Recovery of palladium from spent solutions for manufacture of catalysts

Article abstract of DOI:10.1134/S107042721011011X

Palladium recovery from [Pd(NH₃)₄]Cl₂ solutions (concentration in terms of the metal 1 g l^-1) with flow-through porous electrodes was studied. The conditions of effective electrochemical recovery of Pd were found. Various porous cathodes were compared.

Full text of DOI:10.1134/S107042721011011X

ISSN 1070-4272, Russian Journal of Applied Chemistry, 2010, Vol. 83, No. 11, pp. 1952−1956. © Pleiades Publishing, Ltd., 2010.
Original Russian Text © A.G. Belobaba, A.I. Maslii, A. Zh. Medvedev, 2010, published in Zhurnal Prikladnoi Khimii, 2010, Vol. 83, No. 11, pp. 1820−1824.
APPLIED ELECTROCHEMISTRY
AND CORROSION PROTECTION OF METALS
Recovery of Palladium from Spent Solutions
for Manufacture of Catalysts
A. G. Belobaba, A. I. Maslii, and A. Zh. Medvedev
Institute of Solid State Chemistry and Mechanochemistry, Siberian Branch,
Russian Academy of Sciences, Novosibirsk, Russia
Received July 23, 2009
Abstract—Palladium recovery from [Pd(NH3)4]Cl2 solutions (concentration in terms of the metal ~1 g l–1) with
flow-through porous electrodes was studied. The conditions of effective electrochemical recovery of Pd were
found. Various porous cathodes were compared.
DOI: 10.1134/S107042721011011X
It is known that platinum-group metals (Pt, Pd, etc.)
are widely used for manufacture of catalysts. Because
of their high cost and exhaustion of primary sources of
raw platinum materials, a considerable attention is being
paid to recuperation of platinum metals from catalysts
and technological solutions used in their manufacture [1,
2]. For example, a solution of tetraamminepalladium(II)
chloride [Pd(NH3)4]Cl2 is used in manufacture of glass-
fiber catalysts for purification of automobile exhaust
gases [3]. The impurities appearing in catalyst deposition
make a repeated use of a solution impossible, although
its Pd concentration changes only slightly and remains
at a level of 1 g l–1. At present, Pd is recovered from
these solutions by evaporation, which involves high
electric power expenditure and long process duration.
A possibly more promising and economically efficient
way to recover palladium from these solutions is by
electrolysis with flow-through porous electrodes (PEs).
20A potentiostat, Poland) at a temperature of 20 ± 1°C
on a laboratory installation similar to that described
in [4]. As flow-through porous cathodes served
fibrous carbon materials KNM and VNG-30 (Russia),
Carbopon V-22 (Poland), and metallized Sintepon, i.e.,
polyester fibers coated with a thin layer of silver [5].
Selected characteristics of these materials are listed
in the table [5, 6]. The anode of the electrolyzer was
made of graphite. The geometric area of the electrodes
was 0.3 dm2; volume of the circulating solution being
processed, 0.5 l; and flow rate, 0.1 cm3 s–1 across 1 cm2
of the geometric surface area of the cathode. We used
the circulation mode of PE operation, with the solution
delivered from the front (facing the anode) part of the
cathode.
We identified the cathode processes occurring in the
solutions under study by means of parallel measurements
of the current and electrode mass at a varying potential,
with an IPC-compact potentiostat (Institute of Physical
Chemistry, Russian Academy of Sciences, Russia) and
QCM-200 quartz microbalance (Stanford Research
System, USA). In this case, the role of the working
electrode was played by a gold film deposited in the form
of a circle with an area S = 1.37 cm2 onto one side of
a thin plate of a quartz single crystal (AT-cut, resonance
frequency 5 MHz, mass sensitivity 56.6 Hz μg–1).
The goal of our study was to find conditions for
effective electrochemical recovery of Pd from spent
solutions of a complex ammoniate [Pd(NH3)4]Cl2 and
compare the operation efficiencies of various porous
cathodes.
EXPERIMENTAL
We performed experiments on electrochemical
recovery of palladium in the potentiostatic mode (ER-
The study was carried out on synthetic and real
spent solutions of tetraamminepalladium(II) chloride
1952

RECOVERY OF PALLADIUM FROM SPENT SOLUTIONS
1953
Physicochemical characteristics of the fibrous porous materials used in the study
Layer thickness L, Fiber diameter df,
Electrical conductivity
Material
Porosity ε
cm
μm
æeff, Ω–1 cm–1
Carbonized unwoven material KNM
Carbon felt VNG-30
Carbon felt Carbopon V-22
Metallized Sintepon
0.25
0.37
0.30
1.20
12
11
5
0.98
0.96
0.94
0.99
0.009
0.09
0.12
1.10
20
[Pd(NH3)4]Cl2 with a concentration of about 1 g l–1
in terms of the metal. The Pd concentration was
determined in the course of electrolysis by voltammetry
on a renewable graphite electrode [7] and, at a number
of control points, by the atomic-absorption method. The
cathode potentials were measured relative to a saturated
silver chloride reference electrode. The distribution of
palladium across the thickness of a porous electrode
was studied by processing of micrographs of a PE
cross-section [8]. The micrographs were obtained with
a Neophot-21 microscope (Carl Zeiss, Germany) and
processed with Photo M software.
necessary to preliminarily identify cathode processes
in a solution under study. Figure 1 shows parallel
measured curves of current and cathode mass variation
with the cathode potential in a spent solution. It can
be seen that, at potentials in the range from +250 to
–1200 mV, the voltammetric curve clearly shows
three cathodic peaks with initial potentials of –350,
–450, and –1100 mV. Comparison of this curve with
that describing the variation of the mass shows that
only the second peak is associated with the cathodic
deposition of the metal. The rise in the cathodic current
at potentials more negative than –350 mV is presumably
due to reduction of dissolved O2 on gold, and its new
rise at potentials more negative than –1100 mV, to the
onset of H2 evolution on palladium. Calculation of the
electrochemical equivalent of the metal deposit at E =
–700 mV gives, with the fraction of current consumed
for the oxygen reduction taken into account, a value of
about 53 g, which corresponds to metallic Pd. Figure 2
shows cathodic polarization curves measured on
A starting spent aqueous solution of [Pd(NH3)4]
Cl2 has a high resistivity (About 550–600 Ω cm). The
depth to which the electrochemical process penetrates
into flow-through porous electrodes strongly depends
on the conductivities of the liquid and solid phases
and markedly decreases as these conductivities
become lower [9]. For this reason, it is inappropriate
to subject a starting solution to electrolysis without
preliminarily raising its conductivity. Therefore, we
tested various methods for pretreatment of a solution
before electrolysis. Acidification of a solution to
raise its conductivity was found to be unacceptable
because this leads to precipitation of the poorly soluble
dichloroamminepalladium salt [Pd(NH3)2]Cl2 [10] and
causes a more intense evolution of hydrogen at the
porous cathode in the course of electrolysis.
The best method for pretreatment of the solution is
its alkalization, because this simultaneously raises the
solution conductivity and hinders the side cathodic
reaction of hydrogen evolution. Addition of NaOH to
the starting solution first leads to a sharp decrease in
the solution resistance, but, at alkali concentrations
exceeding 5–6 g l–1, the effect rapidly decays. Therefore,
we mostly performed further experiments with solutions
to which 6 g l–1 of NaOH was added, which corresponded
to pH 12.7 and a resistivity of about 150 Ω cm.
5
0
E, mV
Fig. 1. (1) Current I and (2) change in the cathode mass, Δm,
vs. the potential E. Palladium concentration in solution 1 g l–1,
pH 12.7, potential sweep rate 20 mV s–1.
To choose the working range of PE potentials, it is
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 83 No. 11 2010

1954
BELOBABA et al.
E, mV
from this solution is not anodically dissolved (at least
up to E = 250 mV). This makes unnecessary separation
of electrode spaces in electrochemical recovery of
palladium onto a porous electrode, and the small anodic
current observed in the reverse run at E –500 mV can
be attributed to a possible oxidation of O2 reduction
products accumulated in the near-electrode layer.
I, mA
I, mA
Taking into account the revealed nature of the
cathodic processes, we chose the optimal potential for
potentiostatic deposition of Pd on the porous cathode
on the basis of the following considerations. On the
one hand, it is desirable that, for the metal deposition
process to penetrate deeper into a PE, that the difference
between the potential set on the front side of the PE
and the equilibrium potential of Pd should be as large
as possible [11]. On the other hand, a side process of
hydrogen evolution begins at potentials more negative
than –1100 mV, with the hydrogen-related current
rapidly increasing with the cathodic polarization. The
intense occurrence of this side reaction is undesirable
not only because of the apparent decrease in the current
efficiency, but also due to a possible deterioration of the
PE operation efficiency. The latter is associated with
the rise in the ohmic loss of voltage at increased current
densities [12] and also with an increase in the gas content
of the solution [13]. Therefore, the maximum value of
the PE potential in deposition of Pd from this solution
was chosen to be –1300 mV.
Pd, g l–1
Fig. 2. Cathodic polarization curves measured on a renewable
graphite electrode (S = 0.03 cm2, potential sweep rate
20 mV s–1) in alkaline solutions. (I) Current and (E) potential.
Palladium concentration (g l–1): (1) 0.125, (2) 0.25, (3) 0.5,
and (4) 1. Inset: dependence of the peak current I on the
palladium concentration Pd.
a mechanically renewed graphite electrode in solutions
with various palladium concentrations. It can be seen
that the dependence of the peak height at a potential of
about –750 mV on the metal concentration is linear and
can be used for fast monitoring of the Pd concentration
in the course of electrolysis.
It also follows from an analysis of the reverse run of
the curve describing the variation of the cathode mass
with the potential (Fig. 1, curve 2) that Pd deposited
Figure 3 shows for the chosen electrolysis conditions
how the concentration of palladium depends on the
duration of deposition in solutions with different initial
conductivities (curves 1 and 2) and for various porous
cathode materials and a fixed pH value of the solution
(curves 2–5). At a constant volume and flow rate of the
circulating solution, the slope of the curves in Fig. 3 is
proportional to the effective working surface area of the
PE and to the rate of the mass-transfer-limited recovery
process of the metal. It can be seen from these data
that, in almost all cases, the recovery rate of palladium
noticeably grows in the course of time. This is due to
a decrease in the concentration of the metal in solution,
increase in the slope of the polarization curve (dE/dI),
and the corresponding improvement of the uniformity
of the current distribution within the PE.
t, h
Fig. 3. Dynamics of palladium recovery from a solution under
study (1) without an alkali and (2–5) with addition of 6 g l–1
of NaOH on flow-through porous cathodes made of various
materials. (c, c0) Initial and running palladium concentrations
and (t) time; the same for Fig. 4. Material: (1, 2) KNM, (3)
Carbopon V-22, (4) VNG-30, and (5) Sintepon. For the rest of
electrolysis conditions, see text.
The increase in the working surface area of the PE in
the course of metal recovery is a characteristic feature
of PEs, and the data in Fig. 3 demonstratively confirm
the conclusion that PEs are most effectively used for
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 83 No. 11 2010

RECOVERY OF PALLADIUM FROM SPENT SOLUTIONS
1955
t, h
processing of just diluted solutions [11]. Comparison of
curves 1 and 2 in Fig. 3 clearly shows the importance
of the preliminary treatment of a solution. Alkalization
of the starting solution to pH 12.7 makes it possible
to raise, under the same electrolysis conditions, the
recovery of palladium from 14 to 77% in 2 h. It also
follows from the data in Fig. 3 that the recovery rate
strongly depends on the properties of a porous material
and, on passing from KNM to metallized Sintepon,
grows by approximately a factor of 4. Comparison of
curves 2–5 in Fig. 3 with the data in the table suggests
that the order in which the materials can be arrange in
order of increasing palladium recover rate is for the most
part determined by their conductivity. The noticeable
advantage of metallized Sintepon over Carbopon V-22
can be attributed to its substantially larger porosity.
Fig. 4. Dynamics of palladium recovery from [Pd(NH3)4]Cl2
solutions on a porous KNM cathode in successive processing
of the first three 0.5-l portions of a solution with an initial Pd
concentration of 1 g l–1.
It should be specially emphasized that the above-
discussed operation efficiency parameters of various
porous materials are related to their original state
and initial stage of their filling with the metal. Upon
deposition of palladium, characteristics of the porous
matrices (effective conductivity of threads, their
diameter, porosity) naturally change, which may also
affect the operation efficiency of the corresponding
PEs. These changes are comparatively small for PEs
with high initial conductivity and are more pronounced
for materials with low initial conductivity, e.g., KNM.
Taking into account that this material is now nearly the
only industrially manufactured carbon felt, we studied
the variation dynamics of its operation efficiency in
the course of palladium recovery from successively
processed solution portions with the same volumes
and initial Pd concentrations. In the conditions of
potentiostatic electrolysis with predominance of the
target process of palladium deposition, the variation
of its rate can be characterized by time dependences
of both the palladium concentration in solution and
the overall current. Figure 4 shows as an example data
on the dynamics of palladium recovery in successive
processing of the first three portions of the solution,
and Fig. 5, curves describing the variation of the
overall current from the beginning of electrolysis until
palladium fills the critical cross-section of the PE and
the solution ceases to flow.
I, mA
t, h
Fig. 5. Variation dynamics of the current I in successive
recovery of palladium from ten 0.5-l portions of a solution
with an initial Pd concentration of 1 g l–1 until full clogging of
pores in the KNM cathode. (t) Time. Figures at curves, portion
numbers of the solution being processed.
higher already for the second portion of the solution, and
1.4 times more for the third portion to become almost
the same as that of the most electrically conducting
Sintepon. Further, the rise in the recovery rate is
noticeably decelerated. Comparison of the data in Fig. 4
with the current variation curves in Fig. 5 shows that
the main contribution to the rise in the working surface
area and in the recovery rate of palladium comes from
the increase in the conductivity of the porous matrix.
This is evidenced both by a sharp rise in the initial
current for the first 3–4 portions of the solution and
by a change in the shape of I–t curves. For example,
in the first stage of deposition, the electrolysis current
first steeply grows (from 40 to 220 mA), despite the
It can been from the curves in Fig. 4 that, in the initial
stage of palladium deposition onto KNM, the metal
recovery rate and the PE operation efficiency noticeably
grow. For example, the recovery rate becomes 1.7 times
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 83 No. 11 2010

1956
S, μm2
BELOBABA et al.
geometric surface area of the electrode at a recovery of
no less than 99%.
CONCLUSIONS
(1) It was shown that electrolysis with flow-through
porous cathode is promising for recovery of palladium
from spent catalyst production solutions.
(2) It was found that preliminary alkalization of
the solutions prior to electrolysis is necessary and the
optimal modes of the process were determined.
(3) It was demonstrated that porous materials with
high conductivity and porosity are the most effective for
cathodic recovery of metals from dilute solutions.
L, cm
Fig.6. Final palladium distribution across the thickness L of
the porous KNM cathode. (S) Average area of a metal ring
around a fiber.t
REFERENCES
1. Dawson, R.J. and Kelsall, G.H., J. Appl. Electrochem.,
steady decrease in the palladium concentration (Fig. 5),
passes through a maximum, and starts to decrease only
at the end of the stage (i.e., already at a high degree of
solution depletion. The increase in the conductivity of
the cathode also noticeably affects the shape of the I–t
curve for the second cycle. And only beginning from the
third portion of the solution being processed, the time
dependences of the current take on the steadily decaying
shape characteristic of the equipotential PE. The slight
rise in the initial currents, observed up to the 6th–8th
stage, is presumably due to an increase in the working
surface area because of thread thickening, and their
subsequent decline, to a decrease in porosity in the most
palladium-filled cross-section of the PE.
2007, vol. 37, no. 1, pp. 3–14.
2. RF Patent 2119964.
3. Bal’zhinimaev, B.S., Simonova, L.G., and Barelko, V.V.,
Kinet. Kataliz, 2001, vol. 42, no. 6, pp. 917–927.
4. Belobaba, A.G., Maslii, A.I., Bochkarev, G.R., and
Pushkareva, G.I., Gal’vanotekhn. Obrab. Pov–ti, 2004,
vol. 12, no. 3, pp. 30–34.
5. RF Patent 2178017.
6. Varentsov, V.K., Zherebilov, A.F., and Malei, M.D., Izv.
Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, 1984, no. 6,
no. 17, pp. 120–126.
7. Kletennik, Yu.B., Polyakin, L.Yu., and Tarasova, V.A.,
Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim. Nauk, 1985,
Issue 2, no. 5, pp. 63–67.
The final distribution of palladium across the
thickness of a flow-through porous KNM cathode is
showninFig.6.Itcanbeseenthatpalladiumisdistributed
within the PE rather nonuniformly and mostly fills the
frontal, closest to the anode, part of the electrode. The
maximum relative mass of deposited palladium for the
KNM cathode is small: about 5 g per gram of the starting
porous material. To improve the uniformity of the final
distribution of the metal and raise its limiting mass, it is
necessary to use thinner, more electrically conducting,
and more porous cathode material. The average electric
energy expenditure for Pd recovery onto a KNM PE
were 0.52 W h per gram of the metal, and the average
recovery rate of the metal, 0.83 g h–1 per 1 dm3 of the
8. Medvedev, A.Zh., Zherebilov, A.F., Maslii, A.I., and
Poddubnyi, N.P., Elektrokhimiya, 2008, vol. 44, no. 6,
pp. 818–824.
9. Maslii, A.I. and Poddubnyi, N.P., Elektrokhimiya. 1992,
vol. 28, no. 11, pp. 1650–1655.
10. Ginzburg, S.I., Ezerskaya, N.A., Prokof’eva, I.V., et al.,
Analiticheskaya khimiya platinovykh metallov (Analytical
Chemistry of Platinum Metals), Moscow; Nauka, 1972.
11. Bek, R.Yu., Izv. Sib. Otd. Akad. Nauk SSSR, Ser. Khim.
Nauk, 1977, Issue 6, no. 4, pp. 11–20.
12. Maslii, A.I. and Poddubnyi, N.P., Elektrokhimiya, 1997,
vol. 33, no. 8, pp. 919–926.
13. Zamyatin, A.P. and Bek, R.Yu., Elektrokhimiya, 1984,
vol. 20, no. 3, pp. 351–357.
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 83 No. 11 2010