Control of Autoclave Scaling during Acid

Pressure Leaching of Nickeliferous Laterite Ore

P. B. QUENEAU, R.E. DOANE, M.W. COOPERRIDER, M. H. BERGGREN, and P. REY

An operating problem encountered at the Moa Bay operation in Cuba, where nickeliferous laterite ore
is processed by sulfuric acid pressure leaching, is the formation of alunite and hematite deposits on
the autoclave walls. The AMAX Extractive Research & Development, Inc., metallurgical laboratory
(Golden, Colorado) has made substantial improvements in the Moa Bay process in the area of metal
recovery, energy consumption, and feed versatility. One of the advantages of AMAX's process is its
ability to treat substantial portions of nickel- and magnesium-rich serpentine while maintaining acid
utilization efficiency. Scale formation is minimized by combining staged acid addition with vigorous
agitation and 270 ~ operation. This paper describes how advantage can be taken of MgSO4" XH20
precipitation both to inhibit alunite scaling and to disperse hematite scale within the MgSO4" XH20
matrix. Cooling the autoclave from its 270 ~ operating temperature down to 180 ~ takes advantage
of the reverse solubility of magnesium sulfate. The magnesium dissolves, liberating entrained hema-
tite, thus providing a means for control of autoclave scale with minimum process disruption.

I. INTRODUCTION tides by atmospheric leaching of roasted high-Mg ore. The

clear liquor goes to sulfide precipitation, and the solids
THE AMAX sulfuric acid leaching process involves
residue is pumped to the high-pressure leach to complete the
countercurrent digestion of nickeliferous laterite ore at
dissolution of its metal values. 6
high pressure and at atmospheric conditions.t After solid/
One of the operating problems that must be addressed in
liquid separation, the nickel and cobalt values are pre-
design of the pressure leaching system is scaling of reactor
cipitated by hydrogen sulfide ~5 and refined into nickel and
cobalt metal. 16Research at AMAX began in the early 1970's surfaces.7 The formation of alunite and hematite deposits on
the autoclave walls has been well documented by Russian
at AMAX's Extractive Research and Development Labora-
authors describing the Moa Bay operation. 7's Less recog-
tory in Golden, Colorado, using as a model Cuba's Moa Bay
nized is the potential for formation of MgSO4" U 2 0 and
laterite acid pressure leaching operation. 2 Substantial im-
silica scale. These latter compounds can become trouble-
provements were made over the Moa Bay process, particu-
some when higher grade nickeliferous laterites containing
larly in the areas of metal recovery, energy consumption,
and feed versatility.3'4'5 greater amounts of magnesium silicate are treated.
One of the advantages of AMAX's process is its ability
A block diagram of the AMAX acid leach process is
to process substantial portions of nickel-rich serpentine
shown in Figure I. The ore is first screened into two size
(Mg3Si2OT'2H20) while maintaining acid utilization effi-
fractions. The coarse fraction, being higher in magnesium,
ciency. 91~ The focus of the present work is upon the depo-
is dried and calcined to improve its reactivity.17 It is then
sition of MgSO4"XH20 o n the interior surfaces of the
used for pH adjustment of the pregnant solution before the
pressure leach vessel. The interaction of these magnesium
resulting leach residue is added to the high-pressure leach.
sulfate deposits with the alunite and hematite scale is ex-
The fine fraction is pumped directly to high-pressure leach.
amined. Causes of the deposition are detailed, and means
Optionally, the ore may be mined as separate "L" (limonite)
to minimize the rate of deposit buildup are discussed.
and "S" (serpentine) fractions and handled in separate ore
preparation circuits.
The high-pressure leach consists of a multistage pressure II. AUTOCLAVING OF LATERITE ORES
vessel equipped with turbine agitators. Sulfuric acid is HIGH IN MAGNESIUM SILICATES
added after the slurry is preheated to operating temperature.
The heating of the slurry is carried out in several stages The first commercial operation to utilize sulfuric acid
using steam recovered from the multistage flashing, as well pressure leaching to process nickeliferous laterite ore was
as high-pressure makeup steam. The flashed slurry is sub-
jected to countercurrent decantation (CCD). Neutralized
barren solution is used to displace the pregnant liquor from
the tailings. Residual acid in the pregnant solution is re-
Acid~Leach~~ Slurry Tailings

duced to an acceptable level for precipitation of metal sul-

P. B. QUENEAU, formerly with AMAX Extractive Research and Devel-

opment, Inc., Golden, CO, is now Technical Assistant to the President,
Hazen Research, Inc., Golden, CO 80403. R.E. DOANE, Process En-
gineer, M.W. COOPERRIDER, Mineralogist, and M.H. BERGGREN,
Research Metallurgist, are with AMAX Extractive Research and Devel- CoS
opment, Inc., Golden, CO 80403. P. REY, formerly with COFREMMI,
Paris, France, is now retired, Paris, France.
Manuscript submitted October 7, 1983. Fig. 1- - F l o w pattern of the AMAX acid leach process.

METALLURGICALTRANSACTIONS B VOLUME 15B, SEPTEMBER 1984--433

Freeport Sulfur's Moa Bay plant in Cuba. The Cuban gov- Table I. Autoclave Feed Composition,
ernment expropriated this plant in 1960 shortly after startup. Feed Rate, and Product Composition
The ore contained about 48 pct Fe, 5 pet AI, 4 pct SIO2,
Campaign
1.4 pet Ni, and less than 1 pet Mg. It was wet screened at
20 mesh, thickened to 45 pct solids, and pressure leached in 1 2 3 4
pachucas at 240 to 250 ~ with sulfuric acid. To attain HPL Feed*
95 pet nickel extraction in 60 to 90 minutes, 0.22 pound
of concentrated sulfuric acid per pound of ore was pumped Mg, pct 2.6 3.2 5.0 6.9
Fe, pct 40 38 33 27
into the first autoclave of a four autoclave train. The preg-
A1, pct 4.7 4.8 3.7 3.3
nant liquor assayed approximately 6 grams per liter Ni, Solids, kg/h 400 380 350 280
0.6 grams per liter Co, 3 grams per liter Mg, 2 grams per
liter AI, and 30 grams per liter H2504. The low magnesium ALR Feed*
content of the ore gave little incentive to incorporate an Mg, pct 3.7 4.1 4.1 4.4
atmospheric leaching step, such as is shown in Figure 1. Fe, pct 21 18 17 17
A serious problem at Moa Bay was the buildup of a thick A1, pct 3.4 3.4 2.9 3.1
layer (to 20 cm) of scale within the autoclaves and inter- Solids, kg/h 100 100 95 80
connecting pipelines. A five- to seven-day shutdown was *HPL = high pressure leach; ALR = atmospheric leach residue
required monthly to remove these deposits. The crust con-
tained hydronioalunite, hydroniojarosite, hematite, boeh-
mite, basic iron sulfate, and aluminochromite, s Substantial
Inert Gas
process improvements have since been made by AMAX Bleed
engineers to minimize deposition of these compounds. The
effectiveness of the process improvements was demon- Feed
Steam ALR- Feed
strated during operation of a "prototype plant" (16 tons per
day) at AMAX's Extractive Research and Development fa- T V-301
t---t 1
cility in Golden, Colorado. AMAX's plant processed rela-
tively high-magnesium ore (3 to 6 pct Mg). Acid was added ]E]F
to the autoclave at three locations, rather than all in the first
-270* C
compartment. Turbine impellers assured vigorous agitation,
thus avoiding the steep compositional and thermal gradients
that were so conducive to scaling at Moa Bay. Leaching tem-
H2SO4
perature was also high (270 ~ minimizing the solubility
of troublesome iron and aluminum scaling components. 6 Fig. 2 - High-pressure leach vessel and level-control pot configuration in
Prototype-plant demonstration runs were completed in AMAX's prototype laterite plant.
1980, Table I summarizes typical feed compositions, feed
rates, and product compositions during four campaigns
using various blends of New Caledonian limonite and V306 located immediately upstream from the first pressure
garnierite ores. The duration of these campaigns ranged letdown choke. (Refer to Figure 2.) The deposits were red-
from 12 to 26 days. The HPL (high pressure leach) feed dish brown (no change in color on drying) and fine grained.
was pumped at 25 to 30 pct solids into the first autoclave X-ray diffraction showed that the magnesium compounds
compartment. This limonite fraction was over 90 pct minus were Mg(Ni)SO4"6H20 and Mg(Ni)SO4" H20. The Mg:Ni
325 mesh, and consisted primarily of goethite (i.e., oz- ratio ranged from about 1:1 to 3:1, roughly corresponding to
FeOOH). There was also 3 to 5 pet aluminum, mainly as the ratio of the Mg to Ni concentration in the aqueous
gibbsite (i.e., o~-AI(OH)3). Magnesium was present primar- solution from which the scale deposited. The distribution of
ily as magnesium-iron silicates, although there was also concentrated H2504fed into the first three compartments of
some chromium spinel. Roasted and atmospheric-leached the autoclave was 60 pet, 25 pet, and 15 pet. The addition
residue (ALR) pumped into the fourth or fifth autoclave point for the atmospheric leach residue (ALR) was into
compartment consumed excess acid and solubilized residual compartment E for Run 1, and into compartment D for
nickel values. The ALR solids contained magnesium-iron subsequent runs.
silicates, amorphous silica and quartz, hematite and iron Table II gives typical analyses of scale samples (dried at
hydroxide, and aluminum hydroxide. room temperature) that were collected from the walls of the
For a more detailed description of ore preparation, miner- autoclave compartments after each run. Samples were also
alogy and composition, refer to the Duyvesteyn and Kay taken from the spool piece between the autoclave (V301)
papers. 1"3 The present work is directed toward magnesium and the level-control pot (V306), as well as from the bottom
sulfate deposition phenomena and the relationships between of the level control pot. A scale sample taken from the V301
the magnesium- , aluminum- , and iron-rich deposits. to V306 spool piece following Run 1 consisted of fragments
of alunite-hematite and MgSO4" H20 wall scale (apparently
dislodged from V301) cemented by MgSO4" H20. The
III. CHARACTERIZATION OF T H E piece assayed 4.8 pet Mg, 13 pet Fe, 7.5 pet A1, and 15 pct
S. Particles collected from the bottom of the level-control
MAGNESIUM SULFATE SCALE
pot were subspherical nodules 5 to 10 mm in diameter, and
Magnesium sulfate deposits were found in autoclave were often rimmed with alunite. When a nodule was broken
compartments A through F, and in the level-control pot open, its interior readily dissolved in cold water.

434--VOLUME 15B, SEPTEMBER 1984 METALLURGICAL TRANSACTIONS B

 Table II. Composition of Scale Samples from thick piece of scale sampled from the north wall of com-
the North Wall of the High-Pressure Leach Reactor partment A following Run 4. The bottom layer (1 mm) is
Scale Analysis (Weight Percent) composed of intergrown crystals of alunite and hematite
deposited during startup of Run 4 using low-magnesium
Sample Identification Mg Fe A1 S
ore left over from Run 1. The feed was then switched
Run 1 V301A 0.3 6.9 17 14 over to high-magnesium Run 4 feedstock (Table I), during
B 0.09 24 12 10 which upper 10 mm was deposited. This upper layer con-
C 0.2 16 15 -- sists of MgSO4" H20 containing highly dispersed particles
D 0.2 17 13 20 of hematite.
E 0.6 13 16 --
F 0.2 10 17 -- Figure 5 shows a magnified secondary electron image
(5(a)) of the Run l-Run 4 scale interface photographed in
Run 2 V301A 0.05 11 15 13 Figure 4. Photo 5(b) is an iron Ks X-ray image of the same
B 0.07 36 8.7 8.3
C 0.05 55 3.1 -- field shown in 5(a). The upper layer (MgSO4" H20) has iron
D 3.4 9.5 11 15 uniformly dispersed. The lower layer (alunite + hematite)
E 0.07 6 17 -- has iron dispersed in alunite, and concentrated in hematite
F 0.02 8.8 16 -- (one crystal shown). Photo 5(c) is a magnesium K~ image
V306 11 5.6 1.3 20 of this field showing the expected concentration of Mg in
Run 3 V301A 11 6.8 1.0 21 the upper layer. Photo 5(d) is an aluminum K~ image of
B 0.04 25 11 10 the same field, illustrating the absence of alunite in the
C 0.1 34 8.2 7.3 MgSO4" XH20 layer.
D 11 5.0 1.0 21 Table Ili tabulates the thickness of the scale deposits col-
E 12 4.6 0.8 21 lected from the north wall of each autoclave compartment
F 0.2 41 2.4 2.1 following each of four campaigns. Scale thickness values
V306 14 3.9 0.8 20 have been converted to mm per 30 days to facilitate com-
Run 4 V301A 12 5 0.8 17 parisons among the four campaigns. No compensation was
B 0.3 49 <4 3.9 made in the table for redissolution of magnesium scale when
C 0.2 52 <2 1.9 flushing the autoclave with water during the shutdown per-
D 6.4 11 <6 -- iod. Autoclave flushing was particularly thorough following
E 8.9 9.9 <2 16
F 6.2 16 <4 13 Run 3.
V306 6.6 17 <2 11 Table IV summarizes magnesium and free acid concen-
trations as averaged during each campaign for each of the
six autoclave compartments. Care must be exercised in
A photograph of the scale deposits in autoclave compart- utilizing these numbers, in that only the H 2 S O 4 assays can
ment F following Run 3, and a second photo showing the be taken at face value, assuming one ignores the HSO4
interior of level-control pot V306 following Run 4, appear shift reaction. The magnesium determinations represent
in Figure 3. Figure 4 shows the cross section of an I I mm not only the aqueous Mg soluble at temperature, but also

Fig. 3 - - T h e left photograph shows the interior of autoclave compartment F following Run 3. The right photograph is a shot of the inside of level-control
pot V306 following Run 4.

METALLURGICAL TRANSACTIONS B VOLUME 15B, SEPTEMBER 1984--435

 Mg(Ni}$O 4" H20

ALUNITE +
HEMATITE

Fig. 4--Cross section of an 11 mm thick piece of scale sampled from the wall of compartment A
following Run 4.

Fig. 5 - - ( a ) is a magnified secondary electron image of the interface between scale deposited during Run 4
in compartment A using Iow-Mg feed during startup (bottom portion of each photo) and that subsequently
deposited using high-Mg feed. (b) (c), and (d) are iron K alpha, magnesium K alpha, and aluminum K alpha
X-ray images, respectively. Magnification of (a) 113 times.

436--VOLUME 15B, SEPTEMBER 1984 METALLURGICAL TRANSACTIONS B

 Table III. Millimeters Projected Scale Thickness per 30 25
Days for the North Wall of the Prototype Plant Autoclave

Campaign A B C D E F 20
1 12.5 9.6 1.2 1.1 0,2
2 11.6 2.8 0.8 8.3 1.4 1,4
3 7.4 6.2 1.5 7.3 6.4 3,6
4 4.7 4.8 4.3 7.9 -- 3,7
Io
the often large quantity of Mg that dissolved from the
residue as magnesium sulfate when the sample was cooled
below 200 ~

o ~ I i h
IV. MAGNESIUM SULFATE 190 230 270
SOLUBILITY RELATIONSHIPS tEMPER,#/'ORE {oc}
The solubility of magnesium sulfate decreases with in- Fig. 6-- Solubilityof magnesiumas a functionof temperatureand sulfuric
creasing temperature, as shown in Figure 6. Above 200 ~ acid concentration,tL
addition of sulfuric acid increases magnesium solubility.H
These data, as prepared by W. Marshall, apply to a solution pensity was enhanced by acid added to B and C, which
free of other salts. A number of solubility experiments were increased the solubility of magnesium (Figure 6). Scaling
also performed by AMAX Extractive R&D personnel to recommenced in D not only due to HTN (high temperature
determine the effect of the nickel, iron, and aluminum in neutralization) addition of solids, but also due to the de-
laterite leach liquor on the high-temperature solubility of creasing liquor acidity. This decreased acidity caused a por-
magnesium sulfate. With 20 grams per liter free acid, the tion of the magnesium soluble in C to precipitate in D.
solubility of magnesium in laterite leach liquor at 200 ~ One would have expected magnesium scaling in compart-
was about 25 pct higher than that published by W. Marshall ment A of Run 2 based on the 2.4 grams per liter solubility
(Figure 6). Thus, presence of about 8 grams per liter total differential. Perhaps there was sufficient intermittent
aqueous nickel, iron, and aluminum does not decrease the backmixing from the more acidic compartment B solution
solubility of magnesium. Which hydrate of magnesium sul- to dissolve magnesium scale buildup in compartment A.
fate is the stable phase depends on temperature. The hexa- Or perhaps the water wash prior to shutdown dissolved
hydrate is stable at about 60 to 80 ~ whereas at higher away the compartment A scale. The same water, now loaded
temperatures, the monohydrate becomes the stable phase. with magnesium, would be less effective in dissolving
A comparison of the concentration of aqueous magnesium magnesium scale from compartment D.
at room temperature (in slurry samples collected from the One would have also expected to find more scale after
270 ~ prototype autoclave) was made with Marshall's Run 4 than 3, based on apparent leach liquor magnesium
solubility values. This comparison appears in Table V. Also concentrations. The lower measured scale thickness follow-
shown is the "Mg scale quantity number" obtained by multi- ing Run 4 probably was due to the autoclave being washed
plying the measured scale thickness in mm (Table III) by particularly thoroughly with water during the shutdown
the magnesium analysis of the scale (Table II). For Run 1, procedure that followed this high-magnesium campaign.
the solution concentration in all six compartments never Thus, residual magnesium scale was greater following
exceeded the solubility of magnesium. Refer to Table V. Run 3 than following Run 4 due to the difference in flush-
Therefore, there was very little magnesium sulfate scaling in out procedures.
the prototype plant autoclave during this run.
During Runs 2, 3, and 4, the solubility of magnesium
V. E F F E C T OF MAGNESIUM SULFATE
sulfate was exceeded by a substantial margin. The quantity
DEPOSITION ON ALUNITE AND
of scale was highest in the autoclave compartments where
HEMATITE SCALE FORMATION
magnesium-containing ore was added (i.e., A and D). Mag-
nesium scaling of compartments B and C was nil because The photographs shown in Figures 4 and 5 are striking
the liquor had already been relieved of its excess magnesium evidence of the change in scale morphology that can re-
via scaling in compartment A. This relief of scaling pro- sult from the presence of MgSO4" H:O. During Run 1, the

Table IV. Apparent Concentrations of Aqueous Mg and H2804 in the High Pressure Autoclave Compartments

A B C D E F
Run Mg H2SO4 Mg HzSO4 Mg H2504 Mg H2SO~ Mg H2SO4 Mg H2SO4
1 7.0 33 7.3 42 7.3 47 7.6 46 8.2 40 8.3 40
2 10.0 34 11.7 45 11.1 51 10.8 45 11.4 45 10.4 45
3 12.0 32 12.9 48 13.6 56 13.4 46 12.4 46 12.6 44
4 15.7 33 18.3 49 17.9 58 15.4 44 15.7 42 15.2 40

METALLURGICALTRANSACTIONSB VOLUME15B,SEPTEMBER1984--437
 Table V. Comparison of Apparent Aqueous Mg with Mg Solubility and the "Mg Scale Quantity" (Dimensionless)
Autoclave Compartment
Run Parameter* A B C D E F
1 apparent Mg, grams/liter 7.0 7.3 7.3 7.6 8.2** 8.3
Mg solubility, grams/liter 7.5 9.6 10.3 10.2 9.1 9.1
Mg scale quantity 4.0 0.9 0.2 0.2 0.1 --
2 apparent Mg, grams/liter 10.0 11.7 11.1 10.8"* 11.4 10.4
Mg solubility, grams/liter 7.6 10.1 11.2 10.1 10.1 10.1
Mg scale quantity 0.6 0.2 0.04 28.0 0.1 0.3
3 apparent Mg, grams/liter 12.0 12.9 13.9 13.4"* 12.4 12.6
Mg solubility, grams/liter 7.2 10.6 12.0 10.2 10.2 9.9
Mg scale quantity 84.0 0.3 0.2 83.0 74.0 0.9
4 apparent Mg, grams/liter 15.7 18.3 17.9 15.4"* 15.7 15.2
Mg solubility, grams/liter 7.3 10.9 12.5 9.9 9.6 9.1
Mg scale quantity 38.0 1.0 0.9 50.0 -- 23.0
*ApparentMg is the analysis of the aqueous phase of the slurry sampled from each autoclave compartment. This assay is the sum of magnesium soluble
at autoclave temperature, plus magnesium solubilized from MgSO~ H20 during cooling of the sample. Mg solubility is the solubility of magnesium at 270 ~
as read from Figure 6. Mg scale quantiO, is the product of scale thickness (Table III) multiplied by the magnesium assay of the scale (Table II).
**Compartment to which the ALR feed was added.

Table VI. Effect of Magnesium on Location and Extent of Alunite and Hematite Scaling
Feed Quantity of Scale
(kg/h) Stage A Stage B Stage C
Run Mg A1 Fe Mg AI Fe Mg AI Fe Mg AI Fe
1 10 20 160 4 220 85 0.9 120 230 0.2 22 20
2 12 18 140 0.6 180 130 0.2 5 100 0.1 2 40
3 18 13 120 85 7 50 0.3 70 160 0.2 12 50
4 19 9 75 55 4 30 1.0 <20 240 0.9 <9 230

solubility of magnesium was not exceeded in autoclave increasing acid concentration slows down the rate of alu-
compartment A. See Table V. The scale deposited in com- minum hydrolysis much more than the rate of iron hy-
partment A was composed of refractory intergrown crys- drolysis. ~0.~4It is therefore hypothesized that where there is
tals of alunite and hematite (Figure 5 and Table II). When insufficient magnesium to deposit magnesium sulfate, depo-
the autoclave feed was changed over to high-magnesium sition of hematite and alunite on autoclave surfaces should
material (Run 4; Table I), deposition of alunite essentially increase with increasing magnesium concentration. Increas-
ceased (Figure 5). The scale deposit became MgSO4" H20 ing acid concentration should retard aluminum hydrolysis
containing a fine dispersion of iron plus traces of aluminum. relative to iron hydrolysis, particularly at shorter reten-
Further evidence of the effect of magnesium in changing tion times.
the quantity of aluminum and iron in the autoclave scale is
presented in Table VI. The "quantity of scale" shown in this
table is attained by multiplying the scale thickness on the
VI. C O N T R O L OF M A G N E S I U M
wall (mm/30 days, as shown in Table 1II) by either the
SULFATE D E P O S I T I O N
Mg, A1, or Fe analysis of the scale (Table II). Thus, in
compartment A of Run 1, the scale was primarily alunite; Magnesium sulfate deposition sharply inhibits the rate
magnesium sulfate content was low. The same was true for of alunite scaling. For instance, the highest scaling rate
Run 2. However, in Runs 3 and 4, the compartment A scale measured on the north wall of the autoclave was 12 m m / 3 0
was primarily magnesium sulfate. Alunite content was low. days in compartment A following Run 1 (low magnesium).
Hematite was significant, but lower than in Runs 1 and 2. This scale was primarily alunite assaying 17 pct A1 and
Very little magnesium deposited in autoclave compart- 0.3 pct Mg. Doubling the magnesium content of the feed
ments B and C due to the relatively high acid concentration (Run 3) resulted in the deposition of 7 m m / 3 0 days of pri-
(Table VI). Iron deposition in compartments B and C was marily magnesium sulfate scale assaying 1 pct A1 and 11 pct
generally higher, and aluminum deposition generally lower, Mg. This substitution of magnesium for aluminum is im-
in the high-magnesium runs. Sulfuric acid concentration portant for autoclave descaling. Unlike alunite and hema-
was also higher in compartments B and C during the high- tite, magnesium sulfate hydrate dissolves rapidly in water or
magnesium runs (Table IV). dilute acid at temperatures below 200 ~ Hematite precipi-
Both the extent of hydrolysis of aluminum and iron at a tates on autoclave internal surfaces along with magnesium
given temperature and acid concentration are increased as sulfate, but the hematite is dispersed within the magnesium
the level of magnesium sulfate is increased, x2~3At 270 ~ sulfate matrix. Magnesium sulfate has an inverse solubility

438--VOLUME 15B, SEPTEMBER 1984 METALLURGICAL TRANSACTIONS B

 into this compartment rather than into the gas cap. The
9 *rE.,rE ----i I---7 *'" C*P increase in temperature deposits magnesium sulfate into E
2 I
,1 I
'
~ l I
q while inhibiting co-deposition of amorphous silica. Mag-
nesium solubility decreases with increased temperature,
whereas that of silica increases. A small quantity of air is
added to the gas cap to oxidize the ferrous iron added with
# l the serpentine in the ALR feed.

Fig. 7--Conceptual diagram of ore, acid, and steam sequencing for at-
taining protective magnesium deposition in each autoclave compartment. VII. CONCLUDING REMARKS
1. Sulfuric acid pressure leaching of nickeliferous laterite
with temperature. Cooling the autoclave from 270 ~ to ore containing high levels of magnesium silicates leads to
180 ~ sharply increases magnesium solubility, resulting deposition of magnesium sulfate. Precipitation of mag-
in magnesium dissolution and liberation of the hematite. nesium sulfate is highest where the ore is fed, where acid
The dissolved MgSO4 exits from the autoclave with the concentration is low or decreasing, and where tempera-
pregnant liquor. The liberated hematite follows the leach ture is high or increasing.
residue to tailings disposal. Thus, magnesium sulfate pro- 2. The magnesium sulfate precipitates as a fine-grained,
vides a means not only for inhibiting alunite scaling, but water-soluble Mg(Ni)SOa'H20 binary salt having a
also for facilitating removal of hematite scale. Mg:Ni ratio corresponding to the ratio of Mg to Ni con-
To be effective as a practical alunite-hematite scale in- centration in the leach liquor.
hibitor, the deposition of magnesium sulfate must be reason- 3. Magnesium sulfate deposition sharply inhibits the rate of
ably uniform in all autoclave compartments. Figure 7 shows alunite scaling. Hematite precipitates on autoclave inter-
a profile of ore, acid, and steam additions to the autoclave nal surfaces along with the magnesium sulfate, but is
that is designed to achieve this uniformity. Sufficient ore is dispersed within the magnesium sulfate matrix.
added to compartment A to exceed the magnesium solu- 4. Where there is insufficient magnesium to deposit mag-
bility in this compartment by about 3 grams per liter. As nesium sulfate, deposition of hematite and alunite on
much acid as possible is also added to this compartment, autoclave surfaces appears to increase with increasing
commensurate with the titanium materials of construction magnesium concentration.
constraint. High initial acid concentration should also in- 5. To be effective as a practical alunite-hematite scale
crease nickel extraction, as compared to stage addition. 9 inhibitor, magnesium sulfate deposition must be reason-
Compartment A has the following features: ably uniform. This uniformity is controlled by appropri-
ate selection of the ore- , acid- , and steam-injection
I. Sufficient magnesium, as contained in the ore, is added locations.
to sharply inhibit alunite scaling, and to entrain the hematite
scale in the magnesium sulfate matrix; and
2. High acid concentration in compartment A allows room
for decreasing the acid concentration in compartments B, ACKNOWLEDGMENTS
C, and D, which distributes magnesium sulfate deposition The authors wish to extend their thanks to Compagnie
downstream. Franqaise d'Entreprise Mini~res, Metallurgiques et d'Inves-
In the prototype plant runs, no protective layer of mag- tissements (COFREMMI), a joint-venture company of
nesium sulfate was deposited on compartment B and C sur- AMAX and the Bureau de Recherches Grologiques et Mini-
faces due to increasing H2804 concentration (as compared to ~res (BRGM), Orlrans, France, for permission to publish
compartment A), which maintained aqueous magnesium in this paper, and to Jean-Paul Duterque (Process Engineer,
soluble form. This problem can be corrected by adding COFREMMI) for his assistance during the investigation.
L-ore and/or atmospheric leach residue (ALR) directly to
compartments B and C (Figure 7). Ore distribution is ad-
justed to precipitate the equivalent of 1 to 3 grams per liter
Mg in each compartment. The free-acid concentration in REFERENCES
compartment B is maintained lower than in A; also, acid 1. W.P.C. Duyvesteyn, G.R. Wicker, and R.E. Doane: "An Omniv-
concentration in compartment C is lower than in B. Al- orous Process for Laterite Deposits," International Laterite Sympo-
sium, New Orleans, LA, 1979.
though hematite deposits along with magnesium sulfate, it 2. E.T. Carlson and C.S. Simons: Extractive Metallurgy of Copper,
is dispersed within the MgSO4 H20 matrix. Refer to the Nickel, and Cobalt, Interscience Publishers, 1961, p. 363.
photomicrographs of Figure 5. 3. H. Kay and E. J. Michal: "The AMAX Acid Leach Process for Oxide
Magnesium sulfate scaling in compartment D was quite Nickel Ores," TMS-A1MEA78-36, 1978.
4. S.O. Fekete and G. R. Wicker: "Energy Conservation in Acid Leach-
high during operation of the prototype plant. This deposition ing of Lateritic Ores," 106th Annual AIME, March 6, 1977.
occurred not only due to all the ALR being added at this 5. C.J. Messa, G. R. Wicker, and S. O. Fekete: "Economic Analysis of
location, but also because of the sudden drop in free acid Nickel and Cobalt Recovery from Laterite Deposits Using the AMAX
concentration in passing from compartment C to D. This Acid Leach Process," 107th Annual AIME, Denver, CO, 1978.
problem can be corrected by adding only a portion of the 6. S.O. Fekete, G.R. Wicker, W. E C. Duyvesteyn, and D. E Shieh:
U.S. Patent 4,098,870, 1978.
ALR to compartment D. Magnesium deposition is also 7. S.I. Sobol: Revista Technologica, 1966, vol. 4(4), p. 3; 1969,
maintained in compartment E by injecting makeup steam vol. 7(1), p. 3.

METALLURGICALTRANSACTIONS B VOLUME 15B, SEPTEMBER 1984--439

 8. R.A. Chaves, V, V. Karelin, and B. E Sobolev: Tsvet. Metally, 1968, CIM Hydromet, October 1981.
vol. 9(4), p. 4. 14. U. Kuxmann and M. Landau: Metall., May 1981, vol. 35(5), p. 408.
9. P.B. Queneau and E. C. Chou: U.S. Patent 3,991,159, 1976. 15. C.M. Jha, G. A. Meyer, and G. R. Wicker: J. Metals, 1981, vol. 33
I0. E.C. Chou, P.B. Queneau, and R.S. Rickard: Metall. Trans. B, (I1), p. 48.
1977, vol. 8B, p. 547. 16. M.C. Jha and G. A. Meyer: "AMAX Chloride Refining Process for
11. W.L. Marshall and R. Slusher: J. Chem. Eng. Data. October 1965, Recovery of Nickel and Cobalt from Mixed Sulfide Precipitates",
vol. 10(4), p. 353. Hydrometallurgy, Research, Development and Plant Practice,
12. P.T. Davey and T. R. Scott: Aust. J. Appl. Sci., 1962, vol. 13, p. 229. K. Osseo-Asare and J.D. Miller, eds., AIME, Atlanta, GA, 1983,
13. K. Tozawa, K. Sasaki, F. Hasegawa, and Y. Umetsu: "Effect of p. 903.
Magnesium Sulfate on the Precipitation of Ferric Oxide by Hydrolysis 17. E.C. Chou, C.B. Barlow, and D.K. Huggins: U.S. Patent
of Ferric Sulfate Solution at Elevated Temperatures," 11th Annual 4,097,575, 1978.

440--VOLUME 15B, SEPTEMBER 1984 METALLURGICAL TRANSACTIONS B