphuric, hydro-
chloric, sulphurous and nitric acids. They are also readily soluble in
ammonia and ammonia salts. They are partially soluble in sodium
carbonate and in solutions of potassium cyanide.
-Cupric Nitrate, Cu(N03),3H20. — Cupric nitrate may be obtained by
the action of nitric acid upon cupric oxide, hydroxide, carbonate, or
the metal itself. Copper is soluble in nitric acid, in all of its mineralogical
combinations. Sulphides are decomposed, as solution takes place.
Cupric nitrate is very easily soluble in water.
Cupric Oxide, CuO (Black Oxide of Copper). — Cupric oxide occurs in
nature as the rather rare mineral, tenorite. It may be prepared artifi-
cially by continued ignition of copper in contact with air; by exposing
cupric sulphate to an intense red heat, or the carbonate, nitrate, or
hydroxide to a moderate heat.
When caustic potash or soda is added by drops to a boiling solution
of cupric salts till the acid is saturated the whole of the copper is precipi-
tated as anhydrous black oxide, which may be freed from potash or
soda by boiling with water.
Cupric oxide is a black powder, which rapidly absorbs moisture
from the air. When heated it first cakes together and finally fuses,
giving up part of its oxygen, and leaving a residue consisting of CuO,2Cu20.
When heated with charcoal, or in a stream of carbon monoxide, marsh
gas or hydrogen, it is reduced to the metallic state.
PROPERTIES AND SOLUBILITIES OF COPPER 161
When the cupric oxide is gently heated with metallic copper, it is
converted into cuprous oxide.
A mixture of cupric oxide with excess of sulphur is resolved at a red
heat into cuprous sulphide, sulphur dioxide and a trace of cupric sulphate.
If on the contrary, the cupric oxide is in excess, cuprous oxide and cupric
sulphate are produced, and with only a trace of sulphur dioxide, except-
ing that when the heat is raised to the point at which the cupric sulphate
is decomposed.
Cupric oxide has a strong affinity for acids, dissolving in them easily.
It is soluble in sulphurous acid. It is insoluble in ammonia, but dissolves
on the addition of a few drops of acid or ammonium carbonate. It is
insoluble in dilute, but soluble in warm concentrated caustic soda or
potash. Ferrous chloride converts cupric oxide into cuprous and cupric
chlorides, with the formation of ferric oxide. Ferric chloride converts
cupric oxide into cupric chloride, with the formation of ferric oxide.
Cupric oxide is reduced to cuprous oxide at 1050° C.
Cuprous Oxide, CU2O (Red Oxide of Copper). — Cuprous oxide occurs
native as cuprite, the red pxide of copper. It is formed when finely
divided copper is gently heated in a current of air or when a mixture of
cuprous chloride and sodium carbonate is gently heated in a covered
crucible.
Cuprous oxide is reduced to the metallic stage by gentle ignition with
charcoal or hydrogen.
Cuprous oxide is insoluble in water ; it is converted into cuprous chloride
by hydrochloric acid. Nitric acid converts, it into cupric nitrate with
the evolution of oxide of nitrogen. When acted upon by dilute sulphuric
acid, it is partly reduced to metallic copper and partly oxidized into
copper sulphate. When heated with strong acid it is entirely oxidized
to sulphate.
When copper is oxidized with a considerable quantity of oxygen at
a high temperature, it forms cupric oxide (CuO). If the ignition be
carried further, cuprous oxide, CujO, may be formed from the CuO.
The cuprous oxide is not as readily soluble as the cupric oxide, and it
may be partly for this reason that copper sulphides roasted at a high
temperature do not give a good extraction of the contained copper.
Cuprous oxide fuses at a red heat.
When heated with acids coprous oxide forms a solution of a cupric
salt and metallic copper; for example,
Cu20 + H2SO, = CuSO, + Cu + H20.
However, strong hydrochloric acid does not deposit metallic copper
on dissolving cuprous oxide, which is due to the fact that the cuprous
chloride formed is soluble in strong hydrochloric acid.
Cupric Sulphate, CuSO^jSH^O.— Copper sulphate may be formed
11
162
HYDROMETALLURGY OF COPPER
by applying dilute sulphuric acid to copper oxide, when the sulphate
crystallizes out on cooling; by heating metallic copper with concentrated
sulphuric acid, whereupon sulphur dioxide is evolved, and anhydrous
cupric sulphate is precipitated as a white powder, mixed with a brown
mass of cuprous and cupric sulphides; on digesting this mass with hot
water, the cupric sulphate dissolves, and may be crystallized out of the
solution.
On roasting, the sulphide ores of copper are converted into cupric
oxide and cupric sulphate. When water is applied to the roasted ore,
the copper sulphate is dissolved; by evaporation of the water, the copper
sulphate crystallizes out of the solution.
At 100° C. copper sulphate loses 4 molecules of water, and at 200° it
loses all its water. At a bright red heat it decomposes into copper
oxide and sulphuric acid. When heated with carbon at a dark red heat
the copper is separated, with the formation of carbonic acid and sul-
phur dioxide.
From solutions of copper sulphate, the copper is precipitated by means
of iron, aluminum, and zinc, as metallic copper; with hydrogen sulphide or
the sulphide of the alkali metals, it is precipitated as the cupric sulphide
(CuS) . By electrolysis, copper is deposited from copper sulphate solutions
at the cathode and acid liberated at the anode. If at the anode, ferrous
sulphate is present in the solution, it is converted into ferric sulphate.
The crystallized copper sulphate dissolves in 3 1/2 parts of cold water,
and in much smaller quantities of boiling water.
SOLUBILITY OF COPPER SULPHATE
In 100 parts of water, at the following temperatures
Temperature
Parts CuSOi
c.
F.
0
32.0
14.15
10
50.0
17,50
20
68.0
20,53
30
86.0
24,34
40
104.0
28,50
50
122.0
33,31
60
140.0
39,01
70
158.0
45,74
80
176,0
54.53
90
194.0
64,35
100
212,0
75,22
Per cent, copper
4.00
4,95
5,81
6,88
8,07
9,43
11,04
12,93
14,33
18,23
21,28
All of the chlorides have the faculty of converting copper sulphate, in
solution, into the chloride. Hydrochloric acid dissolves copper sulphate
PROPERTIES AND SOLUBILITIES OF COPPER 163
with considerable reduction of temperature, forming a green liquid,
which when evaporated forms crystals of cupric chloride.
When excess of ammonia is added to a solution of copper sulphate,
a deep blue solution is formed having the composition CuS04,H20,4NH3.
Cupric and ferrous sulphates cannot be entirely separated by crystal-
lization, as a solution of these salts deposits a double sulphate of the two
metals. If, however, the amount of iron present is comparatively small,
the first crop of crystals obtained is moderately pure copper sulphate.
Cupric Chloride, CuClj. — Cupric chloride may be obtained in the
anhydrous condition by the combustion of copper in an atmosphere of
chlorine gas; copper filings or copper foil introduced into dry chlorine
takes fire spontaneously, and burns with a greenish light, producing
a mixture of cupric and cuprous chlorides, and if the chlorine is in excess
the cuprous chloride is slowly converted into the cupric chloride. It
is also produced when compounds of copper are roasted with salt or
other chlorides.
In the wet way it is formed when copper is dissolved in nitro-hydro-
chloric acid (aqua regia), or when cupric oxide, carbonate, or hydroxide
are dissolved in hydrochloric acid. Cupric chloride is readily soluble
in water, forming a deep green solution, which on being largely diluted,
turns blue. The salt crystallizes in green rhombic prisms, with 2H2O,
giving the composition of the crystals as CuCl2,2H20. When heated
to 200° C. it loses its water of crystallization, and at a dull red heat is
converted into cuprous chloride, with evolution of chlorine.
With copper oxides cupric chloride combines in various proportions
to form oxychlorides. From solutions of cupric chloride metallic copper
is precipitated by iron, aluminum, and zinc. With hydrogen sulphide
and the sulphides of the alkali metals and earths the copper is precipi-
tated as cupric sulphide, (CuS). Calcium hydroxide, or lime, precipi-
tates copper as the hydroxide, which on heating, is converted into the
oxide. By passing sulphur dioxide into a solution of cupric chloride, the
copper is precipitated as the cuprous chloride.
By electrolysis, copper is deposited from cupric chloride solutions
at the cathode, while chlorine is liberated at the anode.
SOLUBILITY OF CUPRIC CHLORIDE IN WATER
In 100 parts of water, at the following temperatures
Temperature
Parts CuClj Per cent, copper
C.
F.
0
32
17
62
70 . 6 ! 33 . 6 per cent.
76 . 2 I 36 . 3 per cent.
164 HYDROMETALLURGY OF COPPER
100 parts of water, saturated with CuClj, contains, at the following
temperatures :
Temperature
C.
F.
0
32
17.0 1 62
31.5
88
Parts CuClj Per cent, copper
41.4 i 19.7 per cent.
43.1 20.4 per cent.
44.6 21.2 per cent.
100 grm. of water dissolve 121.4 grm. of CUCI2+2H2O, at 16.1° C.
Cupric chloride is not decomposed by cold sulphuric acid.
It is soluble in solutions of ammonium chloride, and very soluble in
concentrated solutions of common salt. It is less soluble in concentrated
solutions of hydrochloric acid than in dilute solutions.
With ammonia cupric chloride forms a deep blue solution having the
composition CuCl2,4NH3,H20.
Cuprous Chloride, CujClj. — Cuprous chloride may be obtained by
dissolving cuprous oxide in hydrochloric acid. It is more readily pre-
pared by boiling a solution of cupric chloride in hydrochloric acid, with
copper foil or copper turnings. The nascent hydrogen, liberated by the
action of hydrochloric acid upon the copper, reduces the cupric chloride to
the cuprous chloride. The liquid is then poured into water, which causes
the precipitation of the cuprous chloride as a white crystalline powder.
A mixture- of zinc dust and copper oxide added to strong hydrochloric
acid, also yields cuprous chloride, the nascent hydrogen in this case
being derived from the zinc, and this causes the reduction of cupric
chloride formed by the action of the acid upon the cupric oxide.
Cuprous chloride may be formed by heating the cupric chloride to
a dull red heat.
Cuprous chloride melts somewhat below a dull red heat, and when
slowly cooled, solidifies in a translucent yellow mass. In closed vessels
it does not volatilize, even when strongly heated, but if heated in the
air it goes off in white vapor. When exposed to the air in a dry state
it slowly absorbs moisture and turns green; in the moist state it is
quickly turned into a green mass, of oxychloride of copper, CuClj,-
3CuO,4H20. This compound occurs native as the mineral Atacamite.
Cupric chloride, CuClj, when ignited gives cuprous chloride, and
therefore cuprous chloride is always formed when copper enters into
reaction with chlorine at a high temperature. The green solution of
cupric cliloride is decolorized by metallic copper, cuprous chloride being
formed; but this reaction is only accomplished with ease when the solu-
PROPERTIES AND SOLUBILITIES OF COPPER UV>
tion is very concentrated and in the presence of an excess of hydrochloric
acid to dissolve the cuprous chloride. The addition of water precipitates
cuprous chloride.
Many reducing agents which are capable of taking up half the oxygen
from cupric oxide are able, in the presence of hydrochloric acid, to form
cuprous chloride; sulphur dioxide, SOj, acts in this manner. The usual
method of preparing cuprous chloride consists in passing sulphur dioxide
into a strong solution of cupric chloride.
Cuprous chloride forms colorless cubic crystals which are insoluble
in water. Under the action of oxidizing agents,, it passes into cupric
salts and it absorbs oxygen from the moist air, forming cupric oxychloride.
From solutions of cuprous chloride, metallic copper is precipitated
by iron, aluminum, and zinc. Hydrogen sulphide and the sulphides
of the alkali metals and earths, precipitate the copper as cupric
sulphide, CuS.
By electrolysis, copper is deposited at the cathode, while chlorine is
liberated at the anode. If univalent salts are present in the anode
solution, these will be converted into bivalent salts by the action of the
liberated chlorine.
Milk of lime, added to a hot solution of cuprous chloride, precipitates
the copper as cuprous oxide.
Cuprous chloride is insoluble in water, but dissolves in hydrochloric
acid, ammonia, and alkaline chlorides.
SOLUBILITY OF CUPROUS CHLORIDE IN SOLUTIONS OF SODIUM
CHLORIDE
Saturated sodium chloride solution dissolves at
Degrees C.
Degrees P.
Cuprous chloride, CU2CI2
Metallic Cu
90
40
11
194
104
51.8
16.9 per cent.
11.9 per cent.
8 . 9 per cent.
10.76 per cent.
7 . 65 per cent.
5 . 73 per cent.
15 per cent.
VaCl-Aq. dissolves at
90
40
14
* 194
104
57.2
10.3 per cent.
6 . 0 per cent.
3 . 6 per cent.
6.62 per cent.
3 . 86 per cent.
2 . 31 per cent.
166 HYDROMETALLURGY OF COPPER
5 per cent. NaCl-Aq. dissolves at
90
40
194
104
2 . 6 per cent.
1 . 1 per cent.
1 . 67 per cent.
0 . 71 per cent.
Cuprous chloride, when melted, conducts the electric current very
well, copper separating out as fine leaves. The melt cannot be heated
to the melting point of copper and the copper obtained liquid, because
the cuprous chloride vaporizes too easily.
Cupric Silicate, CuSi03 + 2H20. — Silicate of copper occurs native as
chrysocoUa, CuSi03+2H20, and dioptase, CuH2Si04. Chrysocolla is
soluble in dilute hydrochloric acid, leaving a residue of silica. Dioptase
is soluble in nitric and hydrochloric acids, or ammonia, with separation
of gelatinous silica. It is not attacked by caustic alkalies.
Cuprous Sulphide, CujS. — There are two sulphides of copper, corre-
sponding to the two oxides; the cuprous sulphide, CujS, and the cupric
sulphide, CuS.
The cuprous sulphide, when heated at a comparatively low temper-
ature, loses one-half of its sulphur and is converted into the cupric
sulphide.
Cuprous sulphide occurs in nature as copper glance, or chalcocite.
It is produced artificially when copper burns in sulphur vapor, or when
an excess of copper filings is heated with sulphur.
It is not decomposed out of contact with the air; but if air has access
to it, combustion takes place, and sulphur trioxide and cupric oxide are
produced. When heated to redness in a current of aqueous vapor, it is
but slightly decomposed, but at a white heat, it yields large quantities of
hydrogen and hydrogen sulphide together with sublimed sulphur, and
the copper is completely reduced to the metallic state. It is not altered
by ignition in a stream of hydrogen.
It is not decomposed by chlorine gas at ordinarj^ temperatures; very
slowly when heated. It dissolves with difficulty in strong boiling hydro-
chloric acid. In heated nitric acid it dissolves with separation of sulphur,
whereas cold nitric acid dissolves one-half the copper, and leaves the
cupric sulphide. Cuprous sulphide, ignited with cuprous oxide, is
easily converted into sulphur dioxide and copper or cuprous oxide. It
is not dissolved by sulphuric acid. It is slowly acted upon by solutions
of ferric chloride and of ferric sulphate. Cuprous sulphide melts at
1127° C.
Cupric Sulphide, CuS. — Cupric sulphide is met with in nature as the
mineral covelite (blue copper). It is obtained artificially when either
copper or cuprous sulphide is heated with sulphur to a temperature not
beyond 114° C; so obtained, the compound is blue. As a black precip-
PROPERTIES AND SOLUBILITIES OF COPPER 1(37
itate, it is formed when hydrogen sulphide is passed into solutions of
cupric salts.
Treated with hot nitric acid the copper is oxidized, part of the copper
is converted into sulphate and the rest separated, so that the resulting
solution contains both nitrate and sulphate of copper. Hot concentrated
hydrochloric acid slowly converts it into cupric chloride, with evolution
of hydrogen sulphide and separation of sulphur. Cupric sulphide decom-
poses silver salts, the copper dissolving and the sulphide of silver being
precipitated. It is insoluble in dilute sulphuric acid, caustic alkalies, and
fixed alkaline sulphides. It is slightly soluble in ammonium sulphide.
Cupric Hydroxide, Cu(0H)2. — Cupric hydroxide is a pale blue pre-
cipitate produced when sodium or potassium hydroxide is added in
excess to a solution of a copper salt. The compound, when washed,
may be dried at 100° C, without parting with water; but if the liqu
in which it is precipitated be boiled, the compound blackens, apd is con-
verted into a hydrate having the composition Cu(0H)2, 2CuO. Cupric
hydroxide dissolves in ammonia, forming a deep blue liquid. It is very
soluble in acids. It is changed, by standing, to the black compound,
Cu302(OH)2 and by boiling to cupric oxide, CuO.
Ammonium carbonate, like ammonium hydroxide, precipitates the
cupric hydroxide and redissolves it to a blue solution. Carbonates of
the fixed alkali metals, as potassium and sodium carbonate, precipitate
the greenish-blue carbonate, Cu2(OH)2C03, which is converted by boiling
to the black, basic hydroxide, and finally to the black oxide.
From the blue ammoniacal solutions a concentrated solution of a
fixed alkali precipitates the blue hydroxide, changed on boiling to the
black oxide, CuO.
Cupric hydroxide is soluble in a solution of cane sugar in the presence
of an alkali or alkaline earth. It is somewhat soluble in the caustic
alkalies, and very soluble in ammonia.
Copper Cyanides. — ^Potassium cyanide forms, with copper, the yel-
lowish-green cupric cyanide, Cu(CN)2, soluble in excess, with the forma-
tion of the double cyanide, 2KCN, Cu(CN)2, unstable, changing in whole
or in part to cuprous cyanide. The potassium cyanide also dissolves
cupric oxide, hydroxide, carbonate, sulphide, etc., changing rapidly to
cuprous cyanide in solution in the alkali cyanide.
Potassium ferrocyanide precipitates cupric ferrocyanide, reddish-
brown, insoluble in acids, decomposed by alkalies; a very delicate test for
copper (1 to 200,000) ; forming in highly dilute solutions a reddish
coloration.
Solubility of Sulphur Dioxide, SOj. — Sulphur dioxide is largely
used in the hydrometallurigcal methods of extracting copper from its
ores. Lunge gives the percentage of a saturated solution of sulphur
dioxide in water, as follows:
168
HYDROMETALLURGY OF COPPER
Temperature, degrees
c.
F.
20
68
30
86
40
104
50
122
60
140
70
158
80
176
90
194 .
100
212
Percentage
SO,
8 . 6 per cent.
7.4 per cent.
6 . 1 per cent.
4.9 per cent.
.3 . 7 per cent.
2.6 per cent.
1 . 7 per cent.
0 . 9 per cent.
0 . 1 per cent.
The normal quantity of SOj in burner-gas from brimstone burners
is 11.23 per cent, by volume and 8.75 per cent, from burning pyrities.
Sulphur dioxide from roasting furnaces is much more dilute; muflBe
furnaces give a very much more concentrated gas than reverberatories.
SOLUBILITY OF SO^ IN WATER (Watts Dictionary)
Absorbed by 1 grm. of water at 760 mm.
Temperature, degrees C.
Grm.
SO,
C.c. SO,
8
12
16
20
24
28
32
36
40
44
48
50
58.7
49.9
42.2
36.4
32.3
28.9
0.073
25.7
1 0.065
22.8
0.058
20.4
0.053
18.4
0.047
16.4
j 0.045
15.4
One liter of SO^ weighs 2.86336 grm. 1 cu. ft. weighs 0.1787 lb.
With water, sulphur dioxide does not form sulphurous acid proper,
HjSOg. The sulphur dioxide dissolves pretty freely in water, and this
solution behaves in every way as if it contained the real acid, H2SO3.
The solution of SO2 by volume in water at various temperatures is
as follows:
1 volume of water at 0° C— 32° F. dissolves 79.789 volumes SO^
1 volume of water at 20° C— 68° F. dissolves 39.374 volumes SO^
1 volume of water at 40° C— 104° F. dissolves 18.766 volumes SO,
CHAPTER IX
HYDROMETALLURGICAL PROCESSES
Classification and General Consideration. — Hydrometallurgical proc-
esses for the extraction of copper from its ores or matte may be con-
sidered as:
Purely Chemical and
Electrolytic.
In the purely chemical processes the copper is dissolved and precip-
itated by chemical reagents; in the electrolytic processes, the copper is
dissolved chemically but the precipitation is effected electrolytically,
accompanied, usually, by regeneration of the solvent.
Chemical Processes. — These may be classified as follows, based mostly
on the solvent employed:
Alkali processes,
Sulphite processes.
Sulphate processes.
Chloride processes.
Nitric acid, by means of which the copper would be dissolved as the
nitrate or sulphate, has been frequently suggested as a solvent of copper
from its ores. The fixation of atmospheric nitrogen by electricity, offers
a cheap way of producing nitric acid at the mines. There are, however,
inherent difficulties to the use of nitric acid which makes its application
questionable. Nitric acid is the best known solvent of copper, but it is
also an excellent solvent of all the impurities in the ore, so that insur-
mountable difficulties may be expected, both in the solution of the copper
and in its precipitation, if regeneration of the solvent is desired.
The applicability of any solvent process to the extraction of copper,
depends fundamentally on the character of the ore. All acids, likely to
be used in a solvent process, react more or less with other elements;
when so consumed the acids are unavailable for useful work, and fre-
quently bring into solution ingredients which are positively harmful
The elements most detrimental to acid processes are:
Calcium,
Magnesium,
Aluminum,
Zinc,
Manganese.
169
170 HYDROMETALLURGY OF COPPER
To these may be added iron, arsenic, antimony, bismuth; but these ele-
ments need not necessarily be fatal to an acid process, no matter in what
proportion they occur in the ore.
If lime, magnesia, zinc or manganese occur in the ore in large quanti-
ties, acid processes are not applicable. What the limit is, can only be
determined by direct experiment. Chemical analysis of the ore, while
instructive, cannot be relied upon to determine the applicability of an
acid process. Lime, for example, is only detrimental in certain combina-
tions, as the oxide or carbonate. In many ores where sulphuric acid has
been a factor in the deposition or in the oxidation of the vein matter,
much of the calcium will be found as sulphate, which is not particularly
injurious either in a sulphate, sulphite or chloride process. All of this
calcium, however, would usually be estimated as lime (CaO), although it
occurs as sulphate (CaSOJ. Magnesia is highly injurious as oxide and
carbonate, but magnesia is not as widely distributed as lime, in injurious
amounts.
Alumina is widely distributed, but its presence, while undesirable, is
not necessarily particularly injurious. Much depends on its mineralogical
combinations. In Cripple Creek, sulphuric acid has been used for many
years in connection with the chlorination of those ores, which contain
from 15 to 20 per cent, alumina. Zinc, especially as the oxide, is in-
jurious, because it is readily soluble in acids, and as yet no practicable
method has been found for its economic- precipitation. Electrolytic
precipitation offers a plausible way of recovering the zinc, and is in
practical use in several plants, but its general adoption is by no means
assured.
Many oxidized ores are improved by roasting. All sulphide ores, with
the possible exception of certain chalcocite deposits, should be roasted
before chemical treatment, no matter what the nature of the chemical
treatment may be.
The treatment of raw sulphide ores has never met with much encour-
agement, and the cause for this is reasonable enough. The highly oxi-
dized ore is in the best possible condition for the application of any
solvent, and it is difficult to conceive of any oxidizing process, or sub-
stitution for an oxidizing process, cheaper and more satisfactory than
roasting. It is true that low grade ores have been treated in Spain and
Portugal by natural weathering, and for a while with ferric chloride or
ferric sulphate, but the use of ferric chloride has long since been aban-
doned, and the slow process of weathering, in which years are required
to get an adequate extraction, is perhaps nowhere else applicable.
It may be considered, therefore, in hydrometallurgical processes,
that the application of the solvent is to the oxidized ores.
Many of the metal compounds as found either in raw or roasted ore,
have the faculty of reducing the ferric to the ferrous salts, the respective
HYDEOMETALLURGICAL PROCESSES 171
metals being thereby brought into solution. This is notably the case
with ferric sulphate, FejCSOJ 3, and ferric chloride, FeClj. Both of these
substances have been used, and have been extensively experimented with
in the reduction of copper ores, by hydrometallurgical methods. The
leaching method at present used at Rio Tinto, Spain, is based mostly
on the solvent action of ferric sulphate, and the Doetsch process, formerly
used there extensively, was based on the solvent action of ferric chloride.
CHAPTER X
CHEMICAL PROCESSES
ALKALI PROCESSES
The alkali processes have not met with much encouragement in the
hydrometallurgical extraction of copper from its ores. This is due largely
to the slow and low solubility of copper in a solution of the alkalies.
Ammonia and ammonium compounds are about the only alkaline solvents
which have been tried on a commercial scale. The oxides and carbon-
ates of copper are quite readily soluble in ammonia but the solution should
take place in tight receptacles as the volatility of the gas in aqueous
solution is quite perceptible. If ammonium carbonate is used on cal-
careous ores there should be no sulphates present, because they would
be decomposed into ammonium sulphate and calcium carbonate.
After the copper is dissolved the ammoniacal copper solution is
boiled, and the black oxide of copper precipitated. The ammonia vapor
boiled off may be condensed in towers and used on another charge of ore.
The recovery of the ammonia from the salts formed in the boiled-out
solution may be accomplished by means of lime and steam.
According to Schnabel "■ experiments hitherto tried in using ammonia or
ammonium carbonate have failed because ammonia-tight vessels were not
employed, and because the precipitation of the copper, for which iron
cannot be employed, was performed by means of hydrogen sulphide,
calcium sulphide, or barium sulphide. By the use of iron vessels, how-
ever, loss of ammonia may be avoided, but such apparatus has proved
complicated and expensive to operate. The oxides of zinc, nickel, cobalt,
etc., are also soluble in ammonia or ammonium carbonate.
Sodium carbonate has been suggested as a solvent of copper from
oxide and carbonate ores, but the solubility of copper in sodium carbon-
ate is so unsatisfactory that experiments along these lines have not been
encouraging. Copper is also slightly soluble in concentrated solutions
of the caustic alkalies. Schneider ^ purposes increasing this solubility by
the addition of glycerin.
The Mosher -Ludlow Ammonia-cyanide Process.' — This process depends
upon the principle that ammonia, NH3, at the ordinary temperature
forms soluble, stable compounds with the oxides, hj'droxides or carbon-
ates of copper, zinc, nickel, or cobalt, such as Cu(NH3)2.
'Handbook of Metallurgy, Vol.' I, p. 204.
^U. S. patent 932, 643, Aug. 31, 1909.
'Electrochemical and Metallurgical Industry, March 1908.
172
CHEMICAL PROCESSES
173
These ammonia metal compounds are readily dissolved by water
containing a small excess of ammonia over that required to form the
soluble compound. This is the leaching step of the process.
The step of precipitation depends on the fact that those soluble
ammonia-metal compounds break up with great ease at the boiling
point of water into the oxide or hydrate of the metal, which almost
instantly settles as a heavy precipitate, while the ammonia, originally
combined, is set free to be reabsorbed in cold water or boiled-out solution
for use over and over again.
Pump
Fig. 44. — Moser-Ludlow ammonia-cyanide process.^ Diagrammatic sketch.
Where the percentage of copper is large the aim is to first extract as
much of the copper as possible by plain ammonia, and to leave the gold
and silver values to be subsequently extracted with a weaker ammonia
solution containing fractional percentages of potassium cyanide. But
instead of working it in this way it may be preferable in many instances
to add the cyanide at once to the ammonia and to simultaneously ex-
174 HYDROMETALLURGY OF COPPER
tract all the values, including copper, gold, and silver, with an ammoni-
acal solution containing one to several pounds of cyanide per ton. The
object aimed at is to reduce the comsumption of cyanide to a minimum
in the presence of copper, thereby permitting the minute amounts of
potassium cyanide added to the ammonia solution to simultaneously
extract the gold and silver values.
To recover the metallic values from the ammonia-cyanide copper-gold-
silver bearing solution, it is passed through a continuous boiling-out still,
to precipitate the copper as CuO. The boiled-out solution holding the
gold and silver values is agitated with the least amount of zinc dust, or
passed through zinc boxes to recover such gold and silver as the boiled-
out solution may contain.
For treating the ore containing considerable percentages of ammonia-
soluble metals, the chlorination barrel, without lead lining is recommended.
In the treatment of slimes, agitation in a closed conical tank, with subse-
quent filtering with any of the well-known filters.
On account of the powerful oxidizing action of a solution of cupric
oxide (CuO) in ammonia, unoxidized silver minerals may be attacked and
finally dissolved in the raw state, obtaining in this manner a percentage .
of extraction out of this character of mill product entirely impossible by
ordinary cyanide methods.
It is of the greatest importance to have a thoroughly ammonia-tight
equipment, especially in the second half of the process, which comprises
the boiling out of the ammonia solution. This apparatus is designed so
that the boiling out is carried on continuously, and it may in some way be
compared to the artificial ice and cold storage apparatus in which the
ammonia water is boiled in a stUl, the ammonia gas distilled off, liquefied
under pressure by powerful pumps, then permitted to expand, by which
the cooling effect is produced, and finally reabsorbed in cold water to
commence the same cycle of action over and over again.
In the Mosher-Ludlow continuous boiling-out apparatus the in-
coming cold ammonia-copper solution is brought in contact with the
heat of the ammonia-steam vapor in the cooler and condenser and in the
heat exchanger on its way to the boiling-out still, with the overflowing
boiled-out solution. In this way an important part of the heat applied
to boil out the ammonia is passed to the incoming solution, thus saving
steam and fuel.
The ammonia-copper solution is taken from the copper-solution tank
and pumped through the inner coil of a double-pipe counter-current
cooler; thence it passes to the heat exchanger, thence to the boiling-out
still, thence into the exchanger, where part of its heat is passed to the
incoming solution; thence to the settling tanks, where the precipitate,
which is almost pure copper oxide, is allowed to settle, the liquid being
drawn off into the sump tanks. From the sunip tanks the boiled-out
CHEMICAL PROCESSES 175
solution is pumped through a cooler up to the boiled-out solution tank.
From there it is used as needed to make fresh ammonia solution, and as
wash water.
The ammonia and steam vapor from the boiling-out still passes up to
the cooler and condenser into the annular spaces between the two pipes,
parting with a fraction of its heat to the incoming ammonia-copper
solution in the inner pipe. From the cooler and condenser the ammonia
water flows into the ammonia solution tanks where it is diluted and is
ready to be sent to the leaching system. A supply of cold water is kept
flowing over the outside of the cooler and condenser and the boiled-out
solution cooler. The precipitate is removed from the settling tank from
time to time, as required.
It is estimated that the heat required in boiling out 2 per cent, ammo-
nia-copper solution on a 100-ton (24-hour) basis, and precipitating the
copper as black oxide =79.9 per cent, copper, is 51,083,000 b. t. u. of
which, however, according to experience, 60 per cent, is saved by the
heat exchanger, so that the net heat required amounts to 34,333,000
b. t. u. In practice this required heat can be obtained from 1 1/2 to 2
tons of good coal, and with a boiler of from 35 to 40 h. p.
SULPHITE PROCESSES
A solution of sulphur dioxide in water may be regarded as sulphur-
ous acid, (HjSOg). Copper oxide and carbonate are soluble in sulphurous
acid, the sulphide is not.
Various methods have been suggested for the practical use of sulphur-
ous acid as a solvent of copper from its ores. All of these refer more di-
rectly to the precipitation of the copper from the sulphurous acid
solution, rather than the solution of the copper from the ore. Copper
sulphite is not soluble in water, but is readily soluble in excess of sulp-
hurous acid. Copper sulphite is an unstable salt, which is slowly
changed into a mixture of cupro-cupric sulphite and cupric sulphate, as
shown by the following reactions :
(1) 3CuO-F3S02 = 3CuS03.
(2) 3CUSO3 + CuO = Cu2S03,CuS03 + CuSO ,.
The cupro-cupric sulphite is only very slightly soluble in water, but
is quite soluble in sulphurous acid or in a solution of copper sulphate.
Neill Process. — This process consists, first in subjecting the ore to the
action of sulphurous acid to dissolve the copper, and second, in heating
the solution to drive off the excess of sulphurous acid and precipitate the
copper as sulphite.
If the ore to be treated is a sulphide it has to be roasted; if the ore is
an oxide or carbonate it may be treated without roasting. The sulphur
dioxide used in the process may be obtained from burning sulphur or
176 HYDROMETALLURGY OF COPPER
from roasting sulphide ores, which if they are copper ores, may be subse-
quently treated in the oxidized condition with the sulphur dioxide obtained
from the roasting.
There is less sulphur required to extract the copper as sulphite than as
sulphate. It takes 98 lb. of sulphuric acid to dissolve 63 lb. of copper, as
sulphate, while it only takes 32 lb. of sulphur, in the form of sul-
phurous acid, to dissolve the same amount of copper — 63 lb. The sul-
phurous acid may be applied either as a gas or in solution, or a combi-
nation of both.
In practice, the ore is crushed to a suitable degree of fineness and
charged into agitation tanks, and agitating the ore in contact with the
sulphurous acid; or the agitation may be effected in tanks having a conical
bottom, by forcing sulphurous acid mixed with more or less air, through the
pulp. In this way the sulphurous acid is applied to the ore while air
does the agitating. The tanks, or agitators, in which the ore is treated
may be closed at the top, so that the excess of gas issuing from the first
tank may be passed through the second and so on until the sulphur
dioxide is entirely consumed.
The clear solution, after being separated from the ore by either filtra-
tion or decantation, is heated to a temperature sufficient to drive out the
excess of sulphurous acid, thereby precipitating the copper in the form
of a bright-red and very heavy powder (cupro-cupric sulphite). The
cupro-cupric sulphite settles at once to the bottom of the precipitating
tank, and the supernatant liquor may be decanted or siphoned from it,
or the sulphite may be recovered from the solution by filtration. The
excess of sulphurous acid, driven out of the solution by heating, may be
recovered for re-use. The cupro-cupric sulphite after it has been removed
from the tanks, filtered and dried, contains about 50 per cent, metallic
copper. This precipitate may be heated in an oxidizing atmosphere in
a furnace and the cupric oxide produced, or it may be melted in a reducing
atmosphere, producing cuprous sulphide, which m%y then be reduced to
metallic copper by the ordinary converter process.
With ores suited to the process, the copper will pass into solution in
from 1 to 4 hours. When the copper is dissolved by sulphurous acid,
only very small amounts of other metals are dissolved, and the ultimate
product is a very pure copper.
Should any copper exist in the solution as sulphate, due either to
improper roasting, or of sulphur trioxide in the sulphurous acid, this
sulphate of copper will not be precipitated by boiling the solution, but
must be precipitated in some other way. This may be done either with
iron or by electrolysis.
If a solution of cupro-cupric sulphite is heated at a high temperature
(about 200° C.) and subjected to a pressure of about 25 lb., sulphur diox-
ide is liberated, and there is formed copper sulphate and metallic copper.
CHEMICAL PROCESSES 111
About one-half of the copper may in this way be precipitated in the
metallic condition.
Neill Process at Coconino, Arizona' According to Jennings, the orig-
inal method as carried out at Coconino, consisted in treating the ore,
with sulphur dioxide, in a series of upright tanks 8 ft. in diameter and
18 ft. high. About 5 tons of ore were introduced into the tank half filled
with water, and gas was forced, by means of a compressor, into this mix-
ture of ore and water, the excess gas passing from the first tank to a
similar one, also charged with ore and water, and thence to a third tank,
where it was supposed the absorption would be complete. The gas was
not all used up owing to the difficulty of absorbing sulphur dioxide in
water when mixed with large volumes of air. When the ore in the lower
tank was leached, an operation which usually took 10 hours' time, the
solution and the leached ore together were dropped into a pressure tank
and thence passed into a large filter press. The filter press was a con-
stant source of trouble as it was impossible to find a material for the
filters which would stand any length of time.
The solution from the filter press was heated by waste steam from the
crushing plant and 60 per cent, of the copper precipitated as cupro-cupric
sulphite.
Jennings gives the weak points of the Neill process as carried out at
Coconino, as: 1. The attempt to saturate the water by simply blowing
the gas through it; 2. the poor agitation obtained and the consequent
length of time required to leach a comparatively small amount of ore;
3, the dilute solution obtained, 1 per cent, being the maximum amount
of copper which can be held in solution by an excess of SOj; 4. the ease
with which the copper separates from these solutions, both in the leaching
tanks, the pressure tank and the filter tanks, forming the cupro-cupric
precipitate throughout the mass of leached ore, and which it was impos-
sible to redissolve with sulphurous acid, and 5. the difficulty of treating
the remaining 40 per cent, of the copper in solution as sulphate after the
60 per cent, has been precipitated. Scrap iron was not cheaply available.
It is evident that many of the weak points here enumerated by
Jennings should not present much difficulty in a well designed plant.
The effective absorption of gases in liquids, has long been satisfactorily
accomplished in the chemical industry, by some means of subdivision;
agitation is accomplished successfully on an enormous scale in the cyanide
process, where it is necessary to bring air in contact with the ore and
cyanide solution, to effect extraction. In experiments made by Jennings^
on the same, or similar copper-bearing triassic sandstones of northern
Arizona, he succeeded in getting an extraction of 9.5 per cent, of- the
copper, with sulphur dioxide, by leaching the ore 4 hours. The low
copper content of the solutions was due to the small excess of sulphur
'E. P. Jennings, E. and M. J., Jan. 18, 1908.
'M. and J., March 30, 1901.
12
178 HYDROMETALLURGY OF COPPER
dioxide. In later experiments^ Mr. Jennings succeeded in getting as
high as 2 per cent, copper in the solutions.
It is evident that, so far as the operations at Coconino are concerned,
it leaves the process where it was before the plant was erected. It
demonstrated neither technical failure nor success.
Neill, the inventor of the process, in experiments carried out in Salt
Lake City" succeeded in getting a complete extraction of Coconino ore,
having 11 per cent, copper, in 6 hours. He readily obtained a solution
carrying 2.5 per cent, copper, and so long as the solution remained cool
no difficulty was experienced with the copper separating out in the sands,
or in washing the sands.
An experimental plant, installed at the smelter of the Montana Ore
Purchasing Company at Butte, by Neill, gave interesting results. The
material was roasted to about 2 per cent, sulphur content. It was then
placed in a wooden barrel 6 ft. in diameter and 12 ft. long, and SOj gas
from the roasting furnace was blown through the hollow trunions and
brought into more intimate contact with the pulp by means of wooden
paddles arranged on the sides and periphery of the barrel. Two tons of
roasted ore were charged with 5 tons of water, and after passing the
dilute roasting gases through the barrel for 6 or 8 hours the copper was
successfully extracted. The barrel was dumped into a settling tank, the
solution drawn off by percolation, and the sands washed in the same way.
This washing was difficult on account of the ferric oxide, which being
flocciilent, remained in suspension and formed a layer upon the top of
the sands which it was difficult to percolate. The sands after this incom-
plete wash averaged 0.8 per cent, copper, but average samples taken from
the tank and washed by agitation and decantation, gave final tails of
0.31 per cent, copper. The heads averaged 3.15 per cent., showing an
extraction of 90 per cent. The solution was heated in a wooden tank by
a steam coil, and the precipitates, which were slightly contaminated with
alumina, on account of the poor filtration, amounted to 64 per cent, of the
copper extracted. The remainder was precipitated upon iron in a splash
tank and the final solution turned to waste carried only traces of copper.
The amount of iron consumed was exceedingly small and the reaction
very quick, owing to the fact that 'the solution came from the steam
tanks at nearly the boiling point. •
There were 35 tons of material treated, and the figured cost of the
operation compared favorably with anything being done at that time or
now in the Butte district. The process was not adopted because at that
time the silver content of the company's ore was high and could not be
saved by this method, and the space necessary for the plant was not
available.
'E. and M. J., April 18, 1908.
'E. and M. J., March 14, 1908.
CHEMICAL PROCESSES 179
Van Arsdale Process.' — The van Arsdale process consists in precipitat-
ing copper from cupric sulphate solutions and simultaneously producing
sulphuric acid, by adding to solutions of cupric sulphate, sulphur dioxide
and heating with or without pressure. The copper is thrown down in
solid form which may be subsequently treated, while the regenerated
acid is applied to the ore for the extraction of mor^ copper.
The ore must contain the copper as oxide or carbonate. The original
solution for leaching must contain cupric sulphate and should contain
ferrous sulphate. The precipitation will proceed with cupric sulphate
alone, resulting in the formation of a salt of copper and the formation of
sulphuric acid; by the addition of ferrous sulphate the reaction proceeds
better, and by the use of a proper proportion of ferrous sulphate there is
obtained a precipitate of metallic copper. In practice the solutions
will always contain more or less ferrous sulphate dissolved from the ores
treated. Good results are obtained from a solution containing approxi-
mately 10 per cent, each of cupric sulphate and ferrous sulphate. To
precipitate the copper, sulphur dioxide is applied to the solution to
nearly saturation. The solution is then heated to nearly the boiling
temperature, whereupon a reaction takes place, resulting in the precipita-
tion of a part of the copper contained in the solution either as metallic
copper or compounds of copper, or both, together with the formation of
sulphuric acid. The amount of copper precipitated will vary with the
composition of the solution and also with the pressure under which it
is heated. When the solution is heated under pressure an increased
precipitation of copper is obtained, amounting to about 50 per cent, of
the copper originally present, and an amount of sulphuric acid regener-
ated amounting to about double that necessary to redissolve from ore
the amount of copper precipitated. The process being cyclic, there is
no particular harm in returning to the ore a solution containing a con-
siderable amount of unprecipitated copper sulphate.
The chemical reactions involved in the process may be given as:
(1) 3CuS04-F3S02 + 4H20 = CuS03,Cu2S03+4H2SO,.
(2) Cu,S03,CuS03 + 4H2SO, = Cu-t-2CuSO, + 2H2SO,+2S02-F2H,0.
At atmospheric pressure and in the cold, a solution of cupric sulphate
saturated with sulphur dioxide, after standing for some time, deposits a
small amount of cupro-cupric sulphite. On heating such a solution to
boiling and at the same time passing sulphur dioxide through it, a larger
amount of copper is precipitated, resulting finally in a precipitate of
metallic copper, according to equation No. 2. When the solution, how-
ever, is saturated with sulphur dioxide and placed in a closed vessel and
heated under pressure, the yield of precipitated copper is increased to
'E. and M.J., June, 1903; U. S. Patent, March 31, 1903, No. 723,949.
180 HYDROMETALLURGY OF COPPER
40 and 50 per cent, of the copper originally present, and free acid is
formed according to the above equations.
The degree of heat and pressure required for the second operation
are not high, it being only necessary to heat the saturated solution to
nearly 100° C, the pressure produced being about 30 lb. to the square
inch. A lead lined steel tank may be used for this purpose.
Jumau found' the following relations of temperature and proportion
of copper precipitated from a copper sulphate solution saturated with
sulphurous acid, having originally 25 grm. of copper sulphate per liter.
Temperatures Copper precipitated
140° C. 47 per cent.
155° C, 62 per cent.
167° C, 65 per cent.
190° C, 79 per cent.
SULPHATE PROCESSES.
Ordinarily only oxidized ores such as the oxides and carbonates, can
be treated so as to dissolve the copper as sulphate. Sulphides are not
usually amenable, practically, to direct treatment. Roasting is desir-
able. After the ore is roasted the copper should be in the form of oxide
and sulphate, although in improperly roasted ore sulphides may still be
present. The roasted material is then treated the same as naturally oc-
curring oxidized ores. The copper may be dissolved as sulphate either by
Sulphuric acid, or
Metal sulphates,
such as the ferric sulphate. The acid, however, is the solvent most uni-
versally employed.
If iron is used as the precipitant, as it usually is, the sulphuric acid
process consists essentially of applying dilute sulphuric acid to the oxi-
dized ores of copper, which reacts with the oxide of copper as follows:
CuO-|-H2S04 = CuS04-hH20.
The copper sulphate thus formed is filtered from the ore and precipitated
with iron, thus:
CuSO,-f-Fe = Cu+FeSO„
the iron and the copper changing places. The copper is precipitated
while the iron goes into solution as ferrous sulphate. From this reaction
it will be seen that, theoretically, it takes 98 lb. of sulphuric acid to
dissolve 63.6 lb. of copper, and 56 lb. of iron to precipitate 63.6 lb. of
copper; 1.56 lb. of sulphuric acid to dissolve 1 lb. of copper and 0.88 lb.
of iron to precipitate it. These equivalents are for pure acid and pure
'U. S. Patent 930,967, Aug. 10, 1909.
CHEMICAL. PROCESSES 181
iron. Commercial sulphuric acid always contains more or less water and
other impurities, and in iron also, there is more or less foreign matter.
So it may be safely assumed that it will take 1.75 lb. of ordinary commer-
cial sulphuric acid to dissolve 1 lb. of copper and 0.95 lb. of iron to precipi-
tate a pound of copper.
The problem of the commercial treatment of ores by sulphuric acid
is, however, much more complicated than the simple process here outlined.
If the ore contains foreign elements attackable by sulphuric acid, the
acid consumption may be excessive, and if the copper sulphate solution
entering the precipitation tanks contains free acid or ferric sulphate the
consumption of iron for precipitation will also be excessive.
To determine, therefore, whether any ore may be economically
treated by a sulphate process, it is necessary to make a direct test on the
ore, and thus ascertain the consumption of acid. The iron for pre-
cipitation may be taken, in practice, at 1.5 to 2.0 lb. of scrap iron per
pound of copper produced. These factors being known, close approxi-
mations can usually be made as to the commercial applicability of the
process in its simplest form.
Many improvements on the above simple process have for their basis
the cheapening of the solution of the copper, but most of the improve-
ments are based on the precipitation and regeneration of the solvent.
If iron is used as the precipitant in the simple process as outlined, the
acid is lost, so that new acid has to be supplied to the ore at every cycle
of solution. This being the case, sulphuric acid installations may be
made at the mine using the sulphide ore for the production of the neces-
sary sulphur dioxide, for the manufacture of acid. The ore, after roast-
ing, may be treated with the acid so produced, to extract the copper,
which is in the form of oxide or sulphate and is readily amenable to the
process.
Sulphide of copper, in ores, may be converted into sulphate by the
action of ferric sulphate, thus:
xH2SO, + Cu2S+2Fe2(SOj3 = CuSO, + 4FeSO,-FS+xH2SO„
in which x represents an indeterminate amount of the sulphuric acid.
Ferric sulphate has not been used independently for method of extraction
on roasted oxidized ores, owing to its slow action. In imperfectly
roasted ore its presence may be beneficial by promoting the formation of
sulphate from the remaining undecomposed sulphides, should any be
present.
Sulphur dioxide, steam, and nitrous fumes have been applied to ore
to dissolve the copper as sulphate; this amounts, essentially, to the
manufacture of sulphuric acid within the ore mass.
In the application of sulphuric acid for the extraction of copper,
cupric oxide, azurite, malachite, and arsenate of copper dissolve readily;
phosphate of copper with more or less difficulty. Cuprite (cuprous
182 HYDROMETALLURGY OF COPPER
oxide, CujO) , is not readily soluble, but if moistened with acid and left
exposed to the air for some time, it is transformed into the cupric oxide,
and is then readily soluble.
According to SchnabeP at Stadtberg in Westphalia and at Linz on
the Rhine, ores containing 1 to 2 per cent, of copper were sulphated by
means of sulphur dioxide steam and nitrous gases, and then leached.
At Stadtberg the ores were azurite and malachite disseminated through
the quartzose shist, at Linz, at the Stern works, copper carbonates and
phosphates. The leaching vessels were tanks of brickwork 3 ft. 3 in.
deep. Above the bottom proper these had a false bottom of grating,
made of fire brick or other acid proof material, supported by bricks on
edge. The ores were piled up on this grating, the gases being conducted
underneath it. The gases were respectively generated by roasting iron
pyrites in shaft furnaces and zinc blende in muffle furnaces, and by treat-
ing Chili niter with sulphuric acid. The sulphur dioxide, nitrous fumes,
and steam together formed sulphuric acid which converted the copper
compounds into sulphates. After 8 to 10 days the copper sulphate was
dissolved out by means of water or of the acid mother liquor left after
precipitating the copper. The leaching was so conducted that fresh water
or the copper-free mother liquor was allowed to attack the most com-
pletely exhausted ore, while the almost saturated solution was run on to
fresh ore until is was fully saturated (22 to 26° B). The liquor that
drained away ran into receivers, whence it was again pumped on to the
ores. This process, which extracted the copper down to 1/4 per cent.,
has long ago been abandoned. At Stadtberg sulphuric acid was re-
placed by the cheaper hydrochloric acid as long as the supply of oxi-
dized ores lasted.
Acid Plants at the Mine. — Conditions are frequently ideal for acid
manufacture, for leaching purposes, at the mines, if suitable sulphide ore
is available. If the mine produces both sulphide and oxide ores, the
sulphides may be roasted and the sulphurous gases converted into acid
which may then be used to leach both the oxidized and roasted ores.
If the sulphide ore is low in sulphur, concentration will usually be neces-
sary to get a material sufficiently high in sulphur to make a gas suitable
for sulphuric acid manufacture.
The sulphide ore, or concentrates are usually roasted in a furnace of
the McDougal type, and the sulphur dioxide gas passed from the roasting
furnace into a series of leaden chambers, where coming in contact with
gaseous nitric acid and steam it becomes converted into sulphuric acid.
The nitric acid gas is produced by the action of sulphurc acid and nitrate
of soda, and passes along with the sulphurous gases into the lead cham-
bers. The combined gases, together with the seam and air, mix in the
chambers and condense as sulphuric acid. This is known as chamber
'Handbook of Metallurgy, Vol. I, p. 200.
CHEMICAL PROCESSES 183
acid and has a specific gravity of about 52° B. and contains about 65.2
per cent. HjSO^. The chamber acid, in commercial plants, is then con-
centrated to 66°B. containing 93.5 per cent. H^SO^, and this is the ordi-
nary acid of commerce, or oil of vitriol.
In the manufacture of sulphuric acid, for leaching purposes, the process
will be somewhat cheaper than the manufacture of 66° B. acid for com-
merce, because no purification will be required, and the chamber acid
may be used without further concentration. On the other hand, the
manufacture of acid for leaching purposes will usually be conducted on a
small scale, and hence the cost of operation, per unit of acid, will be largely
increased over that of the large commercial plants.
In the manufacture of sulphuric acid, a chamber space cf from 15 to
25 cu. ft. should be provided per pound of sulphur burned in 24 hours, and
nitrate of soda will be consumed in amounts varying from 3 to 5 per cent,
of the sulphur burned.
The cost of manufacturing the acid at the mine will vary within wide
limits. The selling price of 66°B. acid, f. o. b. works, in commerical
plants, is about $18.00 per ton, but this cannot be made the basis of costs
the mine, for the reason that such acid is made under the best possible
economic conditions. On the other hand, if acid is manufactured at the
mine, assuming the ore to be a sulphide or to contain sufficient sulphides
for acid manufacture, there would be no expense for roasting and no
expense for sulphur, for in any event, the suphides would have to be
roasted, and the sulphurous gases would otherwise be wasted.
Sulphuric Acid Leaching of Oxidized Copper Ores, at Clifton, Arizona.
— The Arizona Copper Company have been leaching oxidized surface
ores at Clifton, on a large scale, since 1893. The following description
of the work was prepared by F. N. Flynn, the Company's Metallurgist.'
"Four groups of mines, in the Metcalf District, have contributed this class
of ore, but the Metcalf mine has furnished the principal part of the tonnage.
The occurrence of the ore, and the method of mining have been described by
Mr. Peter B. Scotland, in the Eng. and Min. Jour., July 16, 1910.
" The ores are lowered down the hillside by means of inclined tramways to the
railroad bins at Metcalf. Trains of 40-ton, bottom dump cars are hauled over
the Coronado Railway, 36-in. gauge tracks, to Clifton, 6.6 miles distant. From
the railroad bins, which are common to the various departments of the works,
the ore is conveyed by a 30-in. belt to the "Oxide Mill" bin, located near the
smelting plant.
" The gangue of the ore partakes of the character of the mine formations —
highly altered sedimentary and igneous rocks. The granite-porphyry, quartzite,
shale, and limestone have all been more or less altered, resulting in a mass of
quartz grains, kaolin, serecite, silicious hematite, magnetite, limonite, garnet and
'Private communication from Norman Carmichael, Gen. Manager, Arizona Cop-
per Co., Nov. 1, 1911.
184 HYDROMETALLURGY OF COPPER
various other silicates of alumina. Fortunately the limestone was completely
altered, and the calcium sulphate almost completely removed.
"The copper-bearing minerals, in the order of their importance, are: Mala-
chite, copper-pitch-ore, azurite, impure chrysocolla, cuprite, chalcocite, chalco-
pyrite, native copper and brochantite. Malachite occurs throughout the upper
part of the deposit, and is the all-important mineral. The impure chrysocolla is
neither brochantite nor pure chrysocolla. Specimens of this impure mineral,
containing 22 per cent, of copper, contain, but 0.05 per cent, of sulphur. The
copper is readily soluble in very dilute sulphuric acid without effervescence.
Cuprite occurs in the shale, usually unassociated with other copper minerals,
except native copper in small amounts. It occurs in very thin flakes, and in such
a manner as to suggest its deposit direct from copper sulphate solutions, the
latter having been hydrolized by absorption. Chalcocite is found in the porphyry,
partly and completely replacing the pyrites.
"The milling ores vary between 2.5 and 3.0 per cent, copper, and analyze
about as follows :
SiOj, 59 . 0 per cent.
AI2O3, 18.0 per cent.
Fe, 9 . 0 per cent.
Mn, 0 . 1 per cent.
CaO, 0 . 1 per cent.
MgO, 0 . 05 per cent.
S, 1.0 per cent.
Au, Ag, Pb, and Zn, traces.
About 315 tons of crude ore are treated per day.
The crushing plant consists of :
One 10-in. by 20-in. jaw crusher.
Two sets of 12-in. by 36-in. rolls.
One belt elevator.
One trommel, with 3/4-in. and 5/8-in. holes when new.
Water is fed under the first set of rolls.
The 3/4-in. oversize goes to the second set of rolls.
The 3/4-in. to 5/8-in. size is finished product, also the 5/8-in. undersize.
Nothing is reorushed. The crushing plant and jigs run 17 hours per day.
" Material between 3/4-in. and 5/8-in. goes to one two-compartment Hartz jig,
making top concentrates and tails. The hutch product returns to the ore stream.
"The 5/8-in. undersize, including the crushing water, goes to one five-ccm-
partment Hancock jig, making concentrates and tails.
"The concentrates analyze about as follows:
SiOj, 37.0 per cent.
AI2O3 11.0 per cent.
Fe, 22 . 0 per cent.
Mn, 0.2 per cent.
CaO, 0,04 per cent.
MgO, 0.1 per cent.
S, 5.0 per cent.
Cu, 7.0 to 10.0 per cent.
CHEMICAL PJiOCESSES Ls:,
"The [specific gravity of the minerals entering the concentrate runs between
3. 7(1 and 4.63 on sizes larger than 1/8 in. Eighty-five per cent, of the con-
centrate is larger than eighth inch. From 20 to 2.) per cent, of the values are
recovered as concentrate, at a ratio of about 10 to 1.
"A belt elevator lifts the concentrates to a bin on the smelter charge floor,
from which they are wheeled direct to the blast furnaces.
"The tails bin serves as a dewaterer. The slime water is pumped to settUng
ponds three-quarters of a mile distant. The slime carries 2.4 per cent, copper and
too much soluble alumina to permit of leaching it with the gravel in the present
plant. The tailings vary from 1 in. (due to the wear of the trommel) down to the
finest sands; 75 per cent, is larger than eighth inch. The tailings are hoisted by
inclined skip to a receiving bin, centrally located over the top of the leaching
tanks.
"The Joy Mine at Morenci furnishes the pyrite for the manufacture of sul-
phuric acid. The pyrite analyses:
SiOj, 9.0 per cent.
AI2O3 4.0 per cent.
Fe, 38.0 per cent.
S, 38.0 per cent.
Zn, 2,0 per cent.
Cu, Variable; usually under 1.5 per cent
The pyrite is crushed to 2 in. The fines are roasted in a Herreshoff five-deck
furnace. The coarse material goes to the lump burners. The cinder goes to the
blast furnaces.
"The resulting acid is 52° Baume. The capacity of the acid plant is 10 tons
per day.
"In the leaching department, small circular wooden tanks are used. There
are twelve ore tanks of 13 tons capacity each, or 156 tons total capacity. These
have an inner lining of lighter staves. Between the outer and inner staves an
acid proof cement preparation is used. The tops of the staves are covered with
sheet lead. A false bottom of plank, slightly inclined toward the center, and
perforated with hoels. serves as a filter bottom.
"The ore (jig tails) is charged to the tank by means of fixed launders over-
head, connected with the ore supply bin at the top of the mill.
" The leached tailings are discharged through a circular opening in the bottom
of the tank at the center. The opening is closed bj* means of a wooden plug,
extending up through the ore charge, and suspended from chain blocks. The
tails are flushed from the ore tanks and hoisted to the railroad bins, from which
they are hauled away in 50-ton bottom dump cars. Each ore tank is provided
with a belted centrifugal pump for circulating liquors from bottom to top of
tank, or to any other tank in the series.
" Ordinary circular wooden tanks are used in the precipitating room. Two
square tanks are used for final precipitation. In each of these tanks is a revolving
drum or trommel made of cast copper. The circumference is perforated with
holes like a trommel. The ends are closed and support the axles. A small door
permits charging the trommel with small pieces of wrought iron. The drum is
submerged in the solution up to the bearings, and revolves slowly.
186 HYDROMETALLURGY OF COPPER
The department is arranged on three floors. The top floor for the ore tanks,
the middle floor for the precipitating tanks and trommels, and the ground floor
for cement copper.
"The ore charge (jig tails) analyze:
SiOj, 60.0 per cent.
AI2O3, 16.0 per cent.
Fe, 8.0 per cent.
Mn, 0. 1 per cent.
CaO, 0 . 06 per cent.
MgO, 0 . 03 per cent.
S, 0.4 per cent.
Cu, 2 . 0 to 2 . 6 per cent.
"In practice three tanks are leached as a unit as regards solutions, although
the tanks are charged and discharged singly and at intervals.
"A charge of the solution circulates in each of the three tanks for 4 hours, or
12 hours total time, more or less. In this manner a charge of ore is leached with
three strengths of liquor, approximately as follows :
4 hours with ' Strong Liquor.' High in copper and low in acid.
4 hours with 'Weak Liquor.' Lower in copper and higher in acid.
4 hours with 'Acid Solution.'
0 . 5 hours with Water.
0 . 5 hours Charging and Discharging.
13 hour cycle.
" Sulphuric acid is added to the resulting wash water, to make the 'Acid
Solution,' which in turn makes 'Weak Liquor,' 'Strong Liquor' and 'Copper
Liquor.'
"The 'Copper Liquor' goes to the precipitating tanks with a very small
fraction of a per cent, of free acid. Usually the free acid is too small to
determine.
"The acid consumed in leaching, per pound of pure copper recovered, amounts
to 2.6 lb. of 52° B. acid.
"The 'Copper Liquor' from the various tanks is collected in a distributing
tank, from which it flows continuously in a small regulated stream to the pre-
cipitating tanks. The precipitating tanks are connected in series, and use scrap
iron of all descriptions — all of which comes from the works.
"From the last tank in the series the liquor, low in copper, goes to the 'trom-
mel tank' in 'charges.' Small pieces of wrought iron are charged to the trom-
mel. Usually this consists of ties from cotton bales, cut in strips of a foot in
length.
"A 'charge' of liquor requires from 10 to 30 minutes in contact with the
revolving trommel to complete the copper precipitation. The finished charge
containing its cement copper is flushed to settling tanks.
"The decanted liquor, practically free from copper, either in solution or as
precipitate, is passed through another set of overflow tanks, containing tin cans,
before going to waste.
"The resulting 'Iron Liquor' is pumped to an earthen reservoir, where it
soaks into the ground.
CHEMICAL PROCESSES 187
" The cement copper, after settling to separate the solution, is accumulated in
the patio, and, when sufficiently dry, is moulded into large bricks, by hand.
After sun drying for a month, they are sufficiently dry to be fed to the copper
converters. The cement copper averages 72 per cent, copper.
"The plant is handled by a superintendent and two white foremen with
Mexican laborers. The foremen make frequent mill tests for copper, impurities
and free acid. The working strength of all solutions is governed by Baume
readings.
"No general rule can be followed for the strength of the various solutions,
because the ore is quite variable in composition, both as regards copper and other
soluble salts. When ores are met with which show a readily soluble gangue, the
acid strength of the solution is reduced from that used or on more silicious
ores."
Leaching Plant at the Snowstorm Mine. — At the Snowstorm mine,
Larson, Idaho, there has been in operation for some years a leaching
plant of 250 tons daily capacity. The ore deposit of the Snowstorm
mine consists of disseminations of bornite, chacolcite, and chalcopyrite
in certain beds of Revett quartzite. The greater part of the sulphide
has however, been oxidized to cuprite, malachite, and chrysocolla. The
various prospects are on metasomatic fissure veins carrying chalcopy-
rite, chalcocite or bornite, with quartz, dolomite, or siderite. In the
lower workings of the mine the ore occurs as sulphide, containing only a
very small portion of the copper in the oxidized condition, but no attempt
has been made to leach the sulphide ore.
The oxidized ores average from 2 1/2 to 3 1/2 per cent, copper, 7 oz.
in silver, and $1.00 in gold, per ton. The ore is crushed and run into three
agitators, where it is treated with bleaching powder and a 10 per cent,
solution of sulphuric acid. By this method chlorine is slowly released,
which chlorinates the gold and silver, and to some extent attacks the
small quantities of sulphide in the oxidized ores.
The copper solution goes from the agitators through a series of six
settling tanks, after which it is precipitated with scrap iron. The resi-
dues, containing the silver chloride, are treated with sodium thiosulphate
(hyposulphite) to dissolve the silver, and the solution so obtained is passed
through settling tanks and the silver precipitated from the clear solution
with sodium sulphide. The silver sulphide precipitate is filtered and
shipped for refinement. The process is said to save 90 per cent, of the
assay value of the ore.
Copper Leaching Plant at the Gumeshevsky Mine, Russia.' — The
Sissert property, of which the Gumeshevsky copper mine is a part, is one
of the largest concessions in the Urals and was originally obtained for
working iron ore deposits. As early as 1727 two important copper de-
'Inst. of Min. and Met. Bull., No. 65; Trans. I. M. M., XIX, 212; Min. Ind.,
1910, 210.
188 HYDROMETALLURGY OF COPPER
posits were discovered on the property. These were worked intermit-
tently for more than 100 years.
The Gumeshevesky mine was shut down in 1871. From old data
available it appears that the mine is a contact between limestone and
diorite. Oxidized copper ore occurs in a clay formation along the con-
tact where there are old workings, extending about 2 miles in length
and possibly 1000 ft. wide. The mine was worked for the oxidized ores
only. The deepest shaft is 500 ft. The material raised from the shaft
was evidently hand picked and only the large lumps of oxidized ore
saved.
The dump consists of clay material with the fine ore that escaped
hand picking. It covers an area of about 20 acres, with an average
depth of 17 ft. This was thoroughly sampled and estimated to contain
531,000 cu. yd. of material carrying 23 lb. of copper per cubic yard (about
0.79 per cent).
In brief, the process of treatment consists of leaching the crushed
material with dilute sulphuric acid, then precipitating the copper from
solution with pig iron.
The owners of the property contracted with manufacturers of acid to
erect a plant to use iron pyrite from a deposit about 4 miles from the
dumps, and to sell to them, during a period of 10 years, 53°B. acid at
.'$4.32 per long ton (0.19 cents per pound). The contractors were to ex-
tract the copper from the pyrite paying the estate a royalty of $145 to
$175 per ton of copper produced, and were bound to burn 4000 to 4800
tons of pyrite annually. The pyrite was estimated to contain 3.5 to 8.0
per cent. Cu, and the contractors were required to leach it so as to leave
no more than 0.2 to 0.3 per cent, copper in the tailings.
The method used for the extraction of the copper from the burned
pyrite consists in roasting it in a muffle with the addition of sulphuric
acid at a temperature of 450 to 550^ C. This brings the copper into soluble
condition. The product is then leached in lead-lined wooden tanks,
first with water, then with barren acid solution left after precipitating
the copper on iron plates, and finally with dilute acid. The copper is
precipitated from the soluton at boiling temperature, on cast iron plates.
In this work 2 lb. of acid and from 1 to 2 lb. of iron are used per pound of
copper extracted.
The oxidized dump material is composed of one-third large pieces
needing grinding and two-thirds of fines. It is of the following average
composition; SiOj, 37.0 per cent.; Fe, 19.6; Al^Og, 20; CaO, 0.25; Cu,
0.75 per cent.
The ore is shoveled into side dump cars and hauled by horses to the
end of an inclined troughed belt conveyor which raises it to the top of
the crushing plant. Here it is wet crushed in a breaker and Chili mills
to yield a pulp containing 33 per cent, dry solids and having 50 per cent.
CHEMICAL PROCESSES
189
material fine enough to pass through a 136-mcsh screen. This pulp is
conducted by wooden launders to the leaching vats. Fig. 4.5 represents
a general plan and section of the leaching and precipitating plant. This
consists first of 10 leaching tanks 184X42X6 1/2 ft. deep. These tanks
are erected on a rock foundation, and have masonry walls, covered with
4 in. of concrete, then 1 in. of reinforced asphalt. The asphalt covering
has lasted successfully for 2 years and has been subjected to temperatures
from — 40 to 102° F. The plant is operated in the warmer months only.
Each leaching tank is charged in 8 hours with pulp containing 200
tons dry material. The pulp during this time receives 13.2 tons of 53°B.
Elevated Tallin
Ckiir a.pi^rKsoln. Dleili
TuillDgfl Diflcbarge level to Pump
Section tkrougli G-H
Fig. 45. — Copper leaching plant at the Gumeshevesky Mine, Russia.
sulphuric acid which is run into the tank, after which the stirring arms of
an agitator are lowered into it. This machine is furnished with five
vertical shafts, each having a stirring arm at its lower end. The agitator
is moved from end to end of the tank and stirs the pulp for 9 hours.
The material in the tank is then allowed to settle, the clear solution is
decanted through the launders to the precipitation tanks. "Wash water is
run in to fill the tank, the pulp is again agitated for 4 hours, allowed to
settle, and the solution decanted as before. Four washes and decantations
arc thus made. The tank is provided with a side door through which the
tailings are discharged, while they are kept in suspension by means of
the agitator. Care must be taken to add sufficient water to make at
least 2 1/2 parts of water to 1 of tailings. The tailings go to a centrifugal
pump which raises them to the elevated tailings discharge. It takes
;5 1/2 hours to discharge a tank. The agitator is then transferred by a
crane to another tank. The cycle of the leaching operation requires
5 days, two tanks being charged and two discharged daily.
190 HYDROMETALLURGY OF COPPER
The quantity of solution decanted at one time is 660 tons. The
copper content of the solution varies from 0.121 per cent. Cu at the first
decantation to 0.015 per cent, in the last wash. About 50 per cent, of the
copper is extracted by these leaching operations, the insoluble balance
consisting of copper silicate and native copper.
The accompanying table shows the actual consumption of acid as
compared with laboratory tests.
CONSUMPTION OF ACID
Acid consumed by
Laboratory tests,
Actual practice.
per cent.
per cent.
22.5
25.0
49.9
50.0
12,4
I 15.0
4.2
11.0
I 10.0
Copper
Alumina
Iron oxides
Lime
Organic matter, etc .
There are 20 asphalt lined concrete precipitating tanks arranged in
five rows with four tanks per row. Each tank is 43 ft. long, 19 ft. wide,
and 2 ft. 7 in. deep. The upper tank of each row is filled with 110 to 120
tons of cast-iron plates. The other four have inclined false bottoms,
each covered with about 3 tons of granulated cast iron in a layer 4
in. thick. Thus there are in all the tanks 460 tons of plates and 200 tons
of granulated iron. The solution flows through the system at the rate of
nearly 200 tons per hour. The average entering solution contains 0.04
to 0.07 per cent. Cu, and when leaving it contains 0.002 per cent., indicat-
ing a precipitation of 95 per cent, of the copper. The tanks containing
the plates are charged at intervals of 3 weeks; the other tanks every
4 to 8 days. The plates are cleaned by scraping and washing off the
loosely adherent copper. The cast iron granules are washed with water in
a trommel perforated with 1/8-in. holes. The under size of the trommel,
consisting of cement copper and small grains of iron, goes to a magnetic
separator where the iron is removed. The oversize of the trommel is
returned to the tanks. Granulated cast iron is in every respect a more
convenient precipitant than cast iron plates, bars, or scrap iron.
Twelve tons of granules have the same precipitating capacity as 120 tons
of plates. It takes eight men 45 hours to clean 120 tons of plates, as com-
pared with six men taking care of 12 tons of granules in 8 hours.
The iron consumed is 1.8 to 2.0 tons per ton of copper recovered.
The cement copper, containing 60 to 75 per cent, of the metal, is treated
by melting in a reverberatory furnace, adding to it a small quantity of
matte of the grade of white-metal to remove impurities. The resultant
CHEMICAL PROCESSES
191
blister copper is then rabbled and poled in the usual way to produce a
brand of best-selected metal.
The accompanying table, based on operations for 13 days in June,
1909, gives an idea of working costs under normal conditions.
OPERATING COSTS OF COPPER LEACHING PLANT IN THE URALS
Item of expense
Wages and superintendence
Supplies
Assay office expense
Repairs and renewals
Power
Acid, 351 tons at $4.32
Pig iron, 62.45 tons at $10.48
Refining, 33 tons at $12.50
Tax on 33 tons at $15.96
Depreciation and general expenses
Total
Total cost
Cost per ton of
Cost per lb. of
dump material
copper
$1843.14
$0,238
$0,024
289.65
0.037
0.004
51.51
0.007
0.001
736.06
0.099
0.010
210.56
0.027
0.003
1516.32
0.196
0.020
676.95
0.088
0.009
412.50
0.053
0.006
526.68
0.068
0.007
2218.32
0.290
0.031
$8481.69
$1,103
$0.11.5
From 7735 tons of crushed dump material, 33 tons, or 0.43 per cent,
of copper, was recovered, using 341.1 tons of 53° B. sulphuric acid and
62.45 tons of pig iron. The cost of the plant, not including the acid
works, was $128,700.
Ferric Sulphate, Ye^iSO^g.- — If cupriferous pyrites is treated with a
solution of ferric sulphate, copper goes into solution in proportion to the
quantity of iron that has been reduced from the ferric to the ferrous condi-
tion. If ferric sulphate is applied to a fresh lot of cupriferous pyrites it is
at first quite rapidly reduced, and the richer the mineral in copper, the
quicker and more perfect will be the reduction. As the copper content in
the ore is diminished, the reducing action becomes slower, and in the later
stages of treatment, if the ferric sulphate solution is allowed to remain in
contact with the partially exhausted ore until reduction has taken place, it
will be found, (1) that the free sulphuric acid is on the increase, and (2) that
the copper content of the solution does not increase in proportion to the
quantity of ferric sulphate reduced, as is the case when it is first applied
to untreated ore. On continuing the treatment until the reduced liquor
shows no f urthur increase in its copper content, the proportion of free acid
rapidly increases, due to the action of ferric sulphate on iron pyrites.
At a slightly elevated temperature the leaching action of the mineral is
more rapid.
Before the copper is precipitated with iron, precautions should be
taken to insure as far as possible the reduction of the ferric sulphate, or
the consumption of iron will be excessive and the precipitation retarded.
192 HYDROMETALLURGY OF COPPER
S. R. Adcock carried out at Rio Tinto a series of experiments' to
determine the amount of copper that could readily be extracted from
cupriferous pyrites by washing the mineral with a solution of ferric sul-
phate, and also to note the changes that take place during the opera-
tion. To this end varying quantities of crude smalls, all of which would
pass through a 0.25-in. sieve, were washed in small lead tanks with a
cold solution of ferric sulphate (2 per cent. Fe) ; the liquor was allowed
to remain in contact with the mineral until its color indicated that the
greater part of the iron had been reduced, when it was drawn off and a
fresh quantity of ferric sulphate added.
These experiments show how rapidly ferric sulphate solutions act
upon the cuprous sulphide in the pyrites, and in places where these
liquors are plentiful or can be cheaply manufactured they no doubt
could be used to advantage for extracting say one-half of the total copper
content of mineral running from 1.5 per cent, and upward, before it is
formed into heaps for treatment by the open air or weathering process.
The extraction by this process compares favorably with the (1) open-air
calcination, until recently so extensively used at Rio Tinto and the (2)
"weathering" or air oxidation process.
"The amounts of coppei' extracted at different periods were calculated from
the analyses of the liquors, and the results obtained from these experiments are
given in the following table: \
Weight of mineral
treated
5 kg
Copper content
Days
under treatment
3
S
23
31
Copper extracted
4 . 62 per cent.
31.90 per cent.
44.00 per cent.
61.00 per cent.
66 . 00 per cent.
5 kg 1 .62 per cent. 4
13
66
94
2 kg 6.49 per cent. I 30
' 61
i : 180
23.00 per cent.
38 . 00 per cent.
50 . 50 per cent.
60 . 00 per cent.
50.70 per cent.
63 . 30 per cent.
79 . 00 per cent.
"The open-air calcination as applied to 2.5 per cent. Cu mineral yields three-
fifths of the total copper (60 per cent.) which is at once dissolved out after calcina-
tion.^ The weathering method yields 88 per cent, of the total copper content
of the mineral treated in 6 years. ^ While carrying out the above trials it was
noticed in each case that, after the first two or three washings, the liquors gradu-
ally increased in their free acid content, and further experiments were performed
'Min. Ind., Vol. IX, 1901.
2 J. H. Collins, Trans. Inst. Min. and Met., Vol. II.
'J. H. Brown, Jour. Soc. Chem. Ind., Vol. XIII.
CHEMICAL PROCESSES
193
to determine the chemical action that was taking place. To determine the
action of ferric sulphate on cuprous sulphide, 20 grm. of pure copper glance
(CujS) crushed to a fine powder, was treated with an excess of the ferric solution.
On filtering off the liquor, well washing the insoluble residue, etc., it was found
on analysis that 15.39 grm. of copper had been dissolved, 27.43 grm. iron has
been reduced from the ferric to the ferrous condition, and 3.83 grm. of free
sulphur had been produced. Based on these results it will be seen that the fol-
lowing equation represents the reaction that has taken place :
(1) Cu ^S + 2Fej (SO J , = 2CuS0, + 4FeS0, + S,
but haA'ing noticed on every occasion when carrying out experiments on the
above lines, that the first 50 per cent, of the copper is more easily extracted, it? is
probable that the reaction is more correctly shown as taking place in two stages,
thus:
(a) Cu ^S + Fe, (SO j', = CuS + CuSO, + 2FeS0,.
(b) CuS + Fe^ (SO.) , = CuSO. + 2FeS0, + S.
"With a view to determine the action of ferric sulphate on iron pyrites
FeSj, 100 grm. low grade ore which had been previously ground to a fine powder,
sampled and analyzed,- was washed with a strong solution of ferric sulphate
(2 per cent. Fe) until the analysis of the liquor drawn off from time to time, showed
that practically all the copper had been removed. The washed mineral was then
treated for 63 days with the ferric solution, and at the end of this period it was
noticed that the reduction of the ferric iron was taking place almost as rapidly
as at the commencement of the experiment. From the analysis of the washings,
which were drawn off when the color indicated that reduction had taken place,
it was noted that each successive wash showed a slight increase in its free acid
content until the copper contained by the mineral had been practically exhausted;
at this stage the free acid appeared to have reached its maximum, and from thence
was always found present in quantity, directly proportional to the amount of
ferric sulphate that had been reduced by the mineral.
The mineral at the conclusion of the experiment was well washed with dis-
tilled water, and dried at 100° C. The following are the analyses before and
after treatment :
Components
Before
After
Copper
0.64 per
cent.
0.07
per
cent.
Sulphur
51 . 20 per
cent.
33.09
per
cent.
Iron
43.82 per
cent.
44.60
per
cent.
Arsenic
0.56 per
cent.
0.21
per
cent.
Antimony
0.03 per
cent.
0.02
per
cent.
Bismuth
0.003 per
cent.
0.002
per
cent.
Lead
0.97 per
cent.
1.10
per
cent.
ZiQc
2.00 per
cent.
0.12
per
cent.
Silica
0.39 per
cent.
0.46
per
cent.
Total
99.613 per cent.
99.672
per cent.
Free sulphur
Nil
2.51 per cent.
13
194 HYDROMETALLURGY OF COPPER
"The above analyses show that the copper and zinc originally contained by
the pyrites are almost totally extracted by ferric sulphate, the arsenic to a lesser
extent in the same time, while all the lead remains in the washed material,
probably as insoluble sulphate. It is well to note that there is free sulphur pres-
ent in the material after treatment.
"The results of the analyses of the liquors obtained by washing the mineral
for 63 days, after the copper had been extracted, were as follows: During this
trial the liquors were kept at a slightly elevated temperature; 45.2 grm. of ferric
iron reduced to the ferrous state; 44 grm. free sulphuric acid formed, and 4.2
grm. of iron (from the pyrites) dissolved. Based on these results it is probable
that the following equation represents the action of ferric sulphate on iron pyrites :
(2) 1 IFe^ (SO J 3 + 2FeS3 + UTIfi = 24FeSO, + I2H2SO, + S.
" The amount of pyrites required to effect the reduction is very small, but the
reaction in the cold is slow in taking place, unless a large excess of mineral is
exposed in proportion to the quantity of ferric sulphate to be reduced; at a
temperature of from 50 to 60° C. the reaction takes place much quicker than at
ordinary temperatures."
Ferric sulphate also reacts with cupric oxide, according to the equation :
(3) 3CuO+Fe2(SOj3 = 3CuSO,+Fe203,
and somewhat similarly on the carbonate, Vfith the liberation of CO2.
Zinc in its sulphide combination is readily soluble in a solution of
ferric sulphate, according to the equation:
(4) ZnS-|-Fe2(S04)3 = ZnSO, + 2FeS04 + S.
From equations 1, 3, and 4, it is readily estimated that theoretically,
6.3 lb. of anhydrous ferric sulphate are required to dissolve 1 lb. of copper
from its cuprous sulphide combination, and 2.1 lb. from the cupric oxide:
6.1 lb. of ferric sulphate is destroyed in dissolving 1 lb. of zinc from
sphalerite, and as zinc is more electropositive than iron, it will not be
precipitated by it, but will accumulate indefinitely unless some means is
provided to remove the zinc sulphate, or at intervals renew the solution.
Experiments with Ferric Sulphate at Cananea.— W. L. Austin gives an
account^ of experiments carried out at Cananea, State of Sonora, Mexico,
by the Cananea Consolidated Copper Co. , from the original notes furn-
ished by David Cole, assistant general manager of the company, with a
view to ascertaining the leaching qualities of local material. While these
experiments did not lead to the adoption at Cananea of the method ad-
vocated, the results obtained are instructive and of practical value to
any one contemplating similar experimental work.
"The material treated at Cananea consisted of mill-tailing sands and of flue
dust from the furnaces. The leaching was done by simple percolation without
agitation. The copper was precipitated from the cuprous liquors by means of
' "Mines and Methods," Sept., 1910.
CHEMICAL PROCESSES 195
metallic iron. The spent solutions were regenerated by forcing heated air
through them. The principal difficulties encountered arose in the regeneration
of the liquors.
"There is from 0.27 to 0.66 per cent, zinc in the mill-tailing sands at Cananea
and the copper varies from 0.54 to 0.89 per cent. It was found that the zinc
minerals were more readily attacked than those carrying the copper, for in the
same time that approximately 65 per cent, of the copper was extracted 72 per
cent, of the zinc went into solution. It was thought at Cananea that dilution
of the solutions by wash water introduced to remove the soluble copper salts
from the tails before allowing them to go to waste would suffice to keep the zinc
content in the solution low enough to prevent serious fouling. A certain propor-
tion of the zinc salts would be removed in this manner because the wash water
after flowing over iron to deposit the copper would be allowed to escape, but the
effect of zinc salts in the apphcation of the process in any particular case can only
be determined by experimenting with the ore in question on a considerable scale.
In using 2 lb. solution to 1 lb. of ore it was found that 94 per cent, of the soluble
matter is extracted without increasing the bulk of solution with wash water.
" Only very small amounts of alkalis went into solution in treating the sand
tailings — about 0.1 per cent. — and the quantities of alumina and lime were
negligible. The pyrites in the ore was found to reduce the ferric salts, cau.sing a
waste of the solvent.
"It was found that the average loss of ferric sulphate per pound of copper
extracted was 4.37 lb. This figure was derived by crediting the gain in ferrous
sulphate from both the action of the ore and from the precipitation of the iron,
and 'calculating the oxidation in the tower.'
"It was found at Cananea that a ferric sulphate solution accomplished a very
complete extraction of the copper from material containing the oxides and
carbonates, but that it acted more slowly upon those carrying the metal in the
form of chalcocite. Chalcppyrite was hardly attacked at all. The content of
the solution in ferric sulphate was found to be of comparatively small importance
provided base salts were absent. It worked when it carried as small an amount
of ferric salt as 1 per cent., and the extractions were nearly as complete when
using 2 per cent, solution of ferric sulphate as when using 7 per cent. A freshly
precipitated solution acted more energetically than an old one, even though
relatively weaker. The presence of much basic sulphate was found to greatly
retard the leaching, 'clogging' the action. For this reason the solution was
settled in the oxidizer before using. About 65 per cent, of the copper contents
of the mill-tailing sands could be extracted in about three hours when the liquor
was boiled, or in about 7 hours when it was kept at a temperature of 70°
"Concentrates gave about 40 per cent, extraction.
"After the liquors were thought to contain a sufficient quantity of copper in
the sulphate form they were conveyed to the precipitating tanks where the
metal was removed by being brought in contact with metallic iron :
(5) CuSO,+Fe=FeSO, + Cu.
"The consumption was light because when the solution is hot the precipita-
tion is very rapid and complete.
"After the liquors had passed through the precipitating boxes the iron was
196 HYDROMETALLURGY OF COPPER
practically all in the ferrous condition and it became necessary to regenerate the
solvent — that is, to reoxidize the iron — before it would again be available. It
is precisely in the regeneration of the spent solutions that the weakness of leach-
ing methods based on the use of ferric salts becomes apparent.
" At Cananea the regeneration was effected by pumping the solution into an
oxidizing tank where it was heated by a steam coil and agitated and oxidized by
hot air. The capacity of the oxidizer was 30,000 lb. of solution; the working
height of the column was 11 ft. 8 in. It was thought that 18 ft. would give better
results. From the oxidizer the solution was led to a settler and from which the
clear liquor was drawn off to a feed tank. It was then available for leaching.
"The reactions which take place when an attempt is made to oxidize ferrous
sulphate to the ferric condition, without the presence of free acid, are very
complicated. Basic ferric salts, of which there are many varieties, invariably
form and are precipitated, thereby causing the loss of a large part of the iron
unless free sulphuric acid has been added in the amounts necessary to produce
the neutral ferric sulphate. For the purpose specified (the formation of neutral
ferric sulphate), ten parts of ferrous sulphate require two parts of concentrated
sulphuric acid. The reactions which occur in the transformation are indicated
by the following equation:
(6) 2(FeSOJ+H2SO, + 0=Fe2(SOj3 + H20.
"If a solution of neutral ferric sulphate is heated with ferric hydrate, there
results a deep brown liquor containing a more basic salt — two-thirds as much
sulphuric acid combined with the same amount of iron as in the neutral salt.
This basic salt is also formed when a solution of ferrous sulphate is slowly oxi-
dized by contact with the air, while at the same time a still more basic salt is pro-
duced (with one-sixth as much sulphuric acid as is present in the neutral sulphate) ,
together with other soluble sulphates and the neutral ferric sulphate.
" The basic salt containing two-thirds as much sulphuric acid as is present in
the neutral sulphate, is decomposed by heating, or by dilution of the solution, the
resulting products being neutral ferric sulphate and a yellow precipitate con-
taining the one-sixth salt referred to above. These two last named ferric com-
pounds predominate when a solution of ferrous sulphate is oxidized by exposure
to the air, and are claimed by some authorities to be the final products from the
oxidation described.
.(7) 10(FeSOJ + O,=3(Fe,(SO,)3)+Fe,SO,.
"From the above equation it is evident that in converting a ferrous into a
soluble ferric sulphate by oxidation in the air, with the assistance of heat, without
the addition of free acid, 40 per cent, of the iron is deposited in the form of insolu-
ble basic ferric sulphates and is therefore lost as an active reagent for the required
purposes, in addition to a loss of the acid with which it is combined. Hence in
order to keep up the grade of the solvent in leaching with ferric salt, it becomes
necessary to constantly replenish the acid constituent of the compound, the
latter being the reagent consumed, and the process becomes virtually one of
leaching with sulphuric acid. Where a heavy iron sulphide is calcined before
leaching, the loss of iron becomes of comparatively small importance, but where
the ferrous salt is an item of expense, the cost is considerable. If 10 lb. of ferrous
( 'HEMIC. 1 L PROi 'EsSSES
197
sulphate are required to carry out a certain leaching operation, 4 lb. will be lost
in the process of regeneration by aeration, unless 2 lb. of concentrated (66° B.)
sulphuric acid is added to the solution to prevent the iron from taking the form
of a basic salt. Therefore, in estimating the cost of leaching a given ore, the
relative expense in providing free acid, as against that for ferrous sulphate, has
to be considered.
"From the above remarks, it is apparent that the reactions which take place
in the regeneration of the spent solutions by the methods used at Cananea are too
complicated to be written down in formulae; the actual figures disclosing results
obtained are more illuminating.
"In making the ferric solution employed at Cananea the ferrous sulphate
used contained sulphide impurities which caused some irregularities during the
oxidation, the ferric sulphate being destroyed. The sulphides might have been
removed by dissolving the sulphate in a separate tank and decanting the clear
solution. The ferrous salt was added to hot water into which had been pumped ,
some ferrous solution and the whole was heated by means of a steam coil near the
bottom of the oxidizing tank. Air was forced in by a compressor after passing
through a heater where is attained a temperature of from 200 to 400° F. and
served to agitate and oxidize the solution.
"After it was apparent that the ferrous salt had gone into solution, samples
of the liquor were taken which gave the following results :
Sample
No.
Hours
Ferrous
sulphate
per cent.
Ferric
sulphate
per cent.
Ferrous
sulphate
pounds
Ferric
sulphate
pounds
480
620
Temperature
degrees C.
1
2
3
0
1
2
4
5
6
8
24
26
28
30
32
12.1
11.7
11.3
10.9
10.5
10.5
9.4
6.5
6.1
5.8
5.5
5.1
1,6
2.1
3630
3510
3390
3270
3150
3150
2820
1950
1830
1740
1650
1530
76.0
79.0
79.5
4
80.5
5
80.5
6
81.5
7
81.0
8
9
5.6
1580
79.0
79.0
10
79.0
11
12
6.6
7.0
1980
2100
77.0
77.5
Total ferric sulphate formed in 32 hours was 2100-
formed in 1 hour = 50.6 lb.
-480=1620 lb. Average ferric sulphate
"The table shows no oxidation in the solution between the fifth and sixth
samples. A possible explanation of this feature is that ferric sulphate attacked
the sulphide of iron already referred to as an impurity in the ferrous salt fed to
the oxidizing tank.
"A basic sulphate which formed in the tank during the process of oxidation
is said to have interfered to a great extent with the operation.
"In the test from which the above results were obtained, 300 lb. sulphuric
acid were added after the solution had been oxidizing for 28 hours. Four hours
later a valve leaked and the solution in the tank had to be drawn off, so that it
was thought that the basic sulphate may not have received the full benefit of
198
HYDROMETALLURGY OF COPPER
the acid. The iron sulphides in the solution were considered to have retarded
oxidation to a marked extent.
"At Cananea the oxidation of the ferrous to the ferric sulphate was found to
be a serious problem as the transformation was 'i'ery slow, and it was thought
that a successful application of the process will depend more on a satisfactory-
solution of this feature than on any other.
"In an experiment made previous to the one mentioned, before alterations
had been made in the oxidizing tank, 6.4 tons- of solution were treated for 2 days.
During that interval 1,500,000 cu. ft. of air at a temperature of about 140° C.
(equivalent to about 100,000 lb.) were forced through the solution with the follow-
ing results;
Before blowing.
After blowing. .
Ferrous sulphate
Ferric sulphate
Per cent.
6.80
5.40
Pounds
Per cent. i Pounds
104.9
230.4
"It is stated that only 125.44 lb. of ferric sulphate were formed, which also
included the precipitated basic salts. If all the oxygen in the air had been util-
ized, nearly 200,000 lb. ferric sulphate might have been produced had the
necessary amount of ferrous solution been available. It took 1200 times more
air than was theoretically necessary to produce the results desired.
"In order to increase the efficiency of the oxidizer, perforated discs were
submerged in the solution and heated air from the compressor was forced in at
the bottom. The improved oxidizer gave much better results than were obtained
with the former one.
In making an estimate of the expense of converting the ferrous solution at .
Cananea, no account appears to have been kept of the ferrous salt, nor of labor
and repairs; only the following items are given:
Steam for heating the solution through lead coil.
Power used for compressor.
Coal to heat the air, 262 lb.
Sulphuric acid, 300 lb..
Total,
$3.80
0.85
1.05
2.00
$7.70
"These figures are for 24 hours' run. The sulphuric acid is taken at $0.0066
per pound, at which price it was thought that probably it could be manufactured
at Cananea.
"As the average amount of ferric sulphate formed in the oxidizer was 50.6 lb.
per hour, dividing the 17.70 by 24 to obtain the cost per hour, and dividing this
result by 50.6 to get the cost per pound, gives $0.0063 as the cost of 1 lb. of ferric
sulphate produced in the manner described.
"An estimate made by the engineers who conducted the tests at Cananea is
given below. These figures are said to represent the actual results obtained in
Per ton
Per cent.
$0.50
13.4
0.50
13.4
0.95
2.5 . 5
0.08
2.1
1.00
26.8
0.30
8.0
0.20
5.4
0.20
5.4
CHEMICAL PROCESSES 199
treating a lO-ton lot of Cobre Grande ore. This ore contained 3 percent, copper,
and the extraction is said to have been 96 per cent.
Crushing to six mesh,
Roasting,
Oxidizing,
Acid, 16 lb. at 1/2 cent.
Coal to evaporate wash water.
Iron to precipitate copper.
Heating solution.
Labor,
Total, $3.73 100.00
ex "This is equivalent to $0,064 per pound of copper. It was found in this
perimental work that the best leaching results were obtained when the copper
was in the oxidized condition. A test made on 1735 lb. of flue dust containing
5.63 per cent, copper was satisfactory, as shown below.
"The dust was first calcined to oxidize the suljjhides and then leached with
a hot solution of ferric sulphate. The leaching proceeded rapidly; two applica-
tions of the solution left practically no copper in the tailings. The following
table shows the consumption of ferric sulphate in this operation:
Pound
Ferric sulphate (Fe2(S04)3) in feed solution, 440
Ferric sulphate (Fe2(SO.,)3) in tail solution, 190
Ferric sulphate (FejCSOj),) used to leach, 250
Copper in charge, 97 . 5
Copper n solution, 95 per cent, extraction, 92 . 6
Ferric sulphate consumed per pound of copper extracted, 2 . 7
The time factor was ignored in this test, the perfection of the leach alone being
considered. The dust contained about 30 per cent, iron so that an excess of
ferrous sulphate was found in the tail solution. No attempt was made at esti-
mating the cost in these tests because the expense for calcining, labor, steam to
heat the solutions, repairs, etc., could not be accurately obtained.
"Another test said to have been made on a 10-ton lot of flue dust assaying
7.5 per cent, copper, resulted in an extraction of 94 to 96 per cent. The cost
items were given as follows ;
Per ton
Per cent.
Calcining,
$0.50
12.0
Oxidizing,
1.44
34.5
Acid, 16 lb. at 1/2 cent.
0.08
1.
Iron for precipitating the copper.
0.75
18.0
Steam for heating solufon
0.20
4.8
Coal for evaporating the wash water,
1.00
24.0
Labor,
0.20
4.8
Total, $4 . 17 100 . 0
Cost per pound of copper, $0,029.
200
HYDROMETALLURGY OF COPPER
" The Cananea tests appear to have come to a sudden termination through the
breaking down of the oxidizer, without the cost of the solvent having been con-
clusively established. The extraction was apparently satisfactory.
"The figures given in the foregoing, relative to the cost of producing the ferric
sulphate are either based upon what might have been accomplished with a per-
fect oxidizer, or are incomplete, or are theoretical deductions from imperfect
results. They are not conclusive.
" It is to be regretted that having carried the work along to the point reached,
an effort was not made to definitely ascertain the cost of producing the solvent,
or what is the same thing, the expense of regenerating the solutions.
"An estimate made by an engineer in the early stages of the experiments as
to the probable cost of producing copper by this method is given below. This
estimate was figured before improvements in the oxidizer were made. No
charges are included for mining, transportation, crushing, calcining, general
expenses, etc.
COST OF TREATING SANDS YIELDING EIGHT POUNDS OF COPPER PER
TON
Per ton sands
Per pound
copper
Per cent.
$0.57
0.12
0.08
0.12
0.60
0.40
0.18
$0.0713
0.0150
0.0100
0.0150
0.0750
0.0500
0.0225
27.6
Power, at $0.65 per horse-power day for blowing air
Acid — if made on ground — 2 lb. acid per 1 lb. Cu
5.8
3.9
5.8
28.9
Shipping and refining at $0.05 per pound copper
Added for waste of material, 20 per cent, of $0.89
19.3
8.7
$2.07
$0.2.588
100.0
"A summary of the cost of producing the ferric sulphate used in the experi-
ments at Cananea, derived from independent sources, is given below. The
figures include only steam heat, heat for air, power for compressor, and acid.
COST OF FERRIC SULPHATE
Lb. ferric sulphate consumed
Ferric sulphate cost
Ferric sulphate cost per pound
per pound of copper
of 1 lb.
of copper
Oxides
Sulphides
Oxides
Sulphides
I
2.5
3.3
$0,004
$0.01
$0,013
II
2.7
0.0064
0.017
III
4.37
0,0226
0.099
"The cost per pound of ferric sulphate in No. II is probably the closest
approximation to the actual cost of the pound of ferric sulphate made at
Cananea, but does not include original cost of the ferrous sulphate, losses in
handling, basic salts, nor losses from presence of impurities."
CHEMICAL PROCESSES 201
Thomas' Experiments with Ferric Sulphate on Sulphide Ore.' — F.
Thomas experimented with ore of the following composition:
SiOj, gangue, 44 . 7 per cent.
Cu^S (10.5 per cent. Cu), 13.5 per cent.
Fe^Sj, 25.25 percent.
FejOa, 5 . 76 per cent.
MnO, 0.18 percent.
AljOj, 7.5 percent.
PzOs, 0 . 34 per cent.
CaO, 2 . 83 per cent.
Traces, Sb, Sr, Mg, and K.
The principal results of his experiments are as follows: The double
sulphides of copper, which occur in nature, require for their complete
transformation by means of ferric sulphate such a long treatment and
so fine crushing, that a commercial application of this method of leaching
does not pay under the conditions existing in most copper producing
countries. Copper-iron sulphides, artificially prepared, also resist the
action of ferric sulphate in the same way as the natural ones. Free
copper sulphides and oxides react with ferric sulphate in aqueous solution
easily and quickly. The reason to which the difficulty of the action of
ferric sulphate in the former cases is due must, therefore, lie in the chem-
ical affinity between the ingredients of the natural and artificial sulphur
ores in copper. The presence of a larger quantity of ferrous sulphate in
the solution impairs the solution of copper from cuprous sulphide by
means of ferric sulphate. The method suggested in the Siemens-Halske
process for roasting copper ore, so that the main quantity of the iron is
transformed into oxide while the main quantity of copper remains as
cuprous sulphide, is practically impossible. Neither can satisfactory
results be obtained by means of dead-roasting of sulphide ores, for the
reason that at the temperature required for this purpose basic silicates
are formed by means of combination of the copper oxide with the
silicates of the gangue, and that perhaps also salts of theFe304type are
formed by the combination of the oxides of copper and iron; such salts
are acted upon very slowly by ferric sulphate. The double sulphides
must therefore be destroyed by oxidizing roasting at so low a temperature
that the formation of the compounds just mentioned is impossible. This
temperature is about 450 to 480° C. The product thus obtained contains
the copper essentially in the form of sulphate. The low temperature of
roasting permits the use of simple furnaces, and a crushing of the ore
is sufficient, corresponding to 484 meshes per square centimeter. The
leaching can be so conducted that a solution of copper sulphate is obtained
with only small quantities of iron.
' Metallurgie, Jan. 15, Feb. 8, and 22, 1904.
202 HYDROMETALLURGY OF COPPER
Experiments in Southern Tyrol, Spain.' — A copper plant was erected
in Southern Tyrol, in which an attempt had been made to employ the
old Siemens-Halske process, the first stage of which is leaching with
ferric sulphate solution. This preliminary leaching was a failure, so
that the subsequent electrolytic precipitation of the copper was never
attempted. The reason was that the ore contained the copper in the
form of the compound Cu2S,FeS,FeS2; this was roasted at such a high
temperature that the copper oxide combined with other oxides and
formed combinations like CuO,Fe203 and silicates, which are not amen-
able to leaching. The ore was afterward roasted at a low temperature,
which was then easily leached, and it was decided to work the copper
ore simply for copper sulphate.
Copper Extraction at Kedabeg, Russia.^ — The rich ores at Kedabeg
are smelted. The lean ores, containing less than 5 per cent, copper,
say 3 per cent., and which consequently would not bear the cost of
direct smelting, are treated by leaching. The process is very simple
and well adapted to the local conditions which scarcely permit of the
use of salt or other decomposing reagents. The ore is cheaply roasted
in kilns or Gerstenhofer furnaces without fuel, the copper being thus
brought into soluble form for leaching. This presents no great diffi-
culty, as the copper, originally existing as sulphide, is oxidized par-
tially to sulphate in the furnace, the sulphatization being completed
to a certain extent later during the leaching by the ferric sulphate
formed in the roasting and also present in the residual solutions
from the electrolytic precipitation, the latter being run into leaching
ponds. A comparatively long time is required to obtain a practically
complete change of the copper to sulphate, several years' leaching being
necessary to reduce the copper contents to 0.5 or 0.7 per cent, copper.
However, from 50 to 70 per cent, is obtained at a very small cost in the
first year. The whole plant is in the open, without covering or roofing,
and on sloping ground. The ground, very impermeable to begin with, is
completely hardened by the decomposition of basic salts of iron.
The copper is precipitated in wooden tanks by means of scrap iron, a
rapid circulation being kept up all the time. In this way 409.5 metric
tons of cement copper, with 65 to 75 per cent. Cu, are produced per
annum.
The Millberg Process.^ — Well roasted copper pyrites or copper cinders
contain copper as:
1. Copper sulphate, soluble in water.
2. Copper oxide, soluble in ferric sulphate.
' W. Borohers, Metallurgie, Aug. 8, 1909.
" Gustave KoUe, Min. Ind., Vol. VI.
' Chemicker Zeitung, XXX, 511.
CHEMICAL PROCESSES 203
3. Cuprous oxide, capable of being oxidized by ferric sulphate which
will then dissolve it.
4. Copper sulphide, oxidizable by ferric sulphate which will then
dissolve it.
Millberg's method consists in leaching the roasted ore or cinders with
a solution of ferric sulphate by which the copper is salts are dissolved
out and pass into the filtrate. This filtrate will then contain ferric
sulphate, ferrous sulphate, copper sulphate and sulphate of other metals
when present, such as zinc, manganese, cobalt, nickel, and aluminum.
Ferric sulphate is very effective in bringing the copper salts into solution
so that in burnt pyrites containing 0.8 per cent, to 4.0 per cent, copper,
there is left in the residues 0.05 per cent, to 0.2 per cent. onlj'.
The filtrate contained in a tank is brought to the temperature of 60° C.
and to it is added a little ferric sulphate solution till by testing with
ferrocyanide solution the end point is reached. This may be, for example,
0.66 per cent, of a ferric sulphate solution of 25° B. This oxidizing
action will not exceed 2 days, and for some kind of cinders but a few
hours. Air is then blown into the solution in presence of an alkali base,
such as milk of lime. The ferric sulphate is precipitated as an insoluble
basic sulphate. This is filtered or decanted off and there remains in the
filtrate the sulphates of copper and other metals.
The solution is heated to the boiling point, and to it is slowly added
dilute milk of lime, which precipitates the copper as an insoluble basic
sulphate of a light green color and leaves the other sulphates in solution.
The precipitation must be watched and the treatment stopped as soon as
the copper has been precipitated.
The Elliott Process.' — The Elliott process consists, essentially, in
leaching the ore with a hot non-acid solution of ferrous sulphate, passing
air through the solution during the operation of leaching, precipitating
the copper from the solution with iron, thereby regenerating the ferrous
sulphate for a repetition of the process.
The process is applicable to oxidized ores; if the ore to be treated is a
sulphide, it has to be given a preliminary roasting to convert the sulphide
into the oxide or sulphate. The air, preferably heated, converts the
ferrous sulphate to the ferric sulphate, which then acts on the copper
oxide. The oxidation of the ferrous sulphate to ferric sulphate may be
expressed:
6FeS04 + 03 =2Fe2(SOj3-|-re203
the leaching operation:
3CuO +Fe2(S0 J ^ = 3CuS0, +Fe,03
and the precipitation:
CuS04+Fe = Cu+FeS0,
'U. S. Patent 814,836, March, 1906.
204 HYDROMETALLURGY OF COPPER
so that the leaching solution, is therefore regenerated by the precipitation
of the copper with iron. The ferrous sulphate solution is then again
applied to the ore, and while hot, air is passed through it, thereby re-
generating the ferric sulphate, which attacks the copper and makes it
soluble, thus repeating the cycle. It is claimed that the process is
applicable to ores containing too much lime for acid treatment.
The Laist Process. — This process is based on the use of sulphuric acid
as the solvent, and hydrogen sulphide as the precipitant.
The steps in the process may be summarized as follows :
1. Dissolving copper from the ore with dilute sulphuric acid.
2. Precipitating the copper from its solution as copper sulphide by
hydrogen sulphide, accompanied by the regeneration of acid.
3. Manufacture of the hydrogen sulphide gas: a. Reduction of
gypsum to calcium sulphide with coal; b. Decomposition of the calcium
sulphide by carbon dioxide and water to hydrogen sulphide and calcium
carbonate.
4. Conversion of the copper precipitate to metallic copper.
The process is based' on the following reactions:
(1) CuC03 + H2S04 = CuS04-|-C02 + H20.
(2) CuS04 + H2S = H2S04-l-CuS.
(3) CaS04 + 4C = CaS+4CO.
(4) CaS+H20 + C02=CaC03 + H2S.
The first well-known reaction shows what takes place when copper
carbonate or oxide is dissolved with sulphuric acid. Reaction 2 shows
the effect of the hydrogen sulphide on the copper sulphate, by which the cop-
per is precipitated as sulphide and an equivalent amount of sulphuric acid
regenerated. The third reaction, is the first step in making the hydrogen
sulphide gas. Calcium sulphate (gypsum) is reduced with coal. This
reaction takes place at a bright red heat, or about 1800° F. The reaction
is accompanied by the formation of both carbon dioxide and carbon
monoxide, but the monoxide largely predominates. In the reaction it is
assumed that carbon monoxide only is formed. From the fourth reaction
it is seen that when calcium sulphide is treated Avith water and carbon
dioxide, it decomposes, and calcium carbonate and hydrogen sulphide are
formed.
From the reactions it is clear that 168 parts of gysum will furnish
enough hydrogen sulphide to precipitate 63 parts of copper. For the reduc-
tion of this gypsum 48 parts of carbon is required. In practice there is used
3 parts of coal to 7 parts of anhydrous calcium sulphate, that is to about
8 parts of gypsum. This slight excess of coal is necessary to effect a
quantitative reduction. Hence about 4 lb. of the mixture of coal and
gypsum are used in precipitating 1 lb. of copper, or about 2 2/3 lb. of
gypsum, or 1 1/4 lb. of calcium sulphide precipitate 1 lb. of copper.
CHEMICAL PROCESSES 205
The copper sulphide precipitate is readily converted into metallic
copper, while sulphur dioxide is released.
Most copper ores contain substances besides copper which consume
acid. Acid lost in this way is not recovered in the precipitating tanks,
but is recovered by allowing the acid solution itself to flow down con-
densers up which sulphur dioxide mixed with air from the copper
sulphide furnace, is passed. The hot air in conjunction with iron salts in
the solution oxidize a large part of the sulphurous acid to sulphuric
acid.
Method of Extracting Copper at Rio Tinto, Spain.' — The ore treated at
Rio Tinto is massive iron pyrite containing up to 3 per cent, of copper,
which has been disseminated through the mass by a secondary en-
richment.
The well-known method adopted for the extraction of the copper, and
in use at the present time, consists in allowing huge heaps of the mineral
to oxidize under the influence of moisture and air, and subsequently
washing out the copper sulphate as soon as it is formed, by running water
through the heap.
The application of this system depends largely on the state in which
the copper occurs in the mineral. If it exists as chalcopyrite — CiiFeS2 —
the copper will not oxidize by simple exposure to the air, in one case it
having taken many years to oxidize 10 per cent, of the copper originally
present in the ore. If the copper is in the form of CuS, the oxidation
proceeds very slowly. The best form for solution is CujS, or copper
glance, which constitutes the bulk of the copper in Rio Tinto pyrite.
These statements may be made clearer from the following account of
the reactions that take place during the oxidation.
When the mineral is exposed to free access of air and moisture, some
ferrous sulphate is formed in accordance with the following reaction:
(1) FeS2 + 70 + H2O=FeSO,-hH2SO4
This ferrous sulphate becomes readily oxidized by the air to ferric
sulphate,
(2) 2FeSO, + H2S04 + 0=re2(SOj3 + H,0,
and it is due to the reaction of this ferric sulphate on the copper sulphides
that the copper is rendered soluble, as is shown by the following chemical
equations:
(3) Fe,(SO,)3 + Cu2S = CuS04-F2FeSO^ + CuS.
(4) Fe,(SOj3 + CuS + 03 + H,0 = CuSO, + 2FeSO, + H2SO,
The reaction No.' 3 takes place fairly rapidly and causes half the copper
to go into solution within a few months, while reaction No. 4 proceeds
'C. H. Jones, Trans. Am. Inst. Min. Engs., Vol. XXXV, 1905.
206 HYDROMETALLURGY OF COPPER
much more slowly, and requires, under the most favorable conditions,
about two years to extract 80 per cent, of the remaining half of the
copper.
In the laboratory at Rio Tinto a method has been worked out to
determine the state of combination of the copper in any particular min-
eral. This method depends on the action of the mineral on various
solutions under constant conditions of dilution and temperature, and,
though necessarily somewhat arbitrary, it shows with considerable
accuracy the form of the sulphide in which the copper exists, and conse-
quently whether the copper can be readily extracted by washing in
heaps.
In practice the method adopted to bring about the desired oxidation
is as follows: A site is chosen for the formation of the heap where the
ground is sufficiently concave and sloping to enable the copper liquor
that is formed to collect and run out at the base of the heap. On the
ground is first arranged a net work of air flues, made of rough stones and
having an internal diameter of 12 in. Vertical chimneys, 50 ft. distant
from one another, are built in the same manner and connect with the
ground flues. Care is taken that the mouths of the ground flues are kept
open and not covered by ore. The mineral, the lump portion of which
has been passed through jaw breakers to be reduced to pieces not larger
than from 2 to 3 in. across, is now tipped from side-tip wagons at the
highest part of the selected side over and around the stone flues. Lump
and fines are alternately dumped until the height of the mass at the edge
is about 30 ft., the upper surface of the mineral being kept level. A heap
of this form approximately contains 100,000 tons of ore.
As the mineral is added, the building of the stone chimneys keeps pace,
in order to have a clear opening to the top of the heap. The surface of
the heap is formed into squares by means of ridges of the mineral, the
size of these squares depending on the porosity of the heap. The func-
tion of these ridges is to enable the water to be run on locally over the
surface of the heap in order to insure that all parts are equally washed
and that the water does not run through the heaps in channels. A
system of gutters is also arranged so that the water can be run on to all
parts of the mass. As the heap is being formed, water is run on and the
copper sulphate existing in the mineral is extracted; also, the water
provides the moisture for the oxidation to take place in accordance with
the equations given above. The mineral in the heap is then allowed to
oxidize, which it does pretty rapidly, as evidenced by the heat produced,
the temperature in the chimneys rising to from 170 to 180° F. As the
temperature increases, the surface openings of the chimney may be
closed in order to allow the oxidation to spread through the heap. The
surface gradually shows a brownish coloration, due to the dehydration of
the buff-colored basic ferric salt that forms on top of the mass, and its
CHEMICAL PROCESSES 207
gradual heating up may be noted by this drying action. The greatest
care must be taken not to allow the heaps to fire, for if once started it is
very difficult to extinguish. When the oxidation has proceeded as far as
it is safe to allow it, water is run on at the rate of about 50 cubic meters
per hour, until the soluble copper is leached out. The heap is then allowed
to reoxidize and the washing is repeated. After about a year has elapsed
the surface requires "refilling," and the squares are rearranged so that
the places where the ridges were before are now the middle of the squares.
The gutters also are shifted. At the edge of the heap for a distance of
some yards the mineral, which has become cemented, holds a consider-
able quantity of copper salts and is dug down into terraces in order that
this copper may be extracted by washing.
When the copper is reduced to 0.3 per cent, the heap is considered
washed and the mineral, containing 49.5 per cent, of sulphur, is removed
and exported as "washed sulphur ore" and used for the manufacture of
sulphuric acid.
Successful heap washing depends on the efficient ventilation of the
mass, the trouble usually being a too great excess of fines produced in
mining the ore, which cement hard and clog up the air passages.
The copper liquor as it runs from the heap contains some ferric iron
in solution,which as will be shown later, is very objectionable. In order
to remove the ferric iron the liquor is run over a smaller heap of fresh
mineral known as a "filter bed," which reduces the ferric iron. This
"bed" is laid inside a reservoir formed by a masonry dam across a small
ravine, and the liquor after percolating through the mineral remains in
contact with it until it is required to be drawn off to the precipitating
tanks. When the mineral is fresh the reduction of the ferric iron takes
place rapidly, due to the CujS, as shown by equation 3, but the iron
pyrite itself has an effective reducing action on ferric iron in solution
according to the following equation:
(5) 7Fe2(SOJ3+FeS2 + 8H20 = 15FeSO,+8H2SO,.
The principal constituents of the liquor as it enters the cementation
tanks are as follows, the figures given representing the grams per cubic
meter or units per million parts: Copper 4000, ferric iron 1000, ferrous
iron 20,000, free sulphuric acid 10,000, and arsenic 300. The large
quantities of ferrous iron and free sulphuric acid present are due to the
fact that the waste liquor from the cementation tanks after the copper
has been precipitated, is pumped back and used for washing the heaps in
addition to fresh water, and consequently these solutions tend to become
concentrated. The liquor is then run from the reservoirs at about 300
cubic meters per hour through the precipitation tanks over pig iron in
order to precipitate the copper in the form of "cement copper." These
cementation tanks are arranged in series on the slope of a hill, the liquor
208
HYDROMETALLURGY OF COPPER
CHEMICAL PROCESSES
209
passing backward and forward until it is discharged from the lowest tank
qf the series free from copper.
Each series consists of three tanks in parallel arranged so that the
liquor can be divided and passed along as many tanks as necessary,
depending on the quantity of liquor that is being run through and on the
varying temperature of the liquor with different seasons, the hotter the
solution, which in summer reaches 100° F., the faster the rate of precipita-
tion. Each tank is about 320 ft. long, 5.5 ft. wide by 2.2.5 ft. deep, and
has a slope varying from 2 per 1000 in the first series to 11 per 1000 in
the last, the reason for the increase in slope being, that as the liquor
becomes impoverished in copper the free acid present is more active in
wastefully dissolving the pig iron — an action which is considerably
Fig. 47. — Rio Tinto leaching plant, Spain. General \'iew of mineral licaps, copper
liquor dam, and precipitating tanks.
diminished by increasing the velocity of the liquor by means of the in-
creased slope of the tanks. The tanks themselves are made of 9X3 in.
boards attached to wooden frames set in cement, the space between
parallel tanks being filled with stone and cement, constituting a wall
.supporting the sides of the tanks. Fig. 46 shows the method of removing
the cement copper and Fig. 47 a general view of the mineral heaps,
copper liquor dam and precipitating tanks.
No metal is used in the tank construction, hard wood pegs being em-
ployed to attach the boards to the frames. The spaces between the boards
14
210 HYDROMETALLURGY OF COPPER
are carefully caulked with oakum and pitch in order to render the tank
water-tight. At each end of the tank is an arrangement by which a door
can be dropped in and luted so as to cut out that particular tank, and there
are also wooden plugs that can be removed so that the liquor from that
tank can be run off, thus allowing for the removal of the precipitated
copper. A few old boards are placed on the bottom of the tanks for
their protection and on these are piled up the pigs of iron which are laid
across the tank at the bottom, the next layer being at right angles to the
first, and so on until the tank is filled; each foot-length of the tank con-
tains about one ton of iron. The liquor is allowed to run through the
system of tanks and needs no attention except to remove the precipitated
copper and to add fresh iron. The "sahda" liquor containing from 15
to 20 grm. of copper per cubic meter is allowed to run to waste, for the
reason that about this copper content the amount of iron required to
precipitate the copper equals the value of the copper recovered. Daily
some of the tanks are cleaned out by being closed as described above, the
liquor meanwhile passing down the other tanks of the series; the liquor
is run off into settling tanks, any copper in suspension being recovered;
all of the iron is removed from the tank and piled on to the dividing wall,
at the same time the copper adhering to the iron is knocked off and thrown
back into the tank. The dirty-looking precipitate is then transferred to
the cleaning and concentrating plant, the iron is replaced in the tank
and the liquor again allowed to run through it. This crude precipitate,
containing about 70 per cent, copper, is thrown, a little at a time, on a
perforated copper plate at the head of a long launder or tank and is
washed through the plate by a strong stream of water from a small
nozzle. The material that does not pass through the screen consists of
leaf-copper and small pieces of iron; this material is thrown into a heap
and afterward sorted over by girls who remove the pieces of iron. The
precipitate that passes into the launder is repeatedly turned over against
the stream of water and by this simple means a concentration is effected.
The first few yards of the launder contain a red precipitate known as
"No. 1 precipitate", containing 94 per cent, copper and less than 0.3
per cent, arsenic; following this is "No. 2 precipitate" containing 92
per cent, copper and between 0.3 per cent, and 0.75 per cent, arsenic,
while below is the "No. 3 precipitate"; this is in a state of very fine
division and contains on an average 50 per cent, copper and 5 per cent,
arsenic. This last named portion, which carries all the graphite from
the pig iron, contains the bulk of the antimony and bismuth that is also
precipitated from the liquors. Classes No. 1 and No. 2 are removed to the
drying sheds and bagged for shipment to the refinery; the No. 3 precipi-
tate is removed, moistened with acid liquors, made into balls by hand and
dried in the sun. These balls become cemented hard and can be readily
transferred to the smelter, where they form part of the charge for the
CHEMICAL PROCESSES 211
blast furnaces and are run down to matte to be subsequently bessermerized,
thus effectively removing the arsenic, antimony, and bismuth they con-
tain. The reactions that take place in the cementation tanks are given in
equations 5, 6, and 7.
The first reaction that occurs in the liquor running over metallic iron
is the reduction of the ferric sulphate to ferrous sulphate, the final result
being in accordance with the equation:
(5) Fe2(SO,)3 + Fe = 3FeSO,.
This action causes the consumption of the pig iron without any cor-
responding yield in copper, and consequently should be avoided as far as
possible by having all the iron in the ferrous condition. The second
reaction is the precipitation of the metallic copper, brought about by
galvanic action. The iron becomes coated with copper, and thus the iron
and copper in the acid liquor constitute a galvanic couple with a consider-
able difference of potential. It is due to the electrolytic action that the
copper and all other metals present that are electro-negative to iron will
be precipitated. The ultimate action of the precipitation may be chemi-
cally expressed by the following equation:
(6) CuS0,-|-Fe=FeS04-hCu.
Besides the reactions above mentioned there is one which causes the
liberation of hydrogen, as evidenced by the bubbles of gas that may be
observed to arise in the tank liquor. This action, Avhich causes a waste-
ful consumption of iron, may be expressed as a final result by the follow-
ing equation:
(7) Fe + H,S0,=FeS0, + 2H.
These three equations constitute the main reactions that take place in the
precipitating tanks.
While the liquor is fairly strong in copper, the copper is mostly
precipitated in a coherent form, but in the later stages, as the liquor
becomes impoverished it is precipitated in a powdery state — a condition
which is more effective in its galvanic action with the iron, and thus
unfortunately causes a larger precipitation of arsenic and other impuri-
ties than in the earlier stages. In the later stages also the "solution
reaction, " of iron and sulphuric acid as given in equation No. 7 goes on to
11 proportionately greater extent than does the precipitation of the copper,
and consequently the cost of pig iron in precipitating the copper varies
inversely as the quantity of copper in the liquor.
By keeping careful watch on the reduction, as far as possible, of the
ferric iron before the liquor enters the tanks, and by giving it sufficient
velocity through the tanks, a strongly acid liquor such as given above
during a years' working will not consume more than 1.4 units of pig iron
212 HYDROMETALLURGY OF COPPER
(containing 92 per cent, iron) to 1 unit of copper precipitated. A valu-
able check on the iron being consumed can easily be kept by the labor-
atory, by analyzing the liquors before entering and after leaving the
tanks, and from these analyses the quantity of iron that is being con-
sumed can be calculated.
The following table^ gives the analysis (in grams per cubic meter) of
the entering and outgoing liquors of the precipitation plant at Rio
Tinto:
Entering
Leaving
Entering
i Leaving
!
Copper
Ferrous iron
, ...' 2,710
. . ..i 13,908
610
4,874
70,872
105.818
'■ 1.26 to 1 Cu
19
17,202
4,129 '
69,662
105.718
2,780
14,030
732
4,991
71,980
105.901
1.26 to 1 Cu
19
17,934
Free acid
Total solids
Specific gravity
4,349
72,834
105.972
The best results are obtained when the solution is slightly acid, as it
tends to accelerate precipitation and prevents a falling out of basic salts
of iron while the precipitation is going on.
At Tharsis the inclination of the precipitating tanks is as follows:
For the first 40 per cent, of the copper, 1 in 200
For the next 30 per cent, of the copper, 1 in 150
For the next 20 per cent, of the copper, 1 in 100
For the remainder, 1 in 50
At Rio Tinto and Tharsis, 1.4 units of pig iron containing 92 per cent,
iron are consumed on an average, per unit of copper obtained. Both* at
Rio Tinto and Tharsis the liquor traverses about 3 kilometers before all
the copper is precipitated. Sixty per cent, of the copper in the liquors
is precipitated within the first 700 meters of the tanks. Over 70 per cent,
of the precipitate contains more than 94 per cent, copper.
The following points have been established by the practice at Rio
Tinto:
1. The complete analysis of the solutions both before and after
treatment is essential to prevent undue consumption of iron.
2. Free acid to be eliminated as much as possible.
3. Mechanical contrivances will, in a measure, overcome these diffi-
culties. The inclination of precipitating tanks and consequent velocity
of current should be in inverse ratio to the amount of copper present;
the less the copper and the greater the free acid and ferric iron, the
greater inclination necessary.
'F. H. Probert, Mining and Scientific Press, January 4, 1908.
CHEMICAL PROCESSES 213
4. As large a surface of iron should be exposed to the liquors as
possible.
5. Aeration of the liquor by tumbling through the air is objectionable,
since it has an oxidizing effect and so increases the consumption of iron.
6. In the course of the flow of the liquor through the precipitating
tanks there is a place where the cost of iron exceeds that of the value of
the copper obtained.
7. The warmer the liquors, within certain limits, the faster is the
precipitation of the copper.
M. P. Truchot, chief chemist, Huelva,^ estimates that to extract cop-
per by natural weathering alone, 20 years would be necessary; roasting
destroys the vegetation of the country; there therefore remains the wet
methods where the reactions may be hastened by the use of a regulated
supply of air and water. In preparing the heaps for oxidizing leaching
of the copper, the fines amount to 80 per cent. ; the lumps, 20 per cent.
The two grades are placed in alternate layers and the top of the heap
finished with fines, to prevent too rapid filtration. Practice varies as to
the temperature of the heap, which may be as high as 82 to 90° C. where
it is sought to promote oxidation and to increase the rate of leaching,
while in other cases, the temperature is not allowed to exceed 30 to
32° C. Since the higher temperature is dangerous, one of no more than
45 to 60° C. is recommended.
It takes 6 to 7 years to exhaust one of these heaps of 100,000 tons; the
exhausted ore then contains 0.25 to 0.30 per cent, of copper. The more
permeable copper schists leach more rapidly, however, taking but three
or four years. While the oxidation of chalcopyrite proceeds but slowly,
the rate can be increased by finely pulverizing it and distributing it
throughout the pile.
The copper solution from the heaps varies from 0.015 to 0.5 or even
0.6 per cent. Cu. It is of a reddish-green color containing ferric and
ferrous sulphates, free sulphuric acid, copper sulphate, and other salts.
AVhen the water supply is abundant, and the free acid is conse-
quently low, the consumption of pig iron varies from 1.3 to 1.5 tons per
ton of copper precipitated. This seldom occurs, however, and generally
there is needed 1.75 to 2 tons per ton of copper extracted; that is more
than double the theoretical quantity.
Treatment by Heap Roasting and Leaching. — Instead of extracting
or treating all the material by weathering, the extraction of the copper
may be expedited by first giving the heaps a slow roast, by which process
much of the copper in the cupriferous pyrites is converted into sulphate
and can readily be leached. The production of sulphate of copper l3y
slow heap roasting can only be used on ores that contain proportionately
large amounts of iron pyrites and small amounts of copper p3rrites or
' L' Echo des Mines et de la Metallurgie, 1906, p. 482.
214 HYDROMETALLURGY OF COPPER
other copper sulphides. In the slow and imperfect roasting in heaps or
stalls not all of the copper sulphide will be converted into sulphate;
a part will always remain unaffected, and another part will be trans-
formed into oxides. In order to extract as much copper as possible from
the ores, after the roasting has been finished and the sulphate of copper
leached out, the ores are allowed to weather, as for the fresh or unroasted
ores, already described, whereby in the course of time the copper is con-
verted into sulphate, by natural weathering.
The production of sulphate of copper by roasting, followed by weath-
ering of the leached roasted ore, was used for a long time at Rio Tinto,
in Spain. According to Schnabel,' the ores thus treated were cuprifer-
ous pyrites with 1 1/2 to 2 percent, copper; these were slowly roasted in
heaps of 200 to 1500 tons upon a bed of brushwood, firewood or coals.
The 200 ton heaps were hemispherical, 26 ft. in diameter at the base and
11 ft. 6 in. high. The larger, 1500 ton heaps were elliptical in plan, the
longer axis of the ellipse being 55 ft. 9 in., and the shorter axis 32 ft.
10 'in.; their height was also 11 ft. 6 in. Air was admitted by means of a
system of channels traversing the heaps. The smaller heaps burned for
two months, the larger ones for six. To roast 100 tons of ore required in
the small heaps 27 cu. ft. of wood and in the larger ones 9 cu. ft. The
yield of copper was greater in the small heaps than in the larger ones.
The roasted ores were leached for 50 hours by -which means the copper
present as sulphate was washed out of them. The exhausted residues
still contained 0.4 to 0.5 per cent, of copper, chiefly as sulphide, to extract
which the ores were allowed to weather. With this object they were piled
on a system of horizontal flues built of dry stone, that air could circulate
through the pile of ore. The vertical flues were continued in proportion as
the heaps got higher by the piling on of additional ore. As soon as the
damp heap had reached a certain height, the sulphides began to de-
compose, as was shown by the rising temperature. By checking the air
supply it was kept, if possible, from rising so high that the heap took fire.
From time to time, the heap, or a portion of it, if it was a large one, was
leached out, and the liquor conducted to the precipitating tanks. The ex-
haustion of the heaps, which were continually being increased, and which
may reach 500,000 tons, will not be completed in measurable time, as, in
spite of frequent leaching, the weathering proceeds very slowly. It is
even held that these huge heaps will still be producing sulphate of
copper long after the mines shall have been worked out.
The cost of producing copper at Rio Tinto, Spain, by heap roasting
and leaching, is stated to be about 11.55 per ton, divided as follows:
Mining 89 cents, roasting about 18 cents, including labor in building
heaps, etc., precipitating and collecting about 56.6 cents. Sixty-six tons
of ore were required to produce one ton of metallic copper.
' Handbook of Metallurgy, Vol. I, p. 212.
CHEMICAL PROCESSES
213
The cost of producing one ton of metallic copper by the cementation
process as carried out there is about $144.00.
The first cement copper was produced in Spain at Rio Tinto, in 1752
from heaps of low-grade sulphide ore that had undergone decomposition
through natural processes. The copper was leached out by water, the
metal being precipitated on iron. It was first thought to be merely a
coating of copper on the iron, but it was found if left long enough, the
replacement became complete.
ELIMINATION OF ARSENIC, ANTIMONY, AND BISMUTH.
Atmospheric Oxidation, Without Burning.^ — In atmospheric oxidation
of cupriferous pyrites, and subsequent extraction of copper by leaching,
as carried on in Portugal and Spain, arsenic and antimony are to some
extent dissolved and precipitated with the copper. The proportion in
which they are separated and precipitated is only a fraction of the
amount contained in the ores, as shown by the following tables by
Allen Gibb :
Pyrites
Precipitate
Per cent,
actual
Per cent, relative
Cu = 100per
cent.
Per cent,
actual
Per cent, relative
Cu-100 per
cent.
Total per cent,
of elimination
2.00
0.40
0.03
0.016
100.0
20 0
1.5
0.8
62.0
3.5
0.08
0.06
100.0
6.64
0.129
0.097
Arsenic ...
Antimony
Bismuth
71.8
91.4
87.9
Burning and Subsequent WashiTig. — In this process largely used in
Spain and to some extent in Portugal, for the extraction of copper from
cupriferous iron pyrites, there is a considerable elimination of arsenic,
antimony, and bismuth in burning. The following table, by Gibb, may
be taken as fairly typical of heap roasting, preparatory to leaching, as
conducted in Spain.
Raw Pyrites
Burnt Pyrites
Per cent,
actual
I Per cent, relative '
j Cu = 100per :
I cent.
Per cent,
actual
Per cent, relative
Cu = 100 per
cent.
Total per cent,
of elimination
' Allen Gibb, Trans. A. I. M. E., Vol. XXXIII, p. 667.
Copper
Arsenic
o 45
100 0
3.32 ^
100.0
0.43
17.55
0.133
4.08
76.8
Antimony
0.029
1.18
0.030
0.92
22.0
Bismuth
0.015
0.61
0.017
0.S2
14.8
216 HYDROMETALLURGY OF COPPER
CHLORIDE PROCESSES
The chloride processes have been widely applied to the extraction of
copper from its various ores. Hydrochloric acid, ferric and ferrous
chlorides, are the solvents usually employed.
Hydrochloric acid presents certain advantages over sulphuric acid in
the technical operation, but is usually more expensive, especially in
copper mining districts. Sulphuric acid, in the process of operation,
forms ferric sulphate, which when exposed to the air is decomposed into
basic siilphate of iron and free sulphuric acid, consequently consum-
ing more iron in the precipitation of the copper. Hydrochloric acid is
less apt to form basic salts and therefore yield solutions that contain but
little free acid, and which, accordingly, require less iron for the precipita-
tion of the copper, than do solutions containing ferric sulphate. On
the other hand it attacks oxide of iron more energetically than does
sulphuric acid, but the ferric or ferrous chloride so formed is more likely
to be precipitated out of the solution as the insoluble ferric oxide, than
from a sulphate solution.
Ordinarily only oxidized ores are applicable to treatment by a chloride
process. The copper may be dissolved either by
Hydrochloric acid, or
Metal chloride.
The acid, however, is the solvent usually employed. The chlorides
have been used as solvents on both oxide and sulphide ores but on
sulphide ores the action has been too slow for profitable application,
although both ferric and ferrous chlorides were quite extensively used at
one time.
Hydrochloric Acid. — If iron is used as the precipitant, as it usually
is, the hydrochloric acid process consists essentially of applying dilute
hydrochloric acid to the oxidized ores of copper, which reacts with the
copper oxide thus:
CuO+2HCl = CuCl2 + H20
to form cupric chloride. Some cuprous chloride may also be formed:
The cupric chloride thus formed is filtered from the ore and pre-
cipitated with iron, thus:
CuCl2+Fe=FeCl2 + Cu
the iron and copper changing places. The copper is precipitated while
the iron goes into solution as ferrous chloride. Theoretically the same
amount of iron is required to precipitate a pound of copper from cupric
chloride solutions as from cupric sulphate solutions i.e., 56 lb. of iron to
precipitate 63.6 lb. of copper. In practice, however, it will usually take
less for chloride than for sulphate solutions. If the copper in solution
is in the form of cuprous chloride, only half the amount of iron is required
CHEMICAL PROCESSES '217
as that used when the copper is in the form of either cupric chloride or
cupric sulphate.
If hydrochloric acid is used as the solvent is takes theoretically
approximately 0.6 lb. of acid to extract 1 lb. of copper as cuprous chloride,
and 1.1 lb. as cupric chloride.
In practical operation, much more acid is required, the amount
depending largely on the nature of the ore. The theoretical amount of
acid given to react with the copper, is supposed to be the pure acid, free
from water. Commercial hydrochloric acid, is the HCl dissolved in
water, so that in estimating the commercial acid required, it will be
necessary to first know the acid content, which can be determined from
its specific gravity.
Concentrated HCl + Aq loses HCl, and dilute HCl + Aq loses water
on warming, until an acid of constant composition is formed, containing
20.18 per cent. HCl with a specific gravity of 1.101 at 15° C, which can
be distilled unchanged at 110° C. Concentrated HCl+Aq gradually
gives off HCl on the air until it has a specific gravity of 1.128 at 15° C,
and contains 15.2 per cent. HCl. In leaching copper ores, only very
dilute solutions of hydrochloric acid are usually employed, rarely exceed-
ing 5 per cent. HCl.
In Stadtberg, Westphalia, hydrochloric acid was formerly used to
extract copper from ores containing from 1 to 2 per cent, copper.' The
leaching vessels were rectangular tanks of wood 4 ft. 11/2 in. high,
packed in a lied of clay 1 ft. 3 in. thick, then fitted with a grating of
wooden bars, upon which the ore was piled. The first tanks held 29 tons
of ore, but those erected afterward had a capacity of 90 tons. All the
tanks were situated on a level. The leaching was methodical; ores
nearly free from copper being treated with fresh hydrochloric acid of 12
to 13° B., as obtained from soda works, while frejh ore was treated
with partially saturated solution until the latter was fully saturated,
which took place at 19 to 20° B. The various solutions were allowed
to remain 12 hours in each tank, the saturation point being reached in
10 to 12 days. The solution was circulated by means of pumps and
bucket wheels. After the solution had percolated through the ores, it
ran out thxough a plug hole to which the bottom of the tank was in-
clined, into receivers, whence it was again lifted to its proper tank.
The exhausted ore was allowed to lie for another 12 to 15 hours in water
and was then washed for 12 hours more. Fresh acid was diluted with a
portion of the mother-liquor. For each 100 parts by weight of copper,
550 to 700 parts of hydrochloric acid 12 to 15° B., were employed.
This process replaced the sulphuric acid leaching, previously employed
but was later abandoned when the carbonates of the ore were replaced
with sulphides in depth.
'Schnabel Handbook of Metallurgy, Vol. I, p. 200.
218 HYDROMETALLURGY OF COPPER
At Twiste, in Waldeck, attempts were made to leach malachite and
azurite copper ores occurring in the Bunter Sandstein (Lower Trias) and
containing 1/2 to 1 per cent, of copper, by means of hydrochloric acid,
but had to be abandoned because the ores contained from 1/2 to 1 per
cent, of lime which was dissolved by the acid before it attacked the copper.
Ferric Chloride, FeClj.— Ferric chloride, like ferric sulphate, has the
property of dissolving copper from its oxide, carbonate, and sulphide
combinations. If the sulphides of iron, copper, lead, zinc, arsenic and
antimony are treated with a hot solution of ferric chloride, the minerals
are more or less decomposed, and the respective metals go into solution
as chlorides, usually with the liberation of sulphur. Hydrochloric or
sulphuric acid, added to the solution, aids the reactions. Silver is con-
verted into chloride by the action of ferric chloride. Gold remains
unaffected.
If ferric chloride is used as a solvent for ore containing the copper as
cupric sulphide (CuS) , cupric chloride is produced and the ferric chloride
is reduced to the ferrous chloride:
CuS + 2FeCl3 = CuClj + 2FeCl , + S.
If the copper mineral contains the copper as cuprous sulphide the
reaction will result in the formation of cupric and cuprous chlorides:
Cu,S + 2FeCl3 = 2CuCl + 2FeCl2 + S
Cu,S + 4FeCl3 = 2CuCl2 + 4FeCl2 + S.
With a neutral solution of ferric chloride, the reaction with copper
oxide is:
3CuO +2FeCl3 = SCuCl^ +Fe203
although some cuprous chloride may also be formed by the reduction of
the ferric to the ferrous chloride, and the subsequent reaction of the
ferrous chloride on the copper oxide. If the solution is somewhat acid,
the ferric chloride will be reduced, but the ferrous chloride will not be
precipitated to any great extent. With copper carbonates the reaction
is much the same as for oxides, except that carbon dioxide is liberated
in the reaction.
In reacting with sulphide ores, the ferric chloride is reduced to the
ferrous chloride, but the ferrous chloride does not further react with
the copper sulphides. In any event, the sum total of the reactions may
be considered to be the reduction of the ferric to the ferrous chloride,
with the formation of cupric and cuprous chlorides.
The copper may be precipitated with iron:
CuJlj +Fe =Cu +FeCl2,
2CuCl+Fe =2Cu+FeCl2,
CHEMICAL PROCESSES 219
ferrous chloride being formed. If ferric chloride exists in the solution
during precipitation with iron, it will be reduced to ferrous chloride at the
expense of the iron.
The ferrous chloride, after precipitation of the copper, may be re-
generated back to ferric chloride by the action of air, chlorine, or hydro-
chloric acid. If the neutral ferrous chloride is regenerated by agitation
with air, some of the iron is brought into the condition of basic salts, and
much is precipitated as the ferric oxide.
6FeCl2 + O3 =4FeCl3 +Fe203.
From this equation it is apparent that one-third of the iron is pre-
cipitated from the solution as the insoluble ferric oxide in order to raise
the remaining two-thirds to the ferric condition. In addition to the
precipitated feriic oxide there may be formed, in netural solutions, in-
soluble oxychlorides, and these necessitate a further loss, both of iron and
chlorine. If, however, iron is used as the precipitant, large amounts of
ferrous chloride are produced which may then be brought to the ferric
condition for reuse as a solvent, and the loss of iron as ferric oxide or
oxychloride, is hot a serious matter; in fact an elimination of a certain
amount of the iron is an absolute necessity.
If the ferrous chloride solution is regenerated with chlorine, oxy-
chlorides are not formed, although some iron may be precipitated as fer-
ric oxide. The chlorine may be produced either chemically or elec-
trolytically, by any of the well-known methods. The chlorine combines
directly with the ferrous chloride to produce the ferric chloride;
FeCl2 + Cl=FeCl3.
The solubility of copper from sulphide ores with ferric chloride solu-
tions, depends much on the way the copper is mineralogically combined.
The copper in the form of chalcocite is much more readily soluble than
in the form of chalcopyrite, while gray copper remains quite unaffected.
Careful roasting at a low temperature makes the copper, in any of its
sulphide combinations, readily soluble, but when roasted at a high
temperature, the copper is quite insoluble either in acids or solutions of
ferric salts. This is doubtless due to the formation of silicates and of
ferrites.
Doetsch Process. — Ferric chloride acts on cupric and cuprous sul-
phides to form cupric and cuprous chlorides, while the ferric chloride is
reduced to the ferrous condition. The reactions may be expressed thus:
CuS +2FeCl3= CuCl2 + 2FeCl2 + S
Cu2S+2FeCl3 = 2CuCl +2FeCl2 + S.
On these reactions are based the Doetsch Process, formerly used ex-
tensively at Kio Tinto in Spain. The ore there treated contained on an
220 HYDROMETALLURGY OF COPPER
average of 2.7 per cent, copper. The copper passed into solution while
the pyrite was practically unaffected in the leaching operation. The
process has the advantage of reducing the waste of iron in the precipitating
tanks by avoiding the formation of ferric sulphate. In this way one ton
of copper precipitate was obtained with an expenditure of about an
equal weight of iron.
In the process as carried out at Rio Tinto the ore was crushed to
about 1/2 in., mixed with 0.5 per cent, common salt and a like amount of
ferrous sulphate, and built into large heaps. These heaps were from 12
to 16 ft. high and 50 ft. square at the base. A solution of ferric chloride
was run in a continuous steam upon these heaps.
It took about four months to extract 1.34 per cent, of the copper, or
50 per cent, of the total copper content. After two years 2.20 per cent,
was extracted or 80 per cent, of the contained copper. The final loss
was about 0.48 per cent.
The solutions leached through the heaps, containing cupric and cup-
rous chlorides as also ferric and ferrous chlorides, were precipitated with
iron:
CuCl2+Fe = Cu -hFeClj
2CuCl+Fe = 2Cu+FeCl2
ferruos chloride being formed. The ferric chloride, by its action on the
ore, is changed to ferrous chloride; small amounts of ferric chloride
may remain unchanged in the solution, but this reacts with the iron in the
precipitating tanks and may be reduced there to the ferrous condition.
To make the process continuous, the ferrous chloride is regenerated
back to ferric chloride, which is accomplished by bringing the ferrous
chloride solution in contact with chlorine, in scrubbing towers.
The solution which ran from the ore heaps contained 5 to 7 kilogrm.
of copper per 1000 kilogrm., or 1 cubic meter.
The chlorine gas, for the regeneration of the ferrous chloride to ferric
chloride, was produced by heating salt with ferrous sulphate in a re-
verberatory furnace, holding 500 lb. An abundance of air was admitted
through the working doors. The ferrous sulphate reacting with salt in
the presence of air, produces chlorine, sodium sulphate, and ferric oxide.
2FeSO, + 4NaCl + 03=Fe203 + 2Na2SO, + 4Cl.
The ferrous sulphate used is found in large quantities on the shores
of the Rio Tinto river. The immense heaps of low-grade ore, "Toreros"
being leached by natural cementation, also furnish salts of ferric and
ferrous sulphate. During the reaction in the reverberatory furnace,
some hydrochloric acid is formed; this is converted into chlorine by the
action of manganese dioxide of which there was a certain amount placed
near the flues of the furnace.
CHEMICAL PROCESSES 221
The gases from the furnace, consisting largely of chlorine with pos-
sibly some hydrochloric acid, were conducted to scrubbing towers, where
they came in contact with the ferrous chloride solution from the precipi-
tating tanks, and the ferrous chloride reconverted to the ferric chloride.
The Doetsch process was used both for raw and roasted ore. When
roasted, the roasting was performed in heaps of truncated pyramidal
form, 10 ft. high, with a base 20x26.5 ft. for those containing 800 tons,
and 30X26.6 ft. for those containing 1200 tons.
At the bottom of the heaps, one transverse and three longitudinal
flues, about 20 in. square, were formed by the larger blocks of mineral.
Those were the firing passages, and communicated with vertical chimneys,
of which there were two in the 800-ton, and three in the 1200-ton heaps.
The mass of the heaps was made of lumps a little above nut size. Salt
was added in the proporiton of 14 tons to 800 tons of pyrites ore. When
the fire was started with a little wood, it was kept up by the heat of the
burning sulphur. It was essential that no rich ore was included, as, on
account of the action of kernel roasting, lumps of rich sulphide of copper
were formed which could not subsequently be dissolved. Some of the
roasted ore was at times mixed with the unroasted ore, for the extraction
of the copper with the ferric chloride solution. The reactions between
the salt and ferrous and ferric sulphates in the heaps may be expressed
as follows :
2FeSO,-|-4NaCl + 03=Fe203 + 2Na2S04-l-4Cl.
Fe2(SO,)3-h6NaCH-03 = Fe2034-3Na2SO, + 6CL
The chlorine liberated, acting on the iron and copper sulphides, pro-
duces ferric and cupric chlorides. There are therefore present, FeSO^,
Fe2(S04)3, CUSO4, FeClj, and CuClj, with an excess of salt, which changes
the ferric sulphate into the ferric chloride.
An interesting modification of the Doetsch process was for some time
carried out on a very extensive scale at Naya, close to Rio Tinto. In
this method, the heaps, made in the ordinary way, as soon as they began
to give off sulphurous fumes, were covered up with a fresh quantity of
mineral, partly raw, and partly roasted, to which 2 to 3 per cent, of salt
and a similar proportion of manganese dioxide has been added. The
whole was formed into heaps 26 ft. high with a flat top, which was divided
by gutters into squares of 25 ft. The remaining operations were effected
in the usual way, the heaps being watered at intervals for months and
years, the copper being slowly dissolved, and collected at the bottom of
the heaps. It was necessary to break up the surface with a pick at inter-
vals, to prevent it from becoming impermeable to water. The sulphur-
ous acid gas in this modification of the process, in the presence of steam
formed by the heat developed in the heaps, produces sulphuric acid,
which acts upon the oxides in the crust of the roasted material. The
222 HYDROMETALLURGY OF COPPER
salt and manganese dioxide may, jointly with the sulphuric acid, evolve
chlorine, forming ferric chloride, which decomposes the sulphides of
copper and silver. It is also possible that, under the action of heat and
sulphuric acid, oxygen is evolved which acts directly upon the pyrites.
The precipitation of the copper, in the Doetsch process, was effected
in a series of tanks, 330 ft. long and 33 ft. wide, divided into 10 parallel
series, receiving a uniform supply of copper solution. The total length
was about 1300 ft. with a difference of level of 13 ft., which gave a
sufficiently rapid flow. Pig iron, as run from the furnace, and scrap iron,
were used; the scrap iron was put into baskets. Every 10 days the iron
was removed, and after scraping to collect the deposited copper, it was
returned to the tanks. The consumption of iron was a little more than
one ton of pig iron per ton of copper produced.
The precipitate, as collected, was very impure, containing only
65 per cent, to 70 per cent, copper, the remainder being ferric oxide
more or less arsenical, graphite from the pig iron, silica, etc. After
treatment with water acidulated with sulphuric acid which dissolved
the basic ferric arsenates without touching the copper, it was passed
over 4/10 in. mesh sieves to separate pieces of cast iron. The precipitate
passed through the sieves was washed in a current of water, where it
separated, according to the order of density, into "Cascara," or copper;
"Graphita," or particles of coal and graphite; and "Pucha," a fine black
sand. The copper was smelted to blister copper; the graphita was
smelted with rich ore, and the Pucha was made into balls, dried, and
also smelted with the ore.
The cost of producing copper by the Doetsch process has been esti-
mated by Cumenge, according to the results which he accomplished after
a campaign of four months, during which he obtained 224 tons of cement
copper.^
Frances
Cost of leaoMng, 83 . 81
Precipitation, 181 . 16
General expenses, 28 . 04
Total per ton of cement copper, 293 . 01
Or per ton of pure copper, 345 . frances,
or $67.93
The extraction at the expiration of four months amounted to 1.34
per cent, of the 2.7 per cent, copper contents. 1.12 ton of iron was re-
quired to precipitate one ton of cement copper which is equal to 1.3 tons
of iron to one ton of pure copper.
Dr. 0. Froelich^ made some experiments in dissolving copper from
' Notes sur le Rio Tinto by M. E. Cumenge, Annales des mines, Vol. XVCI.
2 " Metallurgie," 1908, p. 206.
CHEMICAL PROCESSES
223
various sulphide ores by agitation with a hot solution of ferric chloride,
the results of which are given in the following table:
Characler of
material
Chalcocite . .
Chalcocite. . .
Chalcocite. . .
Chalcocite. ,.
Chalcopyrite,
Chalcopyrite
Gray copper.
Extent of
comminution
Powder. .
19-32 in.
13-32 in .
5-32. in
Powder. .
Powder. .
Powder. .
Time leached
hours
15 1/2
24
66 1/2
9
24 1/2
36
Copper cont.
Copper cont.
of ore,
of tails,
per cent.
per cent.
0.94
0.02
0.52
0.13
7.6
0.51
7.8
1.09
7.8
small
16.7
0.53
22.7
15.1
Per cent, of
extract
99
76
93
86
100
97
34
In these experiments the chalcopyrite was first subjected to a tem-
perature of about 200° C. without the admission of air. The chalcocite
was treated without heating.
The Froelich Process.' — In this process the ore is first subjected
in the absence of air to a temperature between 150 and 800° C,
whereby the loose sulphur is driven off. The pyrite is changed by
this operation in its chemical composition and can then be chlorinated
much quicker and better. As chlorinating gases, chlorine, vapor of
hydrochloric acid and of ferric chloride are used. The chlorine and
ferric chloride attack the sulphur compounds of copper, the hydro-
chloric acid and the ferric chloride attack the ox'des of the copper.
The proportion of the gases and vapors in the mixture are adjusted
to the composition of the copper ore. Steam may also be added
to the chlorinating mixture. It is preferable to make the temper-
ature of the chlorination somewhat higher than the boiling point of the
ferric chloride, about 300° C, but it can be lower if the copper in the ore
occurs in combination with sulphur. In this case only chlorine gas is
used. If iron is present in the ore it is chlorinated together with the
copper, but more slowly, and the process is facilitated by a higher tem-
perature. In order to regain the chlorine combined with the iron, the
ore is heated to about 300° C. or more after chlorination, and a certain
quantity of air is introduced during this operation.
The ferric chloride Is then evaporated, and by the air separated into
oxide of iron and chlorine gas, both of which are collected. The chloride
of copper is not changed by this operation. Then the ore is treated
with hot water, and the chloi'ide of copper and, perhaps, a residue of
ferric chloride are extracted. The solution is then introduced into a
revolving apparatus containing pieces of iron, and the metallic copper
is deposited in the form of cement copper. The solution now contains
mainly ferrous chloride, and in order to regain the chlorine from it the
' U. S. Patent, 846,657, March 12, 1907.
224 HYDROMETALLURGY OF COPPER
solution is oxidized in a rotating drum to ferric chloride by means
of an air blast. Then the water, and the wat^r of crystallization
are driven off by heating, and by increasing the temperature over the
boiling point of ferric chloride and introducing a certain quantity of
air, chlorine gas and ferrous chloride are obtained.
W. L. Austin^ suggests the following method for treating with ferric
chloride pyritic copper ores, containing the copper largely as chalcocite.
The results obtained in the laboratory have been very satisfactory.
The novel features introduced are (1) causing the lixiviant to rise through
the ore and removing in an apparatus placed outside of the leaching vat
any slimes which may be carried over; (2) causing the lixiviant to cir-
culate rapidly through the material treated without employing moving
parts within the leaching vat; (3) regeneration of the lixiviant by
treating it with chlorine gas produced by the electrolysis of common
salt in an apparatus specially provided for that purpose; and (4)
cementation with the aid of the coke-iron couple. These features make
it possible to have each succeeding stage in the process under direct
supervision, and avoids complications caused by attempting to carry out
two distinct operations concurrently in the same apparatus. It also
avoids the use of moving parts submerged in a gritty and corrosive
material.
On the basis of a plant treating 100 tons of ore in 24 hours, and
assuming a 2 per cent.' ore (40 lb. copper to the ton), chlorine at $0.02
per pound, arid a 90 per cent, extraction, the following estimated cost of
producing one pound of refined copper by this method is derived:
Cost
per lb.
copper
Milling operations, comminution of the ore to ten-mesh, charging
and discharging the tanks, elevating liquors, etc., total $0.50
per ton of ore treated, $0 . 014
Chlorine, at $0.02 per lb., 0.015
Iron, at $30.02 per ton for pig delivered at works, 0.015
Melting the precipitate into bars, at $8.00 per ton of precipitate, 0 . 005
Freight, refining charges, and selling expenses, 0.014
Repairs and renewals, office expenses, etc., 0.013
$0,076
Mining operations — open pit work, at $0.50 per ton of ore, 0.014
Total operating expenses, $0 . 090
In this estimate no item has been inserted representing interest on
cost of plant, nor amortization; on the other hand, no allowance is made
for precious metals recovered, nor for possible commercially valuable
' Mines and Methods, January 1911.
CHEMICAL PROCESSES 22.-,
bi-products. The figures given for costs of reagents consumed are
liberal estimates.
Ferrous Chloride Process. — The action of ferrous chloride on copper
carbonate was demonstrated by experiments of Schaffner and Unger
in 1862. Some years later Hunt and Douglas based a process on the
action of ferrous chloride on copper oxide and carbonate, which was
for some time working on a commercial scale at Ore Knob, North Caro-
lina, and at Phoenixville, Pa.
Cupric oxide and cupric carbonate are acted upon by solutions of
ferrous chloride, forming cupric and cuprous chlorides and oxide of iron,
and in the case of carbonates, there is also set free carbon dioxide,
according to the following equations:
3CuO + 2Fe( "I^ = CuCl^ + 2(;uCl + FcO^.
3CuC03 + 2FeCl2 = ('uCl2 + 2CuCl+Fe203+3C(),.
The ferric oxide is precipitated, while the chlorides of copper go
into solution. Cuprous chloride, being insoluble in water, is dissolved
by excess of other metal chlorides.
If the ore to be treated is a sulphide, or a matte, the copper must
first be converted into the oxide by I'oasting. Roasting is unnecessary
with oxidized ores.
The ferrous chloride used in the process, may be produced from
common salt and ferrous sulphate, which by double decomposition forms
ferrous chloride and sodium sulphate. After the solution was separated
from the sodium sulphate, which crystallized out, if was ready for use.
When copper is precipitated from a cuprous chloride solution by
iron, ferrous chloride is formed; when both cupric and cuprous chlorides
are present, ferric chloride is first formed,
2CUCI2 + 2CuCl + 2Fe = 2FeCl3 + 4Cu.
Ferric chloride, however, reacts with iron to produce ferrous chloride
2reCl3+Fe=3FeCl2.
The silver in the ore is converted into chloride of silver by the cupric
chloride, and is dissolved in the excess of other metal chlorides.
As some of the copper is in the cuprous condition, relatively less
iron is required to precipitate the copper, than is required in precipi-
tating from a sulphate solution. The objections to the process was the
formation of basic salts, and the difficulty of separating the solution
from the residue, on account of the precipitated oxide of iron, which
clogged the filters. Heat is not necessary in dissolving the copper, but
it hastens the process, and makes the extraction more thorough.
The silver, in solution as silver chloride, may be precipitated witii
the copper, or separately if ilesired, by metallic copper. The precipi-
tation of the silver by metallic copper in the presence of chlorides re-
15
226 HYDROMETALLURGY OF COPPER
quires that the whole of the dissolved copper shoiild be in the cuprous
condition. The clear solution containing the cupric and cuprous chlo-
rides and the silver chloride may be treated with sulphur dioxide to
convert the cupric chloride into the cuprous chloride, and with liberation
of free acid. From such an acid solution any silver present is readily
and completely precipitated by metallic copper, after which the whole
of the dissolved copper may be precipitated with metallic iron, care
being taken to arrest the process before the free acid begins to act on
the iron. In this way a solution is obtained containing, besides the
regenerated ferrous chloride, a considerable amount of free acid.
When copper ores containing lime are being treated, there is the
difficulty to contend with that the ferrous chloride and calcium carbonate,
m the presence of air, are decomposed into calcium chloride and ferric
hydroxide.
According to Hunt' the hydrous silicate of copper (chrysocoUa) is,
like carbonate of copper, completely decomposed by a hot solution of
ferrous chloride with common salt.
The Ferrous Chloride Process as Carried out at Ore Knob, Ashe Co.,
N- C.^ — At Ore Knob, in North Carolina, the ore was crushed to 40 mesh
and roasted. The average composition of the crude ore, as taken from
the two shafts, was as follows:
No. 1
No. 2
Chalcopyrite,
11. 33 per cent.
13. 30 per cent.
Pyrhotite,
37.46
35.74
Ferric oxide,
8.14
16.34
Alumina,
1.84
1.49
Manganese,
0.16
0.50
Lime,
5.32
7.84
Magnesia,
0.35
0.94
Carbonic acid,
4.76
7.19
Zinc,
0.67
0.66
Cobalt,
0.09
0.09
Nickel,
0.71
0.92
Silioious residue,
29.10
13.57
99.93
98.58
Metallic copper.
3.92
4.60
This ore was sorted to bring the average copper content up to about
12.0 per cent. The average composition of the copper in the roasted
ore is represented by the following analysis :
Copper as sulphate, 3 . 76 per cent.
Copper as oxide, 7 . 75 per cent.
Copper as sulphide, 0 . 39 per cent.
11. 90 per cent.
' Trans. A. I. M. E., Vol. X, p. 12,
^ T. A. I. M. E., Vol. II, p. 394, E. E. Olcott.
CHEMICAL PROCESSES . 227
The roasted ore was conveyed to agitating tanks, which were eight
in number, 8 ft. in diameter, and 5 ft. deep, with raised conical bottoms.
These tanks were each charged, once in 24 hours, with 3000 lb. of roasted
ore and 1500 gallons of the solution of ferrous sulphate and common
salt, making about 22° B., and heated by steam to 160° F. This mixture
was kept in agitation for eight hours by means of suspended stirrers,
consisting of a vertical shaft with a horizontal blade at the lower extremity,
while at the top was attached a bevel gear which gave to the stirrers a
speed of 25 revolutions per minute. After eight hours the stirrers were
stopped, and the contents of the tank allowed to settle for four hours,
when the clear liquor was drawn off into the precipitating tanks, and the
remaining portion holding in suspension ferric oxide and particles of
gangue, was drawn into settling tanks. The sands, remaining, were
then washed, first with hot strong solution, and then with weaker solu-
tion. These washing liquors were allowed to settle in the settling tanks,
when the clear portion was drawn off into precipitating tanks containing
iron. The wet sands were then removed from the agitators to leaching
tanks where a portion of the adhering solution, containing copper, was
recovered.
The slimes were allowed to accumulate in the settling tanks till they
were about half full, when they were washed with solution and water till
they contained about 1/2 of 1 per cent, of copper. They were then
washed. The settling tanks were 20 in number, 10 ft. in diameter and
5 ft. deep.
The strong liquors from the agitators were capable of holding 50 lb.
of copper in solution per 100 gallons, but weaker solutions were desired
as they lessened the risk of the deposition of cuprous chloride by cooling.
30 lb. of copper per 100 gallons was found a convenient strength.
The hot and strongly colored liquors were run through launders into
the precipitating tanks. These were 12 in number, 12 ft. in diameter
and 5 ft. deep, containing each, 12,000 lb. of scrap iron. The
temperature of the precipitating tanks was maintained at 160° F. by the
injection of steam. From 12 to 18 hours sufficed to precipitate all but a
trace of the copper from the liquors, which were then drawn off into a
lower tank and from there pumped into the stock tanks to be again
used on a fresh portion of the ore.
For the precipitation of the copper wrought iron was used. The
copper was removed from the precipitating tanks when they contained
from 4 to 5 tons each. The consumption of iron was 70 per cent, of the
pure copper produced. The cement copper, after being washed and
dried, contained generally from 75 to 85 per cent, copper. The impuri-
ties were chiefly ferric oxide and earthy gangue matter.
The cement copper produced at Ore Knob costs a little less than
8 cents per pound of copper, including all expenses for mining, treating
228 HYDROMETALLURGY OF COPPER
and packing. Of this sum nearly two cents was for metallic iron. These
costs were based on a production of 400,000 pounds of copper.
Hunt and Douglas Process. i— In the Hunt and Douglas process the
copper is dissolved as sulphate and precipitated as the insoluble cuprous
chloride. The cuprous chloride is then converted into metallic copper by
replacement with iron.
The process is based on the reaction described by Wohler between
sulphur dioxide and a solution of cupric chloride, in which one half of
the chloride is eliminated to form hydrochloric acid, and with the simul-
taneous formation of sulphuric acid. The reaction may be expressed:
2CuCl2 + SO2 + 2H2O = 2CuCl + 2HC1 + H3SO 4.
The Hunt and Douglas process consists of the following essential
sii'i^s:
1. Roasting, if the ore is a sulphide.
2. -Extracting thecopper from the oxidized ore with dilute sulphuric
acid, regenerated in the operation of the process.
3. Conversion of the cupric sulphate into cupric chloride, by the
addition of some soluble chloride, such as sodium, calcium or ferric
chloride.
4. Conversion of the soluble cupric chloride into the insoluble cuprous
chloride by the addition of sulphur dioxide, and with the simultaneous
regeneration of acid. The acid being used in the second step to dis-
solve more copper.
5. Conversion of the precipitated cuprous chloride into cupric oxide
or metallic copper on the addition of milk of lime or replacement with
iron.
In practically carrying out the process, the matte, or ore if a sulphide,
is roasted. With care in roasting one third of the copper should be
converted into the sulphate, which is soluble in water. The ore is then
leached with dilute sulphuric acid, regenerated in a later stage in the
process. The copper content of the ore is therefore dissolved as the
cupric sulphate, CuSO^. To the neutral solution of cupric sulphate is
then added enough common salt, or other soluble chloride, to convert the
copper sulphate in the solution into that of cupric chloride. The
amount of copper sulphate being determined, salt is added in the pro-
portion of 58.5 parts of sodium chloride to 63.6 parts of copper contained
in the sulphate solution. The salt reacts with the cupric sulphate to
form cupric chloride and sodium sulphate:
CuSO, + 2NaCl = CuCl,-hNa2S04,
so that the copper in the solution, after the application of the sodium
chloride will be in the form of cupric chloride, containing possibly a
little cupric sulphate.
' Trans. A. I. M. E., Vols. X and XVI.
CHEMICAL PROCESSES L'Jl)
Through the clear hot solution of cupric chloride is then driven sul-
phur dioxide, derived from roasting the sulphide ore. The sulphur
dioxide serves to convert the dissolved copper into the form of cuprous
chloride, with the liberation of the amount of acid which was previously
combined with the copper, and the liberation of one half as much more
acid due to the oxidation of the absorbed sulphur dioxide.
The combined reaction between the copper sulphate, salt, and sul-
phur dioxide may be represented by the equation :
CuS0,-t-2NaCl-|-2SO2-h2H2O = 2ru01-|-Na,.SO,-f-2H.,S(),.
It is reasonably certain, however, that the following reaction also tiikes
place, whereby a certain amount of hydrochloric acid is also produced:
2CuCl2-|-S02-|-2H20 = 2CuCl-l-2HCl-t-H2SO,.
The cuprous chloride thus obtained, which is insoluble in water or
in a sulphate solution, quickly settles to the bottom of the tank, as a
white crystalline powder. The clear acid solution, drawn from the pre-
cipitated cuprous chloride, is again applied to the ore, and the cycle
repeated as often as necessary to get the desired extraction.
The resulting cuprous chloride, seperated from the solution, may
then be precipitated with metallic iron, as metalic copper,
2CuCl-hFe = 2Cu-t-FeClj,
or with milk of lime, as cuprous oxide,
2CuCl+Ca(OH)2 = Cu20-t-CaCl2 + H20,
calcium chloride being formed, which may be used to convert the sul-
phate of copper into the chloride instead of the salt, or ferric chloride.
By the use of a solvent containing only a small portion of soluble
chloride, any silver in the ore is converted into the chloride, but remains
in the residue and may be extracted therefrom by solution, amalgama-
tion, or smelting.
The reaction between sulphur dioxide and a solution of cupric chlo-
ride goes on slowly at ordinary temperatures', but is very rapid between
80 and 90° C. (176 to 194° F.). Solutions of sulphate of copper, mixed
with an equivalent of chloride of sodium, and holding 8 per cent, of
copper, after being treated at 90 ° C. with an excess of sulphur dioxide,
retain less than 1 per cent, of the dissolved copper, while in the presence
of an excess of sulphate of copper and sulphur dioxide, the precipita-
tion of the chlorine from chloride of sodium is nearly complete.
The sulphur dioxide, from the roasting furnace, is sufficiently pure
for use. A Knowls pump, connected for the purpose, has proved an
efficient means of injecting the heated gas into the liquid.
The acid liquors, when the reaction with sulphurous acid is com-
230 HYDROMETALLURGY OF COPPER
pleted, have exchanged their bright blue color for a pale green, and now
contain in solution an excess of sulphur dioxide which must be got rid of
before using it to dissolve a fresh portion of copper. This may be effected
by keeping back a small portion of the chlorinated copper solution, and
after the reduction of the gas is complete, as may be shown by the changed
color and the sulphurous odor of the liquid, adding the reserve portion
thereto, by which means the excess of siilphurous acid will be oxidized.
The larger part of the cuprous chloride separates during the passage
of the gas but a furthur portion is deposited on the cooling of the
solution.
The excess of sulphurous acid may also be got rid of by blowing a
current of hot air through the liquid after it has been withdrawn from
the precipitated cuprous chloride.
Cuprous chloride is quickly transferred into cupric oxychloride by
atmospheric oxygen and when dissolved or suspended in an acid liquid
is by this means converted into a cupric salt, which may be again reduced
to cuprous chloride by the action of sulphur dioxide.
If the ore or roasted matte contains silver, the sulphate of copper,
which in well-roasted ore should be about one-third of the copper con-
tent, is first dissolved out with water, taking care, however, to add enough
of some soluble chloride to chlorinate and render insoluble any sulphate
of silver which may be present. From the clear solution thus obtained,
after adding the requisite amount of chloride of sodium, the copper is
precipitated as already described, by the action of sulphur dioxide. The
resulting acid liquor, freed from its sulphur dioxide, is now used to dis-
solve the oxide of copper in the ore, the process being aided by heat, and
if the formation of cuprous chloride is feared a current of heated air
may be injected and made the means of agitating the mixture. If
the ore contains the silver as metal or oxidized sulphide, the chloride of
copper formed is a good agent for bringing it into the condition of silver
chloride. This will be found in the residue after the extraction of the
copper, together with any gold which may be present; lead as sulphate,
oxides of antimony and iron and earthy matters. Cobalt, nickel and
zinc, if present, will however be dissolved, and not being precipitated
by sulphurous acid, will by successive operations, accumulate in the
solutions and may be afterward extracted. From the residues, the
silver may readily be extracted by brine, after which the gold, if present,
may be recovered by chlorination, or the precious metals extracted
together from the residues by amalgamation.
Chloride of silver is soluble to some extent in a solution of cupric
chloride, and is then in part carried down with the cuprous chloride in
the precipitation of the latter.
The cuprous chloride as obtained by the precipitation with sulphur
dioxide is a white coarsely crystalline powder, having a specific gravity
CHEMICAL PROCESSES 231
of 3.376 and is nearly insoluble in water. After being washed from the
acid liquid, it may readily be reduced by placing metallic iron in
the moist cuprous chloride, which should be covered to exclude the air.
The action spreads rapidly through the precipitate, so that a single mass
of iron, within a few hours, will change a considerable volume of cuprous
chloride, around it, into a pure spongy metallic copper. Twice the
amount of copper is, theoretically, precipitated by iron from a cuprous
than from a cupric solution. 45 lb. of iron will suffice to reduce 100 lb.
of copper from cuprous chloride. The ferrous chloride which remains in
the solution may with advantage be used instead of sodium chloride
for chlorinating subsequent solutions of copper sulphate.
Another method of treating the cuprous chloride, consists in decom-
posing it, preferably at a boiling heat, with a slight excess of milk of
lime. The cuprous chloride is by this means converted into a dense
orange-red cuprous oxide, which after being washed from the chloride of
calcium in a filter press or otherwise, and dried, may be readily reduced
to metallic copper, in a reverberatory furnace. For this reaction, 28.0
parts of quick lime are required for 63.4 parts of copper, and the result-
ing chloride of calcium may be used instead of sodium chloride or iron
chloride for chlorinating solutions of copper sulphate. In this case there
will be formed an insoluble sulphate of lime, while the free sulphuric
acid of the solution is replaced by hydrochloric acid.
Later, Douglas proposed electrolyzing the precipitated solid cuprous
chloride, to deposit metallic copper, and use the chlorine again directly
on finely crushed matte. A description of this will be found under
Electrolytic Processes, page 347.
The Hunt and Douglas process was in operation for many years on
a large scale, until quite recently, at Argentine, Kansas, on copper matte.
The description here given, of the practical operation of the process, at
Argentine, is by Ottakar Hofmann.
The Hunt and Douglas Process at Argentine, Kansas.' — The material
treated at the plant of The Kansas City Smelting and Refining Company,
at Argentine, Kansas, was a lead-copper matte, averaging:
Copper, 39 . 55 per cent.
Lead, 12. 26 per cent.
Iron, 19 • 90 per cent.
Zinc, 1 • 88 per cent.
Manganese, 1 ■ 01 per cent.
Sulphur, 21. 43 per cent.
Silver varied from 200 to 300 oz. per ton.
This matte was first crushed by rock breakers, then pulverized in a
Krupp ball mill to pass a 50-mesh screen.
The roasting was done in two Pearce two-hearth furnaces. On the
' Ottokar Hofmann, Mineral Industry, Vol. XVII, 1908, p. 296.
232 HYDROMETALLURGY OF COPPER
upper hearth the temperature was kept as low as the heat developed by
the oxidation of the sulphur permitted. No fire was applied except
after the ore had passed the whole circle of the hearth and came near to
the slot through which it dropped into the lower hearth. There a very
gentle fire was maintained to prevent the temperature from falling too low.
The best results were obtained by regulating the roasting on the upper
hearth so that the material commenced to ignite when it had moved about
8 ft. from the point at which it had entered the furnace. By observing this
precaution the roasting was so much advanced by the time the material
had reached the drop-slot that the oxidation of the sulphur did not
create more heat. This point in roasting was readily observed by
stirring the charge; if the particles thrown to the surface brightened and
remained so for a short while the oxidation still evolved heat; but if
these particles were of a dead red color and began to darken immediately,
it was an indication that, in order to continue the oxidation, heat must
be applied. It was found to be of the greatest importance to have the
roasting well advanced when the material left the upper hearth. When
it was neglected and the speed of the feed increased so that the matte,
after having dropped to the lower hearth still created heat by oxidation,
the finished product was invariably insufficiently roasted. It was endeav-
ored to maintain a gradually increasing temperature in the lower hearth up
to the point of discharge.
In order to regulate the final heat, tests were made at intervals of
the material before and after it passed the last fire. The samples were
sifted and washed in a small dish to determine if any cuprous oxide had
been formed. The presence of cuprous oxide is readily determined by its
pink color. It often happened that although the material was free from
cuprous oxide before passing the last fire place, it could be plainly detected
after passing it. This was always an indication that the fire was too
hot. It was important to avoid this condition because by too high a
temperature cupric sulphate, of which quite a percentage was formed
during roasting, was decomposed into cuprous oxide and sulphuric acid,
and the matte was discharged before the cuprous could be oxidized to
cupric oxide. This test had to be made, not in order to prevent the loss
of acid, because in the Hunt and Douglas process more acid is made than
needed, but for the reason that when cuprous oxide is treated with dilute
sulphuric acid, only one-half of the copper can be dissolved as cupric sul-
phate; the other half changes into metallic copper, which being insoluble,
will remain in the residues.
Even with the greatest care it was impossible to roast a leady matte
free from small lumps. They formed in the very early period of the
process before any additional heat was used, but as a rule, being usually
porous, they were found well roasted. These lumps, however, were very
undesirable in the subsequent operation, as they retarded the solution
CHEMICAL PROCESSES L':5:j
of the cupric oxide in dilute sulphuric acid. The roasted matte, therefore,
had to be crushed. From the roasters the matte was automatically
conveyed to the revolving cooling tables, then fed to a ball mill with
50-mesh screens. From the ball mill it was elevated and convened to
the storage bins.
Solution. — The dissolving of the cupric oxide had to be done in agitat-
ing tanks, it being impracticable to conduct the operation in tanks with
filter bottoms. When roasted copper matte is brought in contact with
dilute sulphuric acid, or even water, it cements and hardens to such an
extent that it cannot be handled with shovels if not previously loosened
with picks or bars. This hardening of the material prevents to a great
extent the free percolation of the solution; this causes much delay and
also makes the discharging of the tank a rather difficult task. Even
while charging the agitating tanks, it was necessary that the acid solution
be kept in lively motion and that the matte be introduced in a gradual
stream and not charged with shovels; otherwise hard chunks were
formed.
The dissolving was done in wooden agitating stir tanks, 12 ft. in
diameter and 6 ft. deep, provided with a strong hard wood propeller
which entered and was driven from above. These tanks were about
two-thirds filled with acid solution, containing 9 to 10 per cent, free acid,
to which some wash water was added. This acid solution, which resulted
in the process, always contained 2 to 2 1/2 per cent, copper. The agitator
was set in operation and a jet of steam introduced through a lead pipe
entering from above and fastened close to the side of the tank. The
roasted matte was then brought from the storage bins in cars, which
were half covered and provided with a slot through which, by tilting the
cars, the matte could be uniformly charged into the tank. The addition
of matte to the dilute sulphuric acid produces considerable heat, which
aids the solution of cupric oxide and diminishes the amount of steam
required to maintain the pulp at the desired temperature.
After a certain amount of matte had been added, the pulp was fre-
quently sampled; these samples were filtered and the filtrate tested for
free acid. When the solution was nearly neutral the charging of the
matte was stopped, but the agitation continued until the solution became
neutral, or almost neutral. This operation was performed with care in
order to avoid an excess of matte, which would have enriched the resi-
dues with copper. In mixing the acid solution with wash water, care
was taken to have enough free acid present so that the resulting solu-
tion would contain from 6 per cent, to 7 per cent, copper.
The neutral solution, together with the residues, was discharged
through an outlet near the bottom of the stir tank, in a large lead-lined
cast iron pressure tank; thence under an air pressure of 40 lb. it M'as
forced through large filter presses with 4x4 wooden plates. When
234 HYDROMETALLURGY OF COPPER
charged, each press was capable of holding 5 tons of residues. When a
press was filled, compressed air was applied to blow out as much as pos-
sible of the strong solution which had been absorbed; then the residues,
while still in the press, were washed with water.
Below the presses there were two rows of tanks, one to receive the
strong liquor, the other the wash water. Some of the tanks were assigned
to the stronger portion of the wash water which went back to the process,
while the remainder were used as collecting and settling tanks for the
weak wash water; this was sent to scrap iron tanks for the precipitation
of the contained copper. The strong liquor and strong wash water
tanks were connected with a pressure tank, placed on a lower level, by
means of which the liquid could be forced to the stir tank level. On
opening the filter presses the washed residue cakes dropped into wooden
push cars; they were then wheeled to an opening in the press floor, through
which they were dumped directly into railroad cars. The residues,
which were rich in silver and lead, were delivered to the lead smelting
department.
The next operation was to chloridize the sulphate solution. For
this purpose the strong solution was elevated, by means of a pressure
tank, into a stir tank used only for this purpose. The solution was tested
for copper, its volume measured, the total copper in the charge calculated,
and as much common salt added as was required to convert the copper
present into cuprous chloride (58 parts of sodium chloride to 63.4 parts
of copper). The solution was agitated, heated, and then discharged
into storage tanks; from these tanks the chloridized liquor was elevated
and charged into so-called reducing towers for the treatment with sulphur
dioxide.
Precipitation of the Copper with Sulphur Dioxide. — The towers in
which this part of the process was carried out were made of steel and
lined with lead; the bottoms were cone-shaped. There were four towers.
The tops were tightly closed, provided with manholes, inlet pipes for
the solution and an outlet pipe for the gas. The cone-shaped bottoms
were provided with gas inlet pipes, a steam pipe and a discharge pipe.
The outlet for the gas was connected with a main pipe which discharged
into a wooden stack. The gas from each tower passed directly into the
stack. This arrangement caused considerable loss of gas. By tests it
was found that a gas which on entering a tower contained 7 per cent,
sulphur dioxide, contained 4 per cent, when discharged, so that there
was a loss of 57.1 per cent. Two towers were then connected up so that
the gas, after passing through the first, was made to pass through the
next tower. By this alteration the loss of gas was reduced to 25.4 per
cent., equal to an increased utilization of 31.7 per cent. This experiment
demonstrated that the precipitation of the copper in the tower under
pressure was correspondingly quicker than in those without pressure.
CHEMICAL PROCESSES 235
The gas was furnished by three revolving cylindrical furnaces, of
which two were kept in operation and one in reserve. These furnaces
were lined and provided with ribs for continually raising the ore and
dropping it in a shower. The front end of the furnace was closed, but
the cover was provided with air registers and two discharge openings
for the roasted ore; the latter opened and closed automatically at each
revolution of the furnace. The back end of each cylinder projected a
few inches into a small dust chamber, which again was connected with a
system of dust chambers. Through the roof of the small chamber, in a
slanting position, entered the feed pipe of the furnace. The feeding was
done by a very short screw conveyor which could be regulated. The
material consisted of iron pyrites concentrates, rich in gold, from
Colorado.
The dust chamber was connected with a heavy lead pipe about 4 in.
in diameter and was strengthened with iron rings, to which the pipe was
fastened. The entire length of this pipe, about 150 ft., was cooled by a
spray of water so that the gas was cooled before it entered the pumps.
There were two double acting gas-pumps of which, however, one was
sufficient to do the work, while the other was kept in reserve. The
cylinder measured 27.5 in. in diameter and the piston had a stroke of
28 in. so that each stroke furnished about 19 cu. ft. of gas. The speed
had to be regulated so as to get a good roast of the concentrates, and at
the same time produce as strong a gas as possible. In order to fulfill
both conditions, it was found that the resulting gas could not contain
more than 5 per cent, sulphur dioxide. Frequent gas tests had to be
made in order to maintain this percentage. Sometimes there was an
increase in strength up to 7 per cent., in which case the roasting was not
satisfactory; but more frequently it dropped below 5 per cent., which
caused a slower precipitation.
The pumps forced the gas through a lead lined receiver, in which
a great deal of sulphuric acid condensed and had to be drawn off daily.
The gas entered the tower under a heavy perforated lead cone which
divided it into small bubbles. Cuprous chloride was precipitated in
white crystals, while sulphuric acid was set free. The reaction was most
energetic in the beginning while the solution was neutral or contained
only a small percentage of acid, became more sluggish in proportion as
the percentage of acid increased, and stopped entirely when the copper
contents of the liquor was reduced to 2 or 2 1/2 per cent. This remain-
ing copper could not be reduced no matter how long the charge was kept
under treatment with the gas. The acid continued to increase slowly,
but the copper did not diminish. It is possible that the formation of
hydrochloric acid accounts for the copper not being precipitated. The
more hydrochloric acid there is formed, the more copper will remain in
solution. The hydrochloric acid dissolves cuprous chloride.
2:;() HYDROMETALLURGY OF COPPER
Sodium chloride as a chloridizer for the sulphate solution is not the
most suitable chemical for the process, as a large quantity of sodium
sulphate is formed which goes into solution. As it is necessary to use the
solution over and over again, on account of the sulphuric acid which is
formed therein, more and more sodium sulphate is formed. In a short
time the solution becomes saturated with this salt so that it- crystallizes
out whenever conditions are favorable. This happens at different
stages of the process, causing much annoyance and lessening the merit of
the process. It was especially aggravating in the operation of the filter
presses. When the matte residues, together with the strong liquor,
were forced into the presses, the filtration in the beginning was free and
satisfactory, but soon grew less so until it finally, stopped entirely,
although the press was not one-quarter filled with residues. On opening
the press it was found that the chambers were partly filled with a sloppy
mass containing many fine crystals, while the filter cloth was densely
covered with them. The only way of cleaning the press was to force
water through it. This not only caused much delay, especially as this
application of water had to be repeated, but caused the making of a
large quantity of wash water, from which the copper had to be pre-
cipitated with scrap iron. Sometimes it happened that the press could
be filled without any trouble, in fact the chamber filling was quite firm;
but as soon as water was used to wash the residues the filling shrunk
in volume, and the frames which previouslj^ were quite full, after washing
were only a little over half filled.
Calcium chloride, formed in converting the cuprous chloride into
cuprous oxide by boiling with milk of lime, was then tried. The result-
ing calcium chloride solution, however, was not strong enough to be used
directly, containing only 9 per cent, chlorine; it was therefore concen-
trated. In using calcium chloride, cupric chloride and calcium sulphate
were formed, the latter being precipitated. Though this chloridizer
made necessary an additional filter press operation, to separate the
calcium sulphate from the solution, it was by far preferable to salt, as
it left a clean solution, free from undesirable salts. It was found, how-
ever, that a sulphate solution, chloridized with calcium chloride, only
about half of the copper in solution could be precipitated as cuprous
chloride with sulphur dioxide. "When the change from salt to calcium
chloride was made, a new acid solution, free from sodium sulphate, was
used. Several attempts were made with the same unsatisfactory
results. It was finally decided to add salt to the new solution and the
results were at once better. From that 'time on chloridizing was done so
that three-fourths of the required chlorine was derived from calcium
chloride and one-fourth from sodium chloride. After adopting this
proportion there was but little trouble with the presses; the filtration
was free.
CHEMICAL PROCESSES 237
The calcium chloride solution obtained in converting the cuprous
chloride into cuprous oxide by boiling with milk of lime, was concen-
trated liy preparing milk of lime with water, allowing the lime to settle,
decanting the clear water and replacing the same by weak calcium
chloride solution. The condition of the slacked lime was not changed
and by repeating this procedure the proportion of calcium chloride in
the solution was increased from 9 per cent, to 24 per cent.
When the precipitation in the towers was completed, the cuprous
chloride, together with the acid solution, was discharged into a system of
seven cone-shaped lead-lined iron tanks. These tanks were so arranged
that the liquor flowed from one to the other, and from the last into
special charge tanks, to be in readiness to dissolve a fresh lot of matte.
These cone-shaped tanks served a double purpose; to give the cuprous
chloride an opportunity to settle, and to cool the solution. This liquor
when hot holds in solution a large amount of cuprous chloride which
precipitates out as the temperature falls. The temperature of the liquor
in the first cone was 50. ,5° C, and in the six following cones it was, 48.0,
47.0, 45.0, 39.5, and 37.0° C, so that the temperature of the last cone
was 19.5° lower than that of the first. The cooling proved to be suffi
cient, as no further precipitation took place in the. storage tanks.
Below the level of the cones were three vacuum filters, into which the
former (iould be discharged by opening the bottom valve. Into these
filters the cuprous chloride was allowed to drain, then receiving a thorough
washing. The washed precipitate was then converted into cuprous
oxide.
Conversion of the Cuprous Chloride into Cuprous Oxide. — The con-
version was done with milk of lime. The lime was slacked in a flat box
and collected in settling tanks. From these tanks the milk of lime of
proper consistency was charged, by means of a steam syphon, into a stir
tank and heated with a jet of steam. The washed cuprous chloride was
gradually charged; its color changed from white to red. Cuprous oxide
and calcium chloride were formed. As calcium chloride dissolves
cuprous chloride, the calcium chloride which was formed in proportion
as the conversion progresses will dissolve some of the freshly charged
cuprous chloride. On this property of calcium chloride was based the
test by which the conversion is conducted. After a certain amount of
cuprous chloride had been added to the milk of lime, charging was inter-
rupted, the stirrer, however, being kept in motion. About 10 minutes
later, a sample was taken in a wide necked bottle suspended by a copper
wire. Part of this sample was filtered and nitric acid and then ammonia
added to the filtrate. If the blue color appeai-ed, some cuprous chloride
was still dissolved in the calcium chloride solution. The agitation was
then continued for 15 or 20 minutes, when another sample was taken.
If the blue color appeared again, more milk of lime was gradually added
238 HYDROMETALLURGY OF COPPER
and at intervals. After each interval a test* was made until the blue
color ceased to appear. The last part of the operation had to be con-
ducted carefully to avoid an excess of lime.
The pulp consisting of cuprous oxide and calcium chloride was
forced, by means of a double acting pump, into a Johnson iron filter
press, where it was washed. The cakes were dumped on a lower floor
and dried on steam slabs. When dry, the cuprous oxide was carted to
the copper smelting department and dumped into bins conveniently
arranged on the charge floor of a cupola furnace. In this furnace it was
reduced to metallic copper.
The cuprous oxide always contained from 4 to 5, and sometimes as
much as 11 oz., silver per ton. Some of the silver undoubtedly came in
with fine particles of matte residues, which were still suspended in the
solution when it was charged into the towers for treatment with sulphur
dioxide, although the resulting cuprous chloride was clear white and
did not show any coloration. However, after more settling tanks were
inserted for the solution, the cuprous oxide contained considerably less
and more uniform amounts of silver. A sample of cuprous oxide con-
taining 6.25 oz. silver per ton, when leached with a solution of sodium
hyposulphite, still contained 5.5 oz. silver per ton, so that only 0.75 oz.
per ton could be extracted by that solution. This test was made with
cuprous oxide produced before the additional settling tanks were in use;
afterward the cuprous oxide did not contain over 2 to 3 oz. silver per ton.
The resulting calcium chloride solution gave with sodium sulphide a
dark precipitate which consisted mostly of lead sulphide, with only a
trace of copper and no silver.
In smelting the cuprous oxide in the cupola, very strong and offen-
sive fumes were formed. These fumes were white, but when very strong
assumed a reddish tinge. They consisted chiefly of volatilized cuprous
and cupric chloride, some hydrochloric acid and flue dust of cuprous
oxide. An investigation showed that the cuprous oxide still contained
1 to 2 per cent, chlorine, notwithstanding the fact that it was subjected
to a very thorough washing in the press. This chlorine could not be
removed or reduced even by an extended washing. This was not due to
the presence of cuprous chloride, caused by an insufficient quantity of
lime being used in the conversion, for, even if for the sake of information
an excess of lime was used and an unusually long time given for conver-
sion, the above stated percentage of chlorine was always found in the
cuprous oxide.
The obnoxious character of the furnace gases was destroyed by
passing them through a shower of milk of lime. A tower was arranged
which was provided at different levels with strong wooden grates. At
the foot of this tower tanks were arranged for making and receiving milk
of lime. Two of these receiving tanks were connected with a force
CHEMICAL PROCESSES 239
pump. The flue was connected with the wooden tower. Coarse lime
rock was placed on the different grates to detain the milk of lime in its
downward course as long as possible. At the top several perforated
pipes, through which the milk of lime was forced by means of the pumps,
were so arranged as to furnish an even spray. The bottom of the tower
was made tight and the outlet made to convey the stream into one of the
other tanks, so that the milk of lime could be passed through the tower
as often as desired.
The effect of the milk of lime was very gratifying. The strong
offensive odor of the gases disappeared entirely. The color of the milk
of lime turned gradually darker and became finally olive green and very
rich in copper. No copper escaped with the gases. When the above
green pulp was filtered the filtrate contained 11/2 per cent, to 2 per cent,
chlorine. The evaporation was great and water had to be added fre-
quently to maintain the same volume of precipitate. This method
proved itself successful and was finally permanently installed.
Treatment of the Wash Water. — In the course of the process a great
deal of wash water was made, principally from washing the copper matte
residues and the cuprous chloride. The latter, which contained from
1 to 2 per cent, copper, was collected in a number of large tanks, from
which it was drawn to be subjected to special treatment for recovering
the copper. Wash water containing 2 per cent, copper and as much as
practicable of the weaker portion, went back to the process and was
used instead of water.
To produce a clean cement copper free from chlorine, a trough was
constructed, in sections, abous 200 ft. long and 12 in. wide and 14 in.
deep. All sections were placed horizontally, but each succeeding one
was placed three inches lower. The outlet of each section was two inches
lower than the inlet. In some of the sections the compartment was made
by inserting across the width of the trough two boards 6 inches wide.
These were placed about 12 in. apart to allow for the insertion of a steam
jet. This part of the trough was tightly covered for 2 ft. on either side
of the jet to prevent the solution from being splashed out by the steam.
The 200 ft. of trough was arranged in U shape to avoid too long a build-
ing, and to make the handling of the material easier. After passing
through the last section, the solution flowed through a few scrap iron
tanks to precipitate any copper which might be present as sulphate.
Each section of the long trough was charged about 4 in. deep with
cement copper, evenly spread. The wash water from the main depart-
ment flowed into a circular tank 8 ft. in diameter and 5 ft. deep, thence
through an overflow into the first section of the long trough. The pur-
pose of the circular tank was to heat the wash water by means of steam
jets before it entered the long trough. It gave an additional, though
not very effective, opportunity for the settling of particles of matte
240 HYDRO METALLURGY OF COPPER
residues which might not have settled in the proper wash water storage
and settling tanks. As the wash water resulting from the different oper-
ations of the process contained cupric sulphate in addition to cupric
chloride, salt water was added to the storage tanks to convert the cupric
sulphate into cupric chloride.
By passing the wash water through and over the cement copper the
copper was precipitated as cuprous chloride. Once a day the cement
copper in the troughs was gently worked in order to bring fresh particles
to the surface. At one side and below the level of the troughs, a re-
volving barrel, 10 ft. long and 6 ft. in diameter was erected; when a large
part of the cement copper in the troughs had changed to cuprous chlo-
ride, it was removed from the troughs and charged into the barrels, to
which small scrap iron, water and salt were added. By means of a steam
pipe, the pulp was slightly heated to start the reaction; then the steam
was turned off. This pipe entered the barrel through one of its axles
and was bent downward to reach into the pulp; it was kept in position
by a stuffing box. Very soon an energetic reaction took place; the heat
developed causing the pulp to boil violently. The steam found an out-
let through another pipe inserted through the opposite axle. The salt
was added to dissolve some of the cuprous chloride; this caused the
reaction to start more quickly. The cuprous chloride changed to cement
copper and the iron into ferrous chloride. The latter had the same
effect as salt and dissolved cuprous chloride, thus assisting the process.
After the steam ceased to escape, the barrel was stopped and a
sample taken. The sample was filtered and the filtrate tested. When
the blue color could no longer be obtained with nitric acid and ammonia,
the pulp was ready to be discharged. The conversion was completed
in from 6 to 10 hours. Below the barrel was a square tank, with a
filter bottom, fitted on two opposite sides with rails which extended
beyond the tank. On this track was an 8-mesh screen fastened to a
wooden frame and provided with four wheels. The screen covered the
whole top of the tank. In discharging the barrel, the copper cover of
the manhole was removed and the barrel gradually turned by means
of a crow-bar. When the charge was out, the inside of the barrel was
rinsed with water. The screen retained pieces of iron, while the cement
copper and solution passed through to the filter tank. By means of a
stream of water, the iron on the screen was separated from adhering
cement copper and returned to the barrel to serve as part of the next
charge. The screen on wheels was pushed away from the tank, the
outlet under the filter opened, and the iron solution allowed to drain off.
Then warm dilute sulphuric acid was permitted to flow in. As soon as
the acid appeared at the outlet the latter was closed. The acid was
applied to remove basic salts, to prevent their formation as far as pos-
sible, and to dissolve any small pieces of iron which had passed through
CHEMICAL PROCESSES 241
the screen. After several hours the acid was removed to a special tank
to be used again. The copper was then thoroughly washed in the tank
to free it from acid. The resulting cement copper was of a very clear color
and unusually pure, containing 99 per cent, copper and but a trace of
arsenic. It was melted in the refining furnace.
This method gave such good results that three more barrels were
erected. The method eliminated the conversion of cuprous choloride
into cuprous oxide by the milk of lime, the smelting of the product in
the cupola furnace, the treatment of the furnace gases, the very unclean
manipulation of the scrap iron tanks, and the additional treatment of
the cement copper to free it from cuprous chloride.
Modification of the Hunt and Douglas Process. — To simplify the op-
erations and to avoid the saturation of the solution with sodium sul-
phate and its attendant disadvantages, Hofmann worked out and
successfully introduced the following modus operandi:
1. The process was started with a stock of dilute sulphuric acid.
By treating the roasted matte in the usual way in the stir tanks, a clean
sulphate solution was obtained which filtered well in the presses.
2. The sulphate solution was chloridized with hydrochloric acid, of
which in starting, a stock on hand was required. By chloridizing the
sulphate solution with hydrochloric acid, cupric chloride is formed and
sulphuric acid set free. No foreign salts are introduced and the solution
remains clean, while by chloridizing with sodium chloride the solution
becomes quickly saturated with sodium sulphate.
3. The cupric chloride solution containing the liberated sulphuric
acid was then treated in a stir tank with cement copper. The cupric
chloride by the action of metallic copper is reduced to cuprous chloride,
while the sulphuric acid remains unchanged. A steam jet is used to
hasten the reaction. An excess of cement copper serves the same pur-
pose. When the filtrate of a sample shows no reaction for copper, the
operation is completed and the pulp is drawn into a pressure tank and
forced through a filter press. The filtrate, which is now a clean sulphuric
acid solution, is elevated to storagfe tanks, whence it is used as required
to dissolve the cupric oxide of a new lot of roasted matte. To produce
as little wash water as possible, the solution absorbed by the matte
residues and by the cuprous chloride is forced out by compressed air; this
works very well, as the filtering capacity of both materials is not
lessened by the formation of crystals. For the same reason the subsequent
Avashing is quickly done, requiring comparatively but very little water
to accomplish it.
4. The washed cuprous chloride was treated in revolving barrels
in the same manner as described above, but care was taken that no more
water was added than necessary, so that as strong a ferrous chloride
solution was produced as practicable.
16
242 HYDROMETALLURGY OF COPPER
5. The ferrous chloride solution was evaporated in the same iron
pans which were formerly used for concentrating the calcium chloride
solution.
6. The solid ferrous chloride was charged into retorts, which were
provided with water for steam and air. When heated heavy fximes of
hydrochloric acid were formed; these were passed through a cooling
arrangement, in which a large portion was condensed. This condensed
acid was strong and contained 35.6 per cent, chlorine. The acid fumes
which were not condensed were made to pass through two towers made
of stoneware pipes and filled with coke. The gas escaping from the
first tower entered at the bottom of the second. To avoid the accu-
mulation of too much water in the stock solution, cupric sulphate solu-
tion was used instead of water as a spray for the coke, thus chloridizing
the solution. The solution, as a rule, after passing through the towers,
contained an excess of hydrochloric acid. This condition, however,
was properly adjusted by adding sulphate solution before the treatment
with cement copper.
7. The resulting cement copper was very pure, containing from
90 to 94 per cent, and more copper. This was smelted in a refining
furnace; no obnoxious fumes were evolved as in the case of cuprous
chloride. There was, of course, a loss of sulphuric acid as well as of
hydrochloric acid which was caused chiefly by the wash water; these had
to be replaced from time to time to keep up the volume of stock solution.
However, this shortage was not great and the loss of replacing it was
far less than that of the eliminated operations.
This modified process was used for some time until Hofmann
received instructions from the company for the necessary alterations of
the works to prepare for the more profitable manufacture of blue vitriol.
This was done by Hofmann's method of producing this material
direct from the roasted copper matte, described in chapter 18.
Copper Extraction at Falun, Sweden. ' — The old copper mine at Falun
produces two classes of minerals known as hard and soft pyrites respec-
tively. The former, consisting of mixtures of quartz and copper pyrites,
contains about 3 1/2 per cent, and the latter, which is mainly composed
of iron pyrites, about 1 per cent, of copper. The hard ore is roasted in
heaps, about 10 per cent, of the sulphur being driven off, while the soft
pyrites is treated in sulphuric acid works, about 30 per cent . of the
sulphur being utilized.
The burnt residues, to the extent of about 45 tons hard, and 12 tons
soft, per day, are mixed with 14 and 10 per cent, of salt respectively,
and ground in ball mills. The hard ore mixture is then subjected to
a chloridizing roasting in a White-Howell roasting furnace, in which
^E. and M. J., Aug. 30, 1902; Berg and Huttenmanisohe Zeitung, 1902.
CHEMICAL PROCESSES 243
15 tons are roasted in 24 hours. The soft ore is roasted in a double bed
roaster in which 7 tons are worked through daily.
When complete chloridization has been effected, the roasted material
is transferred to wooden vats, where the cupric chloride is dissolved out
by weak sulphuric or hydrochloric acid, the latter being obtained by
condensing the waste gases from the roasters.
After the ore has been treated with the dilute acid solution, it is
washed with clear water. The wash water is pumped back and used as
a second liquor on a following charge.
The solution from the ore contains all the copper, bismuth, selenium
and silver, together with a portion of the gold contained in the ore.
The remainder of the gold is extracted later by washing the residues
with chlorine water.
The copper solution is precipitated by scrap iron as cement copper,
which, together with the associated mud and iron salts, is smelted and
granulated for conversion into copper sulphate by means of dilute sul-
phuric acid and air in the ordinary way.
The residual mud from the copper sulphate crystallizers, containing
gold, silver, selenium, and bismuth is dried and smelted with litharge,
soda and sawdust to collect the precious metals into lead. The gold-
bearing solution from the chlorine extraction is reduced by adding a
portion of the original copper extracting solution containing ferrous chlo-
ride, which reduces the chloride of gold, producing metallic gold and
ferric chloride. The gold so obtained is extremely finely divided, and
an addition of lead acetate and sulphuric acid is necessary to obtain a
sufficiently dense precipitate. The gold precipitate is smelted with lead
in much the same as the copper precipitate. The wasted liquor from
the copper extraction is worked up for ferrous sulphate, which by roast-
ing, gives a red ferric oxide, which is used for paint.
The annual production of the works is as follows :
Copper sulphate, 1,600 tons
Ferrous sulphate, 300 tons
Iron oxide, red paint, 1,000 tons
Silver, 400 kilograms
Gold, 100 kilograms
The Bradly Process.' — In the Bradly process the sulphide ore is
carefully roasted, in what he calls an amphidizer, which consists of a
rotary drum with a central heating flue through which heat may
be supplied. The rotation of the drum operates to conduct the ore
and air through the drum in one direction and return it in another
to the same end, so that the copper is more or less completely sul-
phated and the iron oxidized. The roasting is conducted at a tem-
^E. andM. J., Jan. 6, 1912: U. S. Patent 1,011,562, Dec. 12, 1911.
244 HYDROMETALLURGY OF COPPER
perature of from 450 to 550° C, air being blown into the apparatus to
hasten the oxidation.
The roasted ore is then treated with an excess of calci-um chloride
solution in a reaction drum, while the temperature is maintained at
about 100° C. Cupric chloride is produced by the reaction between the
copper sulphate and calcium chloride, while any ferric sulphate in the
ore, due to the roasting, reacts with calcium chloride to produce ferric
chloride. The calcium sulphate resulting from both these reactions is
insoluble and is separated by filtration in the succeeding step. The
production of ferric chloride at this point is advantageous in that it dis-
solves copper oxide, copper sulphide or metallic copper, which remained
unaffected by the roasting, producing copper chloride, and this ferric
chloride also maintains the copper chloride in the cupric condition.
The gold and silver in the ore are brought into solution by convert-
ing all the copper into cupric chloride and then adding a small amount
of chlorine, chlorous, or chloric compounds. The chlorides of silver
and gold being soluble in calcium chloride solutions may afterward be
precipitated with the copper and subsequently separated. After leav-
ing the reaction drum the mass of gangue, solution, and precipitates is
subjected to filtration. The solid matter forms a cake which consists of
the gangue in the ore except a small amount of iron and alumina which
have l^een taken into solution and the calcium sulphate precipitate already
mentioned. The solution comprises a carrier in which has been dissolved
the metals to be recovered, a small amount of iron and alumina and any
zinc which may have been in the ore; the arsenic will have been separated
by filtration, as it has been rendered insoluble. The solution is then
subjected if necessary to a further oxidizing operation in order to be
sure that the metals are all combined at their highest valency.
The solution is then in condition for treatment for the separation
of the dissolved metals. The precipitation of iron and alumina may be
made by cupric oxide, hydrate or calcium carbonate, and as this precipi-
tate will carry some copper it is returned to the amphidizer, or roasting
furnace, after having been removed from the solution by filtration. In the
amphidizer the iron and alumina in the precipitate are rendered insoluble,
while the copper is left in a soluble condition and can be recovered. The
solution from which the iroii and alumina has been removed and which
then contains the bulk of the copper is run into a second tank in which the
copper is precipitated by carbonate of lime as oxide of copper. This
precipitate is filtered from the solution and the copper is recovered by
further treatment such as by reduction in an ordinary smelting furnace.
Any silver and gold in the solution is carried down during the pre-
cipitation of the iron, aluminum, and copper, and finally recovered by
separation from the latter metal. Zinc contained in the ore passes into
solution as chloride of zinc and accumulates. It is therefore necessary
CHEMICAL Ph'OC ESSES
L'-i;
VI
Dust and Fumes
Absorber
Dryer
Heater and (—
Asritatinff ~^
Tank
V
Reaction Drum
Anipliidizer
Heating: Coil
Fig. 48. — Bradly process. Diagrammatic sketch.
246 HYDROMETALLURGY OF COPPER
at stated times to run the solution, or a part of it, after the final treatment
and before returning it to the reaction drum, to a third precipitator in
which the zinc is precipitated by means of caustic lime. The regenerated
solution from which the gangue and all metallic compounds has been
removed and which contains calcium chloride is returned to the reduction
drum for the treatment of additional ore from the amphidizer, thus
completing the cycle.
During the roasting considerable dust and fumes are given off, con-
taining sulphurous and sulphuric anhydride. This is condensed by a
portion of the solution diverted from the main stream, after filtration
from the ore, and is again returned to the reaction drum.
Fig. 48 shows a diagrammatic sketch of the Bradly process.
Longmaid -Henderson Process for Treating Pjrritic Cinders. — (The
descriptions of the early work of this process, and of European practice,
are principally by Clapham, Wedding, Ulrich, Gill and Lunge; the
descriptions here given is taken largely from Lunge's Treatise on the
Manufacture of Sulphuric Acid and Alkali, Vol. 1, 1891.)
Sulphuric acid is largely made from pyritic ores, many of which contain
copper worth recovering, after the sulphur has been roasted off as sulphur
dioxide for the sulphuric acid works. Such ores are quite widely scat-
tered, but the largest and most valuable deposits occur in Spain. From
the Rio Tinto mines in Spain the ore is shipped to many of the commer-
cial and manufacturing nations of the world. During the year 1874,
there existed in Great Britain alone 22 copper works in which 450,000
tons of pyritic cinders were treated annually by wet methods for the ex-
traction of the copper. Two of these works made sulphate, three pro-
duced refined copper, and the rest sold their cement copper precipitate
to copper refineries. In 1882 the quantity of pyritic cinders treated for
copper in Great Britain amounted to 434,427 tons, containing 15,300
tons of copper.
The process used in England at that time is still largely in use to-day,
both in Europe and the United States, and with very little modification
from the original process.
The Longmaid-Henderson Process consists essentially of roasting the
pyritic cinders, from the sulphuric acid works, with salt, leaching out
the chloride of copper so formed, followed by precipitation with iron.
Longmaid obtained a patent dated October 20, 1842, and another in
January, 1844, both relating to the treatment of pyritic cinders by roast-
ing with salt. Longmaid described the principles of the process very
much as it is carried out to-day, certainly with a view of making salt cake
and chlorine as the principal products, and he worked it out on a large
scale; so that he must be regarded as the founder of the wet extraction
of copper. Gossage, in 1850, first employed sponge iron for precipitating
the copper. Henderson carried the process to greater perfection. In
CHEMICAL PROCESSES
247
1865 he erected a plant at Hebburn for the Bede Metal Company, to
extract copper from pyritic cinders, the process of which he had pro-
tected by patent. Henderson's principal improvement was the intro-
duction of absorption towers through which the acid gases from the
chloridizing roasting were condensed and yielded a weak acid solution,
which was then used in leaching the copper.
The most important ores treated by this process, in English works,
were from Spain and Portugal, and contained from 47 to 49 per cent.
sulphur; from 3.50 to 3.80 per cent, copper and from 0.75 to 1.20 oz. in
silver, per ton.
In the German works, at Oker, the ordinary copper ores treated by
this method contained 60 per cent, iron pyrites, 23 per cent, copper
pyrites, 6 per cent, blende, 2 per cent, galena, and 9 per cent, gangue.
The steps in the process may be summarized as follows:
1. Mixing the ore with salt and then grinding the mixture.
2. Chloridizing roasting.
3. Leaching the roasted ore.
4. Precipitating the silver from the argentiferous liquors.
5. Precipitating the copper from the desilverized liquors.
6. To which may be added, the preparation of the residue, rich in
iron, for the iron smelters. This is usually done at the copper works.
The percentage of sulphur in cinders as supplied by the acid works
to the copper extraction plants varies from 2 to 10 per cent. A fair
average may be considered from 4 to 5 per cent. At Oker the cinders
contained from 5 to 8 per cent, sulphur and from 6 to 9 per cent, copper.
The following analyses by Gibb shows the composition of the pyritic
cinders as the copper works received them:
Rio Tinto
Tharsis
San Domingo
Ytteron
(Norway)
Copper
Iron ■ Calculated as CU2S and Fe2S3. . .
Sulphur
1.65
3.64
3.53
2.75
2.02
0.47
0.0037
0.007
0.013
0.20
77.40
6.10
0.24
1.45
1.50
3.23
3.15
2.56
0.55
0.70
0.0023
0.032
0.010
0.25
77.00
5.25
0.17
5.85
1.55
3.76
3.62
2.70
0.47
0.84
0.0023
0.33
0.013
0.28
78.15
5.80
0.25
1.85
1.01
3.33
3.10
0 39
Zinc oxide
6 46
0 06
Silver
Cobaltic oxide
Calcium oxide
2 30
68.06
Sulphuric acid
6 56
0.05
8.74
99.46
100.25
99.31
100.06
248
HYDROMETALLURGY OF COPPER
Philips gives the following as the composition of cinders from San
Domingo ore:
Sulphur,
Arsen'c,
Iron,
Copper,
Zinc,
Cobalt,
Lead,
Lime,
Insoluble,
Moisture,
Oxygen and loss.
3,66
0.25
58.25 (83.0 FC2O3)
4.14
0.37
trace
1.24
0.25
1.06
3.85
26.93
Samples from Widnes and Hebburn showed,
Widnes Hebburn
Copper, 4 . 08 per cent. 5 . 75' per cent.
Sulphur, 4. 12 per cent. 3. 75 per cent.
Of this there was soluble in water,
Copper,
Sulphur,
Soluble in hydrochloric acid,
Copper,
Sulphur,
Copper insoluble in water and HCl,
The ores that used to be treated at Oker, showed, on an average of a
month's run:
46.0 per cent. 26.1 per cent.
43.0 per cent. 37.0 per cent.
22. 2 per cent. 13. 3 per cent.
55 . 0 per cent. 59 . 0 per cent.
31. 8 per cent. 60. 6 per cent.
Copper (principally as CuO),
Iron (principally as FcjOj),
Lead (as PbO),
Silver,
Zinc (as ZnO),
Manganese (as MujOJ,
Sulphur,
Sulphuric acid (corresponding to 3.8 per cent.
Alumina,
Other gangue.
7.83
40.53
per cent,
per cent.
2.09 per cent.
0.008 per cent.
1 . 95 per cent.
per cent.
per cent.
per cent.
per cent.
per cent.
0.40
3.80
9.51
4.43
11.65
The pyritic cinders as received by the copper extraction works,
were first finely ground to about 8 or 10 mesh and at the same time
mixed with a sufficient quantity of salt. In hand furnaces the amount
of salt varied from 10 to 20 per cent, but in mechanical furnaces it was
less — about 7 1/2 per cent. At Oker the cinders were mixed with 15 per
cent, carnallite, which contains chlorides of magnesium, potassium, so-
dium, and calcium; all of which assist in the chloridizing roasting.
The chloridizing was done in reverberatory, and in muffle fur-
CHEMICAL PROCESSES 249
naces. Sometime in a sort of combination between the two, in which
only a part of the furnace was muffled. See Figs. 10, 11, and 12, pages
98 and 99.
The chloridizing roasting of these cinders is in detail essentially
the same as described in the chapter on chloridizing roasting. Whatever
the means of furnace employed, the object to be attained is to sufficiently
convert the copper into sulphate, which owing to the presence of the
chloride salts, at once forms with the cupric sulphate, by mutual decom-
position, cupric chloride and sodium sulphate. The iron should be con-
verted as completely as possible into the ferric oxide, so as to be insol-
uble both in water and dilute acids.
In roasting, the SOg and 0 acting upon NaCl, chlorine is evolved
which greatly aids in chloridizing the copper as well as any other metals
present. At the same time a large quantity of hydrochloric acid is
formed which converts the oxides of copper, silver, zinc, etc., into chlo-
rides, while at the temperature of the roasting furnace ferric chloride
cannot exist. Since chlorides of copper are both unstable and volatile
at very high temperatures, a low red heat ought not to be exceeded;
so that any copper pyrites still present in the cinders is not burned, and
therefore escapes chloridization.
Wedding describes in detail the roasting at Widnes as carried on in
a gas furnace. The charge of 4500 lb. of ore, mixed with 17 per
cent, salt, is spread out evenly on the hearth and slowly heated till a
low red heat has been reached nearest the fire bridge; the charge is
rabbled and left to itself, the gas being shut off but the air allowed to
enter, so that after 2 hours scarcely any glowing can be perceived
at the fire bridge. After 1 hour's and 3 hours' roasting, respect-
ively, the copper of the charge behaved as follows:
1 hour's roasting 3 hours' roasting
Soluble in water, 54 per cent. 51 per cent.
Soluble in hydrochloric acid, 38 per cent. 42 per cent.
Soluble in nitric acid, 8 per cent. 7 per cent.
After 3 hours the charge is quite dark, and is now well rabbled.
There ought to be no necessity for more fire, as the temperature from
first should have been raised to the proper point. The charge is now
rabbled regularly at short intervals, and the temperature of itself rises
in consequence of the chemical reactions. The rise becomes sensible
after 4 3/4 hours, counting from the beginning; so that after 5 1/4 hours
a dark red heat is reached. Up to this point there is a copious evolution
of white vapors and blue flames; from this period there is less of these,
and it is the roasterman's principal task to see that heating of the charge
is uniform, and that some places do not show more flame than others.
After 6 1/2 hours these flames are almost entirely gone; and this fact,
250 HYDRO METALLURGY OF COPPER
along with the greenish-gray color of the charge, are the principal
tests for j udging whether the operation is finished. A sample is now taken,
and if its examination shows the completion of the roasting process,
the charge, which has now been 6 1/2 or 6 3/4 hours in the furnace, is
withdrawn. Of the copper now,
75 per cent, is soluble in water,
20 per cent, is soluble in hydrochloric acid,
5 per cent is soluble in nitric acid.
In roasting, in order to get the best results, the charge is first heated
to a temperature sufficiently high to start the chemical reactions, and
then to maintain these at the lowest possible temperature up to the
finish, and to see that the entire mass of ore is uniformly treated.
It is of great importance not to leave the ore any longer in the fur-
nace than necessary after the roasting is completed. The depth of the
ore is usually from 4 to 5 in., and this depth of charge, while more diffi-
cult to rabble, facilitates the chloridization, since the gas rising in the ore
heated both at top and bottom has all the more opportunity of coming
in contact with all parts of it.
At Oker, where gas-fired furnaces exactly like those at Widnes were
used, each charge of 5000 lb. of ore with 15 per cent, carnallite is brought
to a low red heat in 4 hours; the firing is then discontinued and the
mass rabbled, for 4 or 5 hours with a very low fire or without any, the
air valves being meanwhile kept open in order to allow the air to act on
the charge. The charge is then withdrawn and a fresh charge introduced
as soon as the furnace is empty. Five tons of roasted ore were worked
off in 24 hours, with a coal consumption of 10 to 12 per cent. The com-
position of the roasted ore after chloridizing roasting with 20 per cent,
carnallite was as follows:
Soluble in Wateb
Cu,
Ag,
Fe,
Al,03,
Zn,
Mn,
Ni,
CaO,
MgSO,,
K^SO^ 20.50
NajSOj,
39.066
■ cent.
Per cent.
3 . 86 calculated as CuCl^,
8.17
0 . 005 calculated as AgCl,
0.006
0 . 60 calculated as FeClj,
1.38
0.17 calculated as Al^CSO J ,,
0.56
1 . 64 calculated as ZnClj,
3.42
0 . 75 calculated as MnClj,
1.71
0 . 07 calculated as NiCl^,
0.15
1 . 60 calculated as CaClj,
3.17
CHEMICAL PROCESSES
251
Cu,
Pb,
Fe,
A1,0„
Zn,
Mn-Ni,
CaO,
SO3,
CI,
S,
Insoluble in acids,
Insoluble in Water
Per cent.
2.57 calculated as
1.17 calculated as
34 . 56 calculated as
0,44
0.37
trace
0.49
1.49
trace
0.64
calculated as
calculated as
calculated as CaSO,,
Per cent.
3.18
0.03
1.26
47.91
1.02
0.18
0,44
0.46
1.19
3.69
60.38
In mechanical furnaces less salt was used than in hand roasters —
averaging 7 1/2 per cent, as against 15 per cent, in the hand furnaces.
Sometimes only a portion of the salt was added at the start, and the
remainder added afterward. In the mufHe roasters the ore is first
roasted 9 hours with 12 per cent, salt and another 3 hours with 8 per
cent, more salt. In the combination furnaces with protecting arch the
weight of the ore was 5800 lb., and the time of roasting 8 hours; in the
mechanical furnaces 5 tons were roasted in 9 hours.
Gibb investigated the comparative working of the different furnaces,
which may be summarized as follows:
Cupric chloride. .
Cuprous chloride.
Cupric oxide.
Sodium chloride. .
Sodium sulphate .
Insoluble copper.
Gas furnace
Muffle furnace
Per cent.
4.03 =
0.32
1.26
2.50
13.18
Cu. per
cent.
Per cent.
Mechanical fur.
1.90
0,20
1,00
0,15
3.25
4,25 =
2.00
0,35
0,21
0,88
0,70
3,40
17,40
0,12
3,06
Cu, per _
^ Per cent,
cent.
6.70 =
nil
0,32
0.90
14,03
Cu, per
cent.
3.15
nil
0.25
0.13
3.53
The principal object in roasting, of course, is to get as much copper
as possible soluble in water or dilute acids. In the above comparison
there is a slightly better result in favor of the muffle roaster, but not
252 HYDROMETALLURGY OF COPPER
enough to give a decided advantage, and the advantage is largely over-
balanced by the increased fuel consumption of the mufHe furnaces.
At Oker the average results of the constantly taken samples of the
chloridized ore showed 75 per cent, of the copper was soluble in water
as the cupric chloride and neutral sulphate; 20 per cent, was soluble in
dilute hydrochloric acid, as cuprous chloride and oxychloride, and 5 per
cent, was insoluble in the treatment of the ore for the copper but was
soluble in aqua regia.
In roasting, the sulphur in the pyritic cinders must bear a certain
proportion to the copper. With a 4 per cent, copper ore the sulphur
should not exceed 6 per cent. ; an equal percentage of sulphur and copper
is preferable. If less sulphur is present, raw pyrites must be added.
The test for ascertaining the cpmpletion of the roasting are made by
taking a certain definite quantity of the roasted ore, leaching it with
water and dilute hydrochloric acid just as in the regular leaching; the
residue is then boiled with aqua regia, supersaturated with ammonia,
and allowed to settle; the more or less blue color of the ammonia-cupric
salt gives a sufficient indication of the percentage of insoluble copper.
Condensation of the Furnace Gases. — -In all but the mufHe furnaces
the gas from the roasting ore is mixed with gas from the fire-boxes.
Even in the muffie furnaces the gases from the ore are mixed with air
to such an extent that a condensation of strong acid is not possible.
The acid from the mufHe furnaces is only slightly more concentrated
than from the reverberatories; but this is not a serious matter as the
acids are always used in a very dilute solution for leaching. The furnace
gases contain principally, besides oxygen and nitrogen, sulphur dioxide,
sulphur trioxide, hydrochloric acid, chlorine, and very small quantities
of metallic chlorides. Henderson proposed volatilizing the copper
entirely as cupric chloride and condense the latter in towers; but this has
turned out quite impracticable. The quantity of copper volatilized in
the ordinary process of roasting is not large, about 1/4 per cent, of the
whole, and this is condensed with the acid in the tower acid for leaching
the ore.
The condensation of the gases from the roasters takes place in
towers made of brickwork set in tar and sand (or, better, of stone flags),
and packed with coke, fire bricks, and the like. Large stoneware pipes
are sometimes employed. Coke can be used for the filling material with
the muffle roasters; but the other furnaces require brick, or similar
material and must have larger condensers, as these towers have to serve
for a larger volume of gas. The size of the towers varies with the plant;
for 12 furnaces a tower of 8 ft. square and 40 to 50 ft. high is sufficient.
The gas enters at the bottom, meets a spray of water coming from the
top, which washes the acid out of it, and again leaves the tower at the
top, whence it is taken downward into a flue leading to the chimney.
CHEMICAL PROCESSES 253
The total condensed liquid, which is a mixture of sulphuric and
hydrochloric acids, is used in the succeeding operation of leaching, and
frequently is not even sufficient for dissolving all the copper oxide and
cuprous chloride. The sulphur dioxide in the liquid is oxidized to sul-
phuric acid by the action of the chlorine.
Leaching of the Roasted Ore. — The roasted ore is carried in bogies on
tramways to the leaching tanks. The material used in the construction
of these tanks is wood. It is most difficult to prevent leakage in large
wooden tanks when using a hot chloride solution; on this account the
entire floor of the leaching shed is covered with a thick layer of asphalt
and slopes to one side, so that all liquors Can be recovered in a catch-well.
The leaching tanks are square, about 11X11 ft. wide, and 4 to 5 ft.
deep, made of well seasoned and planed 3-in. planks, secured by corner
pieces, screw-bolts, etc. The joints are tightened by putting a little
redlead between the planks before putting them together; the bottom
joints are best caulked with tarred spun yarn, and the whole tank painted
with hot coal tar. At Oker lead-lined tanks were used, but they were
very expensive and needed frequent repairing. On the bottom of the
tank are placed slats, laid on end; and upon these perforated tiles or
boards; upon this false bottom a layer of sifted furnace cinders is spread
out, and on top of this a layer of sand, coke, or straw from 3 to 6 in. deep.
The ore is charged upon the bed so prepared and is then ready for
leaching. The leach liquors are conveyed in earthenware and india-
rubber tubes of 3 to 4 in. in diameter, which are provided with iron
pinch-clamps. In order to force the liquors from one tank to the other,
or from the catch- well into the tanks, simple stoneware injectors are
provided. Each tank has a steam pipe for heating.
In each tank there is put 10 tons of roasted ore, quite hot, from
the furnace, and is covered with a weak liquor from a previous operation, '
which gets heated by the heat of the mass itself. After one or two
hours the now concentrated liquor is run off by a plug-hole below the
false bottom, and is delivered to the precipitating tanks. The plug is
put in again, and the ore leached with hot water; thus weaker liquors
are produced which are forced to another tank, as described. Generally
three waters are put on, and thus most of the purest copper and 95 per
cent, of all the silver contained in the pyrites are obtained. After the
treatment with the water the dilute acid solution from the tower is
applied, sometimes as many as six applications before all the copper is
satisfact orily extracted.
The liquors obtained by use of acid solutions contain many impuri-
ties, especially arsenic, bismuth, antimony, and lead — according to Gibb,
for each 100 parts of copper, 5.4 arsenic and 0.3 bismuth; these liquors
are usually treated separately in most works because they yield impure
copper.
254
HYDROMETALLURGY OF COPPER
As a rule each solution application is allowed to stand only a few
hours on the ore; the nine washings of each tank, together with the
filling and emptying, takes about 48 hours.
The effect of the leaching is best seen from the following by Gibb,
which at the same time shows the difference in work between mechanical
and hand-worked furnaces:
Soluble in water
Cupric chloride,
Cuprous chloride,
Cupric sulphate.
Ferrous sulphate,
Ferric sulphate,
Zinc sulphate,
Calcium sulphate,
Sodium sulphate.
Sodium chloride.
Mechanical furnace
per cent.
4.16 = 1.96 per cent. Cu
none
1.83 0.80
0.15
0.75
2.01
1.29
9.17
none
19.36 2.77
Soluble in dilute HCl
Cuprous chloride, 0.15 =0.01 per cent. Cu
Cupric oxide, 0.225 0.18
Residue by difference, 80 . 40 0. 08
100.00 3.04
Sodium chloride equiva-
lent to sodium salts
as above, 7.56
Hand-worked furnace
per cent.
3 . 81 = 1 . 82 per cent. Cu
0.19 0.12
none
none
none
1.95
1.39
11.13
2.64
21.11 1.94
0.33 = 0.21 per cent. Cu
1.01 0.81
77.55 0.11
100.00 3.07
11.81
At Oker, the process is carried out as follows: The roasted ore is
leached in charges of 5 tons each, first with the final liquor of a previous
charge, 100 parts of this liquor, of 1.145 specific gravity, contained:
Cu,
0.015
Bi,
trace
FeO,
2.14
Fe,03,
0.15
Al,03,
0.11
ZnO,
0.06
MnO,
0.31
CoO; NiO,
0.01
CaO,
0.12
MgO,
0.52
Alkalies,
2.61
CI,
2.56
SO3
5.89
As; Sb,
trace
Total solids, 14.49 per cent.
This liquor, already heated in pumping by the injector to 50° C. is
further heated, when it comes in contact with the hot roasted ore to
nearly the boiling-point. When the charge is thoroughly saturated
with the liquor, the spigot is opened and the liquor allowed to drain as
long as it shows any blue color. This lasts from 4 to 5 hours and furn-
CHEMICAL PROCESSES 255
ishes a copper liquor of 1.355 specific gravity and of the following com-
position:
Cu,
3.71
CoO-NiO,
9.04
Pb,
0.01
CaO,
trace
Ag,
0.005
MgO,
0.27
Bi,
trace
Alkalis,
10.60
Fe,03
Al,03
1 0.29
CI,
12.56
ZnO,
4.97
SO3
8.95
MnO,
0.58
As-Sb,
0.32
Total solids,
42
.305
per
cent.
"
After the first leaching is over, the dilute condenser acid, first brought
to boiling, is run into tanks and allowed to act for 24 hours; then it is
drawn off, and a third leaching effected by sulphuric acid. For 5 tons
of ore 250 lb. of chamber-acid of 106° Tw., diluted to 12° Tw. and heated
to boiling, is employed and allowed to remain in contact with the ore
for two days, or until the liquor acquires a neutral reaction.
The first copper liquors contain most of the silver, and are therefore
kept apart from the later liquors, which do not contain as much silver.
The cupric chloride is, of course, easily dissolved in the final liquor;
the cuprous chloride in the presence of alkaline chlorides is also dissolved
at a higher temperature without difficulty; lastly cupric oxide is to be
converted into cupric and cuprous chlorides by the ferrous chloride of
the final liquors from previous applications:
2FeCl2 + 3CuO = Feft^ + CuCl^ + Cu^Cl^
but this could only be done by an intimate mechanical mixture of the
liquor with the ore; and it is therefore preferred to dissolve merely 75 to
80 per cent, of the copper by means of the final liquor, and the remainder
by further leaching with dilute acids.
Precipitation of the Copper. — The precipitation of the copper is some-
times preceded by a special treatment for obtaining the silver, separately
from the copper.
The precipitation of the copper is universally done by means of
metallic iron. Gibb proposed precipitating the copper with hydrogen
sulphide, but this method was later given up at the Bede Metal Works
where it was tried on a large scale.
For precipitation with iron, the iron used was either, wrought scrap,
or sponge iron reduced from the residues. Light scrap is better than
heavy scrap, but the copper precipitated, owing to more impurities,
gives a lower grade cement copper. The precipitation takes place in
wooden tanks like those used in leaching the ore, and are furnished with
a steam pipe for heating the copper liquor. The tanks are filled with
scrap iron; copper liquor is run upon it and the steam turned on. The
heating is continued till a bright strip of iron, held in the liquid, no
256
HYDROMETALLURGY OF COPPER
longer indicates the presence of copper in solution. At Oker, according
to the degree of concentration of the liquors, the boiling takes place
two or three times before all the copper is thrown down; the process
lasts from one to three days, and requires as much iron as the weight ot
copper produced, which proves that a large part of the copper must
have been in solution as cuprous chloride. Once a month the precipi-
tated copper is removed from the tanks and washed. If sponge iron
is used for precipitation of the copper, continuous stirring is required,
for which at. some works mechanical agitators are used, at others manual
labor. At the Bede Metal Works an india-rubber hose, through which
a blast of air passes, is moved about in the tank. Perfect mixture is
thus obtained, and. the precipitated copper contains only 1 per cent,
metallic iron.
The composition of the copper precipitated by the various methods
is shown by the following analyses by Gibb :
Precipitated by
Copper
Arsenic
Silver
Lead
Ferric oxide
Carbon
Silica
Sponge iron
Per cent.
67.50
0.137
0.011
1.30
5.15
5.10
3.20
Heavy scrap
Per cent.
72.50
0.3d6
0.046
2.60
4.41
Light scrap
Per cent.
67.50
0.100
0.066
1.74
7.56
At Oker the composition of the copper precipitated by scrap iron
and dried at 100° G., was:
Cu,
Pb,
Ag,
Bi,
As,
Sb,
Fe^Os,
Al.O,
Zn,
Mn,
Co-Ni,
CaO,
MgO-Allialip.s,
CI '
Insoluble in acids,
Oxygen-moisture (by loss),
100.00
77
.45
0
63
0.
10
0
006
0
04
0
15
6
72
0
99
1
02
0
02
0
03
0
1(1
2
71
'1
58
1
19
0
61
3
654
CHEMICAL PROCESSES 257
The copper precipitate from pyritic cinders made at the Witkowitz
works, dried at 100° C, was composed as follows:
Cu,
Cu,0,
Ag,
Au,
Fe,0„
ZnO,
CuC)„
FeCl,
CoCl,,
NiCl„
AsCl^,
PbSO,
Na.SO,
CaSO,,
MgSO,
H.O,
11.
30 1
65.
31 ) =
0,
,521
trace
0
.19
3,
,86
0,
,45
1
,18
0
.20
0,
,32
0
,16
0
.29
0
.07
1,
.32
2,
,19
3,
,39
5.
32
0.
59
2.
89
= 69.45 per cent. Cu
99.641
The cement copper from the leaching works may be smelted to
blister copper, or sent to the smelting works. The copper precipitated
from aqueous solutions, if kept separate from that from the acid solu-
tions, can be smelted direct to blister copper by adding to it lime and
slags; the copper from the acid solutions is frequently so impure that it
has to be mixed with raw ore, and smelted for "coarse metal" which
yields blister copper only after second treatment. At some works both
precipitates are melted together, being charged into the furnace while
moist. The slag produced from this operation, containing from 3 to 10
per cent, copper, are charged into blast furnaces.
The furnace for smelting the copper precipitate used at English wet-
extraction works are reverberatory furnaces of the well-known Swansea
type. After smelting, the slag is skimmed off, and the copper tapped as
blister copper. When sponge iron has been used, the excess of carbon
prevents the copper from being melted directly into blister copper;
therefore about one-half of the precipitate roasted in large furnaces
similar to those used in roasting the ore. Here the carbon is burned
off and the copper partly oxidized; the roasted precipitate is mixed with
raw precipitate and smelted for blister copper. The blister copper is
refined by roasting to oxidize the iron, sulphur, etc., followed by reducing
with charcoal the oxide of copper produced in the roasting, and poling
according to the method of smelting usually employed by the copper
smelter. The copper produced in this way is pure and tough.
17
2.58 HYDROMETALLURGY OF COPPER
Precipitation of the Silver. — Most cupriferous pyrites contain some
silver, and small quantities of gold. The cupriferous pyrites cinders
from Spain, according to Phillips, contain on an average of 0.0027 per
cent, silver, and 0.0001 per cent. gold. On roasting with salt, most of
the silver and much of the gold is converted into the chlorides. Owing
to the solubility of silver chloride in a solution of other chlorides, and
the solubility of gold chloride in water, the silver and gold chlorides are
extracted with the copper by the leaching solvent. At the present
time, when copper is refined electrolytically at a small cost, no attempt
is usually made to recover the gold and silver separate from the copper.
In earlier years, however, this separation was desired, and various
schemes were proposed for their separate recovery.
Claudit Process. — Of the methods proposed for the separate recovery
of the silver and gold, that devised by Claudit was quite generally
employed.
This process consists in precipitating with a soluble iodide the silver
from the liquors in the state of Agl, silver iodide, which is quite insol-
uble in chloride solutions. Only the first liquors, rich in silver, are sub-
jected to the Claudit process for the precipitation of the silver. The
other solutions are returned to the ore, or they are too dilute in sUver
to make its recovery profitable. The liquors, from the leaching tanks,
and before precipitating the copper, are run into settling tanks, where
they are completely settled and their silver contents accurately estimated
by adding to a certain volume of muriatic acid and a solution of lead ace-
tate, and afterward potassium iodide. The precipitate is collected on
a filter, washed, dried and fused with a flux of soda, borax and finely
pulverized carbon. The lead regulus is cupelled, and from the weight
of the silver thus obtained, that contained in the liquors in the settling
tanks is computed. To the liquor in the settling tanks is then added a
solution of potassium, sodium, or zinc iodide of known strength, so that
the quantity is just sufficient to precipitate all the silver; the iodide
solution is diluted to such an extent that it amounts to one-tenth the
volume of the liquid.
The reactions for the precipitation of the sUver with sodium and zinc
iodides are:
AgCH-Nal =NaCl-|-AgI.
2AgCl -F Znl^ = ZnCIj + 2AgI.
The precipitated iodide of silver is allowed to settle for about 48
hours, and tested in the laboratory to see if the precipitation has been
complete. The liquors are then run into the copper precipitating
tanks, where they are treated in the usual way for the precipitation of
the copper.
CHEMICAL PROCESSES 2.59
The quantity of iodide employed for the precipitation is much
larger than that corresponding to the silver present, since a portion
of the lead is also thrown down as PbClj. The silver is probably
precipitated before the lead, but a fractional precipitation is not
possible, so that necessarily a corresponding excess of the precipitant is
required.
The precipitate, consisting principally of Agl, Pblj and PbSO<(which
is deposited on cooling the liquor), is well washed with water; and after
a sufficient quantity of it has been collected, it is treated with metallic
zinc and hydrochloric acid or with sodium sulphide;
2AgI + Zn = Znl2 + 2Ag. or
2AgI + Na^S = 2NaH- Ag^S.
Thus the Agl and Pblj are decomposed completely, the PbSO^ partly,
and liquor containing zinc or sodium iodide is obtained, which is employed
again and the cycle continued indefinitely.
After two or three days the precipitated liquor is drawn off, and
run into copper precipitating tanks where the copper is precipitated, in
the usual way. The liquors may still contain 2 to 3 milligrm. of
silver per liter. The tanks are furnished with two outlets, one at the
bottom and the other about 8 inches higher. The desilverized solution
is drawn off the upper hole while the precipitate remains undisturbed
on the bottom. New liquor from the leaching vats is then let in and
the procedure repeated until there is a sufficient silver slime in the bot-
tom that it has to be recovered. This is usually done every month or
two, depending on the percentage of silver in the ore.
The precipitate obtained, treated with zinc and hydrochloric acid,
ready for melting, contains from 3 to 10 per cent, silver and usually
some gold. An anlaysis shows the following composition:
Ag, 5.95
Au, 0.06
Pb, 62.28
Cu, 0.60
ZnO, 15.46
Fe,0, 1.50
CaO, 1 . 10
SO3, 7.68
Insoluble, 1 • 75
Oxygen and loss, 3 . 62
100.00
About two-thirds of the silver and gold originally contained in the
roasted ore is recovered, with care, by the Claudit method. Under
260 HYDROMETALLURGY OF COPPER
normal conditions, from 10 to 15 per cent, of the iodide is lost in pre-
cipitating the silver.
Disposition of the Residues. — The residues, from the leaching, make
a valuable ore of iron, and is known as "purple ore" or "blue biHy-"
In this way the additional profit may be realized. The following is
the composition of two average samples:
Ferric oxide, 90.61 95.10
Copper, 0.15 0.18
Sulphur, 0.08 0.07
Phosphorous,
Lead sulphate, 1.46 1.29
Calcium sulphate, 0 . 37 0 . 29
Sodium sulphate, 0.27
Sodium chloride, 0. 28
Insoluble, 6.30 2.13
99 .62 99 . 55
Metalic iron 63.42 66.57
This shows an excellent quality of iron ore, entirely free from phos-
phorus, and containing but little sulphur. Its slight percentage of
copper does no harm. The lead contained in the pyrites remains behind
in the residue in the shape of sulphate, and injures its quality as an iron
ore. Schaflfner proposed drenching the residues with a solution of cal-
cium chloride, heated to about 40° C, and acidulated with hydrochloric
acid. By mutual decomposition, gypsum and lead chloride are at once
formed, which remains dissolved in the acid liquor. This is run off and
brought in contact with metallic iron, which precipitates the lead in the
metallic state. After washing, the purple ore is quite free from lead sul
phate. At the same time the CaClj dissolves the last traces of copper
and silver present as CujClj and AgCl; these are precipitated along with
the lead. It should be noted that hydrogen sulphide fails to indicate
the lead in a solution of calcium chloride acidulated with hydrochloric
acid.
The only drawback to the use of pyritic residues in smelting for iron
is their fineness, which militates against their desirability. Unsuccessful
attempts for a long time were made to agglomerate this ore but without
success. At present, however, this is cheaply done by sintering, or by
partially fusing it. It can cheaply be sintered by mixing with it from
7 to 10 per cent, coal or coke dust, the mixture moistened to the proper
consistency, placing it on a permeable hearth such as broken limestone, and
then applying suction (down draft) so that the intense ignition of the
coal dust fuses the ore and agglomerates it into a strong porous mass,
and puts it in ideal condition for smelting.'
'John E. Greenawalt, U. S. Patent, 839,064, Dec. 18, 1906.
CHEMICAL PROCESSES 261
JOxporimcnts made in Denver by this metliod gave excellent results
both as to cost of sintering and product obtained. These experiments
have since been duplicated in large installations in working plants.
In some works the residues are briquetted, and the briquetts, after
being dried, are subjected to a high temperature, so that by partial fusion,
they become coherent. In still other works the residue is sintered and
nodulized by passing the purple ore through a rotary cylinder similar
to those used in revolving roasters, in which the ore is heated to a high
temperature by means of powdered coal blown in at the discharge end.
This product varies in size from that of rice to that of walnuts, and forms
hard balls more or less thoroughly sintered. By the Grondal method
the cinder is formed into briquetts of uniform size by an automatic
plunger press. The briquetts are loaded on flat cars, which are then
slowly pushed through a channel furnace 150 to 200 ft. long, where they
are gi-adually heated until they arrive at a zone of the furnace where the
temperature reaches 2400° F. The finished product is hard and strong,
but porous, so that it is well suited to blast furnace work and open hearth
practice.
The cost for nodulizing or briquetting will vary from $0.75 to $1.25
per ton. Costs for sintering have not yet been established.
Longmaid -Henderson Process at the Helsingborg Copper Works,
Sweden.— In preparing the tanks for the ore, standard bricks are placed
on edge on the bottom. On these special perforated brick, 12 in. square,
are laid as a false bottom for a filter composed of straw or screened lump
cinder, iron ore or something similar, on top of which the ore is charged
and leveled. The tanks are built elliptical in shape, one diameter of
which is 10 ft. and the other 6 ft. The inside depth is 4 ft. and the
thickness of stave 5 in.
After leveling the charge, acid liquors of 5 to 8° B. strength, ob-
tained in previous leaching, are let into the tanks. The ore is still
warm from the roasting. The first liquor issuing from the vats is very
strong — 40° B. It contains about 5 per cent, copper and nearly all the
silver. It is transferred immediately to the silver precipitation de-
partment.
When the strength of the liquor as it is drawn from the vats has
gone down to 22° B., it is taken to the copper precipitators. This con-
tinues until the strength of the outgoing liquor has gone down to 10° B.
The incoming liquor is shut off and in its place acid from the towers is
let into the tank. The acid is allowed to remain in the tanks two hours,
at the end of which time the liquor is exchanged for fresh acid, which in
its turn remains two hours. The liquor so obtained is used in the first
application for freshly filled tanks.
After this leaching, the ore is washed with water, first with wash water
from a previous operation and finally with fresh water. To test whether
2(312 HYDROMETALLURGY OF COPPER
all the copper is extracted, a well-polished iron plate is immersed in the
off coming solution; if there is no coloring of copper, the leaching is
finished.
The wash water ought to have a temperature of about 50° C. The
time of leaching is usually about 40 hours for a 10-ton charge.
Of the copper in the roasted ore about 80 per cent, is soluble in warm
water,
16 per cent, soluble in weak hydrochloric acid,
4 per cent, insoluble, and remains in the purple ore.
The purple ore is briquetted and sintered and then smelted for iron.
The crude purple oie contains- 60.6 per cent, iron and 0.17 percent,
sulphur; the fused briquetts contain 60.6 per cent, iron and 0.023 per cent,
sulphur.
Longmaid -Henderson Process as Carried out at the Works of the
Pennsylvania Salt Manufacturing Co., Natrona, Pa. — A description of a
large modern plant, treating approximately 200 tons of pyritic cinders a
day by the Longmaid-Henderson process is given by Joel G. Clemer.'
The method of operation is as follows:
The pyritic cinders are ground dry in Chillean mills, to about 20 mesh,
and mixed during the operation with 10 per cent, salt, testing high in
NaCl. The mixture is then lifted by bucket elevators to overhead hoppers
from which it is drawn into tram cars, weighed, and dumped into the
furnace hoppers, or pipes set into the floor over the furnaces. Should
the unburnt sulphur in the cinders not be equal to the copper, sufficient
unroasted ore is added in the mill to bring it up to 1.5 times the copper
content.
Then 9600 lb. of mixture is taken as a furnace charge, and is heated
to a very dull-red heat, say 800° F., and well stirred. When properly
worked such a change will be finished in about 8 hours making three
charges every 24 hours per furnace. The charge, when finished, is drawn
from the furnace to the floor to cool, and then transported in tram cars to
the leaching vats.
MuSle furnaces have almost entirely superseded the old reverberatory
furnaces for chloridizing roasting. Contrasting the action of the two
types of furnaces, the muffle possesses the advantage of requiring but one-
half the condensing capacity, as only the gases resulting from the chlo-
ridization pass through the towers while in the reverberatory the gases
of combustion as well must pass through. Apart from this decided
advantage of the muffle furnace, better results are obtained, as both the
gases from the reaction and those from the combustion are under
separate control.
The construction of the modern muffle furnace is well illustrated in
detail in Fig. 49. The arch at the fire-box is made double to prevent the
' Min. Industry, Vols. VIII and IX.
CHEMICAL PROCESSES
263
tTTt-f^
264
HYDROMETALLURGY OF COPPER
cinders from becoming overheated at that end, and the fire flue is placed
beneath the furnace rather than at the end of the passageway, which
is constantly traversed by the workmen. A bridge wall is constructed
beneath the bed of the hearth of the muffle and guides the heat directly
beneath it to the entrance of the underground flue at the fire-box end
where the fire flue damper for controling the draft is located. The
furnace has doors on both sides of the fire and muffle. The furnace
shown in the illustration was designed for the treatment of cinders con-
taining less than 2 per cent, copper, with a normal amount of sulphur,
the charge being of usual weight. The hoppers hold sufficient mate-
rial for an entire charge.
In the construction of the tanks every supporting timber, plank,
and pin should be painted on all sides with hot soft tar before being put
Plug Tap
Fig. 50. — Details of tank construction for leaching cupriferous pyrites cinders.
in place. This is absorbed by ~the wood and protects it against the
destructive action of the acid liquors. The tanks should be constructed
of 3-in. plank, with an inside shell of the same thickness, and a space of
3 in. between the plank and the shell. All should be put together with
wooden pins and bound together as shown in the accompanying drawing.
Fig. 50. The space between the tanks and shells should be filled with a
mixture of hard tar and sand. The bottom should be covered to the
depth of 3 in. with the mixture fused to that between the shell and tank.
This bottom covering should be protected against wear of shovels by
a layer of chemical brick laid in cement.
CHEMICAL PROCESSES 2(55
The holes for drawing off the liquor are bored through wooden
blocks 6X6 in. square set inside of the tank near the bottom, and pro-
vided with wooden spigots or plugs, extending to the weak and strong
liquor launders. All the launders are made preferably of very sappy
yellow pine, dug out, with the ends or joints halved together and caulked
with oakum and red lead.
The proper dimensions for the leaching, settling, and precipitating
tanks are, 12X12X4 ft.; 12x12X6 ft., and 12X12X6 ft., respectively.
In preparing the mixture of hard tar and sand, composed of about
equal parts, care must be exercised not only to produce a homogeneous
mass, but also that all the moisture and air be expelled. The usua
method of procedure is to heat the hard tar in a large kettle, and stir it
until the moisture and air have been expelled, then adding the sand hot
and stirring the mixture until the desired result is obtained. The sand
must of course be first screened, and have all combustible matter burned
out of it before mixing with the tar.
The bottom of the leaching tanks are provided with hard-burned red
brick laid flat side by side, and covered with old hay or small pieces of
refuse coke. This makes a quite durable filter.
The roasted mixture, charged into the tanks, is first leached with weak
liquor from a previous operation, and then with water and dilute hydro-
chloric acid from the condensing towers connected with the furnaces.
After leaching is completed the residue (purple ore or blue billy) is
shoveled directly into gondola cars for shipment to the iron smelters.
All the copper liquors are run into settling and storage tanks; the
weak liquor being pumped or blown back with steam injectors, to the
leaching tanks when required. Any liquors of 18° B. and upward are
left in the settling tanks until all the lead sulphate, etc., has settled out.
The lead sulphate usually contains some gold; indeed an average of
SlOO per month is not an unusual recovery in a plant of the capacity
described, and it is therefore essential for the recovery of this as well as
for other reasons that everything that will settle out of the liquor, be
given time to settle in three tanks. A part of these tanks may also be
used for precipitating the silver with the iodide, in Claudit's process, but
since in practice this process leaves an average of 5 oz. silver in a ton of
precipitated copper, and since electrolytic copper works, as well as blue
vitriol works, handling or using silver- and gold-bearing copper, will
pay for at least 95 per cent, of the silver content and full market value
of the copper and gold contents of the cement copper, this part of the
copper extraction does not pay, especially since the cost per annum of
precipitating the silver by the iodide method in a plant of this size is not
less than $12,000.
After the lead sulphate, etc. has settled out of the strong liquors,
they are run into the copper tanks and the copper, silver, and gold con-
266
HYDROMETALLURGY OF COPPER
tents are precipitated by means of clean thin scrap iron from the rolling
mills. These tanks should, have wooden slats so placed as to form an
open false bottom about two feet above the real bottom, for the support
of the scrap iron. Live steam is let into the liquor during the process
of precipitation. The steam serves to accelerate the process, and also
keeps the liquor in motion, and washes off the copper from the iron as
fast as it is precipitated. The copper naturally finds its way between
the slats to the bottom of the tank, and when it is desirable to remove
it, the remaining scrap iron can readily be removed from the false bottom
practically free from copper. The cement copper, on the bottom of the
tank, after washing through perforated cast-iron plates set in a frame
over the tanks, assays about 90 per cent. Cu; 35 oz. Ag. and 0.15 oz. Au,
per ton. The precipitated chloride solutions are then run into the
sewer, by first passing it through a series of tanks in the ground,
which are filled with scrap iron, to recover any copper and silver
which may have been left in the waste chloride solutions and the wash
waters.
The cost of treating Spanish pyrites cinders, as above described,
at Natrona, Pa., in a works having a daily capacity of say 200 tons of
mixture (cinder and salt) will be about as follows:
2 samplers,
at $2 . 50
$5.00
6 mill men,
at
1.75
10.50
1 mechanic,
at
2.00
2.00
2 engineers,
at
2.00
4.00
3 firemen,
at
1.75
5.25
4 weightmen and furnace chargers,
at
1.75
7.00
28 furnace men,
at
1.75
49.00
1 hoistman,
at
2.00
2.00
27 furnace material handlers, coal, and
cinder
wheelers,
at
1.50
40.50
2 leachers.
at
1.75
3.50
4 copper precipitators,
at
1.50
6.00
Unloading cinders and salt,
20.00
Loading purple ore for shipment.
15.00
21 tons of salt,
at
3.00
63.00
Pyrites fines.
7.00
20 tons of coal.
at
1.00
20.00
5 1/2 tons of sheet-iron scrap,
at
9.00
38.50
Repairs, depreciation, management, etc.
40.00
.25
Or $1.87 per ton of 2000 lb. pyritic cinders, or $1.69 per ton of
2000 lb. of mixture.
Cost of Producing Copper by the Longmaid-Henderson Process in a
Modern Plant, Using Mechanical Roasters.— The cost, per pound of
copper extracted, in a modern Longmaid-Henderson process plant
CHEMICAL PROCESSES 267
located in the Eastern United States, using mechanical furnaces, and
treating from 300 to 400 tons of pyritic cinders per day having a copper
content of 2.27 per cent., and a sulphur content of 2.28 per cent., is as
follows:
Cost, Per Pound of Copper Extracted in Modern Longmaid-Henderson
Plant
Items of expense,
Per pound of copper
Process labor,
$0.0136
Misc. supplies,
0.0010
Coal,
0.0027
Fuel oil.
0.0067
Salt,
0.0166
Scrap iron,
0.0018
Repairs, all labor
and material;
0.0095
Total, $0.0519
The tailings are worked up by agglomeration into an iron ore of the
following composition:
Metallic iron, 68 . 00 per cent.
Sulphur, 0 . 07 per cent.
Copper, 0. 15 per cent.
Silica 3 . 00 per cent.
Phosphorous, 0.012 per cent.
Oxygen, etc., 28.77 per cent.
In the Western United States, under somewhat similar conditions,
in treating material having a copper content of 2.27 per cent., the esti-
mated cost per ton of 2000 lb. is as follows:
Items of expense,
Process labor,
Misc. material,
Coal,
Lubricating oil.
Fuel oil,
Salt,
Scrap iron,
Repairs, labor and material,
Total, $2.96
Elimination of Arsenic, Antimony and Bismuth.* — In the iongmaid-
Henderson process the chloridization of the copper preparatory to its
solution, is accompanied by the three elements, arsenic, antimony, and
bismuth. These chlorides being volatile at low temperatures are carried
off to a greater or less extent, and are largely dissolved in the wash water
and collected with the condensed acids, in the tower liquors. The arsenic
'Trans. A. I. M. E., Vol. XXXIII, p. 667, Allen Gibb.
Cost
per ton of ore
$0.
71
0.
,07
0,
,21
0,
,01
0,
,37
0,
.99
0
.15
0
.45
268
HYDROMETALLURGY OF COPPER
in a notable quantity and the antimony in a less amount mainly in com-
bination as arsenates and antimonates, with some bismuth remain m the
roasted ore. In washing with water, these salts, as well as the bismuth
that remains in the roasted ore, are dissolved only in minute traces, so that
if the copper is precipitated from these solutions it would be practically
free from impurities. The copper, insoluble in water, that is in the
roasted ore, is removed by the use of tower liquors, and under the action
of this solvent a considerable proportion of the arsenic, antimony, and
bismuth that remain in the calcined ore is dissolved. This increases the
proportion of these elements that is already present in the tower liquors.
In the tower liquors, obtained by washing the gases from the fur-
naces, there was found: arsenic, 0.0222 gm. antimony, 0.0005 gm.; and
bismuth, 0.0046 gm. per liter.
The proportion of impurities in the copper from the same ore vary
greatly according to whether the practice of treating the aqueous and
acid solutions separately is followed or not.
The following data is taken from practice in which the two solutions
were not kept separate:
Roasted ore Precipitate
Per cent,
actual
Per cent,
relative
Cu = 100 per
cent.
Per cent,
actual
Per cent.
relative
Cu = 100 per
cent.
Total per cent,
of elimination
Copper
Arsenic
Antimony. . . .
Bismuth
4.65
0.16
0.026
0.018
100.0
3.44
0.559
0.388
73.33
0.974
0.008
0.053
100.0
1.33
0.011
0.072
61.3
98.0
81.4
Extraction of Copper from Atacamite.' — At Chiquicamata, Chile, a
multitude of small fissures, filled with atacamite, or oxychloride of copper,
traverse the country rock, consisting of granites, pegmatites, syenites, in
every direction. This kind of deposit, which appears to be in the nature
of a stock work, a leaching process is usually adopted in order to extract
the metal from the oxychloride.
In treating Atacamite ores the difficulty has been in the filtration of
the liquid containing the dissolved copper. The chemical reactions
which take place between the perchloride and oxide of iron on the one
part and the ferric solution and argilaceous portion of the gangue on the
other, result in the formation of a gelatinous precipitate, which has to be
washed many times in order to get out the dissolved metal.
^London Mining Journal, June 30, 1906, Nioanor Argandona.
CHEMICAL PROCESSES 269
The first part in the treatment of the ore by a new process consists in
converting, by the action of steam, a portion of the chloride of copper in
the ore into hydrochloric acid and black oxide of copper, according to
the equation:
CuCl2 + H20 = CuO+2HCl.
A certain portion of the ore is put into large clay retorts, or into
iron retorts lined with a thin coating of clay, and exposed at a temper-
ature of 230° C, to the action of steam. Theoretically, there is sufficient
water of combination in the ore itself to supply the quantity of vapor
needed. The expenditure of steam is very small, but a certain quantity
is necessary in order to accelerate the reaction. The black oxide result-
ing from this process is reduced by smelting.
The second part of the process consists in submitting the ore, which
contains from 3 to 4 per cent, copper, to the action of hydrochloric acid
obtained as above described. For this purpose cylindrical wooden or
brick vats are used, having special lining of pitch. Ordinary filters can
be employed in this operation, as no gelatinous precipitate is produced.
CHAPTER XI
COPPER PRECIPITANTS
Iron.— Iron is almost universally used in the chemical precipitation of
copper from its solutions. It presents many advantages over other pre-
cipitants; it is usually quite cheap, generally obtainable, and gives the
resultant copper in the metallic condition. Scrap iron is ordinarily em-
ployed, but scrap iron, while cheap enough in industrial centers becomes
prohibitive in mining districts which are located long distances from
railroads and from the source of iron supply.
For the precipitation of copper by iron, the solution should be as free
as possible from acid or ferric salts. The acid and ferric salts act on the
iron and waste it without precipitating any copper. Theoretically 88.8
lb. of iron are required by weight to precipitate 100 lb. of copper from
sulphate solutions. In practice, the consumption greatly exceeds the
theoretical amount, because the solution cannot be kept free from acid
or ferric salts and because scrap iron which is contaminated more or
less with impurities, not available for precipitation is ordinarily used.
With care, the consumption of iron in precipitating copper from sulphate
solutions, in practice, will average about 1.5 lb. of iron per pound of
copper precipitated; under the adverse conditions of free acid, ferric
salts, and impure scrap iron, the consumption of iron may rise to 2 or
even 3 parts of iron to 1 part of copper.
Ferrous sulphate by prolonged contact with air is decomposed into
free sulphuric acid and ferric sulphate; the former dissolves iron, and the
latter combines with it to again form ferrous sulphate. The following
equations explain the principal chemical changes that take place, in
precipitating with iron from a sulphate solution.
(1) CuS0,+Fe = Cu+FeS04.
This represents the theoretical reaction, without any interfering
elements, and if it could be theoretically carried out, only 88.8 parts
of iron would be required to precipitate 100 parts of copper. As both
ferric sulphate and free sulphuric acid are likely to be in the solutions,
the following reactions also take place:
(2) Fe2(SOj3+Fe = 3FeSO,.
(3) H,S0,+Fe=FeS0, + 2H.
The excess of iron consumed may be said in general to be due to the
quantity of ferric iron and free acid in the copper solution. The tendency
270
COPPER PRECIPITANTS 271
of the iron is to reduce the ferric to the ferrous salts, when the theoret-
ical amount will be more nearly approached.
Some of the ferric sulphate may not be directly reduced by the iron,
but may act on the precipitated copper, as shown by the following
equation:
Fe^ (SO ,) 3 + Cu = CuSO , + 2FeS0 „
but as the copper so dissolved has to be precipitated at the expense of
the iron, the ultimate amount of iron consumed in reducing the ferric
salts is the same.
An accurate determination of the iron consumed per ton of copper
produced can be arrived at by testing the ingoing and outgoing solutions
to the precipitators, to determine the amount of the copper precipitated,
of the ferric iron reduced, and of free sulphuric acid neutralized, and
calculating the quantity of metallic iron required to bring about these
changes. S. R. Adcock ' gives the following partial analysis and calcu-
lations, of samples of liquor entering and leaving one of the cementation
tanks at Rio Tinto.
ANALYSIS OF LIQUOR ENTERING AND LEAVING TANKS, IN GRAMS '
PER CUBIC METER
' Entering '' Leaving
Copper I 2.064 j 3
Ferriciron i 1.328 Nil
Sulphuric acid j 1 . 198 | 712
Calculations in Grams per Cubic Meter of Liquor
Cu precipitated, 2.064-3=2.061x8/9 = 1.832 grm. of iron required.
Ferric iron reduced, 1.328 X 1/2 = 644 grm. of iron required.
Sulphuric acid neutralized, 1.189-712 = 486X4/7= 278 grm. of iron required.
Total iron, 2^74 grm.
2.061 grm. of copper precipitated would require 2.774 grm. of iron
or one part of copper precipitated from the liquor would require 1.345
parts of iron. Taking the metallic contents of the pig iron used at 92
per cent, the consumption in this instance works out at 1 part of copper to
1.462 parts of pig iron. In addition, there is a small consumption of
iron, due to impurities in the solution and waste in cleaning up the
copper."
It is desirable, for the best work in precipitation, to have the copper
solution slightly acid. A slight acidity tends to hasten the precipitation,
and prevents the separation of basic iron salts. In general, the acid con-
' Min. Ind., Vol. IX, p. 238.
272 HYDROMETALLURGY OF COPPER
tent of the copper solution for precipitation in a running stream, or if
agitated, should not be more than 0.1 to 0.2 per cent. If the acid much
exceeds this amount the consumption of iron is likely to be correspond-
ingly high.
Every precaution should be taken to reduce the ferric iron to the
ferrous condition before precipitating the copper, for in most instances
the excessive consumption of iron during precipitation, can be traced to
the high ferric iron content of the liquor under treatment.
If there is arsenic in the ore it is likely to go into solution with the
copper, and is to a certain extent precipitated with it. When the liquor
is rich in copper and the precipitation is taking place rapidly, the amount
of arsenic precipitated is comparatively small, but as the liquor gets
weaker in copper, the proportion of arsenic to copper precipitated is
much higher.
If the copper is dissolved as a chloride, either by hydrochloric acid or
metal chloride, the precipitation may take place either from the cupric
or cuprous chloride, as shown by the following reactions:
CuCl^ +Fe=reCl2 + Cu.
Cu,Cl2+Fe=reCl2 + 2Cu.
It is evident, therefore, from these equations, that twice as much copper
is precipitated, theoretically, per unit of iron from cuprous chloride as
from cupric chloride. The precipitation from the cupric chloride is
theoretically the same as that of cupric sulphate — 88.8 parts of iron are
required to precipitate 100 parts of copper — whereas only 44.4 parts
of iron are required to precipitate 100 parts of copper from cuprous
chloride.
In practice, at Stadtberg, 127 parts by weight of iron were required
to precipitate 100 parts of copper from a chloride solution in which
hydrochloric acid was used as the solvent. In the Hunt and Douglas
process, in which the copper is precipitated from cuprous chloride, ohly
from 50 to 70 parts of iron are required to precipitate 100 parts of copper.
If there is free hydrochloric acid or ferric chloride in the solution, iron
will combine with them to form ferrous chloride, and will consequently
be wasted, as in the analogous procedure in the sulphate solution; this
may be shown by the following equations:
2HCl+Fe=FeCl, + 2H.
2FeCl3+Fe = 3FeCl2.
The neutralization of the free acid, and the reduction of the ferric
to the ferrous chloride, is quiet as necessary as with sulphate solutions to
effect the best economy in the precipitation.
The iron for precipitation, whether from a sulphate or chloride solu-
tion may be used in the form of wrought, pig iron, iron bars, or iron spong.
COPPER PRECIPITANTS 273
Pulveront iron in the form of ground sponge acts most rapidly. Bar
iron yields a cgarse grained cement copper, with but little coherence;
gray pig iron, which acts faster than white iron, gives a more pulverent
precipitate, while white iron throws down coherent masses. The
graphite in the pig iron separates out during the precipitation and renders
the cement copper impure.
Sponge Iron. — It is evident that if iron is used as the precipitant of
copper, it would be quite an advantage if the iron could in some way be
cheaply manufactured at the copper reduction works. The purchase of
scrap or pig iron, and its transportation to the reduction works at the
mines, is usually a prohibitive expense. An ordinary iron smelting
plant could not be considered.
The manufacture of sponge iron, in which a high-grade ore is reduced
to metal without the necessity of smelting or fluxing, has in a measure
solved the problem, but notwithstanding that the manufacture of sponge
iron and its application to copper precipitation has long been known, its
use has not met with much encouragement.
The principle of the manufacture of sponge iron is quite simple. If
ferric oxide,' FcjOg, is heated in a highly reducing atmosphere, the
oxygen of the iron oxide combines with the reducing gases, and the
resultant product is finely divided metallic iron, containing more or less
impurities. Great care must be used in cooling the iron to prevent
reoxidation, so that the cooling, as well as the heating must be done in
the presence of reducing gases.
The following description by Lunge* gives the details of early European
practice and its application to copper precipitation.
"The precipitation of copper from its solutions by means of iron takes place
more rapidly by employing spongy iron, as was done at the Beds Metal Works.
This product is made by reducing ferric oxide at so low a temperature that the
iron cannot combine with the carbon and cannot melt, but remains in the finely
divided state as 'sponge.' This method was tried in England for the first time
in 1837. Gossage, in 1859, was the first to use it in the wet method of coppfr
extraction.
"The furnace usually employed in making sponge iron, is a reverberatory in
which the flame, after having passed directly over the charge, returns below the
furnace bed, and thus heats the charge indirectly from below. Figs. .51 and .52
show the essential details of the furnace. It is, in the drawing, 28 ft. 9 in. long;
the working bed has a length of 22 ft. and a width of 8 ft. Dwarf walls, a a, 9 in.
high, divide it into three compartments, which on one .side have two working
doors, 6 b, each. Each compartment is charged and finished by itself. The
working doors are of cast metal and run in air-tight frames ; the same is the case
with the fire doors. The fireplace is constructed for generating a reducing flame ;
the grate has a surface 4 X 3 ft. ; and the bearers d, are 3 ft., latterly even 4 ft. 8 in.
' "Sulphuric .Vcid and .\lkali Manufacture," p. 815.
IS
274
HYDROMETALLURGY OF COPPER
COPPER PRECIPITANTS
275
below the fire bridge ; so that a very deep layer of fuel is obtained, which does
not allow any oxygen to get inside the furnace. The furnace bed is formed by
fire tiles 4 in. thick, with rabbited edges, partly resting upon the walls forming
the divisions of the lower flues, partly upon railway-bars.
"The flame having traveled through these flues, descends in a vertical shaft
along the fire bridge, and thence goes to the chimney. In this descending shaft
there is a fireclay damper, which is closed every time before a working door or
fire door is opened. The 9-in. furnace roof is surmounted by a flat cast-iron
dish, supported by short pillars, for drying the ore and mixing it with coal;
the mixture is charged into the furnace through the 6-in. pipes,/, carried through
Fig. 52. — Sponge-iron furnace. Transverse section.
the arch. The whole furnace rests on brick pillars, g, and the floor on the working
sides must be so much higher than that on the discharge side, that the discharging
boxes can be run underneath the furnace between the brick pillars. The dis-
charging takes place through 6-in. pipes, h, descending in front of the working
doors through the furnace bottom andthe lower flues.
"The discharge boxes are made of sheet iron, of rectangular section, tapering
toward the top. The cover is fast, and has in its center a 6-in. opening with
upright flange, by which the box is connected with the discharge tube. The
bottom of the box is movable, and turns on one side on hinges, while the other
side is fastened by bolts and cotters. The whole is mounted on four wheels in
such a way that they do not interfere with the movements of the bottom. Each
box has a capacity of 12 cu. ft.
" When the furnace is at a bright red heat, it can be charged. Y.Avh. compart-
ment receives a charge of 2000 lb. of 'purple ore' and 600 lb. of coal, which have
passed through a sieve with eight holes to the lineal inch.
"The charging takes place from the cast-iron dish above the furnace roof.
276 HYDROMETALLURGY OF COPPER
The fire and working doors are closed, so that the air enters solely through the
coals on the grate, care being taken that the burning mass does not become
hollow, lest uncombined oxygen should get inside the furnace. The time of
reduction in the compartment nearest the fire bridge varies from 9 to 12 hours;
in the second it is about 18 hours, and in the third about 24 hours. The depth
of the charge lying on the bed is about 6 in. During the time of reduction each
compartment must be turned over twice, or even three times. Although during
this time the damper is closed, a little air always enters the furnace ; but the turn-
ing over is indispensable, as the mass would otherwise cake together. The time
above stated refers to a bright red heat; a low heat is sufficient for reduction and
the iron thus made is even better for the precipitation of copper; but as in this
case much more time is required for reduction (up to 60 hours) this style of
working does not pay. The fire place being so deep, fresh coal need only be
thrown on twice or three times every 12 hours, say 1500 lb. per ton of ore.
"The completion of the reduction is ascertained by testing. A small sample
is taken out, put on an iron plate, covered with a brick until it has become cold,
and 1 grm. of the (unoxidized) central part tested by a cupric sulphate solution
of known strength, which is run from a burette on the spongy, iron with frequent
stirring; from time to time a drop is put on a bright blade of iron, to see whether
any stain of copper is produced upon it. When the reaction in any of the three
compartments is finished, the damper is closed; two of the discharging boxes are
run underneath the furnace, and their openings connected with the discharge
pipes by an iron hoop luted with clay; then the charge is raked down into the
boxes as quickly as possible. The boxes are then closed with the loose cover,
run out again, and allowed to cool for 48 hours. They are then lifted by a crane,
and the cotters are knocked; whereupon the bottom turns on its hinges and the
whole mass of spongy iron readily falls out, owing to the box tapering upward.
The sponge is then finely ground in a heavy chillean mill 6 ft. in diameter, and
passed through a sieve with fifty holes per lineal inch. It is now ready for the
precipitation of the copper.
The following is an analysis of sponge iron produced from purple ore, from
Spanish pyrites.
Purple Ore Sponge Iron
Ferric oxide, 95 . 10 per cent. 8 . 15 per cent.
Ferrous oxide, 2.40 per cent.
Metallic iron, 70 . 40 per cent.
Copper, 0.18 per cent. 0 . 24 per cent.
Lead oxide, 0.96 per cent
Lead, 0.27 per cent.
Sulphur, 0.07 per cent. 0 07 per cent.
Calcium oxide, 0.20 per cent.
Sodium oxide, 0. 13 per cent.
Sulphur trioxide, 0 . 78 per cent
Alumina, . : . . 0 . 19 per cent.
Zinc, 0.30 per cent.
Silicious residue, 2 . 13 per cent. 9 . 00 per cent.
99.55 per cent. 99.62 per cent.
COPPER PRECIPITANTS 277
" When spongy iron is employed for the precipitation of copper, cont'nuous
stirring is required, for which at some works a mechanical agitator is used, at
others, manual labor."
Any good iron ore may be converted into iron sponge, or the con-
centrate residues, after roasting and copper extraction.
Crucible Method of Iron Ore Reduction. — Another method of reducing
iron ore in connection with copper precipitation — that in which the
mixture of ore and coal is reduced in a crucible — might be considered.
This method of manufacturing iron has been in use in China from remote
antiquity, and large quantities are in this way reduced there.
At Shansi, China, the crucibles are about 19 in. high and 6.5 in. in
diameter.. They are filled with a mixture of ore which has been broken
and sorted to walnut size; coal of about the same size, and coal dirt.
The proportions are four baskets of ore, one basket of coal, and one
basket of coal dirt on top. The crucibles are heated in a stall furnace.
It takes about 16 hours to reduce the iron. After cooling the crucibles
are removed and broken and the iron taken out.
Precipitation with a Coke-iron Couple. — ^^'. L. Austin' gives the result
of a 30-day comparative test on a working scale, using ordinary mine
waters, between iron and the coke-iron couple as the precipitant. The
average daily assays showed the following extraction effected within
8 ft. from where the waters entered the precipitation boxes:
Percentage of copper removed in the boxes containing clean wrought
iron 25
Percentage of copper removed in the boxes containing coke-iron
couple 43
It was found that when coke was added to the iron in the precipitating
boxes through which the sulphate mine waters were flowing, by actual
weight from 1.15 to 2.66 lb. copper was precipitated to the pound of iron
consumed. One of the advantages in employing the coke-iron couple
in cementation is that the precipitation of the copper takes place within
a short distance from the point where the liquors enter the boxes, thereby
avoiding to a great extent the formation of basic iron salts which usually
debase the cement copper.
Precipitation with a Copper -iron Couple. — If iron plates or a bundle
of scrap iron in a crate are immersed in a copper sulphate solution and
connected electrically with a copper plate to serve as a cathode, no copper
is precipitated on the iron, but all is precipitated on the copper plate.
The iron acts as a soluble anode, going into solution as ferrous sulphate,
while the copper is deposited in a dense, practically pure condition on the
copper sheet. In this case the iron and copper act as a battery, with a
theoretical difference of potential of 0.81 volt. If more rapid precipita-
' "Mines and Methods," January, 1911.
278 HYDROMETALLURGY OF COPPER
tion is desired than given by this action alone, a current from a dynamo
may also be used in connection with that produced by the iron and copper.
Hydrogen Sulphide. — If hydrogen sulphide is applied to a copper
sulphate solution, the copper is precipitated as cupric sulphide, and an
amount of acid regenerated equal to that combined with the copper
CuS04 + H3S = CuS + H2SO,.
If the solution is a chloride, a similar reaction takes place:
CuCl^ + H^S = CuS + 2HC1.
In either case, the regenerated acid solution is returned to the ore to
dissolve more copper. The precipitated copper sulphide, like copper
matte, may be brought to blister copper in converters. The escaping
sulphur dioxide may be used in other stages of the process, either in the
manufacture of sulphuric acid, or in the preparation of hydrogen sulphide.
Many wet methods of extracting copper from its ores have been based
on the use of hydrogen sulphide as the precipitant. Most of these methods
have as their fundamental idea, the inexpensive generation of the hydro-
gen sulphide, rather than a direct method of copper extraction.
At the Bede Metal Works the hydrogen sulphide was produced by
treating sulphide of sodium with carbon dioxide, generated by burning
coke in a shaft furnace. The sulphide of sodium was made from the acid
mother liquor from which the sulphide of copper had been precipitated,
by evaporating to dryness in a reverberatory furnace, mixing the residue
with coal dust and reducing it in a similar furnace. This product, con-
sisting of sulphide and carbonate of soda, was leached, and the solution
treated with carbon dioxide. The carbonate of soda was produced as a
by-product of the process.
Hydrogen sulphide, according to Schnabel,' is best generated by lead-
ing sulphur dioxide and water vapor over red hot coke or charcoal. For
this purpose the gases produced in roasting sulphides in pyrite burners
or kilns are aspirated by means of Korting injectors, and forced together
with the steam from the injectors through a shaft furnace filled with glow-
ing charcoal or coke. The sulphur dioxide is reduced to sulphur by the
carbon; the water vapor forms with the red hot carbon, hydrogen
and carbon monoxide, and the hydrogen and sulphur combine to form
hydrogen sulphide. The coke or charcoal is kept hot by injecting a
stream of air from time to time as in the production of water gas.
A method devised by Sinding depends upon the action of sulphur
vapor on hydrocarbons and hydrogen; another on the decomposition of
sodium sulphide by carbon dioxide. Sinding generates producer gas with
raw fuel, and leads it over red hot pyrites; by the action of the hydro-
carbons and the hydrogen contained in producer gas upon the sulphur
evolved from the pyrites, hydrogen sulphide is formed, and is made to
^ Handbook of Metallurgy, Vol. I, p. 209.
COPPER PRECIPITANTS 279
traverse a chamber in which the cupriferous solution is dropping down in
the form of rain.
Gill and Gelstharp produced hydrogen sulphide by the action of
carbon dioxide on sodium sulphide.
In the application of hydrogen sulphide to the precipitation of copper
solutions, the precipitation is thorough and complete. If the solution
contains gold and silver, the filtrate will never assay more than a trace;
that is to say, less than 0.0005 oz. or 1 cent per ton. The conditions for
thorough precipitation in a normally working plant, are that the solution
should be slightly acid, or if neutral, contain an excess of free chlorine or
ferric salts. If the solution is alkaline or neutral, the precipitation of the
copper, gold and silver will be complete, but some of the undesirable
elements will be thrown down with them. The precipitate is voluminous
and may be difficult to filter. If the solution is neutral and contains a
large excess of chlorine or ferric salts, the hydrogen sulphide reacting
with these substances, will reduce them, and acidify the solution suffi-
ciently to prevent the baser metals from being precipitated.
C1 + H2S=2HC1 + S,
2FeCl3 + H2S = 2HCl + 2FeCl2 + S,
Fe2(SO,)3-t-H2S=2FeSO, + H2SO, + S,
and thus hydrogen sulphide will be consumed, an equivalent of acid
regenerated, and elemental sulphur precipitated with the copper sulphide.
The metals of the alkalies and alkaline earths are not precipitated from
either the acid or alkaline solutions. Frequently compounds of calcium
and aluminum will be found in the precipitate. Calcium sulphate is
only slightly soluble in water, but is somewhat soluble in chloride solu-
tions, and unless the solution is thoroughly settled or filtered, it is also
likely to be carried into the precipitate by suspension. If the ore con-
tains much alumina and the solution issuing from the leaching tanks is
neutral or alkaline, a white gelatinous substance, probably aluminum
hydroxide, settles out of the clear or milky solution, and may in this way
get into the precipitate. On the addition of acid this precipitate is
redissolved and the solution becomes clear.
Metals which are precipitated by hydrogen sulphide, as sulphide, from
a solution of their salts in the presence of free acid, are :
Platinum, color of precipitate, dark brown.
Gold, color of precipitate, dark brown.
Silver, color of precipitate, black.
Copper, color of precipitate, black.
Lead, color of precipitate, black.
Tin, color of precipitate, yellow-brown.
Antimony, color of precipitate, orange.
Arsenic, color of precipitate, yellow.
Mercury, color of precipitate, black.
Cadmium, color of precipitate, yellow.
280
HYDROMETALLURGY OF COPPER
ifli:t"~"i" --
■■ ■ — ~
_.^_ ___._ _
■
Compressed Air ^
J
Fig. 53. — Hydrogen sulphide generator.
COPPER PRECIPITANTS 281
The hydrogen sulphide employed in precipitation is usually generated
from iron sulphide (FeS), sulphuric acid, and water. A solution having
a large excess of acid will require more hydrogen sulphide to precipitate
the metals than one only slightly acid. A strongly acid solution will
give a cleaner precipitate than a neutral solution, and give less trouble
in the filter presses.
The hydrogen sulphide generator, for producing the gas on a large
scale, from acid and matte, is shown in Fig. 53 and may be made of any
size desired. It consists of an iron cylinder with flat cast-iron bottoms
and tops. If the cylinder is small, it is made of cast iron; if large, of
sheet steel. It is lined with lead, and all connections are made of lead
pipe with burned joints. An equalizing tank, also made of lead, is
located close to the generator.
The materials for generating the hydrogen sulphide are charged in
the following proportions:
Iron sulphide (iron matte) , 1 lb.
Sulphuric acid, 2.5 lb.
Water (approximately), 6 lb.
The total quantity of chemicals charged, will depend on the size of the
generator, and the amount of metal to be precipitated.
In charging the generator, the exhaust valve from the equalizing tank
is left open. The ]-equired amount of water is then introduced. This
may be measured in a small tank over the generator, or it may be run in
through the water pipe, or hose inserted into the manhole, to a certain
depth which has been determined before hand as the right quantity of
water for any required charge. The iron sulphide, broken into pieces
about the size of hens' eggs, is dropped in through the manhole, which is
then closed. The required amount of acid, which had previously been
measured into the small lead tank located over the generator, is then run
in and the valve again immediately closed. The generation of hydrogen
sulphide begins at once. The exhaust valve on the equalizing tank, which
was left open so that the pressure generated by the gas would not prevent
the acid fr^m flowing into the generator, is then closed. The increasing
pressure will soon force the gas into the copper precipitating tanks.
Iron sulphide reacts with sulphuric acid to form hydrogen sulphide:
FeS + H2SO,=FeSO, + H2S.
The gas so obtained always contains free hydrogen, owing to the
presence of uncombined iron in the iron sulphide.
Compressed air is used to force the gas from the generator into the
precipitating tanks and agitate the solutions.
After the precipitated solution has settled long enough to clarify,
usually from 4 to 8 hours, it is decanted through the collar in the bottom
282 HDYROMETALLURGY OF COPPER
of the settling tank, into a filter press, and the clear regenerated solution
may then be returned to the ore to dissolve more copper.
Lime. — Lime, calcium hydrate or milk of lime, is not suitable for
precipitation from sulphate solutions. It has been used in precipitating
copper from cupric and cuprous chloride solutions, calcium chloride being
produced.
Milk of lime precipitates cupric hydrate from cupric chloride, and
cuprous oxide from cuprous chloride. It throws down copper as cupric
hydrate from a solution of the sulphate, but the precipitate is mixed
with the insoluble calcium sulphate, which is formed at the same time,
and gives a mixture which is very voluminous and very troublesome to
smelt. The precipitation of the copper with lime from cuprous chloride
was for a long time used in connection with the Hunt and Douglas
process, in which the resulting calcium chloride was used in another
step in the process, but lime has nowhere been regularly used to pre-
cipitate copper from sulphate solutions.
CHAPTER XII
ELECTROLYTIC PROCESSES
General Consideration of Electrolytic Methods.
Definitions. — Electrolysis may be defined as the decomposition of a
chemical compound by the electric current. The compound may be in
aqueous or igneous solution.
Electrolyte is a chemical compound in aqueous or igneous solution
being decomposed by the electric current.
Electrode is a conductor to convey the current of electricity into or out
of the electrolyte.
Anode is the electrode by means of which the current enters the
electrolyte.
Cathode is the electrode by means of which the current leaves the
electrolyte.
Ions are the constituent elements or radicals which carry the current
of electricity through the electrolyte.
Anions are the elements or radicals which appear at the anode.
Cathions are the elements or radicals which appear at the cathode.
Electrolyzer is the apparatus by means of which or in which the
electrolysis takes place.
Diaphragm is a partition, permeable or impermeable, between the
anode and cathode, which permits the passage of the electric current but
prevents the mixing of the electrolytes in which the anodes and cathodes
are immersed.
Anolyte is the electrolyte in the anode compartment of the
electrolyzer.
Catholyte is the electrolyte in the cathode compartment of the
electrolyzer.
Current Density is the quantity of current, i.e., the number of am-
peres, flowing through a unit of electrode surface.
Current Efficiency is determined by the yield per ampere.
Energy Efficiency is determined by the decomposition per watt.
Watt. — A watt is the product of one ampere multiplied by one volt.
Horse-power is the equivalent of 746 watts, and is approximately
equal to 3/4 kilowatt.
Kilowatt is 1000 watts, and is approximately equal to 1 1/3 h. p.
Columb. — One ampere of electricity flowing for one second.
283
284 HYDROMETALLURGY OF COPPER
Electrolysis has been made the basis of a number of processes of ex-
tracting copper from its ores. Ever since success was achieved in elec-
trolytic refining of blister copper, metallurgists have naturally asked why
similar operations could not be successfully applied in the extraction of
copper direct from its ores. At first thought the matter seems simple
enough, but there are difficulties in the way of ej^tracting copper from its
ores which are not encountered in electrolytic refining. These diffi-
culties, however, do not appear to be insurmountable, and it is quite
probable that electrolytic methods of extraction will be in general use
in the near future. Many of the difficulties, first met with, have been
surmounted, and the remaining ones are gradually being overcome.
Anode. — One of the fundamental differences between electrolytic
refining and electrolytic extraction, lies in the anode. In electrolytic
refining the anode is a slab of blister copper, usually about an inch thick,
and containing less than 1 per cent, of foreign matter. This copper
anode goes into solution and is redeposited on the cathode. Theoretic-
ally, no energy is required to perform this work, since the energy de-
veloped in dissolving the copper anode is equivalent to that consumed
in its deposition on the cathode. The soluble anode presents no serious
difficulty. It has to be replaced at frequent intervals. Theoretically,
no acid is consumed and none generated in electrolytic refining, since the
acid going into combination with the copper at the anode is again re-
leased as free acid by its deposition at the cathode.
When copper is dissolved from the ore, conditions are entirely differ-
ent. The problem then is, to deposit the copper out of solution by
electrolysis while none is going into solution. This manifestly involves
the use of an anode which will conduct the electricity into the copper
solution, while at the same time the anode itself remains unattacked, or
insoluble. This presents the first serious difficulty in precipitating
copper from leaching solutions.
It has been found a most difficult matter to provide a substance
which will be a good conductor of electricity and not be attacked by the
combined action of the current and the solvent. Many substances which
are sufficiently permanent, have too high a resistance, and consequently
the power required to overcome this resistance in the anode, makes the
process so expensive as to be prohibitive; while on the other hand, sub-
stances which have a good electrical conductivity are not sufficiently
permanent. For chloride solutions this problem of suitable anodes has
been satisfactorily solved by the use of graphitized carbon, but for
sulphate solutions no really satisfactory anode has yet been discovered,
although most conceivable substances have been tried. Lead, on the
whole, has given the best results for sulphate solutions. The purity of
the lead is an important factor.
Platinum is too costly for any process operated on a working basis.
ELECTROLYTIC PROCESSES 285
All other commercial metals are attacked, and even platinum is not
entirely unaffected. Ferro-silicon, which is a difficultly attackable
substance, has been repeatedly suggested, but has not proved successful
in practice. Antimonial lead has been tried for sulphate solutions, but
there is no evidence to indicate that it is more permanent than pure lead;
the antimony from the lead, going into solution in the electrolyte, may
be objectionable and its loss by the decomposition of the anode, may be
a serious item of expense. In almost all cases carbon is the only sub-
stance which can be employed in chloride solutions, but for sulphate
solutions it is absolutely worthless.
The quality of carbons for electrolytic work varies considerably, but
even the best are eventually destroyed. Graphitized carbons have given
satisfactory results with chloride, but not with sulphate, solutions.
These carbons, in addition to being durable, are good conductors of
electricity. They possess a specific resistance of but 0.00032 ohm per
cubic inch, which is only one-fourth that of amorphous carbon. The
specific resistance is 0.000813 ohm per cubic centimeter, and since that
of mercury is 0.000094 ohm, the conductivity is 11.75, mercury being
100; or 0.21, copper being 100. In some cases graphitized carbons have
been used for 3 years as anodes in the decomposition of alkali metal
chloride solutions, with a current density of 50 to 250 amperes per square
meter. It is essential, however, that the solution be acid; in alkaline or
neutral solutions, they are not as durable. The following is a list of
standard sizes and weights of Acheson-Graphite electrodes:
Acheson-Graphite Electrodes
Size
Round:
1
n
li
n
2
3
4
5i
6
8
n. diam. X12 in.,
in. diam. X 12 in.,
n. diam. X24 in.,
in. diam. X 24 in.,
n. diam. X 24 in.,
n. diam.X24 in.,
in. diam. X 24 in.,
in. diam. X 24 in.,
in. diam. X 24 in.,
n. diam. X 24 in.,
n. diam. X 24 in.,
n. diam. X 24 in.,
n. diam. X 24 in.,
n. diam. X 40 in.,
n. diam. X 40 in.,
in. diam. Xl9i in.,
n. diam. X 48 in.,
in. diam. X 48 in..
Approximate weight
per piece
0.008 lb.
0.035 1b.
0.15 lb.
0.21 lb.
0.26 lb.
0.40 lb.
0.60 lb.
0,82 lb.
1,00 lb.
1,40 lb.
1.751b.
2,60 lb.
4,50 lb.
16.60 1b.
30.00 lb.
23.50 lb.
75.00 lb.
132.00 !b.
Approximate wei
per piece
7,
.50 1b.
37.
,00 lb.
84,
,50 1b.
2
, 10 lb.
2.94 1b.
4.50 1b.
4.05 lb.
3
.13 lb.
4
.21 lb.
6
.75 1b.
2
.20 1b.
7
. 50 lb.
11
.30 1b.
7
.88 lb.
11
.50 1b.
14.
50 lb.
25
.75 lb.
33
,25 1b.
31
.45 lb.
286 HYDROMETALLURGY OF COPPER
Acheson-Graphite Electrodes. — Continued
Size
Square:
2X2X30 in.,
4X4X40 in.,
6X6X40 in..
Rectangular:
J X 12X12 in.,
iX4 X24in.,
iX6 X24in.,
iX 12X12 in.,
}X Six 19 J in.,
fX5 XlSin.,
1x12X12 in.,
|X2 X21-Hn.,
1 X4 X30in.,
1 X6 X30in.,
liX3 X36in.,
liX5 X30in.,
2 X4 X30 in.,
2 X7 XSOin.,
3 X6 XSOin.,
4 X8fXl5in.,
These electrodes cost approximately from $13.50 to $15.00 per 100 lb.
in lots of 500 lb. for the sizes most convenient for electrolytic use. In
ton lots the cost would be approximately $11.50 per 100 lb.
Cathodes. — The cathodes used in electro deposition are usually
thin sheets of pure copper. Carbon cathodes, in chloride solutions, do
very well but are not advisable for sulphate solutions. If copper cathode
sheets are used, they may be stripped after the deposit has acquired the
desired thickness, or new sheets may be supplied as the old ones are
removed. Lead cathode sheets have been used, and the copper stripped
from the lead, or, the lead may be melted from the copper after removal
from the electrolyzers, and again rolled into sheet lead for new cathode
sheets.
In depositing copper from impure solutions, that is to say, solutions
like those obtained in leaching copper ores, it is not an easy matter to
get a reg-uline deposit of the desired thickness. Unless considerable care
is taken with the electrolyte and the current density, irregular deposition
and sprouting will occur long before the cathodes have acquired the
thickness desired for their removal. This difficulty may be so aggravated
as to make their removal necessary at an early stage of the operation, and
thus adding considerable to the expense. If the cathodes, under such
conditions, are not removed, the difficulty is quickly aggravated and
short circuiting and inefficiency are likely to result. With a reasonably
pure solution and low current density, this difficulty is not likely to occur.
ELECTROLYTIC PROCESSES 2,S7
especially if the solution is agitated or if the cathode is moved through the
electrolyte. The current density, as well as the nature of the electrolyte,
has much to do with the quality of the deposit; the lower the current
density, the more reguline the deposit is likely to be, but there is a mini-
mum practical limit to the current density that can be employed.
Diaphragms.— Most of the electrolytic processes are based on the fact
that the solvent may be regenerated during electrolysis. If the solution
electrolyzed at the cathode requires to be kept separated from that at
the anode, then diaphragms are necessarily inserted between the anode
and cathode, which, while allowing the free passage of the current, will
prevent the solutions from mixing. Or if the electrolyte contains a
large amount of an oxidizable and reducible metal under the anodic and
cathodic influences, it may be desirable to use diaphragms, simply to
overcome the undue loss in electrical efficiency. If copper sulphate
solution, containing much iron sulphate is electrolyzed, considerable
energy may be consumed in the oxidation and reduction of the iron,
which simply renders an equivalent in useless heat.
Various materials may be used for diaphragms, but not many fulfill
the conditions of low electrical resistance combined with sufficient density
to prevent the anode and cathode solutions from mixing too freely. The
material best suited for diaphragms is asbestos, which is not easily attacked
either by acid or alkaline solutions, and when saturated with the electro-
lyte, offers no appreciable resistance at low current densities. At high
current densities, any diaphragm is likely to offer appreciable resistance.
Asbestos, suitable for diaphragms, is manufactured either as cloth, paper,
or mill board. Asbestos cloth will give the best results if it is not neces-
sary to make an absolute separation between the anolyte and catholyte.
When complete separation is necessary it is desirable to use mill board or
asbestos paper in connection with asbestos cloth. In whatever form the
asbestos is used, it is desirable that it be quite free from foreign matter
or admixtures.
While the construction of suitable diaphragms is not an insurmount-
able difficulty, it is desirable to dispense with them wherever practicable.
The insertion of diaphragms in the electrolyzer also presents difficulties,
as well as the construction of the diaphragm itself. In a. copper electrolyte
the ordinary materials of construction cannot be used, and hence to
employ materials to take the place of the ordinary iron nails and bolts is
more or less expensive. Ordinary porous clay, such as is used in battery
jars makes a good diaphram to keep the solutions separate, but its elec-
trical resistance is too high to be used in practice, and it must be used in
sizes too small for economic adaptation.
It is evident that a diaphragm if it is to fulfill its purpose, its solid
particles must not carry any current, since otherwise the diaphragm
would not act as a diaphragm, but as a bipolar electrode. The current
288 HYDROMETALLURGY OF COPPER
passes through the interstices of the solid particles which compose the
diaphragm, and the resistance of the diaphragm is the composite resistance
of all the innumerable passageways of the electrolyte through the pores
of the diaphragm. Diaphragms should be easily permeable to charged
ions but should offer a high resistance to diffusion.
In Denver, at the Greenawalt experimental plant, were used some
years ago, on a large scale, diaphragms made of asbestos cloth, asbestos
paper, or mill board, or a combination of these, sandwiched between
perforated boards from 1/2 to 3/4 in. thick. The perforations were about
1/2 in. in diameter and as close as possible, consistent with strength. The
boards, with the asbestos between, were fastened together with wooden
dowel pins and keyed. These diaphragms were expensive, and as the
perforations did not exceed half the total area, only about half the
diaphragm was effective. Diaphragms constructed in this way, must of
necessity be small. Their use was discontinued.
Later, W. E. Greenawalt constructed diaphrams which were used in
copper electro deposition, and were built 13 ft. 6 in. long by 3 ft. wide,
made of two sheets of asbestos cloth, between which were sandwiched
the desired thickness of asbestos paper, and these in turn, were sand-
wiched between two mullioned oak frames, bolted together with copper
bolts. These diaphragms gave satisfactory results, but were later dis-
carded when it was found that for the process under demonstration, dia-
phramgs could be dispensed with entirely.
Betts states that the best diaphragm he knows of for ordinary weak
acid solutions may be made as follows:' Powdered sulphur is sifted
evenly over a 1/4-in. asbestos mill board and the sheet heated evenly
an hour or so just above the melting point of sulphur. The operation
is then repeated with the other side. It takes from 1/4 to 3/4 lb. of
sulphur to the square foot of mill board. This diaphragm is plastic
when heated and in solutions does not soften at all. It expands
slightly, and should be kept in acid water two or three weeks before
using in any kind of rigid construction. The resistance is somewhat
higher than without the sulphur, but i,he diffusion is smaller.
Current Density. — The current density has much to do with the effi-
ciency of the operation and the nature of the copper deposited. It is
not possible, in electrolyzing impure leaching solutions, to use as high a
current density as in electrolytic refining, owing principally to the im-
purities, and to some extent the leaner copper content. If a reguline
deposit is desired, it will be necessary to u.se a rather low current density.
In electrolytic decomposition high current density causes impoverish-
ment of ions at the electrodes and causes trouble. This impoverishment
is prevented by stirring the electrolyte or moving the electrodes.
The energy efficiency becomes less as the current density is increased,
' " Electrochemical Industry," July, 1908.
ELECTROLYTIC PROCESSES 289
and the theoretical voltage is only approached at very low current
densities. If the voltage for a certain electrolyte, for example, is 1.2 at
5 ampcics per square foot, at a current density of 30 amperes per
square foot it is very likely to be three times that. Or, in other words,
the energy efficiency at a current density of 30 amperes per square foot,
is likely to be only one-third that at 5 amperes per square foot, assuming,
of course, that the amount of copper deposited is the same in both cases,
which may be a gratuitous assumption. It follows, further, that while
the electrolytic deposition of the copper will take three times the power
at 30 amperes per square foot as it would for 5 amperes per square foot,
the size of the plant would be only one-sixth as large, and presumably
cost only one-sixth as much.
The lower current densities will be largely limited to the cost and size
of installation for the same output, and the larger current densities by the
cost of operation.
The cost of power will be a governing factor in deciding the current
density to be employed. If power is cheap it will probably cost no more
to operate at a fair current density than at a low current density; for
while the power, per pound of copper, will cost more, this will be offset
by the decreased cost of installation, maintenance, and attendance.
If high current densities are employed, it will be desirable to agitate
the electrolyte, or employ some form of moving cathode, so as to bring
sufficient copper ions in contact with the cathode and thus avoid useless
expenditure of energy in the decomposition of other substances in the
electrolyte, or even the electrolyte itself.
It was observed soon after practical attempts were made at copper
refining, about the year 1865, that the current density, and consequently
the rate of deposition could be considerably increased by circulating the
electrolyte or moving the electrodes.
Wilde was one of the first to deposit copper on a revolving cathode.
The anodes consisted of copper cylindrical tubes, and the cathode con-
sisted of an iron cylinder which was to be coated with copper. The
cathode was placed in the center of the electrolyzer and rotated on its
axis. This gave an even distribution of the copper over the entire
cathode surface by means of the motion imparted to the solution and
the equal current density resulting from the motion. The current
density used was about 20 amperes per square foot.
Elmore used horizontal mandrels on which copper sheets or tubes
are deposited, while agate burnishers travel continuously over the copper
so as to consolidate it and at the same time prevent the growth of copper
trees or nodules. The current density used was about 30 amperes per
square foot.
Dumoulin introduced a process for burnishing copper, during deposi-
tion, with sheepskin as a substitute for agate. He claimed that the
19
290 HYDROMETALLURGY OF COPPER
process had the additional advantage of insulating any projections that
might be formed on the deposited metal. It was claimed that a current
of from 30 to 40 amperes per square foot of cathode surface could be used
with a difference of potential of 1.6 volts, using a soluble copper anode.
Cowper-Coles gets excellent deposits of any desired thickness by
revolving a cylinder cathode at a speed of from 1500 to 2000 lin. ft. per
minute, with a current density of 200 amperes per square foot.
Attempts have been made at various times to increase the rate of cop-
per deposition by Swan, Elmore, Thofern, Graham, Poore, and others by
impinging jets of the electrolyte against a cathode surface. The quality
of the deposits is likely to be unsatisfactory if impinging jets are alone
employed; it is desirable, therefore, to move the cathode also, in order
that the deposited copper may be uniform over its entire surface.
Whatever the means employed to increase the rate of deposition, the
essential object to be attained is to bring sufficient copper ions in contact
with the cathode, corresponding to the increased current, and to get a
reguline deposit by friction either with the electrolyte or burnisher.
General Laws Governing the Electrodeposition of Copper. — In the
decomposition of copper solutions, work is performed and energy con-
sumed. The factors governing the expenditure of energy, in electrolysis,
are
The electromotive force.
The resistance, and.
The current.
The electromotive force (e. m. f.) is measured in volts; the resistance
in ohms, and the current in amperes. A definite relation exists between
these three factors, whereby the value of any factor may always be calcu-
lated when the value of the other two are known. This relation is known
as Ohm's law, and may be stated thus:
The current strength in any circuit is equal to the electromotive force
applied to the circuit, divided by the resistance of the current. Or more
briefly stated:
Pressure
Current =
Resistance
Volts
Amperes = -—- — or
C =
Ohms
E
R
In other words, the electromotive force (e. m. f.) which may be
assumed to be the electrical pressure by the dynamo, causes the flow of an
electric current. The current is directly proportional to the electromo-
tive force. The resistance (electrical conductors and electrolyte) oppose
ELECTROLYTIC PROCESSES 291
tlic flow of the curjcnl. The current is therefore proportional- to the
electromotive force.
The current strength in any circuit increases or decreases directly
as the electromotive force increases or decreases, when the current is
constant. With a constant pressure the current increases as the resist-
ance is decreased, and decreases as the resistance is increased; or briefly,
the current varies directly as the electromotive force and inversely as the
resistance.
From this it follows, that to double the resistance halves the current,
the electromotive force remaining constant. Or, if with the doubled
resistance the current is to remain constant, the electromotive force,
and consequently the power, must be doubled.
It also follows from Ohm's law that when through a given resistance
the current is required to be doubled, the power must be increased four
times; or in general, the resistance remaining constant the power in-
creases proportionately to the square of the current.
Ohm's law, applied to the deposition of copper, means that, theoretic-
ally, the least power is required to deposit a given amount of copper
when the lowest possible current density is used. This would require in
practice, a plant of unlimited size, so that a well designed plant must of
necessity be a compromise between theoretical conditions and practical
requirements.
The Electrical Power is the product of the number of volts multiplied
by the number of amperes of the current. Its unit is the Watt. 746
watts make a horse power; 1000 watts a kilowatt. Either the horse
power or the kilowatt may be taken as the unit of power; the kilowatt
is the most convenient in electrical and electrolytic work.
The power consumed in depositing a definite amount of copper may
be expressed in horse power, thus:
Volts X Amperes
^- P- = 746
and in kilowatts,
Volts X Amperes
k. w. =
1000
The most important laws in relation to electrolyais are those of
Faraday. His law of electrolysis may be stated thus:
Faraday's Law. — 1. The amount of chemical change produced electro-
lytically by the current is proportional only to the amount of electricity
passing, as measured in columbs, and is independent of the strength or
temperature of the electrolyte, or the size or distance apart of the electrodes.
2. The amount of different elements dissolved or set free by the passage
of a given amount of electricity is proportional to their chemical equivalents.
Or in other words, the amount of electrochemical action produced is
292 HYDROMETALLURGY OF COPPER
directly proportional to the product of ampere hours and the chemical
equivalents.
A given quantity of current, therefore, will always deposit the same
quantity of a given element, and the elements are deposited propor-
tionally to their chemical equivalents; as for example, hydrogen 1,
oxygen 8, chlorine 35.5, cupric copper 31.75, cuprous copper 63.5, etc.
_,, ,, ^, . , ^ . , ,, . , ,. Atomic weight ^,
ihe Chemical Equivalent is the quotient -, ihe
^ ^ valency
electrochemical equivalent of a substance is identical in weight with the
amount of the same substance that would be deposited by 1 ampere
in 1 second (that is, 1 columb). The electrochemical equivalent of any
element may be calculated by multiplying the chemical equivalent of
the element by .000010384, which is the electrochemical equivalent of
hydrogen. For cupric copper it is 0.0003297 and for cuprous copper
it is 0.0006594 (grams per columb).
Theoretical Data for Copper Deposition. — Theoretically 1.1858 grm.
of copper are deposited by 1 ampere-hour from cupric, and 2.3717 grm.
per ampere-hour from cuprous solutions. This for practical purposes
would mean that 1 ampere should deposit 1 oz. of copper from cupric
solutions, and 2 oz. from cuprous solutions, per day of 24 hours. From
this it is easy to compare the current efficiency of any process or opera-
tion with the theoretical amount, which may be taken as 100 per cent.
The following theoretical depositions of copper, referred to various
units, will Ijc found convenient for reference:
One ampere-hour will deposit, theoretically, 1.1858 grm. of copper
from cupric solutions, and 2.3717 grm. from cuprous solutions.
0.8433 ampere-hours will deposit 1 grm. of copper from cupric solu-
tions, while 0.42164 ampere-hours will deposit 1 grm. from cuprous
solutions.
One ampere-hour will deposit 0.02614 lb. of copper from cupric solu-
tions and 0.05228 lb. from cuprous solutions.
382.50 ampere-hours will deposit 1 lb. of copper from cupric solutions;
191.25 ampere-hours will deposit a pound of copper from cuprous
solutions.
746 ampere-hours will deposit 1.9494.1b. of copper from cupric, and
3.8988 lb. from cuprous solutions.
1000 ampere-hours will deposit 2.6143 lb. of copper from cupric, and
5.229 from cuprous solutions.
These statements of deposition do not take into consideration the
voltage at which the current is delivered. It simply represents the
amount of copper irrespective of the voltage. While the current efficiency
of a process may be determined from the above theoretical quantities,
the energy efficiency can only be determined when the theoretical as well
as the practical voltage is known.
ELECTROLYTIC PROCESSES
293
The theoretical voltage is proportional to the heats of combination of
the compounds decomposed, and may be calculated from the molecular
heats of combination as a basis. For cupric sulphate it is 1.22 volts; for
cupric chloride it is 1.35 volts, and for cuprous chloride it is 1.42 volts.
Blount' found by actual experiment in his laboratory that the minimum
■pressure necessary for the deposition of copper from cupric sulphate,
using an insoluble anode, is 1.375 volts.
Knowing the theoretical deposition of copper by a definite amount
of electric current, and the theoretical voltage as determined from the
heats of combination, the theoretical power consumed in electrodeposi-
tion may readily be calculated, and this is taken as the standard of the
energy efficiency, or 100 per cent., to which all practical energy efficiencies
may be referred, since that represents the greatest possible amount of
copper that can be deposited with a given amount of electrical energy.
It represents the greatest amount of copper that can be deposited with a
definite current, and the lowest possible voltage at which it can be
deposited.
The following tabulated statement, gives in convenient form, the
theoretical energy required to deposit copper, with insoluble anodes,
from cupric sulphate, cupric chloride, and cuprous chloride solutions,
based on the molecular heats of combination.
Electrolyte
Cupric sulphate. . .
Cupric chloride. . .
Cuprous chloride..
Grm. per
h. p. -hour
726
656
1246
Grm. per
Lb. per
Lb. per
Lb. per
c. w.-hour
h. p. -hour
k. w.-hour
h. p. day 1
972
878
1670
1.5979
1.4440
2.7451
2.1429
1 . 9440
3.6824
38.35
34.66
65.88
Lb. per
■h. p. day
51.43
46.66
88 . 38
To determine the actual current and energy efficiencies of copper depo-
sition in an electrolytic process, the current is accurately measured both
as to quantity and pressure — amperes and volts — for a certain definite
time. The amperes multiplied by the volts, gives the watts; 746 watts is
equal to a horse-power, and 1000 watts to a kilowatt. The copper de-
posited is carefully collected, dried, weighed and assayed to determine
its purity. The weight of the pure copper deposited by the current and
the power consumed, compared with the theoretical, gives the current
and energy efficiencies.
For Example. — In making a certain test on the electrodeposition of
copper from a cuprous solution, a current of 400 amperes was used foi
12 hours (4800 ampere-hours), at 1.8 volts. The pure copper recovered
was 18.2 lb. The theoretical amount that should have been deposited
at the rate of 5.229 lb. per 1000 ampere-hours, is 25.1 lb. The test, there-
' "Practical Electro-Chemistry," p. 64, 1906.
294 HYDROMETALLURGY OF COPPER
fore, showed a current efficiency of 72.5 per cent. The copper deposited
in the test was at the rate of 2.15 lb. per k. w.-hour; the theoretical
deposition per k. w.-hour is 3.6824 lb.; hence the energy efficiency is
58.2 per cent. As compared with the deposition from cupric solutions,
however, this test would show a current efficiency of 145 per cent, and
an energy efficiency of 116.4 per cent.
In practical electrolytic work it is rarely that 90 per cent, is exceeded
for the current efficiency, and 50 per cent, for the energy efficiency. It
will usually be found profitable to sacrifice electrical efficiency for other
considerations.
In many processes, the secondary anode reactions tend to reduce the
theoretical voltage; this may vitally effect the energy efficiencies as above
given, because the secondary reactions must be taken into consideration
in figuring the eneTgy efficiency for any particular process, when such
reactions are involved. In the Hoepfner process, for example, the
theoretical voltage for the decomposition of cuprous chloride is 1.42
volts, but this is to some extent offset by the recombination of the released
chlorine at the anode, combining with the cuprous chloride to form
cupric chloride, so that the theoretical voltage is only 0.18 volt, instead
of 1.42 volts, or only about one-eighth. The theoretical deposition, tak-
ing into account the secondary reaction, would be 697.2 lb. of copper per
k. w.-day of 24 hours instead of 88.38 as given in the table. A similar
reduction of voltage is theoretically possible in the Siemens-Halse process
as also in processes using sulphur dioxide to combine with chlorine or
sulphion at the anode. It must be evident, however, that these reduc-
tions of voltage due to secondary reactions, are more apparent than real,
largely because an impracticable small current density must be employed
for their realization, or even an approach at realization, although Hoepf-
ner claimed a practical voltage of 0.8 volt in the operation of his process;
but Hoepfner used an exceedingly low current density.
Loss of Energy in Electrolytic Work: Joule's Law. — Much of the
energy consumed in electrolytic work, both in the conductors and electro-
lyte, appears as heat. Joule first discovered that the development of
heat was proportional to:
1. The resistance of the conductor;
2. The square of the current;
3. The time during which the current flows.
When heat is thus produced in the conductors and the electrolyte it
is a waste of power. The current density both in the electrolyte and in
the conductors may seriously affect the heating of the circuit, and con-
sequently the ultimate voltage at the dynamo. The conductors should,
therefore, be amply large to carry the current, and the electrical connec-
tions should be as few as possible, and all connections should be well
ELECTROLYTIC PROCESSES 295
made. At the Anaconda Copper Refinery it was found on careful in-
vestigation, that 20 per cent, of the loss of efficiency was in the conductor
connections.
A considerable percentage of the power used in electrodeposition of
the copper may be lost in the metallic conductors, and contacts, especially
if the conductors are of insufficient cross section and the connections poor.
For electrolytic refineries, which are presumably under at least good
average conditions, Addicks' gives a rough summary of the relative value
of the resistances in practice, as follows:
Metallic resistance, 15 per cent.
Electrolyte, including transfer, 60 per cent.
Contacts, 20 per cent.
Counter e. m. f., 5 per cent.
Contact resistances are met with at the joints in the main bars and at
the connections between the bars and the electrodes. The joints in the
main bars should be equal in conductivity to the bar itself. This stand-
ard can be easily attained if the bars are properly faced. Three hundred
to four hundred amperes per square inch of bearing area will give no
trouble.
The counter electromotive force in copper refining, due to the greater
concentration of the electrolyte at the anode than at the cathode is quite
small, usually about 0.02 volt. Contacts for copper deposition from
insoluble anodes are more likely to be source of loss of energy than in
electrolytic refining, due to the considerably higher voltage between the
electrodes. In copper refining the difference of potential between the
electrodes is from 0.2 to 0.4 volt, while in copper deposition with insolu-
ble anodes, it will vary in practice between 1.5 and 3 volts, depending
largely on the current density employed.
The Electrolyte. — The solvents usually employed for electrolytic
processes have either sulphuric or hydrochloric acid as the basis. The
solvent, to a large extent, determines the details of the process. In the
electro deposition, the solvent should be regenerated, as well as the cop-
per precipitated. In leaching copper ores, therefore, repeatedly with
the same solution all the soluble impurities in the ore are likely to be found
in the electrolyte, and this is true, but not to the same extent, if the solu-
tion is used only once and then wasted, instead of being used repeatedly
in the same cycle. The cycle will ordinarily consist of:
Solution,
Precipitation,
Regeneration;
' The Journal of the Franklin Institute, Dec, 1905.
296 HYDROMETALLURGY OF COPPER
and it is not liliely that any electrolytic process can achieve marked
success on any other basis. The regeneration may be direct; that is to
say, take place in the electrolyzer and under the action of the current ; or,
the anode products may be withdrawn from the electrolyzer and then in
some way combined with the solution.
Whatever method is adopted, the solution, in a cyclic process, is likely
to become charged with impurities and thus reduce the efficiency of the
deposition. Two alternatives present themselves; either to purify the
solution, or waste it at intervals. If the solution is wasted, a small
portion of it may be withdrawn at every cycle, and a corresponding a-
mount of fresh water added. This can be done quite readily if there is
more solvent regenerated in the process than is consumed by the ore.
\'arious methods have been suggested for purifying a foul electrolyte;
these differ somewhat as a sulphate or a chloride solution is used.
The injurious effect of an impure electrolyte is shown more particularly
in the electrolysis. The undesirable metals in the solution may be de-
posited with the copper if the solution has become impoverished, or else
useless energy may be expended in reduction and oxidation, or in depo-
sition and immediate solution, under the influence of the current.
A comparative measure of the energy required to deposit the various
metals, is obtained by taking the heat of combination of the metals with
oxygen, to form salts. Those metals are first deposited which have the
lowest heat of combination, other things being equal. The order of
deposition may be approximately stated as follows: gold, silver, copper,
antimony, bismuth, arsenic, lead, nickel, cobalt, cadmium, tin, iron, zinc,
manganese. The alkali metal salts, either the chlorides or sulphates, are
only decomposed with considerable difficulty, nevertheless, when high
current densities are employed they may frequently be the cause of con-
siderable loss of efficiency.
Gold and silver will be deposited with the copper, and the deposition
may take place either chemically or electrolytically. Neither gold nor
silver are likely to be in sulphate solutions; both may be in chloride
solutions.
The order of precipitation, as given, is dependent upon certain con-
ditions which must be observed. Principally among these conditions
are the strength of current, the nature of the electrolyte, the relative
proportion of the metals in solution, and if the solution is low in copper,
whether or not the electrolyte is agitated. If the current density, and
consequently the voltage, exceeds a certain strength, all the metals, or
several of them, may be deposited together. The more neutral the elec-
trolyte is, the more easily will the more electropositive metals be depos-
ited. The current is always striving to decompose the electrolyte into
metal or oxide and acid or the acid radical; while the liberated acid is
striving to redissolve the liberated metal or oxide. These two forces are
ELECTROLYTIC PROCESSES 297
always opposed to one another, and under varying conditions either may
gain the upper hand. The resolvent action of the acid, in cases where the
components of the electrolyte have a strong chemical affinity, may
overpower the action of a weak current.
In the deposition of copper this secondary reaction may not be of much
importance in a sulphate electrolyte, but it is quite noticeable in the pres-
ence of good circulation of the electrolyte, and more or less access of air
to the cathodes. With a chloride electrolyte this secondary reaction may
be quite pronounced under certain conditions, especially if the electrolyte
is violently agitated. With all of the easily dissolvable metals, it is this
secondary reaction which prevents the metal from making its appearance
at the cathode in metallic form, as the deposition, and solution due to
secondary reaction, are practically simultaneous, and the operation
simply results in a corresponding loss of efficiency.
It is evident that if the current density exceeds in amount that capable
of being supplied with copper ions, other metals in the electrolyte are
decomposed, and under aggravated conditions the electrolyte itself may
be decomposed. The relative amounts of the various metals in solution
also determines the efficiency of the process under certain conditions.
If the electrolyte contains only a small amount of copper and consideral)le
zinc — conditions quite likely to occur — the copper will be deposited to
the exclusion of the zinc with a correspondingly low current density, but
if that density is exceeded, the zinc will be electrolyzed, and if the elec-
trolyte is quite acid it will be redissolved at the cathode as rapidly as de-
posited, with the net result that the copper will be permanently deposited
at the cathode, but at a greatly increased expenditure of energy. If
the metal at the cathode is not readily soluble in the electroh'te, the
copper will be correspondingly impure.
AVhen copper ores are leached with any acid solution, or a solution
having an acid base, any or all of the metals may be in the solution, and
when electrolyzed may be influenced by the current as described. As the
electrolyte remains more acid, purer copper is likely to be deposited,
but when it becomes neutral, or only slightly acid, the undesirable metals
are deposited with the copper and are likely to remain on the cathode.
Such a condition, however, is not likely to occur since a neutral, or
only slightly acid solution, is not an energetic solvent of copper from
its ores.
The principal factors which determine the kind of ion to be deposited
on the cathode is the heat of formation of the different possible reactions;
that reaction which absorbs the least energy occurs, in general, the most
readily; but this statement is only approximately true. The concentra-
tion of the two electrolytes and the current density rrtust be taken into
consideration.
The heat of formation of the most important chlorides and sulphates
298 HYDROMETALLURGY OF COPPER
likely to be in the electrolyte from leaching copper ores, in dilute solutions,
is as follows:
Molecular Heats of Formation of Chlorides and Sulphates in Dilute Aqueous
Solutions
Chlorides
Sulphates
NaCl,
96,600 calories
Na^SO^,
328,500
CaCl^,
187,400 calories
MgSO,,
321,100
MgCl,,
187,100 calories
CaSO,,
321,800
MnCl^,
128,000 calories
Al,S30i„
879,700
AICI3,
238,100 calories
MnSO,,
263,200
ZnCl,,
113,000 calories
ZnSO„
248,000
FeCl,,
100,100 calories
FeSO,,
234,900
FeCl3,
127,850 calories
Fe,S30,„
• 650,500
PbCl^,
77,900 calories
NiSO„
228,700
CuCl,
35,400 calories
PbSO,,
216,000
CuClj,
62,500 calories
CuSO,,
197,500
ASCI3,
71,500 calories
H,SO„
210,200
SbCl3,
91,400 calories
Ag,SO„
162,600
BiCl,,
90,800 calories
AgCl,
29,000 calories
AUCI3,
27,200 calories
SO3,
141,000
HCl,
39,400 calories
S0„
77,600
NiCl,,
93,900 calories
H,0,
69,000
The decomposition values are independent of the solution, in the case
of bases and acids which on electrolytic decomposition evolve oxygen and
hydrogen at the electrode, and this is true for all acids excepting those
whose decomposition values are below the maximum. For these the
value rises with increasing dilution, and finally reaches the maximum.
This is very marked in the case of hydrochloric acid:
Decomposition, point
2n HCl, 1 . 26 volts
n HCl, 1.31 volts
l/2nHCl, 1.34 volts
l/6nHCl, 1.43 volts
l/16nHCl, 1.62 volts
l/32nHCl, 1.69 volts
With l/32n HCl a point is reached where chlorine is no longer given
off, but a large proportion of oxygen.
In leaching ores with sulphate solutions, the chlorides are not likely
to present, but if a chloride solution is used, the sulphates are quite sure
to be present, especially if the ore is roasted. The chlorides will be
decomposed before the sulphates.
If, for instance, the electrolyte contains CuCls ZnClj, and NajSO^,
either the Cu, Zn, or Na may be active in transporting the current
through the solution, but the heat of combination of cupric chloride is
62,500 calories; that of zinc chloride 113,000 calories, and that of sodium
ELECTROLYTIC PROCESSES 299
sulphate 328,500 calories, so that copper will be deposited to the exclusion
of zinc until the minimum voltage for the electrolysis of zinc chloride is
exceeded; then the zinc chloride may be decomposed if there are not
enough copper ions in the solution to transport the current through the
electrolyte, and finally even the sodium sulphate may be decomposed, if
the voltage is sufficiently high. It is evident, however, even with impure
electrolytes, copper may be deposited to the exclusion of other metals
in appreciable amounts.
Effect of Bismuth, Arsenic, and Antimony, in the Electrolyte. — None
of the metals highly injurious to copper, such as bismuth, arsenic and
antimony, are likely to affect the deposited copper. These metals are
largely eliminated in the cyclic operation of the process, even if contained
originally in the ore. If the ore is a sulphide and has to be roasted, these
elements are largely volatilized during the roasting. If they should
accumulate in the solution, they are easily removed and their removal
may be made to render an equivalent in acid by precipitating with
hydrogen sulphide. Bismuth is not deposited on the cathode even when
present in considerable quantities. Neither arsenic nor antimony are
deposited with the copper unless the solution approaches neutrality.
Iron in the Electrolyte. — Iron is most likely to be in the electrolyte,
as it always is associated with copper ores. In sulphate solutions the
iron is likely to accumulate indefinitely, to saturation, unless the solution
is purified at intervals. With chloride solutions, the iron chloride acts
more or less on the copper compounds in the ore to form copper chloride,
while the iron is precipitated as the insoluble ferric oxide. In the
electrolysis of a copper sulphate solution, containing considerable iron
sulphate, the iron passes from the ferrous to the ferric condition at the
anode, and is transformed back again to the ferrous condition at the
cathode, and thus using energy without rendering a useful equivalent.
Similarly the ferrous chloride may be transformed to ferric chloride and
back again, as the solution passes from anode to cathode, but in so doing
the ferrous chloride is likely to be oxidized to the ferric oxide, under the
oxidizing influence of the chlorine.
E. H. Larrison, aptly sums up the effect of iron sulphate in an electro-
lyte of copper sulphate as follows:*
"The process of electrolysis exercises in the main a reducing influence in an
electrolyte of copper sulphate containing iron sulphate. Further reduction, and
probably the retardation of the deposition of the copper, comes about through
some such reaction as the following :
Cu+Fe2(SOj3 = CuSOi + 2FeSO,.
That is, copper already deposited or on the point of being deposited is attacked
by the ferric sulphate and dissolved thereby, also reducing the ferric to ferrous
^E. andM.J., Sept. 7, 1907.
300 HYDROMETALLURGY OF COPPER
sulphate. When the process has gone on for some time the electrolyte becomes
so dilute with respect to ferric sulphate that the last of the copper is able to keep
its place upon the cathode. This accounts for the difficulty of, and the compara-
tive great time necessary for the removal of the last few miligrams of copper from
a solution high in iron. When using a stirrer and high current densities, a very
small reduction in the current density is quickly followed by the re-solution of
much of the deposited copper. It seems as if the copper is able to keep its place on
the cathode only with a certain electrical or reducing tension.
"Circulation of the electrolyte not only allows the use of higher current
densities, and thereby more rapid work, but it gives the current greater efficiency.
This is not only due to the fact that the solution is kept uniform and the copper
deposits more rapidly, but also the reducing effect of the iron is much greater
and consequently the dissolving power of the solution is decreased much sooner.
A very rapid method of reducing the iron before applying the current would save
time.
"The influence of iron on copper electrolysis may be summed up as follows:
" 1. High percentages of iron in the ferric condition materially retard copper
deposition. Ferrous salts have little or no effect.
"2. The effect of the electric current in the usual solution is to reduce ferric
salts to ferrous salts. This reduction must proceed to a certain point before all
the copper will remain on the cathode. This point is about the same whatever
the iron and copper percentages, but it varies some as the solution is circulated
or stationary.
".3. Rapid circulation of the solution by a stirrer permits the use of high
current densities without sponging, thereby making faster deposition. The
current also is more efficient for both depositing copper and reducing iron."
Purification of the Electrolyte. — Various methods have been suggested
for purifying the electrolyte, and these depend to a large extent on the
nature of the solvent. Many are based on the practice of the electrolytic
refineries.
The solution may be treated electrolytically by passing a current of
electricity through it with a high current density at the cathodes, and
employing either copper or lead anodes for sulphate solutions, and
carbon anodes for chloride solutions. By this means many of the
metallic impurities can be thrown down on the cathode, especially if the
solution is not too highly acid.
At Anaconda a process was used which consisted in passing the impute
electrolyte repeatedly through a layer of oxidized copper, so as to partially
precipitate the antimony and bismuth. By this treatment the solution
becomes nearly neutral, and saturated with copper, and was then oxidized
by passing air through it, so that the iron is partially precipitated as
ferric oxide.
Ulke states that one of the best methods for purifying old solutions is
that in which it is electrolyzed in special vats, the anodes being of lead,
and the cathodes of copper. A current density is employed sufficiently
ELECTROLYTIC PROCESSES 301
great to deposit the arsenic and antimony, but not strong enough to
deposit the iron. The solution thus freed from the arsenic and antimony
is returned to the copper depositing vats to be used in the ordinary way,
and this is repeated until the bath contains so much iron that it is
necessary to remove it by crystallizing out the ferrous sulphate.
Ottaker Hofmann gives the following method of purifying copper
sulphate solutions having in addition to the cupric sulphate, salts of iron,
arsenic, antimony, bismuth, cobalt, nickel, etc'
"The crude neutral cupric sulphate solution is forced into towers, about 6 ft.
in diameter and 16 ft. high. These towers are lined with lead. A lead pipe,
connecting with an air compressor, is led into the tower through the funnel-
shaped bottom. Near the bottom is a lead steam coil for heating the solution.
"When the tower is filled with impure solution, obtained by treating roasted
copper matte with dilute sulphuric acid, steam is allowed to enter the coil and
heat the solution and at the same time air is forced through the pipe at the
bottom. The ascending air imparts a violent boiling motion to the liquor.
Part of the iron is precipitated as basic salt, by the action of the air. More than
half the iron was never precipitated although the treatment was extended many
hours. When the solution is hot (75 to 80° C.) roasted matte is added. The
violent boiling motion of the solution keeps the matter in suspension, and after
3 or 4 hours the solution will be entirely free from iron, arsenic, antimony, bis-
muth, etc. To observe and regulate the progress of the operation the solution
is tested from time to time for iron by taking samples through a small coc'k
inserted in the sides of the tower. It is not necessary to test the solution for
other impurities, because the iron predominates, and by the time all of it has been
precipitated, no trace of the other impurities will be found."
"The chemical reaction of this process may be expressed as follows:
2FeS0i + 0 + CuO =Fe203 + 2CuS0<.
This shows that the cupric oxide, which together with air, is used to precipitate
the iron, combines with the sulphuric acid of the ferrous sulphate, and goes in
solution as cupric sulphate; a decided advantage, as the precipitant is converted
into cupric sulphate, and thus enriches the copper solution."
At the Kalakent Copper Works, Russia,^ when the impurities in the elec-
trolyte had accumulated to such an extent as to endanger the quality of
the electrolytic copper, the foul solutions were withdrawn from the in-
dividual or group of tanks and regenerated, this regeneration being ac-
complished with considerable difficulty at first, but finally it was done
advantageously as follows:
"The foul solutions were heated and passed over dead roasted matte fines,
heaped in loose-bottomed trays arranged in series in upright rows. By this
method all the sulphuric acid in solution, down to about 2 grm. per 100 c.c, was
neutralized and combined with copper and iron. The solution was then allowed
'Mineral Industry, Vol. VIII, p. 192.
^ Titus Ulke, "Modern Electrolytic Copper Refining," p. 145.
302 HYDROMETALLURGY OF COPPER
to trickle through heaps of roasted low-grade copper ores, which resulted in the
almost complete neutralization of its acid contents. It was then diluted with
washwaters down to 10 to 12° B. and heated, to free the solution from. iron,
arsenic, tin, and bismuth, in lead pans, which were provided with an appliance
for injecting compressed air into the solution. Anode scrap was suspended in
the pans so as to neutralize any remaining portion of acid, and to take up any-
new acid set free through the separation of iron hydrates, and thereby form
copper sulphate. By blowing compressed air into the solution heated to about
50° C, the copper in the anode residue or scrap was quickly dissolved, and the
separation of the iron, arsenic, antimony, tin, and bismuth, which occurred when
the solution had been nearly or completely neutrahzed, accomplished. The
solution was then concentrated up to 14 to 15° B. and clarified in special reser-
voirs. It now contained 3.5 to 4 grm. of copper per 100 c.c, and of impurities
only 0..5 to 1 grm. iron, besides traces of zinc, nickel, and cobalt, and was there-
fore pure enough for reuse as electrolyte. The excess of purified electrolyte
which gradually accumulated with this method of regeneration was eventually
withdrawn from the circulating system and worked up into blue stone. The
electrolytic copper possessed a purity of at least 99.9 per cent, by analysis, and
averaged 99.93 per cent, copper."
If cupric sulphate is crystallized out of impure solutions and redis-
solved in water, it may be electrolyzed to deposit the copper and liberate
the combined sulphuric acid. The acid solution may then be again
applied to the ore, with all or 'most all of the impurities eliminated.
Hoepfner' proposed precipitating the impurities from a chloride
solution with oxychloride of copper. In doing this, some of the cuprous
chloride in the solution is converted into the oxychloride of copper by
blowing air or oxygen into the electrolyte:
CU2Cl2 + 0 = CU2Cl20.
"This is most conveniently done by cooling the solution, or a part thereof to
precipitate the cuprous chloride, which is then converted into the oxychloride by
contact with air. This solid precipitate is a most efficient reagent for iron and
to enrich solutions poor in copper so as to make it more suitable for electrolysis.
The reaction may be expressed as :
3CU2CI.O + 2FeCl2 = 4CUCI2 + Cu^Clj -l-FejOs.
Lime, the caustic alkalis or alkaline earths, or their carbonates, may be used as
precipitants ; or the oxides or carbonates of metals, as, for instance, of copper,
may be employed in the separation of the undesirable metals from the electrolyte,
the metals being precipitated according to the precipitant used in the form of
oxide or in the form of arsenate and antimonate of iron or copper, the arsenic
and antimony being present in the solution in the form of arsenious and antimon-
ious acids, or arsenic and antimonic acids which are converted by the precipitant
into insoluble arsenic and antimonic salts, while if an oxide or carbonate of
copper is used the arsenic and antimony are converted into insoluble arsenate or
lU. S. Patents No. 507,130, Oct. 24, 1893, and No. 704,639, July 15, 1902.
ELECTROLYTIC PROCESSES 303
arscnite of copper and the corresponding salts of antimony. Inasmuch as the
electrolyte contains copper it may readily happen that the salts last referred to
will be formed without the use of a copper salt when lime or an alkali is used as
the precipitant. In either case, but a comparatively small proportion of the
copper goes over with the precipitant, which quantity is in no ease greater than
t!ie quantity of the arsenic and antimonic salts precipitated."
Greenawalt purposes purifying chloride solutions by electrolyzing
common salt to form caustic soda and chlorine; the caustic soda is used
as a precipitant of the impurities, and the chlorine to produce acid.
In purifying the electrolyte, a certain portion is withdrawn from the
cycle of operation, and the caustic soda applied to it;
RCl2 + 2NaOH = 2NaCl + R(OH),
in which R may represent any or all the base metals.
In this way the base metals are precipitated and the sodium chloride
regenerated. The chlorine liberated by electrolysis of the salt is combined
with sulphur dioxide, in the presence of water, or the solution, to form
acid;
2Cl + S02+2H20=2HCl + H2SOi.
In this way all the undesirable impurities are eliminated from the solu-
tion, and a corresponding amount of acid solution regenerated.
In the purification of the electrolyte it is generally assumed that a
pure deposit of copper is necessary. If the impurities in the electrolyte
do not materially interfere with the efficiency of the process it would
be better to work with impure solutions even though an impure copper
is deposited. There is no reason why the purity of the deposited copper
should be a matter for serious consideration.
The relative efficiency of the process, due to impurities in the electro-
lyte is more serious than the relative purity of the copper. Some ele-
ments cause serious loss of efficiency if present in certain combinations,
while in other combinations the loss cannot be considered of much
consequence. Ferric salts, for example in the electrolyte, may cause a
serious loss of efficiency, while ferrous salts are comparatively harmless.
And this applies generally to the elements which have different valencies
for different combinations, and which are capable of oxidation and
reduction in the electrolyte due to electrolysis.
Depolarizers. — Various depolarizers have been suggested in connection
with the electrodeposition of copper; among the most important is sul-
phur dioxide. The use of reducing gases, and particularly sulphur
dioxide, for the depolarization of insoluble anodes in the deposition of
copper, was described as long ago' as 1878 by Cobly. Later Luckow
used sulphur dioxide in the electrolysis of zinc solutions, and its applica-
tion in more recent years has been quite common and well understood.
304 HYDROMETALLURGY OF COPPER
If cupric sulphate is electrolyzed:
CuSO 4 + electric current = Cu + SO 4.
The resulting products are copper at the cathode and sulphion at the
anode. The liberation of the SO 4 at the anode immediately results in a
secondary reaction,
S04 + H20 = H2S04 + 0
whereby water combines with the sulphion to form sulphuric acid, while
oxygen is released. If, however, sulphur dioxide is in the anode solution,
the SO4 and SOj combine, with water, to form two atoms of sulphuric
acid,
S04 + S02 + 2HjO = 2H2S04
and, theoretically, no oxygen is released. This reaction develops a
certain electromotive force at the anode, in the same direction as the
current, and thus reduces the necessary voltage in the decomposition of
the copper.
If sulphur dioxide is used in connection with chloride solutions, the
electric current decomposes the copper chloride into copper and chlorine,
2CuCl + electric current = 2Cu + 2C1.
The chlorine combining with the water,
2C1 + H20 = 2HC1 + 0 + 10,800 calories,
to produce hydrochloric acid and release oxygen, and thus chlorine may
become an oxidizing agent. This reaction takes place very slowly, but
the oxygen can immediately exert a further chemical action in the
presence of sulphur dioxide, when a rapid decomposition of water takes
place,
2Cl + SO2 + 2H2O = H2SO4 + 2HCl + 74,400 calories
in which case the chlorine released at the anode is by secondary reaction,
converted into sulphuric and hydrochloric acids with the development of
74,400 calories, which tends to reduce the voltage in the operation.
Ordinarily, chlorine will be given off at the anode in the absence of sulphur
dioxide, because the oxidizing action of chlorine is exceedingly slow; in
the presence of sulphur dioxide the amount of chlorine released will
depend upon the current density. If the current density is sufficiently
low to permit of the released chlorine coming in contact with the sulphur
dioxide before either can escape, no chlorine will appear, as such.
Other reactions have been suggested in the decomposition of copper
compounds, acting on the principle of depolarizers. In Body's process,
ferrous chloride is converted into ferric chloride by the chlorine released
ELECTROLYTIC PROCESSES 305
at the anode; in the Siemens-Halske process ferrous sulphate is recon-
verted to the ferric sulphate in the deposition of the copper from a solu-
tion of cupric and ferrous sulphates; and in the Hoepfner process, in
which the copper is deposited from a cuprous chloride solution and a
certain amount of the cuprous chloride converted back to the cupric
chloride by the chlorine released at the anode.
In all of these reactions, the electromotive force developed works in
the direction of the current and thus reduces the necessary voltage. The
approximation of the theoretical to that realized in practice, depends on
the conditions of operation, principally among these conditions is the
current density. If the current density is so low as to permit all of the
anode gases to combine with the depolarizer, the theoretical efficiency
may be quite closely realized.
Rapid Deposition of Copper. — To rapidly deposit a metal from solution
by the electric current, it is necessary that the metal ions be present in
sufficient number at the cathode. If a comparatively high cathode density
is used, the danger is that the electrolyte in proximity to the cathode
becomes poor in the ions deposited, and other processes start, especially
the development of hydrogen. To counter this tendency, it is necessary
to artificially bring fresh electrolyte to the cathode surface, as for instance,
by stirring, or rotating the cathode.
In an ordinary electrolyzer, if the electrolyte is not agitated or cathode
rotated, as soon as the current is switched on, copper is deposited on the
cathode and a thin layer of electrolyte touching the cathode becomes in
consequence impoverished in copper. Before more deposition can take
place this thin layer of exhausted electrolyte has to be removed. If left
to itself, removal and changing will take place quite slowly, nevertheless
it will be fast enough to supply sufficient copper ions to the cathode if
the current density is correspondingly small. No loss of efficiency would
therefore result. If, however, the current be greater than the correspond-
ing removal of the impoverished electrolyte from the cathode, the current
will sieze upon the next available material in the electrolyte, and the
result will be a corresponding loss of efficiency and impure and rough
copper deposited. To prevent this action from taking place it is neces-
sary to remove the film of exhausted electrolyte as rapidly as it is formed,
either by violent circulation of the electrolyte, friction, or rotating cath-
ode. Consequently the process which most effectually removes the
impoverished electrolyte from the cathode, and which at the same time
will give the most uniform distribution of current and the most uniform
strength of electrolyte over the whole surface of the cathode, should be
the one to produce the highest efficiency, the best deposits of copper, and
allow of the highest current density.
Various methods are employed to accomplish these results. One
generally used, is to place the electrolyzers in cascade series, that is to
20
306
HYDROMETALLURGY OF COPPER
say, in which there is a slight difference in level between one electrolyzer
and the one next to it, so that the electrolyte pumped into a tank above
the highest electrolyzer is fed into it, and gradually flows through the
entire series and finally issues from the lowest in the series and is again
pumped back to the tank, or to the leaching vats. The rapidity of the
circulation is governed by the difference in level between any two tanks
and the slant of the tank itself. If the tank has not sufficient slant, the
electrolyte will overflow if the circulation is at all rapid.
Fig. 51. — Coffin revolving cathode apparatus.
Another method of accomplishing the same purpose is to force air into
the tank and electrolyte while the electrolysis is in progress. Care
should be used if the electrolyte is agitated with air, that the agitation is
quite uniform, in all parts of the electrolyzer. Sometimes stirrers,
working reciprocally between the electrodes are used to better advantage
than air, and the agitation is likely to be more uniform. The electrolyte,
impinging against the cathodes in jets has given good results, but is now
ELECTROLYTIC PROCESSES
307
03
a
o
fin
308
HYDROMETALLURGY OF COPPER
Fig. 553. — Hoepf ner revol ving cathode
apparatus Section.
nowhere used. A rotating cathode gives excellent results, but a very
different type of cell is required than when any of the other methods are
employed. If diaphrams must be used, a rotating cathode may offer
some constructional difficulties, which, however, need not be insur-
mountable. Fig. 54 shows one form of revolving cathode apparatus
devised by Coffin' in which A is the electrolyzer, lined, preferably, with
sheet lead, fi is a post or pedestal secured in the center of the tank and
surmounted by a metallic head D, or socket, E is the cathode cylinder,
and 0 the anodes. Figs. 55 and 55i represent another form of apparatus
devised by Hoepfner' in which d indi-
cates the carbon anodes, D the dia-
phragms, and a the cathodes, revolving
on a horizontal shaft c. In addition
to the revolving cathode, brushes are
sometimes used and adjusted so as to
cause friction on the deposited cathode
material. Cowper-Coles found that
copper possessing the advantages of
hard rolled copper of high tensile
strength and free from porosity can be
deposited to any thickness desired at '
a rapid rate by revolving the cathode
at a peripheral speed of from 1500 to 2000 ft. per minute when employ-
ing current densities of 200 amperes per square foot of cathode surface
and an electrolyte containing 12.5 per cent, of copper sulphate and 13
per cent, of sulphuric acid at a temperature of 40° C
In experimental work in the laboratory of applied chemistry at the
University of Wisconsin'' it was observed that the critical current density,
that is to say the current density at which powdery deposit occurs, is
approximately proportional to the speed of rotation, or better, to the
linear feet per minute, as the cathode speed depends upon the speed of
rotation and the diameter of the revolving cathode.
Betts^ uses anodes preferably in the form of rods, and gives to them
a reciprocating motion in a direction perpendicular to their length,
whereby the layer of electrolyte touching the anode is rapidly changed.
In electrolyzing a solution of ferrous and cupric sulphates, with a dia-
phragm, depositing copper on a cathode, and converting ferrous to ferric
sulphate at the anode, and with a circulation of electrolyte that was pre-
viously considered amply sufficient, the electromotive force required to
work the cell was considerably reduced, the evolution of gas at the
' U. S. Pat. 415,024, Nov. 12, 1889.
= U. S. Pat. 598,180, Feb. 1, 1898.
' U. S. Patent 895,163, Aug. 4, 1908.
* J. G. Zimmerman, N. Y. Meeting Electrochemical Society, 1904.
= U. S. Pat. 803,543 Nov. 7, 1905.
ELECTROLYTIC PROCESSES 309
anode was entirely stopped, and the current efficiency raised from about
50 per cent, to 100 per cent, on giving the anodes a reciprocating motion
of about one hundred complete cycles per minute, with an amplitude of
about 1 in.
ELECTROLYTIC SULPHATE PROCESSES
The electrolytic sulphate processes may be divided into two general
classes, based on the solvent, as,
Sulphuric Acid,
Ferric Sulphate,
but it is evident that neither of these processes can be carried out to the
exclusion of the other. All solutions resulting from leaching roasted or
oxidized ores will have more or less iron sulphate either in the ferric or
ferrous condition, and the utility of a neutral solution of ferric sulphate,
as compared with an acid solution, is questionable. Sulphuric acid is an
energetic solvent of copper from roasted or oxidized ores, and is to be
generally preferred. On certain chalcocite ores ferric sulphate has given
good results without roasting.
In a copper sulphate solution containing ferrous sulphate, both
• sulphuric acid and ferric sulphate may be regenerated during electrolysis,
and to the formation of ferric sulphate much of the difficulties in the elec-
trolysis of impure copper sulphate solutions may be attributed. If a solu-
tion of copper sulphate, as for example that derived from leaching copper
ores, is electrolyzed without a diaphragm, then in addition to the reaction;
CuSO 4 + H^O + Electric C '.rrent = Cu + H^SO ^ + 0,
there may also take place,
CuSO 4 + 2FeS0 4 + Electric Current = Cu + Fe2 (SO 4) 3,
3FeS0 4 + Electric Current =Fe +re2(S0 4) 3.
The ferric sulphate, finding its way back to the cathode, and under the
influence of the current, gives rise to reversible reactions, thus:
Fe2(S04)3-^Cu =2FeS04 + CuS04,
Fe2(S04)3+Fe =3FeS04,
thus nullifying the previous reactions, and resulting in a loss of effi-
ciency. This loss will depend largely on the amount of iron in the solu-
tion. If the iron is excessive, the copper may be dissolved as rapidly as
precipitated, and the sum total of the energy expended will be nil, so
far as any useful effect is concerned.
If a suitable diaphragm is interposed between the electrodes, then
both sulphuric acid and ferric sulphate are regenerated, but the deleterious
reactions at the cathode are avoided, and the current efficiency may
very closely approximate the theoretical, but the energy efficiency will
be governed somewhat by the resistance of the diaphragm. The relative
310 HYDROMETALLURGY OF COPPER
amounts of sulphuric acid and ferric sulphate regenerated at the anode,
will depend mostly on the relative proportions in the solution. If the
ferrous sulphate is highly concentrated then ferric sulphate will be
largely regenerated, and the process becomes one in which ferric sulphate
is used as the lixiviant and in which the resulting ferrous sulphate is re-
generated to the ferric sulphate.
Sulphuric Acid Process. — The simplest form of the electrolytic sul-
phate process is when sulphuric acid is used as the initial solvent to
extract the copper from its ores and the sulphate solution thus obtained
electrolyzed to deposit the copper and again liberate the combined acid.
Copper in its sulphide combinations is not soluble in sulphuric acid; it is
therefore necessary to roast the ore if a sulphide before the solution of
the copper can be effected. After roasting, the copper will usually be in
the form of oxide or sulphate, depending upon the nature of the roast.
In any event, the sulphuric acid may be regarded as acting on the oxide
of copper:
CuO + H^SO, = CuSO, + H^O.
The copper sulphate so produced, in addition to that soluble in the
ore, is then electrolyzed:
CuSO 4 + Electric current =Cu + S04.
SO, + 11,0 = 11,80, + 0.
The current decomposes the copper sulphate, depositing the copper on
the cathode, and the acid radical is liberated at the anode. The acid
radical, combining with water, is converted into an equivalent amount
of sulphuric acid to that from which the copper was deposited, and oxygen
makes its appearance at the anode as a result of the secondary reaction.
For every pound of copper deposited 1 . 54 lb. of acid is regenerated.
If there were nothing to consider but the solution of the copper from
the ore with sulphuric acid, and the electrolytic decomposition of the
resulting copper sulphate solution into metallic copper and the acid
radical, this would be a perfect process, requiring nothing more than the
expenditure of a certain amount of energy to carry it on indefinitely.
The difficulties, however, in practically carrying out this simple process
are considerable.
Much of the acid consumed in treating copper ores reacts with the
base elements, and hence there is not sufficient acid regenerated to treat
the next charge of ore; so that the deficiency has to be made up in some
other way. The difficulty, therefore, of the solution of the, copper is
only partially solved, nevertheless it is a great step in advance of iron
precipitation, where all the acid is irrecoverably lost. If the ore is a
sulphide, and has to be roasted, some of this difficulty may be overcome
by roasting as much of the copper as possible to sulphate rather than to
oxide. The sulphate of copper thus formed is the same as that produced
ELECTROLYTIC PROCESSES 311
by the action of sulphuric acid on copper oxide, so that while less acid is
used in dissolving the copper from the ore more acid is regenerated in
the electrolysis by the amount of copper soluble in the ore as sulphate.
The deficiency may also be supplied by installing a small acid works
in coimection with the other metallurgical operations, or it may be pur-
chased from acid manufacturers, but both of these methods add con-
siderable to the expense, and hence do not ingeniously solve the difficulty.
Further, the impurities in the ore, which cause an irrecoverable loss
of acid, also contaminate the electrolyte, and cause difficulty in the
electrolysis.
No really satisfactory anode for sulphate solutions has yet been
discovered. Lead is ordinarily used for the insoluble anode in depositing
copper from sulphate solutions, and while lead makes the best anode for
sulphate solutions, it is far from being satisfactory. In depositing
copper from sulphate solutions, oxygen is released at the anode; this
oxygen is not entirely harmless, but attacks the lead and converts it
into the peroxide, PbOj. The peroxide, in addition to destroying the
anode, offers considerable resistance to the electric current, and thus
necessitates an excessive consumption of power.
In a very careful test made by Thomas P. Hughes, in Denver, to
determine the peroxidation of antimonial lead anodes in electrolyzing
copper sulphate solution produced by leaching Arizona carbonate ore,
the following results were recorded:
Duration of test, 75 . 78 hours.
Copper deposited, 6 . 56 lb.
Average current, 30.0 amperes.
Anode area, 3 . 5 sq. ft.
Cathode area, 3.5 sq. ft.
Current density, 8 . 5 amperes per square foot.
Average voltage, 2 . 0 volts.
Watts, 60.0
Kilowatt-hours, 4 . 547
Copper per k. w.-hour, 1 . 4 lb.
There were 15 sheet copper cathodes, and 14 antimonial lead anodes,
each 6X6 in. The anodes increased 14 oz. in weight. There was a
black coating of peroxide of lead on the anodes, which could easily be
scraped off. The increased weight of the anodes was evidently due to
the oxygen combined with the lead to form the peroxide. As 86.6 per
cent, of peroxide of lead is lead, and 13.4 per cent, oxygen, it follows that
5«6 lb. of lead was peroxidized in depositing 6.56 lb. of copper; or for
every pound of copper deposited, 0.85 lb. of lead was peroxidized. The
peroxide of lead, unless closely watched, is likely to drop to the bottom
of the electrolyzer and short circuit the current. In this experiment
the current was not operating continuously; it was turned on in the
morning and shut down in the evening.
312 HYDROMETALLURGY OF COPPER
In experiments made by Greenawalt to get more information on the
rate of oxidation of lead anodes, one test of 1000 ampere-hours, continu-
ous run, with a current density of 20 amperes per square foot resulted
in depositing 37 oz. of copper and in producing 14 oz. of lead peroxide.
Another run of 500 ampere-hours was made in which 20 oz. of copper
was deposited and 3 oz. of peroxide recovered. While still another run
of 1458 ampere-hours, with a higher current density produced 60 oz. of
copper and only 3 1/4 oz. of peroxide of lead.
There is no difficulty in collecting the peroxide and again reducing it to
metallic lead to be reused for anodes, but it necessitates an extra expense
in the operation of a plant which, however, is largely compensated for by
the fact that with the exception of a small loss in reduction there is no
cost for material for replacements.
The suggestion naturally occurs that if the peroxide of lead is the
ultimate product of the lead anode, why not make an anode of peroxide,
and thus overcome entirely the difficulties due to oxidation? This has
been tried repeatedly, and Hughes tried it in various ways, but the
resistance to the electric current, even with a good conducting skeleton,
was so high as to put it beyond further consideration.
If sulphur dioxide is used as a depolarizer in the electrolysis of copper
sulphate solutions, then, theoretically, twice the amount of acid com-
bined with the copper is regenerated:
S04 + S02-h2H20=2H2SO, + 21,320 calories,
but it is difficult to carry this out in practice, for the reason that the
sulphur dioxide and the acid radical, at the moment of liberation, cannot
be Ijrought into sufficiently intimate contact to make it effective, espe-
cially if a reasonably large current density is used.
If sulphur dioxide is used as a depolarizer, some of the lead in the
anode will be converted into the sulphate, but the action of the forma-
tion of sulphate of lead is very much slower than in the ordinary elec-
trolysis where the lead is converted into the peroxide. The lead sulphate,
however, is more difficult to reconvert back into metallic lead than
the oxide.
The energy required to decompose an aqueous solution of copper
sulphate is theoretically, 1.22 volts. In practice it will usually vary
between 1.5 and 3 volts, depending principally upon the current density
used. The theoretical output of copper is therefore, 38.35 lb. per h. p.-
day, or 51.43 lb. per k. w.-day, of 24 hours. If a depolarizer is used, such
for example as sulphur dioxide, then the acid radical combining with
the sulphur dioxide to form sulphuric acid, will develop an electromotive
force working with the current, and thus reduce the theoretical voltage,
but the amount of this reduction is limited largely by the current density
employed, and the completeness with which the two substances are
ELECTROLYTIC PROCESSES 313
brought in contact witli one another at the moment the acid radical is
released at the anode. Tossizza' ascertained by experiment that the
transformation of the sulphur dioxide into sulphuric acid at the anode
gives rise to an electromotive force which diminishes the necessary volt-
age and lowers it to 0.2 volt.
While the theoretical voltage gives much desirable information, it is
always best, indeed necessary, to get the voltage by direct experiment
because there are so many factors which occur in practice that do not
occur, and cannot be taken into account in the theoretical determina-
tions. The practical voltage, once determined, is always constant, and
is independent of the magnitude of the operation. This is a factor which
can be just about as accurately determined on a small scale, on a labo-
ratory basis, as in a large working plant. Similarly, the amount of copper
deposited from a certain electrolyte, on a small scale under the conditions
obtaining in practice, will always be constant, no matter what the scale
of operations may be. For this reason, the factor of efficiency, and of
voltage, can be quite accurately determined beforehand for any par-
ticular process and made the basis of calculations of a large plant of any
sized unit.
For sulphate solutions, in practice, with a current density of 5 to 10
amperes per square foot, the voltage will usually be found to vary from
1.5 to 2.5 volts. The practical energy efficiency can only be determined
by weighing the copper deposited in a certain definite time, under an
observed voltage and current. It will be found, in making such tests,
that frequently the results will be far from the theoretical, and that the
temperature, purity of the electrolyte, and current density have much
to do with the efficiency. The best that can be done in practice is to
approach the theoretical efficiency although it does not follow that
the most efficient process in the electrolysis is the most economical in
operation.
Electrolytic Extraction of Copper From Ore at Medzianka, Poland. —
At Medzianka, a copper mining locality in Russian Poland, about 50
miles from Cacrow, and 140 miles east of Breslau, explorations by the
Laszczynski brothers, led to the discovery of ore bearing limestone,
about 11/4 miles long and 150 ft. wide, containing copper ore inter-
spersed in strips 1/2 in. thick, with calc spar and some quartz. The
mineral is almost entirely copper glance, but mixed with it is a little
azurite and malachite.
The produce of the mine has been divided into ore with 50 per cent, of
copper which is separated underground, and mixed ore with 16 to 20
per cent, copper containing calcite and pieces of limestone, which is
improved by hand picking, at the surface. The ore as brought from the
'U. S. Pat. 710,346, Sept. 30, 1902.
314 HYDROMETALLURGY OF COPPER
mine is crushed in rolls, mixed with 5 per cent, of damp brick earth and
moulded into blocks, which when dried by the waste heat of the furnace,
are subjected to a partial roasting in a kiln fired from the outside,
with free access of air, which converts the copper into sulphate and
oxide.
The roasted blocks are then crushed fine and leached in lead lined
wooden tanks, with the electrolyzed solution from the electrolytic cells
containing about 5 per cent, of free sulphuric acid. A liquor containing
about 5 per cent, of copper and 1 per cent, free sulphuric acid is obtained.
This solution is passed through a filter press, to thoroughly clarify it,
and then electrolyzed in tanks of about 35 cu. ft. capacity. Insoluble
anodes of lead plates enclosed in cloth bags, and thin copper cathodes
are used. A current of 1000 amperes at 2.5 volts, corresponding to a
current density of about one ampere per square decimeter of cathode
surface (10 amperes per square foot) is used, producing metallic copper
free from sulphuric acid or oxygen. The deposited copper, about 1.1
grm. per ampere-hour, is nearly equal to the theoretical amount. The
power consumed per kilogram of copper is 2.28 k. w. -hours or 3 1/2 h. p.
(1.3 k. w. -hours or 1.6 h. p. -hours per pound of copper). Between the
anodes and cathodes there are wooden stirrers, which agitate the solution
during the entire electrolytic process. The liquor is exhausted in from
36 to 40 hours, and then again containing about 1. per cent, copper, and
from 1 to 7 per cent, free acid, is applied to a fresh lot of ore. The
cathodes remain in the bath for about a month, when the deposit, from
1 to 1 1/4 in. thick, is removed and sold. It is of greater purity than the
ordinary electrolytically refined copper. The four baths used in the
process are served by a Siemens dynamo of 1000 amperes at 12 volts.
The daily output of copper is from 225 to 500 lb. The entire process is
supervised by one man in the mill without any other trained assistance.
A vital point in the success of the process is in the employment of
closely fitted bags or envelopes of thick cotton duck for the lead anodes.
The bags, soaked with sulphuric acid, exclude the iron salts and thus
overcome much of the difficulty from that source. Its function is about
the same as a diaphragm. These cotton bags are renewed about once a
year.
The Laszcynski process, used at Medzianka, Russia, is described by
the inventor as follows:^
"If a solution of sulphate of iron, FeSO^, is electrolyzed, at the cathode the
bivalent ferro-ion is metallically deposited at the same time the said cathion
comes in contact with the insoluble anode, there being no diaphragm, and is there
oxidized into trivalent ferri-ion. The latter, however, before it is deposited as
metallic iron has to be reduced at the cathode to ferro-ion. In this manner
there is soon set up a state of equilibrium in which the same quantity of ferro-
' U. S. Pat. No. 757,817, April 1, 19, 1904.
ELECTROLYTIC PROCESSES
3 If)
ions arc reduced at the cathode as are produced at the anode. The chemical
action of the current, therefore, is nil."
"In the present process the detrimental side action is avoided by wrapping
around the insoluble anode a cover or envelope of porous fabric. The wrapping
being permeable, there is before closing the circuit no difference between the
chemical composition of the anode and cathode bath. As soon as the current
is turned on, however, a layer of pure sulphuric acid will form around the
anode, since there the SO^ ions are discharged. Consequently new sulphuric
acid is generated which can only drain off into the close-fitting envelope, dis-
placing in this manner the solution of iron sulphate, so that in a sliort time
a second process takes place similar to the one described. Since the ferro-ions
and the ferri-ions are cathions, they travel at the closing of the circuit from
the anode to cathode. The envelope aro>ind the anode forms a layer of quiet
liquid, no matter if the electrolyte is in circulation, so that the traveling of the
cathions can take place without being dis-
turbed. The result of the two actions is
that no ferro-ions can be oxidized, since
none come in contact with the anode."
"Referring to the drawing, Fig. 56, in
the application of the process to cojjper
ores, a is the electrolytic cell, preferably
made of wood and tightened with asphaltum
or the like, b represents the copper sheet
cathodes, and c represents the anode which
consists of refined lead and is provided with an envelope of thick cotton stuff;
for example, fustian."
All copper ores without exception contain iron, which when treated with
sulphuric acid dissolves, together with the copper. By the electrolysis the iron
is oxidized at the anode to ferric sulphate, which salt dissolves the copper
deposited on the cathode equal to the action of dilute nitric acid :
Cu-l-Fe,(SO,)3 = CuSO, + 2FeSO,.
In this way the amount of copper deposited is not only reduced to one-half or
even less, but also a brittle and inferior metal is obtained.
"The present invention now prevents the oxidation of the iron salts and
makes possible the direct electrolysis of copper baths containing iron, even if
they contain twice as much iron as copper, with a useful effect differing but
slightly from the theoretical, because the iron remaining in the state of
wholly inoffensive ferrous sulphate, FeSO^, there is no corroding action of
any kind."
Fig. 56.
The Laszcynski process for electrolytically obtaining metals, especially
copper and zinc, out of the ores by means of insoluble anodes, consists in
tightly wrapping the insoluble anode in a porous and perfectly permeable
envelope, of fabric or other material, the thickness of which is in inverse
proportion to the applied density of current, for the purpose of preventing
anodic oxidation of the cathions.
316 HYDROMETALLURGY OF COPPER
Plant of the Intercolonial Copper Co., N. S. Canada. — The plant of the
Intercolonial Copper Co., in N. S. Canada, was designed by Henry Car-
michael, and according to Johnson' this plant for some time produced
one ton of electrolytic copper daily, which was sold to brass founders
as equal to the best brands of electrolytic copper.
The copper in the ore varied from 2 to 4 per cent. The very small
values in the precious metals were lost. The ore was crushed to 20
mesh, which was fed by screw conveyors into a battery of 15 revolving
roasters. These consisted of long tubes of cast iron passing through a
firebrick muffle, heated by a flame from the fiire-box. The first part of
the roaster was a brick lined drum. All revolved, and the ore was first
partially desulphurized in brick lined drums. It was then passed to the
iron tube, where it was dead roasted. The lime was sulphated and the
iron was changed to the ferric state. The roasted ore from the revolv-
ing drums was carried by a chain conveyor directly to lead lined vats
of 20-tons capacity. The roasters had a capacity of 2 to 3 tons each
per day.
The hot ore falling from the conveyors dropped into a 5 per cent, solu-
tion of sulphuric acid. The solution of the copper took place rapidly.
The solution from the leached ore contained 2.5 per cent, copper and
considerable free acid, which was drawn into a storage vat. The tailings
assayed less than 0.10 per cent, copper.
The solution, as drawn from the ore and pumped to a storage tank,
was impregnated with sulphur dioxide gas, made by burning brimstone
in an iron pot under blast. The copper solution in the storage tank,
inipregnated with sulphur dioxide, was then flowed into electrolyzers,
which were arranged in cascade series so that the solution could flow
from one to the other. Sulphur dioxide was also blown through the
electrolyte in the electrolyzers by means of perforated hard rubber tubes,
which in addition to supplying the necessary sulphur dioxide, agitated
the electrolyte, and thus gave the desired circulation.
The sulphur dioxide protected the lead anodes from peroxidation.
The anodes were gradually converted into lead sulphate, but the sul-
phatization was much slower than the peroxidation which would have
occurred without the introduction of the sulphur dioxide.
Large quantities of sulphuric acid were regenerated by the use of
sulphur dioxide. The sulphur dioxide also acted as a depolarizer, thus
reducing the necessary voltage, and consequently the amount of power,
in the electrodeposition.
The copper was precipitated at the Intercolonial plant at 1.5 volts,
with a current density of 6 amperes per square foot, and electrodes about
1 1/2 in. apart. The current eflaciency was about 90 per cent. The
' Electrochemical Industry, April, 1903.
ELECTROLYTIC PROCESSES 317
cathodes were greased and graphitized. The electrolysis was conducted
until the copper contents was reduced from 2.5 to 1 per cent. The
electrolyzed solution, regenerated in acid by the secondary anode reac-
tions, and still containing 1 per cent, copper, was returned to the leaching
vats, and the cycle continued indefinitely.
Keith Process.— In the Keith process the electrode area is increased
in the different cells, as the electrolyte becomes improvished in the
metal being deposited. The electrodes of each cell are in multiple in
Fig. 57. — Keith process. Tanks arranged so that there is a, gradual reduction of
current density, as the electrolyte becomes impoverished in copper.
the cell, but in series in their relation to all other cells. The strength
of current is the same in all the cells.
It is evident that a solvent, strong in the metal being deposited,
entering the first cell, will admit of a greater current density in producing
reguline metal on the cathodes, than will the weaker electrolyte entering
the succeeding cells in the series. The deposition of metal in each cell
of the series impoverishes the electrolj^e which enters the succeeding
cell, and therefore the current density, must be correspondingly
318 HYDROMETALLURGY OF COPPER
less in order to insure a reguline deposit of copper. To effect this Keith
increases the number of electrodes, or the surface area of the electrodes,
in a progressive order in each succeeding cell from the first to the last of
the series so that the current density will be approximately proportional
to the strength of the solution in the metals being deposited. Fig. 57.
Keith Process at Arlington, New Jersey/ — The ore, containing the
copper as chalcocite, malachite, azurite, and cuprite, was crushed to
30 mesh, and roasted in a mechanical furnace with a hearth 200 ft. long
by 10 ft. wide and fired by coal in seven fireplaces arranged along its sides.
The capacity of the furnace was about 125 tons per day.
From the roasting furnace the ore was conveyed to four leaching vats,
30 ft. in diameter and 6 ft. deep, holding from 125 to 150 tons of the
roasted ore. In the bottom of each tank there was a filter bed of coarsely
crushed rock, covered with canvas, which was caulked tightly around the
edges, and around the outlets through which the tailings were sluiced
after the copper was extracted.
From the center of the bottom of each tank, under the filter bed, a
pipe with a stop cock served through which to draw off the solvent to the
electrolyte cells. The first solvent, consisting of sulphuric acid and largely
of ferric sulphate, took up nearly all the copper and run out as cupric
sulphate, and decreased in amount as the ore became improvished
and the extraction completed. The tanks were then sluiced out and
recharged.
The deposition vats were rectangular boxes of wood with a suitable
lining. They were arranged so that the electrolyte flowed from No. 1
through the series and finally out of No. 128, depleted of its copper, into
a large sump tank, from which is was pumped back into the stock tank
for reuse.
The deposition vats were so set at different elevations above the floor
building that the electrolyte run by gravity from one to the next, and so
on'through the series to the end. The vats were arranged in six rows,
and one of each row at greater elevation than the end next to it of the
preceding row, it was necessary to raise the electrolyte from one row to
the next at those points. This was done by means of air lift pumps.
The electrical connections were as follows: the electrodes in each vat
were connected in multiple, and the several vats of the electrodes con-
nected in a series of 128. But for the purpose of insuring a retrogressive
decrease of current density at the electrodes of each vat, after the first of
the series, the number of electrodes was progressively increased, from the
first to thp last of the series. The current was the same for each vat of
the series, but the current density was less and less in decremental order
from the first to the last of the series, to compensate for the decrease
' American Inst. Electrical Eng., 1902 Meeting, S. N. Keith.
ELECTROLYTIC PROCESSES 319
of copper in the electrolyte in its course through the vats. The current
density was from 15 to 20 amperes per square foot of cathode area with
a 6 per cent, copper solution, provided proper circulation and sufficiently
rapid movement of the electrolyte was kept up between the anode and
cathode surfaces.
During the electrolysis there was a counter-electromotive force of 1.6
volts in each of the series. The generator was operated at approximately
249 volts, which gives, for 128 cells, 1.87 per cell. The voltage, per cell,
required was 1.87—1.60 = 0.27 volt for resistance of conductors, etc.
Primarily both the anodes and cathodes were of sheet lead, but under
electrolytic action the anodes became coated with PbOz and the cathodes
with copper. As soon as the copper deposit reached a thickness of card-
board it was stripped off each lead cathode, which was then replaced in
its cell, and the two copper sheets thus produced had connections rivited
on them, and were then rehung as cathodes in some of the vats, where
they remained a sufficient length of time to receive the desired
thickness.
The Siemens -Halske Process. — In this process the ore is finely ground
and roasted at a moderate temperature in such a way that the iron is
almost completely oxidized, while the copper is contained in the roasted
material, partly as copper sulphate, and partly as copper oxide, but
principally as cuprous sulphide, CujS. It is stated by the inventors that
roasting is not necessary with all ores. The roasted ore is then treated at
a temperature of about 90° C. (194° F.), with a solution of ferric sulphate,
Fe2(S04)3, to which is added a little sulphuric acid.
On dissolving the copper, the ferric sulphate is reduced to ferrous
sulphate, the copper going into solution as cupric sulphate. The solution
of copper and ferrous sulphate is then led into the cathode compartment
of an electrolytic cell, and in which the anodes and cathodes are separated
by a permeable diaphragm. Part of the copper is deposited on the cath-
ode, and the solution then circulates through the diaphragm to the anode.
This consists of carbon rods. At the anode the ferrous sulphate is reoxi-
dized to the ferric sulphate, which is then again used to dissolve copper
from new charges of ore.
The chemical reactions, which take place during the electrolysis, and
the leaching of the ore, are clearly shown by the following equations:
(1) xH2SO, + 2CuSO, + 4FeSO, = 2Cu + 2Fe2(S04)3+xH2SO„
in which the copper is electrolytically precipitated, and the ferrous
sulphate reconverted to the ferric sulphate, at the anode. The electro-
lyzed and regenerated solution is then returned to the ore and the
copper dissolved;
(2) xH3SO, + Cu2S + 2Fe2(SO,)3 = 2CuS04-F4FeSO, + S-FxH2S04.
320 HYDROMETALLURGY OF COPPER
If there is cupric oxide in the ore, it may be acted upon either by the
sulphuric acid or ferric sulphate;
(3) CuO + H2S04 = CuSO, + H20.
(4) 3CuO+re2(S04)3 = CuS04+FeA-
A comparison of equations 1 and 2 shows that if the copper in the
ore is in the form of cuprous sulphide, the electrolyte, after passing
through the leaching vats, will contain exactly the same quantity of
copper sulphate, ferrous sulphate, and free sulphuric acid as it did prior
to electrolysis; and that it is, therefore, completely regenerated, and may
be used again for the electrolytic decomposition. But if the copper is
present in the ore partly as oxide, it is evident from equations 3 and 4
that in this case the solution will be richer in copper, but poorer in
respect of iron and sulphuric acid than it was before electrolysis. These
equations do not take into account the , possible reactions with base
elements in the ore.
If a solution containing cupric and ferrous sulphates is electrolyzed
in the presence of sulphuric acid, copper is deposited in preference to the
iron. If the electrolysis is performed without a diaphragm between the
electrodes, the ferrous sulphate is oxidized at the anode to ferric sulphate,
and reduced again at the cathode to ferrous sulphate. This represents
a waste of energy, which appears as heat, and the electrolyte, as a solvent
for copper, is not much improved. Ferric sulphate is a solvent of copper
from its oxide and sulphide combinations; ferrous sulphate is not; it is
therefore desirable to have as much as possible of the iron in the solution
in the ferric condition, before again applying it to the ore, and this is
brought about by interposing diaphragms between the electrodes, and
then passing the solution from the cathode to the anode compartment,
or by permeating the solution through the diaphragm from the cathode
to the anode compartment.
The scheme in the Siemens-Halske process, at first contemplated
roasting the ore containing the copper sulphides, at a low temperature, so as
to oxidize the sulphide of iron which it contains, to ferric oxide, and to free
the cuprous sulphide originally forming a constituent of copper pyrites
in the ore. In the course of the roasting some of the cuprous sulphide
is converted into sulphate, but this only advances the process by the
amount of copper sulphatized.
Copper-iron sulphide, as occurring in nature in the ore, is not readily
soluble in ferric sulphate so that an impracticable long time is required
to effect the solution of the copper, even with exceedingly fine grinding,
so that a commercial application of this method of leaching sulphide ores
is not feasible under existing conditions. It is impossible to perform
the delicate roast necessary to release the copper sulphide and oxidize
the iron sulphide.
ELECTROLYTIC PROCESSES
321
Free copper sulphides and oxides react with ferric sulphate easily and
quicldy. The presence of a large quantity of ferrous sulphate in the
ferric sulphate solution impairs the solution of the copper from cuprous
sulphide by means of ferric sulphate.
The method suggested in the Siemens-Halske process for roasting
sulphide ore, so that the main quantity of the iron is transformed to
oxide while the larger portion of the copper remains as cuprous sulphide,
is quite impossible. Neither can satisfactory results be obtained by
dead-roasting, for the reason that at the temperature required, basic
silicates are formed by means of a combination of the copper oxide with
the silicates of the gangue, and perhaps also salts of the type FegO^ are
formed by combination of the oxides of copper and iron; such salts are
acted upon very slowly by ferric sulphate. Cuprous oxide is also formed
by roasting at fairly high temperatures, and the cuprous oxide so formed
is not readily soluble in a solution of ferric sulphate.
There is no difficulty in making the copper soluble by roasting at a
low temperature. This temperature is about 450 to 480° C. The
copper in the ore, at these temperatures, is largely converted into sulphate
and some into oxide. There would appear to be no reason why the
roasting should be attempted to be carried on as originally proposed by
the inventors, when it is more satisfactory and just as cheap to give the
ore a thorough roast at a low heat.
X '
Fig. 58. — Horizontal diaphragm cell. Used in the Siemens-Halske process.
Various difficulties were encountered in the practical operation of the
Siemens-Halske process, principally among these, was the indifferent
nature of the solvent and the inability to obtain suitable anodes and
diaphragms. The anode difficulty has not yet been overcome, as this
process, in common with all other sulphate processes is still laboring
under the disadvantage of not being able to find a suitable insoluble
anode. In the Siemens-Halske process, as experimentally carried out
some years ago, carbon was used as the anode, but it was not at all
satisfactory.
In the later form of apparatus used in the electrolysis, the diaphragms
and electrodes were placed horizontally, and constructed as shown in
Fig. 58. The electrolyzer, E, is divided horizontally into two compart-
21
322
HYDROMETALLURGY OF COPPER
ments by an asbestos diaphragm, D. In the upper compartment is the
cathode C, and in the lower compartment the anode A. The cathode may
be made of a thin sheet of copper and the anode of carbon or of sheet
lead. The solution from the ore is introduced at K, and drawn off at Y,
the rate of flow being adjusted, so that it passes slowly and continuously
through the permeable diaphragm D, and is in contact with the electrodes
successively for a sufficient time to allow the deposition of most of the
copper in the upper compartment, and of the oxidation of the ferrous
sulphate to ferric sulphate in the lower compartment. The electrolyzed
solution as withdrawn from the anode compartment is returned to the
ore.
General arrangement of a Siemens-Halske plant.
Fig. 69 shows a complete outline plant for the Siemens-Halske pro-
cess. A is the storage tank for the solution to be electrolyzed. The
solution passes through the pipe B into the bath C, flows first into the
cathode division k, and then through the filter into the anode division a,
from which the escape pipe D leads it to the pipe G, which conducts it into
the solution tank H. Here it comes in contact with the ore to be leached,
which has been ground in the ball mill E. After the copper has been
dissolved, the mixture of exhausted ore and liquor runs into the vacuum
filter K. The solution aspirated through the filter is again conveyed
to the storage tank A by the pipe M.
The theoretical voltage required in the Siemens-Halske process, to
precipitate the copper and convert the corresponding amount of ferrous
sulphate to ferric sulphate is 0.36 volt. The inventors considered that
ELECTROLYTIC PROCESSES 323
0.7 volt would give a current density of the required strength for practical
operations. In the experimental tests it varied from 0.75 to 1.8 volts.
The process is nowhere now in practical use.
M. DeKay Thompson Jr.'s Experiments on the Siemens -Halske
Process.
Interesting experiments on the Siemens-Halske process were made
by M. DeKay Thompson Jr.' to determine the solvent action of ferric
sulphate on copper compounds likely to occur in raw or roasted ore, and
also to determine the efficiency of the electrolytic precipitation. His con-
clusions may be summarized as follows:
Cupric Oxide, CuO. — 1. The investigation of the action of ferric sul-
phate on cupric oxide leads to the conclusion that the reaction between
the two is probably represented by the equation:
3CuO -l-FejCSO,) 3 =3CuS04 +Fe203,
when the two are present in equivalent quantities. When this is not the
case, basic salts are formed to a considerable extent.
2. Copper sulphate is precipitated by copper oxide. When the
amount of oxide is equivalent to the amount of copper in the solution,
the precipitation is only partial; in the presence of a large excess of copper
oxide it is complete.
3. A metathesis takes place between copper oxide and ferrous sul-
phate, analogous to that with ferric sulphate. Ferrous oxide and copper
sulphate are the resulting products.
4. Under certain conditions all iron and copper salts could be thrown
out of solution.
Cuprous Oxide, CujO. — 1. The results of the experiments with cuprous
oxide are:
1. Cuprous oxide reduces ferric sulphate to ferrous sulphate according
to the equation,
CujO +Fe2(S0,) 3 + H^SO, = 2CuS0, +2FeS0, + H^O,
and both ferrous and ferric iron is precipitated.
2. Cupric sulphate does not act on cuprous oxide.
3. Ferrous sulphate does not act on cuprous oxide.
Cuprous Sulphide, CujS. — The results of the experiments with cuprous
sulphide showed:
1. The verification of the equation,
Cu,S -F 2Fe2 (SO J 3 = 2CuS0 , + 4FeS0 , + S.
2. The FeSO^ formed has no effect on cuprous sulphide.
3. The CUSO4 formed has no effect on cuprous sulphide.
' Electrochemical Industry, June, 1904.
324 HYDROMETALLURGY OF COPPER
4. Sulphuric acid dissolves cuprous sulphide in the presence of
oxygen very slowly. The presence or absence of sulphuric acid in the
solvent is therefore of very little consequence as far as the CujS is
concerned.
Cupric Sulphide, CuS.— The following seems to be the simplest reac-
tion that can take place between copper sulphide and ferric sulphate:
CuS+Fe,(SOj3 = CuSO, + 2FeSO, + S.
In the tests some sulphur was set free, but the amount of copper dis-
solved was in excess of that called for by the above equation. This
excess was evidently due to the oxidation of the copper sulphide to
sulphate. To show this, tests were then made with sulphuric acid of
1.2 sp. gr., with a 1.15 per cent, solution of cupric sulphate, and with a
0.55 per cent, solution of ferrous sulphate. In all cases about the same
amount of copper was dissolved, which shows that it was due to the
oxygen present. Copper sulphide is therefore not dissolved by any of
these reagents.
Many copper ores contain iron, which is changed over to oxide on
roasting. The reaction that takes place between ferric oxide and
ferric sulphate is to precipitate iron from the solution as a basic salt.
Experiments were then made with a copper ore having the following
analysis:
Cu, 29 . 99 per cent.
Fe, 27 . 89 per cent.
S. 33 . 32 per cent.
SiOj, 9.75 per cent.
100.95 per cent.
The composition of the mineral in this ore corresponds closely to that of
copper pyrites.
The first experiment showed that the ore was scarcely attacked by
a 5 per cent, solution of ferric sulphate. The powdered ore, crushed to
80 mesh, was then roasted for several hours. After roasting it was found
to contain 28.8 per cent, copper, and 18.9 per cent, sulphur, and the iron
was computed to be 27.0 per cent. Some of the iron and copper was
oxidized, so that 7.6 per cent, of the ore was copper sulphate, and 2.3
per cent, iron sulphate. Solution tests showed that not much more than
75 per cent, of the copper in this roasted ore could be dissolved from this
sample with a ferric sulphate solution containing 2.833 grm. of iron in
50 c.c. of solution.
A second set of experiments were made with a smaller amount of the
ore, and the same volume of solvent. In this case a little more copper
was dissolved. These results confirmed those previously obtained.
Experiments were then made to see if more of the copper could be dis-
ELECTROLYTIC PROCESSES 325
solved out of the ore already treated by using fresh portions of the solvent.
After from three to five treatments in this way, the ultimate extraction
was 81.0 per cent. The residues from these experiments were collected
and roasted again. It was found that by this means 91.8 per cent, of
the copper still remaining in the ore was dissolved by the solvent in one
hour.
For this reason the ore used in the above experiments was roasted
again. The copper then amounted to 30.3 per cent. This ore, when
agitated for 1 hour with a 6 per cent, solution of ferric sulphate showed
an extraction of 93.9 per cent, of the copper. In 5 hours, under similar
conditions, 98.9 per cent, of the copper was dissolved.
These experiments show that copper can be dissolved by ferric
sulphate pretty completely as soon as the ore is sufficiently roasted.
The question is what are the compounds formed by roasting, which are
so much more soluble than the unroasted ore. It seems that this is due
to the formation of copper oxide. If this is the case, just as much copper
would be dissolved by sulphuric acid. To test this conclusion some of
the ore was similarly treated with sulphuric acid of 1.2 sp. gr. for 5 hours.
This showed an extraction of 97.7 per cent, of the copper. It was found,
however, in the study of cuprous sulphide, that this was easily extracted
by ferric sulphate, and this is exactly what was under investigation in
these experiments, combined with sulphide of iron. It must be this
double combination, therefore, that prevents the copper from being ex-
tracted by the ferric sulphate.
The results of the experiments with the ore may be summarized as
follows :
1. Copper pyrites is not appreciably attacked by ferric sulphate.
2. Roasting so changes the ore that nearly all the copper can be
extracted by either sulphuric acid or by ferric sulphate. This makes it
seem probable that the roasting changes the copper largely over to oxide.
The Electrolysis. — The electrolysis comprises two reactions, the reduc-
tion of the copper at the cathode and the oxidation of the iron at the
anode. These were investigated separately. The vessel used was
a copper voltameter jar divided into three narrow compartments by two
clay diaphragms made fast to the glass sides with paraffine. Theanode
and cathode were in the outer compartment, while the inner one pre-
vented diffusion from one electrode to the other.
In the experiments on the deposition of copper a lead plate was used
as anode, and the anode compartment was filled with sulphuric acid.
The cathode compartment was filled with a solution of the following
composition:
5 per cent, ferrous iron.
3.5 per cent, copper.
2.5 to 3.0 per cent, sulphuric acid
326
HYDROMETALLURGY OF COPPER
The same solution was contained in the middle compartment. During
electrolysis the liquid in the cathode compartment was stirred by a cur-
rent of carbon dioxide.
The object in these experiments was to determine how poor the solu-
tion may become in copper without affecting the character of the copper
deposited at a given current density. This was to be determined for
the different current densities. For this purpose the gain in weight of
the copper cathode was compared with the copper deposited in a copper
voltameter. The accompanying table gives the results obtained with a
current density of 0.98 ampere per square decimeter. (9.1 amperes per
square foot.)
Time in
hours
Cu deposited in voltameter
for 1/2 hour
Cu deposited in cell for 1/2 hour
Per cent, yield
1/2
0.3346
0.3201
98.65
1
0.3346
0.3328
99.45
1 1/2
0.3377
0.3356
99.40
2
0.3385
0.3384
100.00
2 1/2
0.3474
0.3469
99.97
3
0.3397
0.3399
100.00
3 1/2
0.3321
0.3312
99.75
i
0.3340
0.3341
100.00
4 1/2
0.3399
0.33.52
After 4 hours the copper commenced to be spongy. The cathode solution
was then analyzed and was found to contain 0.72 per cent, copper. In
another similar experiment with the same current density, the copper
did not begin to be spongy till the strength of the solution had reached
0.38 per cent, copper.
In still another experiment with a current density of 1.8 amperes per
square decimeter (16.8 amperes per square foot) the copper began to get
spongy at the end of the second hour. The electrolysis was therefore
discontinued and the solution in the cathode compartment analyzed.
It contained 0.98 per cent, copper. The current density was then reduced
to 0.47 ampere per square decimeter (4.4 amperes per square foot) and
the electrolysis continued. The copper came down in good form till the
concentration of copper was 0.05 per cent, in the solution. Two other
experiments were carried out with current densities of 3.4 amperes and
2.6 amperes per square decimeter, respectively. (31.6 and 24.2 amperes
per square foot.) In the first case the copper was deposited in a spongy
form immediately. The same was true in the second case though not in
so marked a manner. Some of the spongy copper was tested for iron,
but this impurity was not present.
Oxidation of the Iron at the Anode. — In all of the experiments to
determine the oxidation of the iron at the anode, the yield for the first
ELECTROLYTIC PROCESSES 327
few hours amounted to more than 100 per cent. It seems probable, there-
fore, that if the disturbing causes were removed, the yield would not fall
much below 100 per cent. At the end of the first period the solution
contained 2.1 per cent, ferrous sulphate. Therefore the conclusion may
be drawn that starting with a solution containing 5 per cent, ferrous iron,
this may be oxidized with approximately 100 per cent, yield, till the
solution contains only 2 per cent, of ferrous iron, using a current density
of from 0.3 to 0.5 ampere per square decimeter (2.79 to 4.64 amperes per
square foot) . When the solution contains between 2 per cent, and 1 per
cent, oxidizable iron, the yield would still be about 90 per cent. When
the concentration is still more diminished, the evolution of gas becomes
strong and the yield falls off correspondingly.
Siemens-Halske Process in Spain. — The Siemens-Halske process was
installed for practical operation in Southern Tyrol in Spain, but the results
were unsatisfactory, owing mostly to the indifferent nature of the solvent.
The ore contained the copper in form of the compound CujS.FeS.FeSj,
which was found difficult to dissolve, either raw or roasted, to encourage
further operations. Later, the ore was roasted at a very low tempera-
ture, and leached with sulphuric acid, and the copper in the solution
crystallized to copper sulphate, for which there was a good local market.
Experiments at the Ray Mines, Arizona. — W. L. Austin gives an ac-
count of experiments made at the Ray Mines, by W. Y. Westervelt in the
summer of 1905, which are interesting.'
"Westervelt, after making preliminary wet concentration tests which gave
unsatisfactory results, instituted leaching experiments on the Ray ore. These
were sufficiently encouraging to warrant the construction of a small plant on the
property for the purpose of carrying out further investigations.
" The ore from the Ray mines consists of disseminated sulphides in a porphyry
—or schistose — gangue, and is thought to average about 2.22 per cent, copper,
accompanied by little or no precious metals. Immense quantities of this ore
are known to exist.
"At Ray, leaching with sulphuric acid was found to remove a small percent-
age of the copper from the crude ore but did not attack that which was shut up
in the sulphides. However, practically all of the copper present could be readily
brought into solution by treatment with a hot-acid-solution of ferric sulphate,
and investigations along t 
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Books to Borrow
Open Library
TV News
Understanding 9/11