Kinetics and mechanism of palladium(II) acetate reduction by hydrogen on the surface of a carbon support

Article abstract of DOI:10.1134/S0023158411020029

The kinetics of the reduction of Pd(II) compounds by dihydrogen on the surface of a carbon support has been investigated for palladium acetate as an example. A kinetic model has been constructed for this reaction. An autocatalytic mechanism is suggested, in which the key role is played by Pd(0) compounds and their hydrides. The reaction occurring on the support surface is compared with the same reaction in solutions of palladium phosphine acetate complexes, where a similar mechanism is observed. One of the most important features of the surface reaction is the relatively slow reduction of the Pd(I) compounds to Pd(0). This makes it possible to obtain materials with a high Pd(I) content of 5% and above.

Full text of DOI:10.1134/S0023158411020029

ISSN 0023ꢀ1584, Kinetics and Catalysis, 2011, Vol. 52, No. 2, pp. 296–304. © Pleiades Publishing, Ltd., 2011.
Original Russian Text © A.S. Berenblyum, Hussein Ali AlꢀWadhaf, E.A. Katsman, V.R. Flid, 2011, published in Kinetika i Kataliz, 2011, Vol. 52, No. 2, pp. 305–313.
Kinetics and Mechanism of Palladium(II) Acetate Reduction
by Hydrogen on the Surface of a Carbon Support
A. S. Berenblyum*, Hussein Ali AlꢀWadhaf, E. A. Katsman, and V. R. Flid
Moscow State Academy of Fine Chemical Technology, Moscow, 119571 Russia
*
eꢀmail: ayb@go.ru
Received August 30, 2010
Abstract—The kinetics of the reduction of Pd(II) compounds by dihydrogen on the surface of a carbon supꢀ
port has been investigated for palladium acetate as an example. A kinetic model has been constructed for this
reaction. An autocatalytic mechanism is suggested, in which the key role is played by Pd(0) compounds and
their hydrides. The reaction occurring on the support surface is compared with the same reaction in solutions
of palladium phosphine acetate complexes, where a similar mechanism is observed. One of the most imporꢀ
tant features of the surface reaction is the relatively slow reduction of the Pd(I) compounds to Pd(0). This
makes it possible to obtain materials with a high Pd(I) content of 5% and above.
DOI: 10.1134/S0023158411020029
Supported palladium catalysts are widely used in priate method for monitoring the conversion of surꢀ
largeꢀscaleꢀprocesses of the petrochemical and petroꢀ face compounds and, thereby, the reaction kinetics.
leum refining industries, including the hydrogenation
Here, we report the kinetics and mechanism of the
of unsaturated compounds, dehydrogenation, isomerꢀ
reduction of supported palladium(II) acetate with
ization, and hydrogenolysis [1, 2]. In recent years,
hydrogen, a widespread reductant in the synthesis of
much attention has been given to developing the sciꢀ
supported metal catalysts [7, 8].
entific foundations of the preparation of these cataꢀ
lysts, particularly their synthesis, structure, and physiꢀ
cochemical properties [3–5]. The sustained striving
towards better catalysts, primarily more active and
more selective ones, prompts researchers to study catꢀ
alyst preparation chemistry in greater detail.
EXPERIMENTAL
Chemicals, Materials, and Their Purity
The following chemicals were used: palladium aceꢀ
including palladium ones, is by loading a support with tate (reagent grade) as received, acetone (specialꢀ
the initial palladium compound followed by the purity grade), hydrogen gas (brand A, 99.99%), and
A conventional way of preparing metal catalysts,
reduction of this compound [6]. Therefore, for obtainꢀ
ing advanced, highly efficient catalysts, it is necessary
to investigate the processes occurring on the support
surface, including the reduction of the palladium preꢀ
cursor, in which palladium is usually in the oxidation
state +2, into its active state.
nitrogen gas (specialꢀpurity grade, 99.999%).
The support was Vulcan XCꢀ72 carbon with a speꢀ
2
cific surface area of about 250 m /g [9]. Prior to cataꢀ
lyst synthesis, the support was annealed in an oxygen
atmosphere at 450°C for 4 h [10].
Although there have been lots of publications dealꢀ
ing with heterogeneous catalysts, the reduction mechꢀ
anisms of the catalysts precursors on support surfaces
are still poorly understood. Obviously, knowledge of
these mechanisms is quite essential for correctly
selecting a precursor, obtaining finely dispersed metal
catalysts, etc. In addition, investigation of the chemisꢀ
try of these complicated solidꢀphase surface reactions
is of fundamental interest as a way to elucidate the
effect of the surface on these reactions.
Synthesis of CarbonꢀSupported Palladium(II) Acetate
Acetone (50 ml) was added to carbon (500 mg),
and the mixture was ultrasonicated at room temperaꢀ
ture for 20 min to obtain a homogeneous suspension.
An appropriate amount of a palladium acetate disꢀ
solved in 15 ml of acetone was added to the suspenꢀ
sion, and the mixture was stirred for 15 min. Next, the
entire solvent was evaporated in a rotary evaporator
When investigating reactions on support surfaces, under careful heating. Samples containing 2.0, 4.0,
researchers encounter difficulties typical for surface and 9.1 wt % Pd (relative to the palladium + support
studies, including the problem of choosing an approꢀ weight) were thus obtained.
296

KINETICS AND MECHANISM OF PALLADIUM(II) ACETATE REDUCTION
297
Kinetic Experiments
Direct kinetic problem solver [11, 12]. This proꢀ
gram employs the subprogram STIFF [13] in the
numerical integration of sets of ordinary differential
equations. It enables one to calculate not only the
reactant concentrations, including the acetic acid
concentration, at preset points in time, but also other
Palladium acetate reduction by hydrogen takes
place according to the following overall chemical
equation:
Pd(OAc) + H
2
Pd + 2HOAc.
2
A sample (0.1 g) was placed in a tubular reactor, components of the observation model (soꢀcalled
and hydrogen at a preset partial pressure was passed responses) for a given kinetic model [11, 12, 14]. Here,
through the reactor. The total absolute gas pressure the concentrations of palladium in the oxidation states
was 1 atm, and the desired hydrogen partial pressure +2, +1, and 0 are meant. Analyzing the responses, we
was obtained by diluting hydrogen with nitrogen. The evaluated the goodness of fit between the calculated
acetic acid resulting from the reaction was carried and observed data. The measure of the goodness of fit
away by the hydrogen stream and then by a nitrogen was the objective function (functional) defined as the
stream and was absorbed by a solution of an excess square root of the average sum of the weighted squares
amount of an alkali. After the run, the alkali was of the differences between the calculated and observed
titrated with hydrochloric acid. The experimental values. The weighting factors were the inverse squares
setup, the reduction procedure, and the quantification of the standard measurement error [11] for the correꢀ
of acetic acid by titration were described in detail in an sponding responses. In other words, the goodness of fit
earlier publication [10]. The error of acetic acid deterꢀ was nothing else but the standard deviation between
mination did not exceed 7 rel. %. Specialꢀpurpose the calculation and the experiment in terms of stanꢀ
experiments demonstrated that, at the hydrogen feed dard response measurement errors. If a kinetic model
rate used in our experiments (0.72 l/h), there are no is in good agreement with the experimental data, the
significant external diffusion limitations [10]. In order goodness of fit is close to unity; if the goodness of fit is
to see whether there are internal diffusion limitations, well above unity, the model is considered to be invalid
we examined a support sample in which ~80% of the [11, 14].
particles were 0.40–3.20 mm in diameter, much larger
than those used in standard kinetic runs (<0.01 mm).
It was found that the acetic acid evolution rates for the
standard catalyst sample and for the sample syntheꢀ
sized using large carbon particles are equal within the
experimental error. Therefore, for the carbon particles
sizes below 0.01 mm, the internal diffusion limitations
in the reaction between hydrogen and palladium aceꢀ
tate can be neglected.
Inverse kinetic problem solver (numerical estimaꢀ
tion of kinetic model parameters) [11, 12]. This proꢀ
gram is based on the leastꢀsquares method [11]. It
minimizes the functional by the configuration method
[
15]. The search step size in the minimization proceꢀ
dure is varied automatically so as to maintain the highꢀ
est convergence rate at a preset accuracy.
Program for numerical analysis of the identifiability
of kinetic model parameters. This program is run to
calculate the sensitivity matrix for the responses of the
model to variations of model parameters [16] and to
factorize this matrix [14]. Next, an analysis is perꢀ
formed to single out the insignificant parameters, such
as the rate constants of the rapid or slow reactions
steps that are not to be determined precisely [16]; to
singleꢀout the constants that cannot be calculated
alone, but only in combination with other parameters
Spectroscopic Methods
Xꢀray photoelectron spectra were recorded on an
XSAMꢀ800 spectrometer (Kratos, United Kingdom).
The Xꢀray source was a magnesium anode with a charꢀ
acteristic Mg
K radiation energy of 1253.6 eV. The
α
maximum power dissipation in the anode during specꢀ
trum recording was 90 W. The background due to the
secondary electrons and the photoelectrons that had
lost energy was approximated by a straight line. Meaꢀ
(
soꢀcalled parameter complexes [11]); and to separate
the parameters to be determined independently. For
the latter parameters, (and, if necessary, for the
parameter complexes), the determination error is calꢀ
culated using the Fisher matrix [11, 16].
–8
surements were taken at a pressure of ~5 × 10 Pa. The
samples were secured on a sample holder with a conꢀ
ductive doubleꢀsided adhesive tape. Spectra were
obtained at room temperature. The spectrometer was
calibrated against the Au
4f7/2 and Ni
2p3/2 peaks, whose
RESULTS AND DISCUSSION
binding energies are 84.0 and 852.7 eV, respectively.
IR spectra were recorded for KBr pellets on a Digꢀ
ilab Scimitar 1000 IR Fourier transform spectrometer
The reaction between dihydrogen and pallaꢀ
dium(II) acetate was studied using the above method
of determining the amount of acetic acid resulting
from the reduction of supported palladium(II) comꢀ
pounds [10]. Figure 1 plots the percentage yield of
acetic acid (actual yield divided by theoretical yield),
(
Varian).
Kinetic Data Processing
Y, as a function of the reaction time at two different
The following three programs were used in the conꢀ temperatures (50 and 120°C) and two hydrogen partial
struction of the kinetic model.
pressures (1.0 and 0.052 atm). Clearly, the kinetics of
KINETICS AND CATALYSIS Vol. 52
No. 2
2011

298
BERENBLYUM et al.
tively slow reduction to palladium(I) and the latter is
reduced rapidly to Pd(0). A radically different situaꢀ
tion is observed on the support surface. In this case,
Y
, %
90
80
70
60
50
40
30
20
10
(a)
the reduction rate falls markedly at Y > 50%. It can be
hypothesized that, on the support surface, the pallaꢀ
dium(I) compounds result from the comparatively
rapid reduction of palladium(II) acetate and are then
reduced slowly to Pd(0). This is the likely reason why
the maximum acetic acid yield attainable in 3 h for the
3
2
6
1
sample containing 9.1% Pd at 50°C is Y = 68%. As the
5
reaction temperature is increased to 120 and 300
increases to 77 and ~100%, respectively.
°C, Y
The assumption that the Pd(I) compounds are
reduced by hydrogen at a low rate and, as a conseꢀ
quence, accumulate on the support surface was fully
verified by XPS studies. Figure 2 presents the Xꢀray
photoelectron spectra of the samples containing 9.1%
Pd that were reduced with hydrogen at 50 and 120°C
for 30 and 180 min, respectively. The spectra show
Pd(II), Pd(I), and Pd(0) peaks at binding energies of
4
0
20 40 60 80 100 120 140 160 180
Y, %
90
80
70
60
50
40
30
20
10
(b)
3
6
5
1
3
38.2, 336.4, and 335.2 eV, close to those reported in
the literature [18]. The Pd(II), Pd(I), and Pd(0) perꢀ
centages in the total palladium content were deterꢀ
mined by comparing the intensities of the correspondꢀ
ing Pd
centages in the sample reduced at 50
be 25.3, 49.1, and 25.6%, respectively. Raising the
reduction temperature to 120 caused a decrease in
3
d5/2 peaks. The Pd(II), Pd(I), and Pd(0) perꢀ
°C
were found to
°C
the Pd(II) and Pd(I) contents and an increase in the
Pd(0) content, just as was expected, and the correꢀ
sponding percentages appeared to be 11.4, 34.8, and
0
20 40 60 80 100 120 140 160 180
Time, min
5
3.8%.
From XPS data, we derived the theoretical amount
of acetic acid that should result from the reaction: for
50 and 120° should be 50.2 and 71.0%. According
to titration data, for these samples is 52.0 and 77.0%,
respectively. Thus, the acetic acid yields measured by
the two independent methods coincide within the
experimental error.
Fig. 1. Acetic acid formation kinetics at (a) 50 and (b)
1
20 C and different palladium acetate concentrations on
°
the support and different hydrogen partial pressures:
(
1
) 2% Pd, 1 atm; (
2
) 4% Pd, 1 atm; (
3
) 9.1% Pd, 1 atm;
) 9.1% Pd,
С, Y
(4
) 2% Pd, 0.052 atm; (
5) 4% Pd, 0.052 atm; (
6
Y
0.052 atm. The points represent experimental data, and
the lines represent the data calculated using the kinetic
model.
All of the kinetic data (Fig. 1) and the XPS data for
this reaction is rather complicated and the averaged the samples reduced at the two different temperatures
orders of the reaction with respect to the palladium were used in the construction of a kinetic model.
and hydrogen concentrations are not integer. Furtherꢀ
more, under certain reaction conditions (e.g., 4% Pd,
50°C), the kinetic curves are Sꢀshaped and indicate an
induction period. The induction period almost disapꢀ
pears as the palladium concentration and temperature
are raised. Apparently, the induction period is shorter
than 15 min under these conditions.
The hypothetical mechanisms of the reaction
included reaction step combinations corresponding to
different states of the initial compounds and different
intermediates, products, and conversion routes. About
1
00 hypotheses were considered. Using the procedure
described in the experimental section, for each
hypothesis we determined the goodness of fit between
Similar kinetics were observed for the reduction of the model and the experimental data. Those descripꢀ
Pd(II) triphenylphosphine acetate complexes in soluꢀ tions were taken to be adequate for which the standard
tion [17]. The observed Sꢀlike shape of the kinetic error did not exceed 7–10 rel. % and there were no
curves is due to the autocatalytic action of the comꢀ systematic deviations between the calculated and
pounds of palladium in the low oxidation state and experimental data in the entire ranges of experimental
palladium hydride complexes in the reduction of the conditions (temperature, reaction time, hydrogen
initial Pd(II) complexes [17]. However, there are sigꢀ partial pressure, initial palladium(II) acetate concenꢀ
nificant distinctions between the reduction kinetics of tration). When more than one hypothesis was adeꢀ
Pd(II) in solution and that of Pd(II) on the surface. In quate, we chose the simplest one, guided by the prinꢀ
solution, the Pd(II) compounds undergo comparaꢀ ciple of simplicity (Occam’s razor) [19].
KINETICS AND CATALYSIS Vol. 52
No. 2
2011

KINETICS AND MECHANISM OF PALLADIUM(II) ACETATE REDUCTION
Intensity, arb. units
299
500
400
300
200
100
0
(
a)
1
2
4
5
3
−
100
345
340
b)
335
3
3
2
2
1
1
50
00
50
00
50
00
(
1
2
5
4
50
0
3
−50
345
340
335
Binding energy, eV
Fig. 2.
at (a) 50 and (b) 120
) Pd(0).
(
1
) Observed Xꢀray photoelectron spectrum of palladium(II) acetate on the carbon support after reduction with hydrogen
°
C. ( ) Spectra obtained by deconvolution: ( ) Pd(II) + Pd(I) + Pd(0), ( ) Pd(II), ( ) Pd(I), and
2–5
2
3
4
(5
It is quite obvious that the reduction of Pd(II) comꢀ can conventionally be divided into seven groups each
pounds is a multistep reaction both in solution and on containing a number of reaction steps (Table 1).
the surface. Based on data available from the literature
The first group includes the reactions of pallaꢀ
[
17, 20] and on our experimental data, we suggested
dium(II) acetate on the carbon surface before interacꢀ
and analyzed the main reduction routes. These routes tion with hydrogen.
KINETICS AND CATALYSIS Vol. 52
No. 2
2011

300
BERENBLYUM et al.
Table 1. Rate constants of the reactions steps of the model at 50°C
Rate constant**
Step no.
Chemical equation*
Group no.
reverse
reaction
forward reaction
1
/2
/4
Pd (OAc)
Pd (OAc) + Pd(OAc)
2
1
2
3.17E+03
1.16E+00
9.70E+08
2.20E–01
3
6
2
4
3
Pd (OAc)4
2Pd(OAc)2
Pd + 2HOAc
Pd (OAc) + 2HOAc
2
5
6
7
8
9
Pd(OAc) + H2
2.50E–03
2
Pd (OAc) + H
2
4
2
2
2
2
2
Pd (OAc) + H
Pd(OAc) + Pd + 2HOAc
2
2
4
Pd (OAc) + H
Pd (OAc) + Pd(OAc) + 2HOAc
2.29E–03
1.41E+01
3
6
2
2
2
Pd (OAc) + H
3
Pd (OAc) + Pd + 2HOAc
2
6
2
4
1
0/11 2Pd (OAc)2
Pd (OAc)4
2.96E+03
7.42E+01
2
4
1
1
1
2/13 Pd + H2
PdH2
3
4
2.95E+08
1.79E+10
2.44E+06
4/15 2Pd
Pd2
6/17 Pd + H2
Pd H2
2
2
1
1
2
2
2
2
2
2
2
2
8
9
0
1
2
3
4
5
6
7
Pd(OAc) + PdH2
2Pd + 2HOAc
8.63E+00
4.90E+04
5.63E+11
2
Pd(OAc) + Pd H
2
Pd + Pd + 2HOAc
2
2
2
Pd (OAc) + PdH
Pd (OAc) + Pd + 2HOAc
2
4
2
2
2
Pd (OAc) + PdH
2
Pd(OAc) + 2Pd + 2HOAc
2
4
2
Pd (OAc) + Pd H
Pd (OAc) + Pd + 2HOAc
2
4
2
2
2
2
2
Pd (OAc) + Pd H
2
Pd(OAc) + Pd + Pd + 2HOAc
2 2
2
4
2
Pd (OAc) + PdH
Pd (OAc) + Pd(OAc) + Pd + 2HOAc
2.54E+02
1.65E+03
3
6
2
2
2
2
2
Pd (OAc) + PdH
Pd (OAc) + 2Pd + 2HOAc
2 4
3
6
Pd (OAc) + Pd H
2
Pd (OAc) + Pd(OAc) + Pd + 2HOAc
2 2 2 2
3
6
2
Pd (OAc) + Pd H
Pd (OAc) + Pd + Pd + 2HOAc
7.13E+05
3
6
2
2
2
4
2
2
8
9
Pd (OAc) + H
2
2Pd + 2HOAc
Pd (OAc) + 2Pd + 2HOAc
5
6
2
2
2
Pd (OAc) + H
4
4
2
2
2
30
31
32
33
Pd (OAc) + PdH
2
3Pd + 2HOAc
2
2
Pd (OAc) + PdH
Pd (OAc) + 3Pd + 2HOAc
2.12E+07
4
4
2
2
2
Pd (OAc) + Pd H
2
Pd + 2Pd + 2HOAc
2
2
2
2
Pd (OAc) + Pd H
4
Pd (OAc) + Pd + 2Pd + 2HOAc
2
4
2
2
2
2
3
3
3
4/35 Pd(OAc) + Pd
Pd (OAc)2
7
4.09E+11
2.09E+11
2
2
6/37 Pd (OAc) + Pd
Pd (OAc) + Pd(OAc)
2 2 2
2
4
8/39 Pd (OAc) + Pd
Pd (OAc) + Pd (OAc)
4
9.42E+13
3
6
2
2
2
*
*
The steps making a significant contribution to the overall reaction are set off in bold type.
*Units of measure: monomolecular reactions, min ; bimolecular reactions not involving hydrogen gas, g mol min–1; bimolecular reacꢀ
±
±
–
1
–1
tions involving hydrogen gas, atm– min (g is the weight of the support in grams).
1
–1
E n stands for ×10 .
n
The IR spectroscopic study of the sample obtained symmetric stretching band shifts from 1429 to
–1
by loading carbon with palladium(II) acetate without 1407 cm and weakens. These data are evidence in
subsequent reduction with hydrogen (9.1% Pd) demꢀ favor of the assumption that the observed changes in
onstrated that both the spectrum of the initial pallaꢀ the IR spectrum are due to the comparatively weak
dium(II) acetate, which is a trimer [21], and the specꢀ interaction between the palladium(II) acetate trimer
trum of the supported acetate show two absorption and the support surface. Of course, it is not impossible
bands due to the C–O stretching vibrations characterꢀ that this interaction can initiate the reversible converꢀ
istic of bridging structures [22]. The position of the sion of the trimer into dimeric and monomeric comꢀ
–1
antisymmetric stretching band (1606 cm ) is almost plexes stabilized by the carbon support, just as the solꢀ
unchanged by the supporting of the salt, while the vent causes palladium(II) acetate, which is trimeric in
KINETICS AND CATALYSIS Vol. 52
No. 2
2011

KINETICS AND MECHANISM OF PALLADIUM(II) ACETATE REDUCTION
301
the crystalline state, to decompose to dimeric and yielding a Pd(I) compound, as in the case of palladium
monomeric species in solution [23].
Simulation of the reaction demonstrated that
attempts to exclude the palladium acetate dimer and
complexes in solution [17]. This group consists of
steps 34/35, 36/37, and 38/39.
It was established that the reaction network preꢀ
monomer (which differ in reactivity toward hydrogen sented in Table 1, which consists of 30 reactions,
from the trimer) from the reaction mechanism lead to including 9 reversible ones, provides a good descripꢀ
an inadequate description. Inclusion of the tetramer tion for the observed kinetics. The error of this
into the reaction mechanism does not decrease the description does not exceed 5–7 rel. % for all reactions
goodness of fit. It follows from the simulation data conditions examined. Figure 1 illustrates the good fit
that, before reduction with hydrogen, the entire pallaꢀ between the experimental data and the data calculated
dium(II) acetate is in the form of the trimer. The other using this model.
assumptions lead to an inadequate description. Note
Tables 1 and 2 list the numerical values of the rate
that attempts to include reaction steps in which pallaꢀ
constants for the reactions involved in the model at 50
dium(II) acetate changes its structure without changꢀ
and 120°С, respectively.
ing its nuclearity, such as a change in the proportions
of bridging and terminal acetate groups, dramatically
reduce the goodness of fit. For this reason, reaction
steps 1/2 and 3/4 were included in the first group.
The second group of reaction steps is made up by
the noncatalytic reactions of different palladium aceꢀ
tate species with dihydrogen. The simplest step is the
reaction between the palladium(II) acetate monomer
with hydrogen yielding Pd(0) and acetic acid (step 5).
This group also includes the reduction of the pallaꢀ
dium(II) acetate dimer and trimer to Pd(0) and Pd(I)
compounds (steps 6–9, 10/11). As a rule, the Pd(I)
A detailed analysis of the results of the simulation
using the above programs revealed the following.
Reaction steps 1/2, 12/13, and 39 are very rapid and
steps 3/4, 36, and 16 can be considered as relatively
rapid. Since steps 6, 7, 9, 15, 17, 19, 21, 23, 26, 28–30,
3
2, 33, 35, 37, and 38 are very slow, they can be
excluded from the reaction network without causing
any significant changes in the result of simulation (this
is true for the reaction steps whose rate constants are
not included in Tables 1 and 2).
Because of the complexity of the reaction, the rate
compounds are in dinuclear and tetranuclear forms constants of most of its steps can be determined only
17, 24]. Hereafter, it is assumed that the reaction in combination with other constants. The rate conꢀ
[
includes the interactions of palladium acetate trimers stants of only two reaction steps (24 and 25) can be calꢀ
and dimers with hydrogen and with palladium culated unambiguously (Table 1). These steps make
hydrides or Pd(0) and that these interactions take the determining contribution to the reaction rate and
place as follows. Part of the palladium of the initial palladium(II) acetate conversion, particularly in the
complexes retains its initial oxidation state in the region where the maximum acetic acid formation rate
resulting compounds, and the other part undergoes is observed.
reduction to Pd(I) or Pd(0) (steps 7–9, 21, 23–27, 29,
At 120°С, steps 4, 11, and 18 are also very slow, like
3
1, 33, 36/37, and 38/39). Inclusion of intermediate
the above 17 steps. For this reason, it was possible to
estimate the activation energy only for reaction
steps 24 and 25 (Table 3).
palladium(II) hydride–acetate complexes (whose forꢀ
mation was assumed for the reduction of pallaꢀ
dium(II) acetate complexes in solution [17]) and the
Pd(I) monomer and trimer into the reaction network
does not improve the goodness of fit.
The kinetic parameters thus determined enabled us
to evaluate the roles of particular steps and palladium
species in this complex reaction. The reaction begins
with the slow reduction of the palladium(II) acetate
trimer with dihydrogen (step 8), which yields the palꢀ
ladium(I) acetate dimer and palladium(II) acetate
monomer. The reversible conversion of the pallaꢀ
dium(II) trimer into the palladium(II) acetate dimer
and monomer occurs simultaneously (steps 1/2
and 3/4).
The third group consists of the reactions of comꢀ
pounds containing one or two Pd(0) atoms with
hydrogen, which yield the corresponding hydrides
(
steps 12/13, 14/15, and 16/17). These hydrides can
efficiently reduce palladium compounds, making the
reaction autocatalytic [17].
The fourth group includes all reactions of the
Pd(II) compounds with palladium hydrides, which
result in the reduction of Pd(II) to Pd(I) and Pd(0).
Specifically, the reduction of Pd(OAc) , Pd (OAc)4,
The reduction of the palladium(II) acetate monoꢀ
mer to Pd(0) (step 5) ensures passage to the rapid
phase of the overall reaction. The Pd and Pd2 particles
generated in step 14 turn comparatively rapidly into
2
2
and Pd (OAc)6 by the monomeric and dimeric pallaꢀ
3
dium hydrides are meant here (steps 18–27).
the hydrides PdH2 (step 12) and Pd H2 (step 16). Both
2
The fifth and sixth groups include the interactions
of Pd(I) compounds with hydrogen (steps 28 and 29)
and palladium hydrides (steps 30–33), respectively.
hydrides reduce, more rapidly than dihydrogen, all
forms of palladium(II) acetate—trimer (steps 24, 25,
and 27), dimer (steps 20 and 22), and monomer
The seventh group includes disproportionation (step 18)—to palladium(I) acetate or Pd(0). These
reactions between a Pd(II) compound and Pd(0) reactions regenerate Pd(0), and this is among the facꢀ
KINETICS AND CATALYSIS Vol. 52
No. 2
2011

302
BERENBLYUM et al.
Table 2. Rate constants of the reactions steps of the model at 120°C
Rate constant**
Step no.
Chemical equation*
Group no.
reverse reacꢀ
forward reaction
tion
1
/2
/4
Pd (OAc)
Pd (OAc) + Pd(OAc)
2
1
2
3.36E+03
1.16E+00
5.37E+15
3
6
2
4
3
Pd (OAc)4
2Pd(OAc)2
Pd + 2HOAc
Pd (OAc) + 2HOAc
2
5
6
7
8
9
Pd(OAc) + H2
1.85E–02
2
Pd (OAc) + H
2
4
2
2
2
2
2
Pd (OAc) + H
Pd(OAc) + Pd + 2HOAc
2
2
4
Pd (OAc) + H
Pd (OAc) + Pd(OAc) + 2HOAc
2.29E–03
1.41E+01
3
6
2
2
2
Pd (OAc) + H
3
Pd (OAc) + Pd + 2HOAc
2
6
2
4
1
0/11 2Pd (OAc)2
Pd (OAc)4
2
4
1
1
1
2/13 Pd + H2
PdH2
3
4
2.99E+09
1.80E+10
1.00E+07
2.03E+03
4/15 2Pd
Pd2
6/17 Pd + H2
Pd H2
2
2
18
19
20
21
22
23
24
25
26
27
Pd(OAc) + PdH
2
2Pd + 2HOAc
2
Pd(OAc) + Pd H
2
Pd + Pd + 2HOAc
2
2
2
Pd (OAc) + PdH
Pd (OAc) + Pd + 2HOAc
1.06E+08
1.06E+12
2
4
2
2
2
Pd (OAc) + PdH
2
Pd(OAc) + 2Pd + 2HOAc
2
4
2
Pd (OAc) + Pd H
Pd (OAc) + Pd + 2HOAc
2
4
2
2
2
2
2
Pd (OAc) + Pd H
2
Pd(OAc) + Pd + Pd + 2HOAc
2 2
4
2
2
Pd (OAc) + PdH
Pd (OAc) + Pd(OAc) + Pd + 2HOAc
4.99E+03
5.34E+03
3
6
2
2
2
2
2
Pd (OAc) + PdH
Pd (OAc) + 2Pd + 2HOAc
2 4
3
6
Pd (OAc) + Pd H
3
Pd (OAc) + Pd(OAc) + Pd + 2HOAc
2 2 2 2
6
2
2
Pd (OAc) + Pd H
Pd (OAc) + Pd + Pd + 2HOAc
8.41E+05
3
6
2
2
2
4
2
2
8
9
Pd (OAc) + H
2
2Pd + 2HOAc
Pd (OAc) + 2Pd + 2HOAc
5
6
2
2
2
Pd (OAc) + H
4
4
2
2
2
30
31
32
33
Pd (OAc) + PdH
2
3Pd + 2HOAc
2
2
Pd (OAc) + PdH
Pd (OAc) + 3Pd + 2HOAc
1.06E+09
4
4
2
2
2
Pd (OAc) + Pd H
2
Pd + 2Pd + 2HOAc
2
2
2
2
Pd (OAc) + Pd H
4
Pd (OAc) + Pd + 2Pd + 2HOAc
2
4
2
2
2
2
3
3
3
4/35 Pd(OAc) + Pd
Pd (OAc)2
7
1.24E+18
2.09E+11
2
2
6/37 Pd (OAc) + Pd
Pd (OAc) + Pd(OAc)
2 2 2
2
4
8/39 Pd (OAc) + Pd
Pd (OAc) + Pd (OAc)
4
1.06E+16
3
6
2
2
2
*
*
The steps making a significant contribution to the overall reaction are set off in bold type.
*Units of measure: monomolecular reactions, min ; bimolecular reactions not involving hydrogen gas, g mol min–1; bimolecular reacꢀ
±
±
–
1
–1
tions involving hydrogen gas, atm– min (g is the weight of the support in grams).
1
–1
E n stands for ×10 .
n
tors responsible for the autocatalytic character of the
reaction as a whole.
As was noted above, unlike the palladium(II) aceꢀ
tates, the palladium(I) acetates are practically irreducꢀ
Note that the reduction of palladium(II) acetates, ible by dihydrogen and, at 120
particularly that of the monomer and dimer, occurs reduced slowly by PdH (step 31). The active species in
not only under the action of dihydrogen or palladium this reaction is the Pd(I) tetramer. Kinetic data demꢀ
hydrides, but also under the action of Pd(0) (steps 34 onstrate that the Pd (OAc) , Pd (OAc) , Pd, Pd , and
and 36). The latter results both from the reduction of Pd concentrations are low throughout the reacꢀ
palladium(II) acetates and from disproportionation tion; that is, they are shortꢀlived intermediates.
°C and below, are
2
2
4
4
4
2
H
2
2
(
step 39). These steps also make a certain contribution
The kinetic data obtained in this study make it posꢀ
sible to compare the mechanisms of Pd(II) reduction
to the autocatalytic character of the overall process.
KINETICS AND CATALYSIS Vol. 52
No. 2
2011

KINETICS AND MECHANISM OF PALLADIUM(II) ACETATE REDUCTION
303
in solution (for triphenylphosphine acetate complexes Table 3. Rate constants and activation energies of the main
steps of the reaction
as an example) and on the support surface. The reacꢀ
tion is autocatalytic in both cases, but its mechanisms
are somewhat different. In both cases, the initial pallaꢀ
dium(II) compounds are reduced relatively slowly by
dihydrogen to palladium(I) compounds. The trimeric,
dimeric, and monomeric forms of palladium(II) aceꢀ
tate on the surface show different reactivities. Signifiꢀ
cant distinctions between the reaction in solution and
the reaction on the surface are observed in the reducꢀ
tion of the palladium(I) compounds. In solution, the
Pd(I) compounds are reduced by dihydrogen or pallaꢀ
dium hydrides much more rapidly than the Pd(II)
compounds. Just the reverse situation is observed on
the surface, and palladium(I) reduction in this case is
almost entirely due to palladium hydrides.
Rate constant, g mol–1 min–1
Ea
kJ/mol
,
Step no.
50°
С
120°С
2
4
5
2.54
1.65
×
×
102
103 ± ±30% 5.34
±
30% 4.99
×
×
103 ± ±30%
103 ± ±20%
45 ± ±
8
7
2
18 ± ±
forms of palladium. This can be very helpful in the
synthesis of catalysts with uncommon properties, priꢀ
marily materials with a high Pd(I) content (5% and
above).
The resulting Pd(0) reacts with the initial Pd(II)
both in solution and on the surface to yield Pd(I) comꢀ
pounds, which are then reduced to Pd(0). In both
cases, the palladium hydrides reduce the initial Pd(II)
compounds more rapidly than dihydrogen does. The
difference is that, on the surface, the reduction of the
monomeric form of Pd(II) to Pd(0) takes place
ACKNOWLEDGMENTS
The authors are deeply grateful to Yu.Z. Karasev
for assistance in taking IR spectra and to
A.V. Naumkin for XPS analyses.
(
step 5), as is demonstrated by kinetic modeling. It is
REFERENCES
most likely that this process occurs without intermediꢀ
ate formation of Pd(I) compounds:
1
. William, H.J., Catalysis in Organic Synthesis, New York:
Academic, 1980, p. 314.
2. Maitlis, P.M., The Organic Chemistry of Palladium, New
Pd(OAc) +H
2
PdH(OAc) + HOAc,
Pd + HOAc.
2
PdH(OAc)
York: Academic, 1974, vol. 2, p. 216.
3
. Stakheev, A.Yu., Mashkovskii, I.S., Baeva, G.N., and
Telegina, N.S., Ross. Khim. Zh., 2009, vol. 53, no. 2,
p. 68.
For solutions of palladium(II) triphenylphosphine
acetate complexes, the formation of monomeric speꢀ
cies is not assumed [17].
Thus, if there were no step 5, Pd(0) accumulation
on the surface would proceed much more slowly and
this would lead to a significant decrease in the overall
reaction rate.
4. Semikolenov, V.A., Usp. Khim., 1992, vol. 61, p. 320.
5
. Lisitsyn, A.S., Parmon, V.N., Duplyakin, V.K., and
Likholobov, V.A., Ross. Khim. Zh., 2006, vol. 50, no. 4,
p. 140.
6
. Toebes, M.L., van Dillen, J.A., and de Jong, K.P.,
The complexes of both Pd(II) and Pd(I) can react
with various reductants, such as carbon monoxide,
hydrogen, etc. [17, 20], and it was demonstrated that
the redox properties of Pd(I) vary in a very wide range,
depending on the nature of the ligands [20]. It can,
therefore, be expected that the ratio of the rates of the
Pd(II)ꢀtoꢀPd(I) and Pd(I)ꢀtoꢀPd(0) reductions by
hydrogen will depend on the reaction conditions, on
the composition of the complexes, and on the properꢀ
ties of the ligands (support). Our observation that
Pd(I), as compared to Pd(II), is reduced less rapidly
on the surface than in solutions of phosphine acetate
complexes may have at least two causes. Firstly, the
triphenylphosphine ligands present in these solutions
stabilize the Pd(0) complexes [25] and can thereby
favor Pd(I) reduction to Pd(0). Another possible cause
is the interaction of the Pd(I) compounds with Pd(0)
particles on the surface, which leads to a decrease in
the reduction rate.
J. Mol. Catal. A: Chem., 2001, vol. 173, p. 75.
7
. Xia, W., Schulter, O.F.K., Liang, C., van den
Berg, M.W.E, Guraya M., and Muhlerl, M., Catal.
Today, 2005, vols. 102–103, p. 34.
. Sung Hwa Jhung, Bull. Korean Chem. Soc., 2005,
vol. 26, p. 563.
. Antolini, E., Appl. Catal., B, 2009, vol. 88, p. 1.
0. Berenblyum, A.S. and Hussein Ali AlꢀWadhaf, Vestn.
MITKhT, 2010, vol. 5, no. 5, p. 43.
1. Gorskii, V.G., Planirovanie kineticheskikh eksperimenꢀ
tov (Design of Kinetic Experiments), Moscow: Nauka,
8
9
1
1
1984.
1
2. Katsman, E.A. and Sobolev, O.B., Informatsionnyi
byulleten’ po khimicheskoi promyshlennosti SEV (Inforꢀ
mation Bulletin on the Chemical Industry of the
CMEA Countries), 1982, vol. 68, no. 5 (98), p. 47.
3. Zakharov, A.Yu. and Turchaninov, V.I., STIFF—proꢀ
gramma dlya resheniya zhestkikh system obyknovennykh
differentsial’nykh uravnenii (STIFF: A Solver for Stiff
Sets of Ordinary Differential Equations), Moscow:
Inst. Prikl. Mat. Akad. Nauk SSSR, 1977.
1
Thus, this kinetic study of the interaction between
carbonꢀsupported palladium(II) acetate and hydrogen
demonstrated the possibility of obtaining materials 14. Gorskii, V.G., Katsman, E.A., Klebanova, F.D., and
with controllable proportions of different valence
Grigor’ev, A.A., Teor. Eksp. Khim., 1987, vol. 23, p. 181.
KINETICS AND CATALYSIS Vol. 52
No. 2
2011

3
1
1
04
BERENBLYUM et al.
5. Wilde, D.J., Optimum Seeking Methods, Englewood 21. Lyalina, N.N., Dargina, S.V., Sobolev, A.N., et al.,
Cliffs, N.J.: Prentice Hall, 1964.
Koord. Khim., 1993, vol. 19, p. 57.
6. Polak, L.S., Gol’denberg, M.Ya., and Levitskii, A.A.,
Vychislitel’nye metody v khimicheskoi kinetike (Compuꢀ
tational Methods in Chemical Kinetics), Moscow:
Nauka, 1984.
2
2. Nakomoto, K., Infrared and Raman Spectra of Inorꢀ
ganic and Coordination Compounds, New York: Wiley,
1986.
1
7. Berenblyum, A.S., Grigor’ev, A.A., Katsman, E.A., 23. Stoyanov, E.S., J. Struct. Chem., 2000, vol. 41, p. 440.
Mund, S.L., Knizhnik, A.K., and Moiseev, I.I., Kinet.
Katal., 1982, vol. 23, p. 1494.
8. http://srdata.nist.gov/xps/index.htm.
9. Filho, J.E.C., Rev. Bras. Saude Matern. Infant. (Recife)
24. Moiseev, I.I., Stromnova, T.A., Vargaftik, M.N.,
Mazo, G.Ja., Kuz’mina, L.G., and Struchkov, Yu.T.,
J. Chem. Soc., Chem. Commun., 1978, no. 1, p. 27.
1
1
,
2
004, vol. 4, p. 225.
0. Temkin, O.N and Bruk, L.G., Usp. Khim., 1983,
vol. 52, p. 206.
2
5. Berenblyum, A.S., Knizhnik, A.G., Mund, S.L., and
Moiseev, I.I., J. Organomet. Chem., 1982, vol. 234,
p. 219.
2
KINETICS AND CATALYSIS Vol. 52
No. 2
2011