Cobalt powder from Co(OH)2 by hydrogen reduction

Article abstract of DOI:10.1016/S0025-5408(02)00939-X

A study was performed to prepare spherical cobalt powder of about 400 nm from aqueous cobaltous hydroxide slurry under hydrogen reduction conditions using palladium chloride as a catalyst. The reduction kinetics showed good agreement with a surface reaction core model equation, and the activation energy of 35.6 kJ/mol was obtained from the Arrhenius plot at the temperature range of 145-195°C. Additionally, a relationship between the half-life of the reduction, the induction period, and the hydrogen gas pressure was found to be as t_1/2 - t_i = aP_H2^1-n.

Full text of DOI:10.1016/S0025-5408(02)00939-X

Materials Research Bulletin 37 (2002) 2067±2075
Cobalt powder from Co(OH) by hydrogen reduction
2
a,*
Dong-Jin Kim , Hun-Saeng Chung , Kening Yu
a
b
a
Korea Institute of Geoscience and Mineral Resources, Daejeon 305-350, South Korea
b
Institute of Process Engineering, CAS, Beijing 100080, China
(
Refereed)
Received 27 July 2001; accepted 23 August 2002
Abstract
A study was performed to prepare spherical cobalt powder of about 400 nm from aqueous
cobaltous hydroxide slurry under hydrogen reduction conditions using palladium chloride as a
catalyst. The reduction kinetics showed good agreement with a surface reaction core model
equation, and the activation energy of 35.6 kJ/mol was obtained from the Arrhenius plot at the
temperature range of 145±1958C. Additionally, a relationship between the half-life of the
reduction, the induction period, and the hydrogen gas pressure was found to be as
1
H
Àn
.
2
t1=2 À ti ˆ aP
#
2002 Elsevier Science Ltd. All rights reserved.
Keywords: A. Metals; A. Interfaces; B. Chemical synthesis
1
. Introduction
Fine metallic materials with high purity offer new physical and mechanical
properties. Considerable efforts have been put into the chemical preparation of these
powders from aqueous phases. Among various wet chemical methods, the precipita-
tion of metals from metallic salt solutions by hydrogen reduction at elevated pressure
and temperature has been widely used as a convenient technique to prepare ®ne
metallic powders [1,2]. On the other hand, studies on metal powder preparation from
aqueous metallic salt slurry systems have been rare even though the heterogeneous
method is favorable for maintaining equilibrium conditions in controlling the particle
size and shape [3]. Their complex heterogeneous reactions are very complicated,
*
Corresponding author.
E-mail address: djkim@kigam.re.kr (D.-J. Kim).
0025-5408/02/$ ± see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 2 5 - 5 4 0 8 ( 0 2 ) 0 0 9 3 9 - X

2068
D.-J. Kim et al. / Materials Research Bulletin 37 (2002) 2067±2075
because of dif®culties in controlling the system's pH values and in transferring
hydrogen from gas to liquid phases [4,5].
Some studies on the hydrothermal reduction of nickel hydroxide slurry suggest that
the slurry ®rst dissolves under hydrothermal conditions, and then the nickel ions are
reduced to metallic nickel particles by hydrogen [6,7]. The reduction kinetics of their
experiments was in good agreement with a core model equation of solid±gas surface
reaction, and the rate determination step differs in low and high temperatures [8].
The domain of thermodynamic stability of cobalt has a small area in common with
that of water [9]. In this respect, it is very similar to nickel. Like nickel, cobalt metal
ions in a cobaltous hydroxide slurry system can be reduced by hydrogen gas only
when the pH is higher than a certain value, which depends on the speci®c conditions
in the reactor, i.e. temperature, pressure and ion concentrations. The reduction
‡
reaction generates H ions lowering the pH of the system resulting in the cessation
of the reduction if no neutralizing agent is present.
A prerequisite for cobalt powder production is to initially form ®ne particles in the
nucleation stage. Different cobalt metal powders can be produced by varying the
composition of the feed solution and by adding inorganic salts or organic compounds
which catalyze or modify the reduction mechanisms. Known cobalt-reduction
catalysts such as ferrous sulphate or palladium chloride behave similarly in increasing
the number of active sites on the basic cobalt slurry system.
The present work investigates the reduction kinetics of Co(OH) using palladium
2
chloride as a catalyst under hydrothermal conditions to produce ®ne cobalt powder.
The effects of various parameters on the reduction conditions in the formation of sub-
micron sized cobalt powders were studied. The parameters were pH of slurry,
hydrogen pressure, temperature and the use of a catalyst.
2
. Experiment
The hydrogen reduction experiments were carried out in an autoclave equipped
with a 2 l glass lined reactor (10 cm  25 cm H) and an agitator having a 5.8 cm
diameter six-blade marine type impeller rotated at 800 rpm. The reactor is charged
with 500 ml Co(OH) slurry prepared by mixing solutions of CoSO and NaOH
2
4
stoichiometrically, and then a small amount of PdCl was added as an activator and
2
acetic acid as a buffer solution. Table 1 summarizes the standard hydrogen reduction
conditions adapted in the present work. The chemicals used were reagent grade in
purity. Sodium hydroxide was also used as a neutralizing agent.
Once the reaction temperature reached a predetermined level, hydrogen gas
was introduced into the reactor after ¯ushing a few times. The change of hydrogen
pressure with reaction time was recorded to check the reduction fraction. After
completion of the reaction, the cobalt powder was ®ltered, washed and dried under
nitrogen gas atmosphere to prevent the oxidation of cobalt metal. The collected
powder was characterized in chemical composition, particle size, and morpho-
logy while the residual cobalt ion concentration in the ®ltrate was measured

D.-J. Kim et al. / Materials Research Bulletin 37 (2002) 2067±2075
2069
Table 1
Experimental conditions for hydrogen reduction
Factor
Conditions
Temperature
PdCl
145±1958C
2±12 mg/l
1±4 MPa
3.3±9.8
2
PH2
pH of the end solution
2‡
Co concentration
NaOH (stoichiometrically)
Acetic acid
20±60 g/l
110%
0.8 v/o
Fig. 1. Experimental procedure.
colorimetrically. Fig. 1 shows a brief overview of the experimental procedure for
the present work.
3
. Results and discussion
3
.1. Effect of catalyst
In the cobaltous hydroxide slurry system, a catalyst is required to initiate the
hydrogen reduction of cobalt ions at an appreciable rate. Palladium chloride was used
and its effects on the reduction rate were studied. The concentration of palladium
chloride varied in a range of 2±12 mg/l, and its effect on the reduction fraction with
time is shown in Fig. 2. The reduction fraction herein essentially indicates the mol
ratio of precipitated Co metal to starting Co ions. It is seen that the increase in
palladium chloride highly accelerates the reduction of Co ions after an induction
period which is a time lag between the injection of hydrogen gas and the initiation of

2070
D.-J. Kim et al. / Materials Research Bulletin 37 (2002) 2067±2075
2‡
Fig. 2. Reduction fraction at various amount of PdCl
PH ˆ 3 MPa).
2
(Co ˆ 30 g/l, pH ˆ 6, T ˆ 1808C,
2
the reduction reaction. An induction period of about 5±8 min is needed under the
present operating conditions. Fig. 3 shows the reduction rate of Co ions with the
increase in the concentration of palladium chloride. The rate increases linearly up to
<
8 mg/l, and becomes constant at palladium chloride concentrations above it. In other
2
‡
Fig. 3. Reduction rate as a function of PdCl (Co ˆ 30 g/l, pH ˆ 6, T ˆ 1808C, PH ˆ 3 MPa).
2
2

D.-J. Kim et al. / Materials Research Bulletin 37 (2002) 2067±2075
2071
words, the catalytic reaction with an excessive addition of the activator does not show
any signi®cant contribution to the increase in the reduction rate or the decrease in the
induction period.
A serious problem in the hydrogen reduction of solutions is the tendency for the
precipitated metal to plate out on the surfaces of the reactor and to form agglomerates.
A considerable amount of cobalt metal was often observed to deposit on the wall of
the reactor at palladium chloride concentrations under 2 mg/l. This may be because
the reduction in the heterogeneous process is too slow to form cobalt metal
precipitates in solution due to the insuf®cient addition of the catalyst; instead, the
internal surfaces of the reactor act as a catalyst. On the other hand, virtually no
deposition on the wall were found at palladium chloride concentrations above 8 mg/l,
but some cobalt particle agglomerates existed in solution, which likely resulted from
the rapid nucleation and growth of the cobalt precipitates by the excess addition of
palladium chloride.
3
.2. Effect of temperature
The experiments were carried out in the temperature range of 145±1958C and the
kinetic values are plotted in Fig. 4. The initial reduction rates against time at different
temperatures are in good agreement with the rate equation of the shrinking core model
dominated by chemical reaction control for small spherical particles [10]:
1
=3
1
À ꢀ1 À x† ˆ kꢀt À ti†
where x is the reduction fraction, k the rate constant, and t is the induction period.
(1)
i
1=3
2‡
Fig. 4. 1 À ꢀ1 À x†
PdCl2 ˆ 4 mg/l).
as a function of reduction time (Co ˆ 30 g/l, pH ˆ 6, PH ˆ 3 MPa,
2

2072
D.-J. Kim et al. / Materials Research Bulletin 37 (2002) 2067±2075
2
‡
Fig. 5. Arrhenius plot of reduction rate (Co ˆ 40 g/l, pH ˆ 6, PH ˆ 3 MPa).
2
The Arrhenius plots of the reduction rate in the temperature range of 145±1958C
are given in Fig. 5. The apparent activation energies at two different concentrations, 4
and 8 mg/l, of PdCl are 35.6 kJ/mol, and it can be concluded that the reaction rate is
2
chemically controlled.
Therefore, it is certain from the results of both Figs. 4 and 5 that the chemical
reaction is the dominant factor in determining the hydrogen reduction rate in the
cobaltous hydroxide slurry system to produce metallic cobalt powder in the tem-
perature of 145±1958C.
3
.3. Effect of hydrogen pressure
The partial pressure of hydrogen gas in the reactor is directly involved in the
reduction reaction and will affect the rate. The tests were performed with an initial
hydrogen pressure in the range of 1±4 MPa, and with different temperatures, amount
of catalysts, and cobaltous ions concentrations. Fig. 6 shows the experimental values
of t1=2 À t as a function of the initial hydrogen pressure, P , where t is the half
i
H2
1/2
decay period resulting in half of the initial cobalt concentration and t is the induction
i
period in which the initial concentration stays the same. Based on the log(t1=2 À t )
i
versus log PH2 plots which give straight lines, the following relation was found to
be obeyed:
logꢀt1=2 À t † ˆ ꢀ1 À n†log P ‡ log a
(2)
(3)
i
H2
or
1Àn
t1=2 À t ˆ aP
i
H
2

D.-J. Kim et al. / Materials Research Bulletin 37 (2002) 2067±2075
2073
Fig. 6. Reduction time as a function of hydrogen pressure at pH ˆ 6.
where PH2 is the initial hydrogen partial pressure and n and a are constants. The value
of n, the slope of a line, varies between 1.6 and 1.8 at 30 g/l of cobaltous ion
concentration regardless of changes in catalyst concentration and reaction tempera-
ture, but it decreases down to about 1.3 at 15 g/l.
For irreversible reactions in a constant-volume batch reactor, it is known that the
half-life of the reaction, t , as the time needed for the concentration of reactants to
drop one-half the original value can be written as [11]:
1/2
1
Àn
t1=2 ˆ bC0
where C is the initial concentration and b is a constant. Therefore, one can notice that
(4)
0
the experimental Eq. (3) expressing the half reduction period as a term of the initial
partial pressure of hydrogen gas is analogous to Eq. (4) in which a plot of log t1/2
versus log C gives a straight line of slope 1 À n.
0
3
.4. Effect of pH
Fig. 7 shows the reduction rate and the recovery ratio of cobalt metal plotted as
a function of pH of the end solution after completing the Co(OH) reduction.
2
The reduction rate linearly decreases from 2.7 g/(l/min) at pH ˆ 3:2 down to
0
.27 g/(l/min) at pH ˆ 10. On the other hand, as the pH increases, the recovery
ratio rapidly increases up to nearly 100% at pH values of 6±7. Therefore, referring to
the Pourbaix diagram [9], the following reduction processes can be contemplated
depending on pH values.
‡
À
For pH > 6À7; CoꢀOH† ‡ 2H ‡ 2e ! Co ‡ 2H O
(6)
2
2

2074
D.-J. Kim et al. / Materials Research Bulletin 37 (2002) 2067±2075
Fig. 7. Reduction rate and Co recovery at different pH of end solution (Co ˆ 40 g/l, PH ˆ 3 MPa,
2
T ˆ 1808C, PdCl2 ˆ 4 mg/l).
and
‡
2‡
2‡
À
For pH < 6À7; CoꢀOH† ‡ 2H ! Co ‡ 2H O;
Co ‡ 2e ! Co
7)
2
2
(
In the high pH range, the direct reduction from Co(OH) to Co proceeds at a slow
2
rate, but the recovery becomes high. This is a necessary condition to slow down the
metal particle growth at maximum recovery. On the other hand, Co ions which are
Fig. 8. SEM micrograph of Co powder.

D.-J. Kim et al. / Materials Research Bulletin 37 (2002) 2067±2075
2075
dissolved in low pH is reduced with a high rate at ®rst and then the lowered pH of the
end solution limits the reduction resulting in low cobalt metal recovery.
A scanning electron microscopic photograph shown in Fig. 8 shows the cobalt
powder produced from the cobaltous hydroxide particles generated from sulfate
solution. As observed, the cobalt particles are spherical in shape with an average of
4
00 nm in size.
4
. Conclusion
Hydrogen reduction was performed in the preparation of ®ne cobalt powder from
cobaltous hydroxide slurry under hydrothermal conditions. It was found that the
reduction rate was mainly affected by the amount of catalyst, hydrogen pressure,
temperature, and the resulting pH. The cobalt powder obtained was spherical in shape
with a particle size of about 400 nm.
The reduction kinetics in the temperature range of 145±1958C were well ®tted to a
1
=3
core model equation, 1 À ꢀ1 À x† ˆ kꢀt À t †. The activation energy obtained from
i
the Arrhenius plot was 35.6 kJ/mol at the temperature of 145±1958C. In addition, the
half-life for the reduction started after the induction period was proportional to a
fractional power of 1 À n of the initial hydrogen gas pressure.
References
[
1] W. Kunda, et al., in: H.H. Hausner (Ed.), Modern Developments in Powder Metallurgy, vol. 1,
Plenum Press, New York, 1966, pp. 15±49.
[
[
[
2] K.M. Sista, C.M. Sliepcevich, Metall. Trans. B. 12B (1981) 565±568.
3] F. Habashi, Principles of Extractive Metallurgy, vol. 2, Science Publishers, 1969, pp. 252±274.
4] R. Derry, R.G. Whittemore, in: Proceedings of the 2nd International Symposium on
Hydrometallurgy, Chicago, USA, 1973, pp. 42±62.
[
[
[
5] T.R.N. Kutty, et al., Mater. Sci. Lett. 1 (2) (1982) 67±70.
6] K. Matsuda, H. Majima, J. Min. Met. Jpn. 49 (1118) (1981) 257±262.
7] K. Yu, et al., in: Proceedings of the 1st European East±West Symposium on Materials and
Process, Finland, 1990, pp. 10±18.
[
[
8] N. Yamasaki, H. Liang, Metall. Trans. B 24B (1993) 557±561.
9] M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous solutions, National Association of
Corrosion Engineers, 1974, pp. 322±329.
[
[
10] F. Habashi, Kinetics of Metalllurgical Process, Metallurgie Extractive Quebec, 1999, pp. 149±
53.
11] O. Levenspiel, Chemical Reaction Engineering, Wiley, New York, 1972, pp. 53±54.
1