Sie sind auf Seite 1von 14

Hydrometallurgy 72 (2004) 59 – 72

www.elsevier.com/locate/hydromet

A comparative study of leaching kinetics of limonitic laterite and


synthetic iron oxides in sulfuric acid containing sulfur dioxide
G. Senanayake *, G.K. Das
A.J. Parker Cooperative Research Center for Hydrometallurgy, Department of Mineral Science and Extractive Metallurgy,
Murdoch University, Perth, WA 6150, Australia
Received 4 January 2002; received in revised form 2 September 2002; accepted 7 June 2003

Abstract

Limonitic laterite ore of particle size 90 – 125 Am containing goethite, magnetite and hematite was leached for 6 h at a pulp
density of 10% (wt/vol) in sulfuric acid in the absence or presence of sulfur dioxide at atmospheric pressure and 90 jC in a glass
reactor vessel. The sulfur dioxide flow rate was kept at 0.6 L min 1 L 1 of slurry to maintain a constant SO2 concentration of
c 0.3 mol L 1 in solution, and the sulfuric acid concentration was varied between 0 and 0.72 mol L 1. The relative percentage
extractions of Fe, Ni, Co and Mn indicate that the Fe and Ni extractions are inter-related at a ratio of Ni/Fe = 0.7 – 0.9 and
suggest the possibility of catalysis of manganese dissolution by solubilized iron(II). This leads to a Mn extraction of over 90%
in less than 30 min compared with 20 – 40% Fe extraction in the same period, depending on the acid concentration. The initial
rate of leaching of iron shows first-order dependence with respect to H+. Whilst the synthetic iron oxides leach according to the
shrinking particle/sphere kinetic model, the results obtained in the first 4 h of laterite leaching can be described by a shrinking
particle model with an insoluble product layer that retards the diffusion of H+ to the reaction sites at the interface. The
heterogeneous rate constants for both models increase with the increase in H+ concentration. The effective diffusion coefficient
of H+ (DH +
) through the product layer (0.5  10 9 to 4  10 9 cm2 s 1), determined in the present study, is in the magnitude
range of the reported data for DH +
in polycrystalline Fe3O4 and MnO2, but lower than DH +
in aqueous media, 9  10 5 cm2 s 1.
D 2003 Elsevier B.V. All rights reserved.

Keywords: Limonitic laterite; Iron; Nickel; Cobalt; Manganese; Leaching; Heterogeneous kinetic models; H+ diffusion

1. Introduction choice of acidic leaching of these metal oxides in


nickeliferous limonite under atmospheric conditions
Sulfur dioxide is an efficient leaching agent for compared to the acid pressure leaching commercially
minerals containing oxides of iron, nickel, cobalt and practiced in Cuba and Western Australia (Chou et al.,
manganese (Byerley et al., 1979; Miller and Wan, 1977; Kyle, 1996). The kinetics of the acid dissolution
1983; Abbruzzese, 1990; Grimanelis et al., 1992; (Majima et al., 1985; Cornell et al., 1976) and
Kumar et al., 1993; Das et al., 1997). It offers the reductive dissolution of natural or synthetic iron
oxides in the presence or absence of sulfur dioxide
* Corresponding author. Fax: +61-8-9360-6343. (Warren and Hay, 1975; Kumar et al., 1993; Byerley
E-mail address: gamini@central.murdoch.edu.au et al., 1979; Chiarizia and Horwitz, 1991) as well as
(G. Senanayake). the reductive dissolution of manganese dioxide by

0304-386X/$ - see front matter D 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0304-386X(03)00132-4
60 G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72

sulfur dioxide or by Fe(II) sulfate (Miller and Wan, with respect to the concentration of H+ (Cornell et al.,
1983; Tekin and Byramoğlu, 1993) have been studied 1976) due to the slow desorption of FeOH2 + from the
in detail. The relevant kinetic data have been corre- surface, followed by aqueous reaction as shown by
lated with one or more of the physico-chemical factors Eqs. (3) – (5) (Warren and Hay, 1975):
that include the ionic activity of H+, complexation
with OH or SO32 , electrochemical surface reaction FeOOHðsÞ or 0:5Fe2 O3 H2 OðsÞ þ3Hþ ¼ Fe3þ þ 2H2 O
and changes in surface area. ð1Þ
Along with manganese, which may also be present
as pyrolusite, MnO2, nickel and cobalt in laterite Fe3 O4 ðsÞ or Fe2 O3  FeOðsÞ þ 8Hþ
deposits are enriched in weathering products such as
¼ 2Fe3þ þ Fe2þ þ 4H2 O ð2Þ
goethite and limonite by ion replacement (Kyle,
1996). A proper understanding of the kinetics and
mechanism of reductive leaching of oxide minerals by pOFe  OHðsÞ þ Hþ
acid and sulfur dioxide is useful in developing meth- ¼ pOFeþ ðsÞ þ H2 O ðfast equilibrationÞ ð3Þ
ods for selective leaching of valuable metals from
nickeliferous ores. The reaction mechanisms for pure pOFeþ ðsÞ þ Hþ ¼ FeOH2þ ðaqÞ ðslowÞ ð4Þ
components proposed by previous researchers are
important in the discussion of the reductive leaching FeOH2þ ðaqÞ þ Hþ ¼ Fe3þ þ H2 O ðfastÞ ð5Þ
of limonitic laterite ore containing multi-valent oxides
of important metals. However, the interpretation of Despite the fact that the rate-determining step (4) is
kinetic behavior of such ores is complicated due to the first order with respect to H+, the formal reaction order
fact that they contain more than one component that of H+ can show different values in the range 0.5– 2
can react with reductive leaching agents such as sulfur depending on the hydration, complexation, redox
dioxide and/or iron(II) sulfate. For example, the reactions and other factors (Azuma and Kametani,
limonitic laterite used in one of the previous inves- 1964; Gorichev and Kipriyanow, 1984; Majima et
tigations (Das et al., 1997) and the subject of this al., 1985). For example, Table 1 lists some of the
study consists of goethite a-FeOOH and magnetite relevant complexes and their equilibrium constants (K)
Fe3O4 as major components and hematite a-Fe2O3 as showing that K increases with increasing temperature,
the minor component along with quartz. Whilst the but decreases with increasing ionic strength. Potenti-
Fe(III) oxides are dissolved by H+ and/or by SO2 to
produce Fe(III) and Fe(II), the SO2 as well as Fe(II)
produced during reduction can subsequently act as Table 1
reductive leachants for Mn(IV), Fe(III) and Co(III). Equilibrium constants (logK) for protonation and sulfate complex-
This paper describes the analysis of leaching results of ation at 25 jCa
limonitic laterite in sulfuric acid and in perchloric acid Reaction Ionic strength logK
in the absence or presence of sulfur dioxide on the H+ + SO24  = HSO
4 0 1.99 (3.07at 100 jC)
basis of surface chemical reaction with H+ and SO2 as 1 1.07
well as Fe(II) and the heterogeneous mechanisms H+ + HSO
3 = H2O + SO2 0 1.77
1 1.37
described by the shrinking particle models with or Fe2 + + SO24  = FeSO04 0 2.3
without insoluble product layer. 1 1.0
Fe2 + + HSO 4 = FeHSO4

1.2 0.78
3+ 2 +
Fe + SO4 = FeSO4 1.2 2.23
2. Iron chemistry and kinetic models Fe3 + + 2SO24  = Fe(SO4)2 1.2 4.23
Ni2 + + SO24  = NiSO04 0 2.40
1 0.57
2.1. Acid dissolution of iron oxides Mn2 ++ SO2
4 = MnSO4
0
0 2.28 (3.0 at 45 jC)
Co2 ++ SO2 0
4 = CoSO4 0 2.36
The initial rate of dissolution of hematite or mag- 1 1.06
a
netite (Eqs. (1) and (2)) follows first-order kinetics Data from Sillen and Martell, 1964.
G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72 61

ometric measurements and predicted Eh –pH and spe- Fe3 O4 ðsÞ or Fe2 O3  FeOðsÞ þ 3H2 SO4 þ 2SO2
cies distribution diagrams have shown that FeSO40 and
Fe(SO4)2, respectively, are the predominant Fe(II) and ¼ 3FeSO04 þ H2 S2 O6 þ 2H2 O ð9Þ
Fe(III)– sulfate complex species in acidic sulfate sol-
utions (Senanayake and Muir, 1988). The Eh – pH H2 S2 O6 ¼ H2 SO4 þ SO2 ð10Þ
diagram shown in Fig. 1 highlights the need to rewrite
the stoichiometry of acid dissolution shown in Eqs. (1) FeOOHðsÞ or 0:5Fe2 O3  H2 OðsÞþ0:5SO2 þ 0:5H2 SO4
and (2) to accommodate actual complex species such ¼ FeSO04 þ H2 O ð11Þ
as FeSO40 and Fe(SO4)2 in solution:
Fe3 O4 ðsÞ or Fe2 O3  FeOðsÞ þ 2H2 SO4 þ SO2
FeOOHðsÞ or 0:5Fe2 O3 H2 OðsÞ þ 2H2 SO4
¼ 3FeSO04 þ 2H2 O ð12Þ
¼ FeðSO4 Þ
2
þ
þ H þ 2H2 O ð6Þ
2.3. Kinetic models
Fe3 O4 ðsÞ or Fe2 O3 FeOðsÞ þ 5H2 SO4
The initial rates of leaching have been used to
¼ 2FeðSO4 Þ þ 0
2 þ 2H þ FeSO4 þ 4H2 O ð7Þ model the kinetics of the dissolution of hematite on

2.2. Effect of SO2

The catalytic effect of a reducing agent is the result


of a requirement that the surface of the mineral be
composed of a stoichiometric compound that will
dissolve into the solution (Nicol, 1983). For example,
the proposed mechanism for the dissolution of mag-
netite and goethite in SO2 involves slow desorption of
FeHSO3+ from the oxide surface (Byerley et al., 1979).
The presence of reducing agents capable of lowering
the [Fe(III)]/[Fe(II)] concentration ratio will accelerate
iron oxide dissolution (Gorichev and Kipriyanow,
1984). This shows the importance of considering the
reaction between Fe(III) and SO2 in solution.
Solution chemistry studies of Fe(III) and SO2 have
shown a stoichiometry of 1:1.2 for the reversible
reaction between Fe(III) and SO2 to form Fe(II) and
the monomer HSO3 free radical or the dimer H2S2O6
of dithionic acid at pH 0.5 (Higginson and Marshall,
1957). Despite the experimental evidence for the
formation of both S2O62  and SO42  at pH 0– 3 (Sato
et al., 1978), IR spectroscopic studies have indicated
that SO42  is the major species formed during the
leaching of goethite with SO2 (Kumar et al., 1993).
Considering the formation of both dithionate and
sulfate and the sulfate complex of Fe(II), the reductive
dissolution of goethite by SO2 may be represented as:

FeOOHðsÞ or 0:5Fe2 O3 H2 OðsÞ þ SO2 þ H2 SO4 Fig. 1. Eh – pH diagram for Fe(III)/Fe(II)/SO2/SO24  system at 25
jC, based on thermodynamic data in Table 1 at ionic strength 1 and
¼ FeSO04 þ 0:5H2 S2 O6 þ H2 O ð8Þ equimolar species in solution.
62 G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72

the basis of surface chemical reaction with H+ in the the increase in thickness of the solid product layer
absence of SO2 (Cornell et al., 1976; Majima et al., with time, whilst the spherical nature and the radius of
1985) or surface electrochemical reaction in the pres- the particle remains unchanged.
ence of SO2 (Kumar et al., 1993). For example, the The leaching of pure MnO2 has been studied in
initial rate of dissolution of a-FeOOH, h-MnO2 and SO2/Na2SO3/pH 1 – 2 (Miller and Wan, 1983) and in
g-MnO2 has been found to be proportional to [SO2]0.5 FeSO4/H2SO4 (Tekin and Byramoğlu, 1993) at tem-
indicating a half order reaction with respect to SO2 peratures up to 50 jC under controlled hydrodynam-
(Miller and Wan, 1983; Kumar et al., 1993). In the ic conditions. The stirring speed was sufficiently
case of relatively fast surface reactions, it is essential large to confirm that the rate of leaching was not
to consider the decrease in surface area in a shrinking governed by the transport process. The decrease in
sphere model with no product layer formation (Leven- the rate of leaching with time has been related to the
spiel, 1972; Ray, 1993). For the first-order 1:1 molar shrinking particle/sphere kinetic model with a value
heterogeneous rate-controlling step for the dissolution of kss = 4.4  10 3 s 1 and k = 1.2  10 3 cm s 1
of FeOOH (Eq. (4)), this model can be represented by for the leaching of MnO2 in 0.5 M SO2 at pH 2 at a
the mathematical relationship: stirring speed of 650 rpm. The value of k was less
than the predicted mass transfer coefficient for
1  ð1  X Þ1=3 ¼ ½Hþ bulk kr1 q1 t ¼ kss t ð13Þ suspended particles, 2.4  10 2 cm s 1 in the leach-
ing reactor, showing that the leaching rate was con-
where, [H+]bulk (mol cm 3) = concentration in the
bulk solution, k (cm s 1) = intrinsic rate constant of
the surface chemical reaction having the units of a Table 2
mass transfer coefficient, r (cm) = initial particle radi- Experimental conditions, apparent rate constants (kss and kpl) and
us of solid, q (mol cm 3) = molar concentration of Fe proton diffusion coefficients (DH+) for iron dissolution from
synthetic iron oxides and limonitic laterite ore
in the solid, kss (s 1) = apparent rate constant in the
shrinking sphere kinetic model and X = fraction of Fe Material Particle Pulp T SO2 H2SO4 106 106 109
size density (jC) (M) (M) kss kpla DH +b
reacted in time t. The relationship between q and wt.% (m (%, (s 1) (s 1) (cm2
Fe in the material and its density (qm, g cm 3) is: mean) wt/vol) s 1)
FeOOHc 17 2.5 80 nil 3 111
q ¼ qm  %Fe in material=ð100  molar mass of FeÞ 0.1 3 157
ð14Þ Fe3O4d 75 0.4 50 0.54 nil 1.7
Lateritee
Test 1 (T1) 106 10 90 0.3 nilf 0.06 0.12
If a porous solid product layer is formed on the
Test 2 (T2) nil 0.72 3.3 0.51
surface during the reaction, the slow diffusion of H+ Test 3 (T3) 0.3 0.18 1.5 0.87
through the product layer becomes the rate-controlling Test 4 (T4) 0.3 0.36 3.3 1.0
step. The shrinking core with product layer model in Test 5 (T5) 0.3 0.54 30g 10 2.0
such cases is represented by the equation: Test 6 (T6) 0.3 0.72 43g 27 4.2
Test 7 (T7) 0.3 nilh 18 2.2
a
1  3ð1  X Þ2=3 þ 2ð1  X Þ Rate constant for shrinking particle model with solid product
layer for 4 h (Fig. 10).
b
¼ 6½Hþ bulk Dþ 2 1
H r q t ¼ kpl t ð15Þ Proton diffusion coefficient based on Eqs. (14) and (15):
r = 5.3  10 3 cm, qm = 3.8 g cm 3 (Berkman, 1995) and %Fe = 39.
c
Synthetic goethite (Chiarizia and Horwitz, 1991).
where DH +
(cm2 s 1) = diffusion coefficient of H+ d
Synthetic magnetite,1 g solid in 250 mL liquid (Byerley et al.,
through the product layer and kpl (s 1) = apparent rate 1979).
e
constant. Eq. (13) assumes that the decrease in rate This work.
f
No H2SO4 added, natural [H+] = 0.062 M, based on mass
with time is due to the decrease in particle size and
balance for Eq. (16).
thus the surface area. Eq. (15) assumes that the g
Rate constant for shrinking particle model for the first 60 min
diffusion of H+ through a solid product layer is the (Das et al., 1997).
h
rate-determining step and the decrease in rate is due to No H2SO4 added, 1 M HClO4.
G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72 63

trolled by the surface reaction rate (Miller and Wan, 4. Results and discussion
1983).
4.1. Time and acid dependence of Fe and Ni
extraction
3. Experimental
Fig. 2 plots %Fe extractions from limonitic laterite
A water-jacketed 1-L Pyrex glass vessel fitted with at time intervals 0.5, 1, 2, 4 and 6 h and compares
a Teflon-coated stirrer, baffles, gas inlet tube, sampling with some of the data reported previously (Das et al.,
tube, thermometer and water condenser was used as 1997). The leaching with SO2 in the absence of added
the reactor for leaching; other details of dimensions acid (Test 1) shows the lowest iron extraction but a
were similar to those reported by Tekin and Byramoğlu linear increase in % extraction with time. The %Ni
(1993). The limonitic nickel laterite of particle size extraction presented in Fig. 3 also shows the lowest in
range 90 – 125 Am from Bulong, Western Australia, Test 1, but %Ni remains at 20% and independent of
was used in all experiments. The standard leaching time. The highest Fe and Ni extraction at 6 h is shown
conditions were: pulp density = 10% (wt/vol), temper- by SO2 + 0.72 M H2SO4 (Test 6) but the rate decreases
ature = 90 jC, agitation speed = 650 rpm, leaching time with time. In Test 3, Ni extraction reaches only 25%
6 h. The sulfur dioxide flow rate was maintained at 0.6 but Co and Mn reach 100% in 30 min.
L min 1 L 1 of slurry and the acid concentration was The %Fe and %Ni extraction after a given
varied between 0 and 0.72 M H2SO4 or 1 M HClO4 in leaching time generally increases in the order of
seven tests T1 – T7 (Table 2). These conditions were acid concentration: Test 1 < Test 3 < Test 4 < Test
consistent with the previous investigation (Das et al., 5 < Test 6. This can be mainly attributed to the
1997). Samples were collected at 0.5, 1, 2, 4 and 6 h, initial rate of leaching with SO2, represented by
and the solutions were analysed for Ni, Co, Mn and Fe the initial slopes in Figs. 2 and 3, which also
by atomic absorption spectrophotometry. increase in the same order. However, the rate

Fig. 2. Kinetic plots for iron extraction from limonitic laterite. T1, T2 and T6 from Das et al. (1997); other data from the present study, see Table
2 for conditions.
64 G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72

Fig. 3. Kinetic plots for nickel, cobalt and manganese extraction from limonitic laterite. Solid lines represent Ni extraction in T1 to T6 (Table 2),
dashed line represents Co and Mn extraction in T3.

decreases with time, except in Tests 2 and 3 carried behavior can be further examined by considering the
out with no SO2 or low acid concentrations (0.18 relative extractions of the four metals.
M), respectively, where the rate after 2 h remains
fairly constant. In general, the decrease in rate with 4.2. Relative extraction of Fe, Ni, Co and Mn
increasing leaching time can be related to one or
more of the following reasons: (i) the decrease in 4.2.1. Ni vs. Fe
acid concentration with the consumption of H+ in Despite the higher iron content compared to nickel
Eqs. (6) –(12), (ii) the change in equilibrium SO2 in the starting material, 39.05% Fe, 1.17% Ni, 0.124%
concentration in Eq. (16) caused by the decreasing Co, 1.25% Mn (Das et al., 1997), Fig. 4 shows linear
acid concentration and changing ionic strength (Ta- correlations between the %Ni and %Fe extractions at
ble 1), (iii) the decrease in surface area of the different time intervals up to 6 h. These linear rela-
particles or (iv) the increase in thickness of a solid tionships of slopes ranging from 0.7 to 0.9 in the
layer, which retards the diffusion of reactants or range 20 – 85% Ni extractions can be related to the
products. incorporation of nickel in the iron oxide fraction of
laterite and thus the release of nickel(II) into the
HSO þ
3 þ H X SO2 þ H2 O ð16Þ solution is associated with the leaching of iron.
However, the nickel and iron extraction by sulfur
Unlike in the case of Fe and Ni, the % extraction of dioxide in the absence of added acid (Test 1) was
Co and Mn rapidly increases and reaches 85 –100% low, < 20% Ni and < 10% Fe, and the results do not
during the first 30 min in the presence of SO2 due to show a correlation with the other data in Fig. 3. In
the reactivity of SO2 and the occurrence of Co with contrast, 0.72 M H2SO4 alone with no SO2 (Test 2)
Mn (Das et al., 1997). Consequently, despite the extracts c 45% Fe and Ni; SO2 + 0.72 M H2SO4
high acid concentration of 0.72 M in Test 2, the (Test 6) extracts c 85% Fe and Ni in 6 h. These
extraction of these metals remains low at 12 –28% results also show the necessity of breaking up of the
Co and 20 –40% Mn in 0.5 – 6 h of extraction time goethite structure by SO2 and/or H+ to release both
due to the absence of SO2 (see later in Fig. 5). This iron and nickel into solution.
G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72 65

Fig. 4. Correlation between nickel and iron extraction at different time intervals in Tests T1 – T7.

4.2.2. Ni, Co, Mn vs. Fe presence of SO2 in Tests 4 –6, which were cooled to 25
Fig. 5 plots the % extraction of Ni, Co and Mn jC, were lower (0.56 V) than in the case of Test 2. This
against Fe in selected Tests 2, 6 and 7. The lower Eh falls in the stability region of FeSO40 in Fig. 1 and
extraction of 45% Fe in Test 2 in 6 h is clearly due to shows the reductive dissolution of ore according to
the fact that the dissolution of iron oxide is caused Eqs. (8) –(12) forming FeSO40, causing the [Fe(III)]/
only by the acid attack according to Eqs. (6) and (7) in [Fe(II)] ratio to drop to values below unity. The main
the absence of SO2. This is further supported by the feature in Fig. 5 is that the Co and Mn extractions
measured Eh of the leach liquors in Test 2, which remain 85 –100% even at lower Fe extractions; this is
were cooled to 25 jC. The measured Eh in the range much higher than the Co and Mn extractions ( < 40%)
0.60 –0.68 V at pH c 1.5 in Test 2 corresponds to the in the absence of SO2 in Test 2. Higher and rapid
stability region of Fe(SO4)2, indicating a ratio of extraction of Mn and Co in the presence of SO2 is a
[Fe(III)]/[Fe(II)] greater than unity (Fig. 1). Although result of the reductive leaching of the type:
the %Ni vs. %Fe extraction corresponds to the line of
slope 0.9, Co and Mn extraction corresponds to lower MnO2 ðsÞ þ SO2 ¼ MnSO04 ð17Þ
slopes of 0.5 –0.8 due to the lower extraction in the
absence of SO2 in Test 2. The mechanism of this reaction has been well docu-
Fig. 5 also shows the % extraction of Ni, Co and mented (Miller and Wan, 1983; Abbruzzese, 1990;
Mn against Fe in the three tests to compare the effect of Grimanelis et al., 1992). High leaching rates at pH 1 –2
SO2 in 0.72 M H2SO4 (Tests 2 and 6) and the effect of have been related to the formation of complexes of
changing 0.72 M H2SO4 to 1 M HClO4 (Tests 6 and 7). Mn(II/III) of the type Mn(SO3)22  and Mn(SO3)2,
The measured Eh of the leach liquors produced in the respectively.
66 G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72

Fig. 5. Correlation between nickel, cobalt, manganese and iron extraction at different time intervals. Comparison between the effect of SO2 (T2
and T6), 0.72 M H2SO4 and 1 M HClO4 (T6 and T7).

Fig. 5 shows that the Ni vs. Fe extraction in Tests 6 hydrolysis and/or the increase in Fe(II) concentration
and 7 corresponds to a line of slope 0.7, lower than in solution:
the slope 0.9 for data in Test 2, irrespective of the
different acids, 0.72 M H2SO4 (in Test 6) and 1 M MnO2 ðsÞ þ 2FeSO04 þ H2 O
HClO4 (in Test 7), but remains above the line of slope
¼ MnSO04 þ Fe2 O3 ðsÞ þ H2 SO4 ð18Þ
0.9. This indicates a higher Ni extraction, compared to
Fe, in the presence of SO2. However, the Ni extraction
MnO2 H2 OðsÞ þ FeSO04 ¼ MnOOHðsÞ þ FeSO4 OHðsÞ
in Tests 6 and 7 starts to drop back to the line of slope
0.9 after the complete extraction of Co and Mn. It is ð19Þ
possible that the decrease in Fe extraction is due to the
MnOOHðsÞ þ 0:5SO2 þ 0:5H2 SO4 ¼ MnSO04 þ H2 O
involvement of the dissolved Fe(II) in the reductive
dissolution of MnO2, and the cobalt associated with it, ð20Þ
in acidic media causing Fe(II) to re-precipitate as
Fe(III) species MnOOHðsÞ þ FeSO04 ¼ MnSO04 þ FeOOHðsÞ ð21Þ

4.3. Catalytic effect of Fe(II) FeSO4 OHðsÞ þ H2 O ¼ FeOOHðsÞ þ H2 SO4 ð22Þ

The catalysis of reductive leaching of MnO2 by 2FeSO4 OHðsÞ þ H2 O ¼ Fe2 O3 ðsÞ þ 2H2 SO4 ð23Þ
H+/Fe(II) takes place via the intermediate MnOOH
(Tekin and Byramoğlu, 1993) leading to the partial 2FeSO4 OHðsÞ þ FeSO04 þ 2H2 O
precipitation of iron as the basic sulphate FeSO4OH or
the oxides Fe2O3, Fe3O4 or FeOOH due to the ¼ Fe3 O4 ðsÞ þ 3H2 SO4 ð24Þ
G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72 67

Some evidence for these conclusions and chemical 4.5. Kinetic models
reactions also comes from the previous studies (Gri-
manelis et al., 1992) with the sulfur dioxide leaching 4.5.1. Initial rate of iron leaching
of a manganese rich pyrolusite ore containing 40.74% Results from the previous studies on initial rate of
MnO2 and 0.84% Fe2O3 at ambient temperatures. The dissolution of goethite failed to agree with the rate
extraction of Mn and Fe at 30 jC in the first 2 min equation:  {dnFeOOH /dt} = constant [H+]n[SO2]m
was 75% Mn and 20% Fe, respectively. In 20 min, Mn with the proposed values of n = m = 0.5 based on an
extraction increased to 85%, which remained constant electrochemical reaction model (Kumar et al., 1993).
over time, whilst the Fe extraction decreased to about The experimental values for m for leaching in 0.3 –
10% during the same time interval due to the re- 0.5 M SO2 at pH 1.1– 2.2 varied between 0.50 and
precipitation of hydrated Fe(III) oxide or an insoluble 0.67; but n showed a much larger variation from 0.17
basic sulfate. Moreover, the ratio of %Mn/%Fe in- to 0.40. This leads to the suggestion that leaching
creased with the decrease in pH indicating that high may also be taking place via acid attack in addition to
acidity led to more Mn dissolved relative to Fe, due to the reductive leaching (Kumar et al., 1993). However,
the catalytic effect of Fe(II). the literature data for the dissolution of hematite
It is also of interest to note that the XRD patterns (Majima et al., 1985) in sulfuric acid in the absence
showed peaks indicating the presence of FeOOH, of sulfur dioxide at 50 –55 jC plotted in Fig. 6 show
Fe3O4, Fe2O3 and SiO2 in the starting material used a slope of 0.67 with respect to acid concentration.
in the present study, but the presence of only Fe3O4 and Therefore, it is important to establish the order of the
SiO2 in the leach residue (Das et al., 1997). This initial leaching reaction with respect to H+, before the
observation supports the formation of Fe3O4 Eq. (24), effect of change in surface area and/or the formation
in contrast to the hydrothermal conversion of goethite of insoluble solids at the reaction interface can be
to hematite in the acid pressure leaching of limonitic considered.
laterite (Briceno and Osseo-Asare, 1995) where there The solubility of SO2 in water and in 1 mol L 1
were no peaks corresponding to Fe3O4 in the leach H2SO4 at 1 atmospheric pressure of SO2 is c 0.3 mol
residue. The formation of FeSO4OH(s) as a solid L 1 and fairly independent of the acid concentration
species in sulfuric acid pressure leaching of laterite (Linke and Seidell, 1958). Therefore, it is reasonable
has not been confirmed due to the extremely fast ki- to assume that the SO2 concentration remains constant
netics of Eq. (23) (Rubisov and Papangelakis, 2000). in the tests carried out in the present study under a
constant and continuous flow of SO2. Moreover the
4.4. Effect of changing H2SO4 to HClO4 Eh– pH diagram predicts that SO2 is the predominant
species (Fig. 1) and that the equilibrium concentration
When the acid was changed from 0.72 M H2SO4 to of HSO3 in Eq. (16) is negligibly small at pH < 2
1 M HClO4 in the present study (Test 6 to Test 7), the (Abbruzzese, 1990). Thus, the slope is close to 1 for
Mn and Co extraction decreased from 100% to 85– the plot of log{d[Fe]/dt} against log[H+] in Fig. 6 for
95% in Fig. 5, although the Ni extraction was unaf- the initial (t = 0 – 0.5 h) leaching in Tests 2– 7, reflect-
fected. This shows the influence of background sulfate ing first-order kinetics with respect to H+. This is in
on Mn and Co extraction, compared to background agreement with the previous studies for the dissolution
perchlorate. It is possible that the formation of stable of a-FeOOH in perchloric acid (Cornell et al., 1976).
sulfate complexes of MnSO40 and CoSO40 (Table 1) as
well as the intermediates such as FeSO4OH(s) and 4.5.2. Shrinking particle (sphere) model for iron
FeSO40 formed in sulfate solutions as indicated in Eqs. leaching
(18) – (24) would favor the thermodynamics and kinet- The kinetic studies for the dissolution of synthetic
ics of leaching of these metals. Thus, the results iron oxides in H2SO4 in the absence or presence of
summarized in Figs. 2 – 5 clearly indicate that the SO2 carried out by previous researchers have provided
overall leaching behavior of laterite is largely con- evidence for a shrinking particle model (Chiarizia and
trolled by the kinetics and mechanism of the dissolu- Horwitz, 1991). Some of the literature data for goe-
tion of iron. thite and magnetite dissolution are summarized in Fig.
68 G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72

Fig. 6. Effect of acid concentration on initial rate of iron leaching. Laterite: Das et al. (1997) and this work (Table 2). Hematite: Majima et al.
(1985) (74 – 104 Am 95.4% Fe2O3).

7 as plots of 1  (1  X)1/3 against time and the rate On the basis of the dissolution behavior of the
constants kss are summarized in Table 2. The value of synthetic goethite and magnetite summarized in Fig.
kss for the dissolution of synthetic FeOOH in 3 M 7, it was expected that the dissolution of iron in the
H2SO4 at 80 jC increases by a factor of 1.4 with the forms of goethite, magnetite and hematite in the
addition of 0.1 M SO2. The dissolution of synthetic limonitic laterite ore investigated in the present study
Fe3O4 in 0.54 M SO2 at 50 jC corresponds to a kss would follow the same trend. However, as reported
value that is 1/100 times less than that for FeOOH at previously (Das et al., 1997), the iron dissolution
80 jC with 3 M H2SO4, due to the lower temperature obeys the shrinking particle kinetic model only during
and absence of added acid (Table 2). the first hour of leaching and the relevant kss data are

Fig. 7. Comparison of kinetic plots for iron dissolution from goethite, magnetite and limonitic laterite: shrinking sphere (particle) kinetic model.
References and conditions described in Table 2.
G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72 69

listed in Table 2. Limited data (Tests 2 and 6) shown reactions with SO2 and Fe(II) described previously.
in Fig. 7 confirm this behavior. Fig. 10 shows that all the results obtained for the
As a matter of interest, the value of kss for the dis- dissolution of iron during the first 4 h in the present
solution of h-MnO2 in 0.5 M SO2 (Tekin and Byra- investigation fit with a shrinking particle model with a
moğlu, 1993) and g-MnO2 in FeSO4 (Miller and Wan, solid product layer, with R2>0.99 for Tests 3 and 6 and
1983) at pH 1– 2 seem to be 1000– 5000 times greater R2 c 0.98 for the others. The kpl values obtained from
than the values for iron reported in this work leading the slopes are listed in Table 2. The product layer may
to faster leaching of Mn compared to Fe. This consist of basic iron sulfate and/or iron oxides (Eqs.
explains why the %Mn and %Co extraction reached (18) –(24)) and quartz, in addition to any other hydro-
100% in Tests 3 in 30 min compared to 7% Fe and lysis products from the gangue.
25% Ni (Figs. 2 and 3).
4.6. Effect of H+ concentration on rate constants
4.5.3. Shrinking core model with product layer for
iron leaching Both H+ and SO2 should be available for the
The results from the two tests T2 and T6 in 0.75 M reductive leaching reactions represented by Eqs. (8) –
H2SO4 in the presence or absence of SO2, plotted in (12). The value of the heterogeneous reaction rate
Fig. 7 do not show a linear correlation for the shrinking constant, kpl increases eightfold from 3.3  10 6 s 1
particle model over the longer leaching period of 6 h. in 0.72 M H2SO4 to 27  10 6 in SO2 + 0.72 M H2SO4
Figs. 8 and 9 compare the plots of 1  (1  X)1/3 and (Table 2), mainly due to the change from acid leaching
1  3(1  X)2/3 + 2(1  X) against t, respectively, for (Eqs. (6) and (7)) to reductive leaching (Eqs. (8) –
the dissolution of Fe, Ni, Co and Mn in 0.72 M H2SO4 (12)). Since the SO2 concentration was maintained
in the absence of SO2 (Test 2) during the first 2 h. constant in the present study, it is reasonable to assume
Clearly, the Fe and Ni extraction in Fig. 9 show an that the increase in kpl with increasing concentration of
excellent linear relationship with correlation coeffi- H2SO4 is mainly due to the change in H+ concentration
cients of R2 c 0.99 and the rate constant kpl = 4.2  in the bulk (Eq. (15)). The value of pKa for Eq. (26),
10 6 s 1 for iron and 6.1  10 6 s 1 for nickel. The
fact that manganese and cobalt do not behave in the H2 SO4 ! Hþ þ HSO
4 ð25Þ
same way as iron and nickel supports the view that the
dissolution of manganese (and cobalt) is a result of the HSO þ 2
4 X H þ SO4 ð26Þ

Fig. 8. Two-hour kinetic data: shrinking sphere (particle) kinetic model for dissolution of Fe, Ni, Co and Mn from limonitic laterite with 0.72 M
H2SO4 in the absence of SO2 (Test 2).
70 G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72

Fig. 9. Two-hour kinetic data: shrinking core model with a solid product layer for dissolution of Fe, Ni, Co and Mn from limonitic laterite with
0.72 M H2SO4 in the absence of SO2 (Test 2).

depends on the ionic strength and temperature and dissolved metal ions and their complexes. If it
(Table 1). The ionic strength of leach liquor changes is assumed that the ionic strength remains close to
with time due to the formation of sulfur species 1, the mass balance for equilibrium in Eq. (26)

Fig. 10. Four-hour kinetic data: shrinking core model with a solid product layer for dissolution of Fe in Tests 1 – 7.
G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72 71

Table 3 approximately first order with respect H+ but


Effect of medium and material on proton diffusion coefficient DH+ appears to obey the shrinking particle model due
Medium/material DH+(cm2 s 1) Reference to the decrease in surface area.
Aqueous solutiona 9  10 5 Bockris and Reddy, 1977 This is consistent with the leaching behavior of
Single-crystal ironb 8  10 5 Bockris and Reddy, 1977 synthetic iron oxides in the absence or presence of
g-MnO2 8  10 8 Allen et al., 1979
SO2, which also follows first-order kinetics and the
Product layer in 0.5  10 9 to This work (Table 2)
laterite leaching 4  10 9 shrinking particle kinetic model.
Fe3O4 8  10 10c Allen et al., 1979 Subsequently, the rate of leaching appears to be
NiO 10 10 – 10 12 Allen et al., 1979 controlled by the slow diffusion of H+ through an
a
At infinite dilution and 25 jC. insoluble solid layer produced during leaching. The
b
Diffusion of eloctronated H+ (atomic hydrogen) into the metal. magnitude of the diffusion coefficient of H+
c
Minimum possible value based on electrochemical studies through the insoluble solid product (0.5109 to
with polycrystalline magnetite.
4109 cm2 s1) is much less than that in aqueous
media, but in the same order as that in solids such
based on the pKa value at I = 1 leads to the linear as Fe3O4 and g-MnO2.
relationship: The dissolution of nickel follows the same trend as

½Hþ ¼ 1:085½H2 SO4 þ 0:0178 ð27Þ iron with an extraction ratio of Ni/Fe = 0.7– 0.9 and
obeys the shrinking core model with a solid product
This can be used to examine the effect of H2SO4 layer. However, the dissolution of manganese (and
concentration on the rate constants. A higher con- cobalt), which reaches 90 –100% in the first 30 min
centration of [H+]bulk would favor the adsorption before the product layer sets in, does not follow the
equilibrium: Solid + H+ X SolidH+(ads) and thus the same trend as iron and nickel due to the direct
rate controlling surface chemical reaction (Eq. (4); reaction with SO2 and the catalytic effect of Fe(II)
Cornell et al., 1976) occurs to improve the rate of produced by the reductive leaching of iron oxide.
dissolution. This would increase kss according to
Eq. (13), as reflected in the data listed in Table 2
Acknowledgements
for Tests 5 –6. Likewise, the increase in [H+]bulk
would increase the rate of diffusion of H+ through
The authors thank Profs. David Muir and Pritam
the product layer to the reaction interface and hence
Singh for valuable discussion during the Targeted
increase kpl according to Eq. (15).
Institutional Research Links Program.
The calculated values of D+H , based on Eq. (15)
listed in Table 2 range from 0.5  10 9 to 4.2  10 9
cm2 s 1. The variation of D+H with the change in acid
References
concentration largely reflects the non-validity of
the assumptions of constant values of r and [H+]bulk
Abbruzzese, C., 1990. Percolation leaching of manganese ore by
used in the calculation of D+H using Eq. (15). Never- aqueous sulfur dioxide. Hydrometallurgy 25, 85 – 97.
theless, the increasing order of D+H shown in Table 3 Allen, P.R., Hampson, N.A., Bignold, G.J., 1979. The electrodisso-
indicates that D+H calculated in the present study is lution of magnetite. J. Electroanal. Chem. 99, 299 – 309.
c 1/10 000 times smaller than D+H in aqueous solutions Azuma, K., Kametani, H., 1964. Trans. Am. Inst. Min. Eng. 230,
853 – 862.
but within the magnitude range reported for the diffu-
Berkman, D.A., 1995. Field geologist’s manual. Australas. Inst.
sion of H+ through solids such as Fe3O4 and g-MnO2 Min. Metall., 21 – 34.
based on electrochemical methods (Allen et al., 1979). Bockris, J.O’M., Reddy, A.K.N., 1977. Modern Electrochemistry.
Plenum, New York, pp. 1328 – 1330.
Briceno, A., Osseo-Asare, K., 1995. Particulates in hydrometal-
5. Summary and conclusions lurgy: Part I. Characterization of laterite acid leach residues.
Metall. Trans. 26B, 1123 – 1131.
Byerley, J.J., Rempel, G.L., Garrido, G.F., 1979. Copper catalysed
At constant [SO2], the rate of leaching of iron from leaching of magnetite in aqueous sulfur dioxide. Hydrometal-
limonitic laterite in the first 30 – 60 min is lurgy 4, 317 – 336.
72 G. Senanayake, G.K. Das / Hydrometallurgy 72 (2004) 59–72

Chiarizia, R., Horwitz, E.P., 1991. New formulations for iron oxides Linke, W.F., Seidell, A., 1958. Solubilities of Inorganic and Metal-
dissolution. Hydrometallurgy 27, 339 – 360. Organic Compounds, 4th ed. Van Nostrand, Princenton, NJ.
Chou, E.C., Queneau, P.B., Rickard, R.S., 1977. Sulfuric acid pres- Majima, H., Awakura, Y., Mishima, T., 1985. The leaching of hem-
sure leaching of nickeliferous limonites. Metall. Trans. 8B, atite in acid solutions. Metall. Trans. 16B, 23 – 30.
547 – 554. Miller, J.D., Wan, R.-Y., 1983. Reaction kinetics for the leaching of
Cornell, R.M., Posner, A.M., Quirk, J.P., 1976. Kinetics and mech- MnO2 by sulfur dioxide. Hydrometallurgy 10, 219 – 242.
anisms of the acid dissolution of goethite (a-FeOOH). J. Inorg. Nicol, M.J., 1983. The non-oxidative leaching of oxides and sul-
Nucl. Chem. 38, 563 – 567. phides. In: Osseo-Asare, K., Miller, J.D. (Eds.), Hydrometal-
Das, G.K., Anand, S., Das, R.P., Muir, D.M., Senanayake, G., lurgy—Research Development and Plant Practice. TMS-
Singh, P., Hefter, G., 1997. Acid leaching of nickel laterites in AIME, Warrendale, PA, pp. 176 – 196.
the presence of sulfur dioxide at atmospheric pressure. In: Coop- Ray, H.K., 1993. Kinetics of Metallurgical Reactions. International
er, W.C., Mihaylov, I. (Eds.), Hydrometallurgy and Refining of Science, New York, pp. 17 – 51.
Nickel and Cobalt, vol. 1. Canadian Inst. Min. Metall. and Rubisov, D.H., Papangelakis, V.G., 2000. Sulfuric acid pressure
Petrolium, Montreal, pp. 471 – 488. leaching of laterite-speciation and prediction of metal solubil-
Gorichev, I.G., Kipriyanow, N.A., 1984. Regular kinetic features of ities ‘‘at temperature’’. Hydrometallurgy 58, 13 – 26.
the dissolution of metal oxides in acidic media. Russ. Chem. Sato, T., Simizu, T., Okabe, T., 1978. The formation of dithionate
Rev. 53 (11), 1039 – 1061. by the reaction of Fe(III) – edta with sodium sulfite Nippon Ka-
Grimanelis, D., Neou-Syngouna, P., Vazarlis, H., 1992. Leaching of gaku Kaishi (3), 361 – 366.
a rich Greek manganese ore by aqueous solution of sulphur Senanayake, G., Muir, D.M., 1988. Speciation and reduction poten-
dioxide. Hydrometallurgy 31, 139 – 146. tials of metal ions in concentrated chloride and sulfate solutions
Higginson, W.C.E., Marshall, J.W., 1957. Equivalence changes in relevant to processing base metal sulfides. Metall. Trans. 19B,
oxidation – reduction reactions in solutions: some aspects of the 37 – 45.
oxidation of sulphurous acid. J. Chem. Soc., 447 – 458. Sillen, L.G., Martell, A.E., 1964. Stability constants of metal com-
Kumar, R., Das, S., Ray, R.K., Biswas, A.K., 1993. Leaching of plexes. Chemical Society Special Publication, vol. 17. Chemical
pure and cobalt bearing goethites in sulphurous acid: kinetics Society, London.
and mechanism. Hydrometallurgy 32, 39 – 59. Tekin, T., Byramoğlu, M., 1993. Kinetics of the reduction of
Kyle, J.H., 1996. Presure acid leaching of Australian nickel/cobalt MnO2 with Fe2 + ions in acidic solutions. Hydrometallurgy 32,
laterites. In: Grimsey, E.J., Neuss, I. (Eds.), Nickel’96 Mineral 9 – 20.
to Market, Kalgoorlie, W.A., 27 – 29 Nov. Australas. Inst. Min. Warren, I.H., Hay, M.G., 1975. Leaching of iron oxide with aque-
Metall., AUSIMM, Melbourne, pp. 245 – 250. ous solution of sulphur dioxide. Trans. Inst. Min. Metall. Lond.
Levenspiel, O., 1972. Chemical Reactions Engineering. Wiley, New 84, C49 – C53.
York.

Das könnte Ihnen auch gefallen