War. P.es.Vol.27, No. 6, pp. 985-992, 1993
Printed in Great Britain.All rightsreserved
0043-1354/93 $6.00+ 0.00
Copyright © 1993 PergamonPress Ltd
COBALT(II) REMOVAL FROM WATER BY CHEMICAL
REDUCTION WITH SODIUM BOROHYDRIDE
C. GbMEZ-LAHOZ, F. G A X C i A - H ~ u z o , J. M. RODRIGUEZ-MAROTO and
J. J. ROD~OUF.Z. 0
Departamento de Ingenieria Quindca, Unlversidad de Mklaga, 29071-M~laga, Spain
(First recewed March 1992; accepted in revised form December 1992)
Almract--The stoichiometry and kinetics of Co '+ reduction with NaBH4 in aqueous solution have been
studied. A sutntantial excess of borohydride is needed to achieve complete Co'+ removal from water,
mainly due to borohydride consumption in the BI-14-/H,O reaction unless the pH is snlf~ently alkaline.
In thl, case, cobalt precipitation occurs mainly as Co(OH),. Absence of dis~lved oxygen in the reactor
is required to avoid rapid reoxidation of reduced cobalt. The elemental analysis of the reduced g ~ e s
is cottsistent with Co2B. The ease of sludge reoxidation and the final boron concentration in the effluent
resnlfing from the needed excess of borohydride are serious disadvantages of this technique for Co2+
removal. A kinetic model has been proposed to describe the behaviour of the system. The model
reproduces reasonably well the experimental results, although some discrepancies were observed in the
early ~ages of the process. These can be mainly attributed to mass-transfer fimitations.
Key words----cobalt(ll),borohydride, cobalt boride, chemical reduction, precipitation, kinetic model, redox
reactions
NOMENCLATUME
G = growing factor of ks [defined in equation (10)]
/ ~ .ffirate of H + production from boric acid dissociation
h,, •rate of H + production from water dissociation
/fffiffifirst dimociation constant for boric acid [defined in
equation (1 l)]
K, = solubility product of Co(OH), [defined in equation
(12)]
K,, =water ionic product [defined in equation (13)]
k ~ - kinetic comtant [defined in equation (6)]
k,-kinetic comtaat [de41,~d in equation (6)]
k~ = kinetic constant [defined in equation (7)]
k, -kinetic oomUmt [defined in equation (7)]
k s = kinctk constant [defined in equation (8)]
ks - kinetic mau tmmfer coefficient for oxygen [defined in
equation (9)]
k~ = kinetic mare tmmfer coefficient for oxygen in alnence
of chemical reaction [defined in equation (10)]
oh. = rate of OH- production from Co(OH)2 dissociation
Greek letters
= etfu:iencyof borohydride use (mole of Co '+ removed
per rook of NaBI-I~ added)
,= rate of cobalt boride reoxidation following equation
(5)
~b=ffirate of borohydride consumption following equation
(2)
f~ = rate of cobalt(lI) removal following equation (4)
dards. The subsequent sludge handling and disposal
involves some problems due to the high leachabifity
of these substances. Alternative methods are being
developed which focus more on detoxification of
sludges or on waste avoidance or minimization by
metals recovery, rather than on achieving a higher
quafity effluent. Of course, this latter would also be
welcome. Two of these recovery technologies are ion
exchange and reverse osmosis, both of which can
reach high quality cfltuents and a concentrated
stream that in some cases may be reused in the
production cycle. Also, reductive electrolysis is
used to recover spent metals from wastewatc~, but
capital costs may be high and high concentrations are
required for the process to work at acceptable
cfflciencies.
Chemical reduction and precipitation has low capital costs and is easily done. It results in ion reduction,
usually to the metallic state, with a marked improvemerit in physical and chemical sludge characteristics.
Sodium borohydride is a mild but very effective
reducing agent, with a low equivalent weight (eq.
weight ffi 4.7 g/equiv.) and a high reduction potential
(E e ffi - 1 . 2 3 V). The general reaction for reduction
of divalent metallic ions to the metallic state is:
INTIODUCTION
BH~" + 4 M e a+ + 3 H~O--, H3BO3 + 4 Me° + 7 H +
Conventional alkaline precipitation as an end-of-pipe
metala control technology for industrial wastewaters
treatment before effluent discharge performs reasonably well in most cases in achieving regulatory stan-
(1)
Table 1 shows the theoretical weight ratios that
may be achieved from equation (!). When other
oxidants are present borohydrid¢ will also be consumed in competitive reactions involving these
species. Most of these, like dissolved oxygen, may be
*Author to whom all correspondence should be addressed.
~/~--z
985
C. ~ - L A H O Z
986
Table 1. Theoreticalweight ratios of reduced
metalobtainablefromionicspecieswith sodium
borohydrkiebased on reaction (!)
Oxidationstate
Weight ratio*
Cd2+
' 12
Co'+
6
•Cu 2+
7
AS+
Pb2+
Hg2.
Hg+
23
22
21
42
Ni 2+
6
Cobalt(ID does not follow equation (I) when reacting with NaBH4, but forms a boride of mixed stoichiometry, generally expressed as Co2B (Heizman and
Ganem, 1982; Osby eta/., 1986). The reaction may be
written as:
BH4- + 2 Co2+ --, Co, B + 1/2H2 -I- 3 H +
EXP~M~
previously removed with cheaper reducing agents
(i.e. sodium dithionite). However, in aqueous systems
one major problem will arise from the reduction of
water by borohydride given by reaction (2). The
kinetics of this reaction, at p H values below 9,
follows equation (3) where the preexponential factor
is g i v e n i n (l.mol -L s-I), according to literature data
(Levine and Kreevoy, 1972; Kreevoy and Hutchins,
1972).
d[BHi ]
d ~ -- 6.62 × 1012e,<'~°/~[H+][BH~" ]
(2)
(3)
Work on heavy metals removal from water e/fluents using sodium borohydride has been reported in
the literature, and includes industrial case histories.
According to those reports the technique is reliable at
controlled neutral/alkaline p H values for precious
and noble metal ions like silver, mercury, copper and
lead (Rosenzweig, 1971; Cook et aL, 1979; Cushnie,
1983; Lindsay e t a / . , 1985; Patrick et al., 1987)
although thcre is still some lack o f fundamental
information regarding the borohydride/metal systems. This lack is more substantial in the case of other
heavy metals with higher oxidation potentials. Ying
et al. (1987) has reported interesting work on nickel
and C o o k st a/. refers to the removal o f cadmium in
a silver cadmium effluent. Nevertheless, further investigation is required with regard to these types of
metals to establish the applicability o f the method.
The present focuses on the study of the stoichiometry
and kinetics of Co 2+ reduction with sodium borohydride. The experimental conditions are restricted to a
range o f initial pH values in which no precipitation
of Co 2+ as hydroxide takes place.
The redox potential of the Co2+/Co ° pair
(E ° : - 0 . 2 8 V) makes a general knowledge of the
behaviour of the Co2+/BH,- system of interest. Previous work on Cu 2+ (EO=0.34V) revealed some
limitations of the borohydride reduction technique
(Gomez-Lahoz, 1991; Gomez-Lahoz et al., 1992),
although this technique is effective with Cu '+. On the
other hand, when this method was tested to reduce
Zn 2+ (Ee - - 0 . 7 6 ) a rapid reoxidation o f the reduced
species of zinc by water was observed which prevents
the use o f the technique with zinc (Gomez-Lahoz,
1991).
(4)
and the theoretical weight ratio for cobalt reduction
by this reaction is one half that given in Table 1.
*Weightratio = maximumg of metal redaced/g
of NaBH4.
BI-14- + 3 H20 + H+ --, H3BO3 + 4 H2
et al.
Simulated wastewaters were prepared using deionized
water with conductivity values ~ 5/~mhce/cm. The necessary amount of a previously prepared solution of
COC1,.6H20 containing 1000mgCo2+/l wag added to obtain the desired initial concentrations of cobalt(lI). Dilute
HCI or NaOH were added to fix the initial pH in each
experiment.
Merck (powder, for synthesis) sodium borohydride was
used. Since sodium borohydride is highly hygroscopic, it
was kept in a desiccator, taking, for the experiments that
were performed with solid borohydride, amounts of approximately 10g. These were placed in a second desiccator,
from which was weighed the necessary amount for each
experiment. Stabilized water solutions of sodium borohydride were prepared with different concentrations of NaOH,
and maintained at temperatures below 10°C. The reducing
capacity of the borohydride solution was measured immediately before use by the Lyttle method (Lyttle et al., 1957).
This analysis was also used for solid bomhydride. Borohydride with a reducing capacity ~ 97.5% the theoretical was
always used.
The experimental runs were carried out batchwise in a
2.351 cylindrical bottom-rounded glass reactor provided
with a glass anchor-like .stirrer. Redox potential and pH
were continuously registered by means of Cryson probes
and pH meters, connected to a 2-channel resister. All the
redox values reported in the following sections are referred
to the Aa/AgCI/KCI 0 M) electrode and ex_pee,_~__~_____as
oxidant potentials. Dissolved oxygen was measured by
means of a Hanna Instruments HI8543 oxymeter. The
reactor was immersed in a thermostatic water bath where
temperature was controlled within +0.1°C by means of a
Tectron apparatus. Stirring velocity was set at 450 rpm after
preliminary experiments. Higher stirring velocities were
found to produm instabilities in the pH probe.
Sample, were taken by means of a Munmtat three-way
valve syrinse, or a peristaltic pump, connected to a Millipore filtration system, with membranes "GSWPO2500" of
0.22 #m pore size. This system was chosen after previous
experiments with cellulexic filters, where Co2+ adsorption
was observed, and with fritted-glass filters of 10-20 #m pore
width, which showed filtering rates sub6tantially lower than
needed to obtain the desired 10-20 samples per minute.
Cobalt analyses were performed by atomic absortion spectrophotometrY (AA) with a Varian AA-475 instrument. The
precipitate was dissolved in HNO3 and analyaxl for cobalt
and boron by AA and the cumumine method (APHA,
1975), respectively.
ItESULTS AND DISCUSSION
The Pourbaix diagram for cobalt at 10 and 40 ms/l
is presented in Fig. 1. As indicated before, the
r e d u c e d species o f cobalt with NaBH4 has been
reported in the literature as a boride of mixed stoichiometry, generally expressed as Co2B. This stoichiometry is in agreement with the cobalt and boron
Cobalt(II) reduction with NaBH4
c~om,
M
CO"
-
2
4
6
8
,
10
.
.
12
-
14
pH
Fig. I. Pourbalx d~a~-am for cobalt: [Co2+] of 40 (. )
and 10 (...) mg/l. Oxidation potentials referred to the
Ag/AgCI/KCI(3M)electrode.
analyses of our precipitate. Nevertheless, thermodynamic data are not available and this species has not
been included in the diagram. More recently Kim and
Brock (Kim and Brock, 1987) studied the precipitate
by ICP-AES and XPS and concluded that it consists
most probably of a metallic cobalt/boron admixture
rather than a metal boride. The redox potential
values for precipitate formation observed in our
experiments fails in the vicinity of that corresponding
to the formation of metallic cobalt. Thus the Pourbaix diagram given in Fig. I may be reasonably valid
for practical purposes. Accbrding to it a redox potential of - 6 0 0 m V is needed to accomplish Co2+
reduction. The H2/H + equilibrium line shows that
Co o may be reoxidized by H20 at pH lower than 7.
Other oxidant species, such as dissolved oxygen, may
undergo reaction with borohydride as well as with
the reduced cobalt species. Previous experiments
allowed us to neglect the contribution of the
NaBH4/O 2 reaction due to its slow rate (GomezLahoz, 1991; Gomez-Lahoz et al., 1992). Nevertheless, the reoxidation of reduced cobalt by dissolved
oxygen has to be considered. This reaction can be
written as:
987
hausted, Co 2+ reduction takes place rapidly, giving
rise to a sharp decrease in CO2+ concentration, pH
and potential values. A blackish precipitate was
clearly observed in the reactor at this stage. The Co 2+
concentration reaches a minimum and subsequently
an increase is observed, coincident with a similar
trend in the pH and potential curves. Borohydride
exhaustion in the aerated reactor leads to a predominance of cobalt reoxidation. The Co2. concentration
stabilizes finally, due to insoluble Co(OH)2 formation. Analogous trends were observed in a run
performed under similar operating conditions but
using higher Co2+ and borohydride doses (760 and
1420/~mol/1), respectively). The initial induction
period for Co 2+ removal was substantially shorter,
consistent with the higher cobalt and borohydride
concentrations used in this experiment.
An experiment fairly similar to the one reported in
Fig. 2 was carded out in an oxygen-free stoppered
reactor. The results are shown in Fig. 3. In the
absence of dissolved oxygen Co 2+ reduction takes
place rapidly. The initial stages of the pH curve,
which were verified to be oscillatory, can be explained
as a result of mass-transfer limitations occurring
during the early period after the borohydride was fed.
This was confirmed by flow-pattern experiments
using NaOH and solid NaBH4 (Gomez-Lahoz,
1991; Gomez-Lahoz e t a / . , 1992). The initial pH
increase in Fig. 3 is attributed to a high local excess
m.
.10
s.
3 so.
s
m\
~t" m.
~ zo
*
.'i"
•
""
4~
2
m
-
.
-
~0
.
~11
.
.
1100
-
a00
..F.
.
400
s
-
.
S01}
-
.
~0
t(.)
Co2B + 7/4 02 + 4 H +
2CO2+ + H3BO3 + 1/2 H2O (5)
S
Figure 2 shows the variation of CO2+ coucentration, dissolved oxygen, pH and redox potential
with time in an experiment performed in an unstoppered stirred reactor using excess borohydride. There
is an initial period in which no cobalt precipitation
takes place, Most likely, reduced cobalt is undergoing
immediate reoxidation, as suf&ested by the dissolved
oxygen decrease during this period. The increase of
pH durin8 this stage can be explained by the borohydride/water reaction and cobalt reoxidation
[equatiom (2) and (5), respectively]. This initial
period is followed by a smooth decrease in dissolved
Co '+, most probably due to Co(OI-l)2 formation,
given the pH values. As dissolved oxygen is ex-
s.
aoo
!00
..100
g~ 7,
-aoo ~
~o..o-"
s.
6
..600
~t
**-'g
| ..'
-
.
zo
-
.
m
.
,....
m
t(,)
j,,
a
~0
~o
.700
Fi$. 2. [Co=+], [02], pH (
) and redox potential (...)
evolution with time in emtoppe~ reactor at 20°C
and 4.S0rp~" p i l e - 6.0;, [Co=+]e-- 47$ #tool/l; [NaBH4]o
900 Izmol/l, -___,'],~J_as solid.
C. C_rbM~-LAI~Z et al.
988
pHs would be needed, but in that case a Co(OHh
precipitate would be obtained.
99.
Figure 5 shows the results obtained in an experiment carried out under operating conditions fairly
similar to those reported in Fig. 4 but in which
NaBH4 was added as an alkaline stabilized solution.
The efficiency of borohydride use (e = mole of Co 2+
removed per mole of NaBI-I4 added) after the two first
s.
additions is now e = 0.55 vs e ffi 0.45 when one uses
• OHiO HD
•
•
sofid N a B I L . This improvement is substantially less
0
e~
significant than the observed for copper (Gomezt (s)
Lahoz, 1991; Gomez-Lahoz et al., 1992). There the
use of stabilized alkaline solution showed an important advantage over solid NaBH,, allowing one to
9oo
increase ~ from 0.84 to 2. The NaOH concentration
-o
in the stabilized borohydride solution was found to
f
be the main factor in efficiency improvement, due to
7'
the inhibition of the borohydride/water reaction in
the early stages following NaBI-I4 feed. In the case of
,.4oo r~
cobalt we used NaOH/BI-14- molar ratios in the
stabilized solution that were 10 times lower than
..e0o
those used for the copper experiments to avoid
Co(OH), formation prior to Co 2+ reduction. This
4110
problem did not arise in the case of Cu 2+ due to the
o
~
,~
~
~o
higher initial rate of copper reduction.
t (8)
Although pH is a very important variable with
Fig. 3. [Co2+], pH (- ) and redox potential (...) evolregard
to the process studied, initial pH showed no
ution with time in oxygen-free reactor at 20°C and 450 rpm;
pH0 ffi6.0; [032+]0 ffi420 pmol/l; [NaBH4]0= 875/~mol/l, significant effect on borohydride efficiency within the
added as solid.
~ 4 D ~
. . . . . . . . . . . . . . . . . . .
e . . . .
. . . . . . . . . . . . . .
ad
laad.
of borohydride leading to predominance of the
BI-I4/I-IsO reaction. As the local BH4- concentration
decreases and pH increases this reaction slows and a
sharp drop in the pH curve is registered consistent
with Co '+ reduction. Further water reaction with
excess borohydride once the Co '+ concentration
reaches a low level leads to a pH increase.
To establish the BI-I;/Co'+ molar ratio needed to
achieve complete Co 2+ reduction we performed several experiments at the operating conditions described in Fig. 3 but using different NaBH4 doses. In
these runs the reactant fed was partitioned into
several additions. We previously confirmed that the
final results were not dependent on the number of
additions of borohydride but only on the total
amount fed. A representative example of these experiments is reported in Fig. 4. The results obtained
allow one to conclude that a molar ratio close to 2.4
was necessary for complete Co '+ removal; this ratio
means about 9.5 times more borohydride than the
theoretical dose calculated for Co '+ reduction to
metallic cobalt given in Table 1. Nevertheless, as
indicated before, Co 2+ reduction follows equation (4)
forming COsB. With respect to the stoichiometry of
this latter reaction the required BFL-/CO'+ molar
ratio represents a borohydride excess of 380% above
the stoiehiometric requirement. This is due to the
occurrenos of the borohydride/water reaction to an
important extent. To avoid this, rather alkaline
rd
acL
SKL
961 t
o~0
m"
5'
.
L
-
0
t (.)
400
9
~0
s'
o
7¸
.....,/--
8
5
,1
L........
'.4111 M
'.~10
-1~
o
t (a)
FiB, 4. [CoZ+], pH (
) and n~dox pot~t~d (...) ~oiution with time in oxygen-free reactor at 20°C and 450 rpm
for 3 additions of solid NaBIL; pile-6.0;
[Co2+]0 m 415 pmoi/l; [NaBH4]0-ffi325, 355 and 285 pmol/l.
Cobalt(II) reduction with NaBH4
30
~md.
989
this, the rate of Co 2+ consumption from reaction (4)
may be expressed by:
-[dtC°2+] 7
= k, [BH~-liCe2+]
L dt _]~4)
fi----
+ k2[BH~-][Co2+][Co2B]
",-
,
o
.
,
e0
.
,
~
(6)
The rate of borohydride consumption from the
borohydrid©/water reaction might be written as:
.
,
tm
.
,
~O
~
.
~0
380
f
[dimh-] ]
n
a0
480
= k, tBH_]IB÷ ]
¢ -L~j=,~2>
=
t (S)
+ k4tBH~'][I-I+][Co2 B]
In an unstoppered reactor, Co2B reoxidation by
dissolved oxygen was found to be significant, causing
o
9
the induction period seen in Fig. 2. The rate of Co2B
consumption by reaction (5) may be expressed by:
'-~00
S¸
1J ~ , ~ = kdCo2B][Od
. = - l ~
m
?
(7)
(8)
..49oo
Oxygen is absorbed from air at a rate given by:
e"
..800
-
0
,
e0
.
•
1,510
.
•
.
m0
•
~0
.
,
N)0
-
°
~0
-
•
~
-
d[O2] 1
-looo
480
t (s)
Fig. 5. [Co2+], pH (
) and rcdox potential (...) evolution with time in oxygen-free reactor at 20°C and 450 rpm
for 3 additions of 310 ~mol/l NaBH4 and 64 pmol/1 NaOH:
each, as stabilized water solution; pHo ffi 6.0;
[Co2+ ]o = 485 pmol/l.
:
dt _1~,
= k6([O2Lt - [02])
(9)
An experimental vahm of 1.25 x l0 -3 s -~ has been
obtained for ke at the stirring velocity used in all the
It
30
i
nd
mt
rd
2aKi.
3..,~
P.
35"
mildly acidic v a l e s investigated in this work, as
can be seen from Figs 5 and 6. In fact when one
starts at pH 5 the expected additional BH~- consumption to reach pH 6 should be no more than
10 ~mol/l.
The effect of temperature can be observed from
Figs 5 and 7, where all other operating conditions
are identical. A decrease in the initial Co 2+ reduction
rate can be noticed when temperature is lowered
from 20 to 10°C, giving rise to an induction period
of about 40s after which the results obtained
become fairly similar to those at the higher
temperature.
Increasing the initial Co 2+ concentration to
~,40m8/1 showed little effect on borohydride
efficiency, as can be seen from Figs 5 and 8. The
resulting ~ values for the first addition are 0.60 and
0.66 respectively. The pH curve in Fig. 8 shows
somewhat lower stabilization values consistent with
the higher cohalt(II) concentrations.
Model proposal
The reduced cobalt species, Co2B, has been
reported in the literautre to catalyze the borohydride/
water reaction (Schlesinger, 1953; Kanfraan, 1985).
Our experimental resulta suggest as well an autocatalyti¢ effect of Co2B on Co 2+ reduction. According to
tin
lull
G'
0
. . . . . . . .
e0
120
180
~0
E-300
M0
t (z)
9'
"~0
8'
.0
7'
-40o
6'
"-~0
G'
"-800
G0
1~
tOO
~0
$)0
M0
420
t(z)
Fig. 6. [Co2+], pH (
) and redox potential (...) evolution with time in oxygen-free reactor at 20°C and 450 rpm
for 3 additions of 310/Amol/l NaBI-I~ and 64 pmol/l NaOH
each, as stabilized water solution; pHe ffi 5.0;
[Co2+ ]o = 485/~mol/l.
99O
C. G O t ~ . I . 4 c ~ z et al.
ad
The concentration values in equations (!1)-(13)
are calculated from the following relationships:
~0
~.==.
[H+],+a~ = [H+],+ {2tf~- ~ - 4 / 4 + h . + h,,} At
(14)
q,
16'
[OH-I,+~ = [OH-l, + {~, + oh,} At
(15)
[H2BO~"],+at = [I-I2BO~-]t + h.~ At
(16)
alto
5'
(I0
110
18)
~0
m0
am
[H3BO3l,+at = [H3BO3], + {# + p - h . } ~ t
(17)
[Co2+],.a~ffi [Co2+], + {2~ - f~ + ½oh.} At
(18)
4m
t (s)
in,
":i00
9'
"0
where h~ and h. represent the rate of H + production
following equifibria (11) and (13) respectively and oh,
is the rate of O H - production according to equilibrium (12).
The following values were obtained for the kinetic
constants used in the model:
S,
kl = 51 mol-I s-I
~7,
"-tOO M
k2-- 7.5 × 1081Zmol-2s-J
6'
'-e00
k3 = 7.7 x l0 s I tool -I s -l
5'
4100
k 4 ---- 10s 12tool -2
-1~00
k s - - 5 x 10~lmol-I s -t
4
0
eO
1,t0
~i0
~
~
~
4IS0
s-I
k* ffi 1.25 x 10 -3 s -1
t (a)
G-- I(P I tool -I
Fig. 7. [Co2+], pH (
) and redox potential (...) evolution with time in oxygen-free reactor at 10°C and 450 rpm
for 3 additions of 310#mol/l NaBH4 and64/~molp NaOH
each,
as
stabilized water solution;
[Co2+ ] 0 : 485 pmol/l.
pHo ffi 6.0;
runs reported in this work (450rpm), using previously N2-stripped distilled water. As oxygen is being
consumed in a fast chemical r e a s o n , an increase in
the mass coefficient should be expected and it may be
represented by:
k6 : k~(l + G[Co2BD
m
v l 0w
~
to
(10)
1~
where k* is 1.25 x 10-3s -~, as determined experimentally and G represents a growth factor.
The model described by equations (6)-(10) has
been integrated using the fourth-order Runge--Kutta
method. A Turbobasic program was written and run
on a microcomputer equipped with an 80286 microprocessor and running MS-DOS. A variable time
increment is implemented in the program and at each
of the four integration steps in the Runge--Kutta
method the following equilibrium equations are
satisfied:
pl2aO; ][I-I+]
/~ffi
P/3BOd
(11)
Ks = [ C o 2 + ] [ O H - ] 2
(12)
/ ~ : [H+][OH -]
(13)
where the corresponding values of K~, K, a n d / ~ are
well known.
I
•
m0
:1 ].
m
ao
t(s)
g
8"
'0
7~
8~
m
i+
5"
o
,~
~
~o
lO0O
t (s)
FiS. 8. [Co:+ ], pH (
) and redox potential (...) evolution with time in oxypn-free reactor at 20°C and 450 rpm
for 3 additions of 310/4mol/i NaBH, and 664 p mol/l NaOH
eael~ as stabilized water solution; pHo ffi 6.0;
[Cu2+]offi 705 #mol/l.
991
Cobalt(ll) reduction with NaBH4
I0
It
8
'4 O"
nd
9ad
lui
rd
sad
g5
~
10.
'2
0
100
99O
5"
0
l
0
t(s)
~)
0
120
la0
~g0
t (s)
9'
7 ¸
t=
s~
7
6
6-
S"
4
o
4
-
•
-
"
.
"
•
"
•
"
t (s)
Fig. 9. Results obtained using the model at the experimental
conditions of Fig. 2.
Figures 9 and 10 show the results obtained from
the model using the initial conditions of the experiments reported in Figs 2 and 4, respectively. As can
be seen, the model reproduces fairly well the Co 2+,
pH and dissolved oxygen curves obtained in the run
performed in an unstoppered reactor (Figs 2 and 9).
Regarding the experiments carried out in an oxygenfree stoppered reactor, good predictions of the final
Co 2+ concentration and pH values are obtained. The
discrepancies observed during the early stages after
each borohydride addition may result from the
assumption of instantaneous homogeneity of concentration made in the model, whereas mass.transfer
limitations were found in practice as indicated before.
Thus, more sophisticated experiments would be
required to follow the behaviour of the system within
t h e initial stages of the process.
CONCLUSION
Some conclusions of practical interest can be
pointed out from the results obtained: Complete
reduction of Co 2+ with this technique requires a
substantial excess of borohydride. Absence of dissolved oxygen in the reactor is needed to avoid
reoxidation of the reduced cobalt. Optimization of
borohydride efficiency in Co 2+ removal would require
a control of pH above the neutral point. Although a
falHy well settleable sludge is obtainable, it becomes
easily oxidizable giving rise to lixiviation problems.
This, together with the final boron concentration in
t (s)
Fig. 10. Results obtained using the model at the experimental conditions of Fig. 4.
the effluent, derived from borohydride excess, limits
seriously the utility of this method as a competitive
technique for Co 2+ rer,-1oval.
Acknowledgements--The authors want to acknowledge the
Spanish CICYT which provided financial support for this
work through Project PA 86/0088 and was granted to
C. Gomez-Lahoz. We appreciate the helpful advice of
Professor David J. Wilson of Vanderbilt University in the
revision of the mamtseript. We also acknowledge Ventron
Thiokol for kindly providing technical information.
APHA, A W W A and W P C F (1985)Standard Method~ for
the Examination of Water and Wastewater, 16thedition,
pp. 274-275. American Public Health Association,
American Water Works Association and Water Pollution
Control Federation, Wuhington, DC.
Cook M. M., Lander J. A. and Littlehale D. S. 0979) Case
histories: reviewing the use of sodium borohydride for
control of heavy metal discharge in industrial wastewaters. Proc. 34th Industrial Waste Conference, Purdue
University, 8-10 May 1979, pp. 514-519.
Cmhnie O. C. Jr (1983) An lrwestigation of Technologiesfor
Hazardous Sludge Reduction at AFLC Industrial Waste
Treatment Plants; Vol. I: Sodium Borohydride Treatment
and Sludge Handling Technologies. CENTEC Corporation, Reston, Va.
Gomez-Lahoz C. (1991) Eliminacion de Metnies Resados de
las Aguas pot Reduccion Qulmica con Borohidrudo de
Sodio. Ph.D. thesis, Univereidad de Malaga.
Gomez-Lahoz C., Garcia-Herruzo F., Rodrigucz-Maroto,
J. M. and Rodrigucz J. J. (1992) Copper removal from
water by chemical reduction with sodium borohydride.
Selm Sci. Technol. 27, 1449-1468.
992
C. GbUEZ-LAHOZ et al.
Heizman S. W. and Ganem B. (1982) The mechanism of
sodium bomhydride-cobaltous chloride reductions.
J. Am. Chem. Soc. 104(24), 6801-6802.
Kaufman C. M. lind Sell B. (1985) Hydrogen generation by
hydrolym of sodinm tetrahydroborate: effects of acids
and transition meta~ and their salts. J. Chem. Soc. Da/ton
Trons. 2, 307-313.
Kim S. G. and Brock J. R. (1987) Growth of ferromagnetic
pastic,les from cation reduction of borohydride ions.
J. Colloid Interface Sci. 116, 431-443.
Kreevoy M. M. and Hutchins J. E. C. (1972) H2BH 3 as an
intermediate in tetrahydridoborate hydrofisis. J. Am.
Chem. Soc. 94, 6372-6376.
Levine L. A. and Kreevoy M. M. (1972) Solvent isotope
effects on tetrahldridoborate hydrolysis. J. Am. Chem.
Soc. 94, 3346-3349.
Lindsay M. J., Hackman M. E. and Tremblay R. J. (1985)
Sodium borohydride reduces hazardous waste. In Proc.
40th Industrial Waste Conference, Purdue University,
14-15 May 1985, pp. 477-482.
Lyttle D. A., Jemen E. H. and Struck W. A. (1957) A simple
volumetric assay for sodium borohydride. Analyt. Chem.
1843-1844.
Osby J. O., Heinzman S. W. and Ganem B. (1986) SUglies
on the ~
of tranmtion-metal-uaisted radium
borohydride and lithium hydride reductions. J. Am.
Chem. $oc. I I 67-72.
Patrick G. C., Hockenbury M. R., Thompmn W. T.
and Begavich J. M. (1987) Development and dmign
criteria for treatment of metal-.cont_ainins Wllstewgtergat
Oak-Ridge National Laboratories. In Proc. 4~¢d Industrial Waste Conferences, Purdue University, 12-14 May
1987, pp. 819-830.
Rosenzwei8 M. D. (1971) Paring mercury pollution. Chem.
Eng., 78 : S (Feb. 22), 70-71.
Sclde~inger H. I., Brown H. C., Finhoit A. E.,
Gilbreath J. R., Huekstra H. R. and Hyde E. K. (1953)
Sodium borohydride, its use as a reducing agent and in
the generation of hydrogen. J. Am. Chem. Soc. 75,
215-219.
Ying W. C., Bonk R. R. and Tucker M. E. (1987)
Precipitation treatment of spent electrodelem nickel
plating baths. In Proc. 42~i Industrial Waste
Conference, Purdue University, 12-t4 May 1987,
pp. 831--846.