Supported palladium nanomaterials as catalysts for petroleum chemistry: 2. Kinetics and specific features of the mechanism of selective hydrogenation of phenylacetylene in the presence of carbon-supported palladium nanocatalyst

Article abstract of DOI:10.1134/S0965544115020048

The selective hydrogenation of phenylacetylene (PhA) into styrene (St) in the presence of a palladium nanocatalyst has been investigated. Salient features of this reaction have been revealed, such as independence of the PhA hydrogenation and St hydrogenat

Full text of DOI:10.1134/S0965544115020048

ISSN 0965ꢀ5441, Petroleum Chemistry, 2015, Vol. 55, No. 2, pp. 118–126. © Pleiades Publishing, Ltd., 2014.
Original Russian Text © A.S. Berenblyum, H.A. AlꢀWadhaf, E.A. Katsman, 2015, published in Neftekhimiya, 2015, Vol. 55, No. 2, pp. 125–133.
Supported Palladium Nanomaterials as Catalysts for Petroleum
Chemistry: 2. Kinetics and Specific Features of the Mechanism
of Selective Hydrogenation of Phenylacetylene in the Presence
of CarbonꢀSupported Palladium Nanocatalyst
A. S. Berenblyuma, H. A. AlꢀWadhafb, and E. A. Katsmana
aLomonosov State University of Fine Chemical Technologies, Moscow, Russia
eꢀmail: ayb@go.ru
bHajjah University, Hajjah, Republic of Yemen
Received July 15, 2014
Abstract—The selective hydrogenation of phenylacetylene (PhA) into styrene (St) in the presence of a palꢀ
ladium nanocatalyst has been investigated. Salient features of this reaction have been revealed, such as indeꢀ
pendence of the PhA hydrogenation and St hydrogenation (after exhaustion of PhA) rates from the PhA and
St concentrations, respectively; their nearly firstꢀorder dependence on the hydrogen pressure; and the first
order in catalyst loading. The experimental results obtained according to a special plan have been processed
using computer programs for various reaction mechanisms. It has been found that adequate description is
achieved for the only scheme involving the adsorption one to two PhA and/or St molecules and one to two
hydrogen molecules on the active site of the same type. The presence of one hydrogen molecule on the active
site leads to the hydrogenation of PhA to St or St to ethylbenzene (EB). In the presence of two adsorbed
hydrogen molecules, PhA is selectively hydrogenated to St and nonselectively hydrogenated directly to EB.
It has been shown that the nature of the selectivity of the palladium catalyst is determined by the thermodyꢀ
namics of competitive adsorption of PhA and St. The activation energies of the individual steps of the process
have been determined.
Keywords: kinetics, mechanism, selective hydrogenation, phenylacetylene, kinetic modeling, palladium
nanocatalysts
DOI: 10.1134/S0965544115020048
The catalytic reaction of selective hydrogenation of role in the performance of a selective hydrogenation
catalyst.
acetylene and its derivatives has been of continuous
interest to researchers for decades [1–3]. The interest
is mainly due to the fact that the quality of ethylene
produced by pyrolysis and used for polymerization
very strongly depends on the concentration of trace
The development of nextꢀgeneration catalysts and
improvement in the technology of selective hydrogeꢀ
nation of acetylene compounds with no practical loss
of ethylene hydrocarbons are impossible without
acetylene. The latter is typically removed by catalytic knowledge of the hydrogenation kinetics and mechaꢀ
nism. There are a number of publications on the kinetꢀ
ics of the heterogeneous catalytic hydrogenation of
acetylenes to olefins; see, for example, [1, 2, 11,
hydrogenation in the presence of a palladium catalyst
supported on a carbon or an oxide carrier, with ethylꢀ
ene loss being at the minimal level [4, 5]. The purificaꢀ
tion of styrene (St), a highꢀproductionꢀvolume monoꢀ
mer, by selective hydrogenation for the removal of
impurity phenylacetylene (PhA) present in styrene is a
practically important process as well [6]. This reaction
is also important from the theoretical point of view for
understanding the nature of selectivity of processes in
the presence of palladium catalysts [7].
15⎯17]. Unfortunately, the available published data on
the kinetics and mechanism of this reaction important
from both practical and theoretical viewpoints are
obviously insufficient for understanding the similariꢀ
ties and dissimilarities between various catalysts, subꢀ
strates, and conditions of hydrogenation.
The purpose of this work was to create a complete
kinetic model that would adequately describe the
kinetics of phenylacetylene and styrene hydrogenation
over a palladiumꢀonꢀcarbon nanocatalyst with indeꢀ
pendent varying the reaction conditions (PhA and St
Recent studies of palladium nanocatalysts on both
traditional supports [8, 9] and nanowires or nanotubes
[10–13] have stirred particular interest. As shown in
[14], the particle size of palladium plays an important concentrations, hydrogen pressure, catalyst loading,
118

SUPPORTED PALLADIUM NANOMATERIALS AS CATALYSTS
119
temperature) within a relatively wide range and with a flame ionization detector and an HPꢀultra2
thereby to get new information about the underlying fusedꢀsilica capillary column of a 50 m length and a
features of the reaction mechanism of selective hydroꢀ 0.2 mm inner diameter, coated with polymethylsiloxꢀ
genation of acetylenes and olefins.
ane and 5% phenylsiloxane (column oven temperaꢀ
ture, up to 250 ; evaporator temperature, 270 ),
°C
°C
was used. The results of the analysis were processed
using the normalization procedure run by the software
program Chromatec Analytic. The uncertainty of the
experimental data (PhA, St, and ethylbenzene (EB)
EXPERIMENTAL
Reagents and Materials
concentrations) was 7–8%
.
Reagentꢀgrade dimethylformamide (DMF) and
highestꢀpurity ethanol were distilled prior to experiꢀ
ments, chemically pure palladium diacetate Pd(OAc)2
was used without further purification, and HCl and
NaOH solutions were 0.1 N titration standards. Gasꢀ
eous hydrogen of grade A with a purity of 99.99% and
nitrogen of the special purity grade (1st grade) with a
purity of 99.999% were used without purification. Gas
mixtures of hydrogen with nitrogen were prepared by
admitting an appropriate amount of hydrogen and
then nitrogen into a cylinder. Thereafter, the gases
were mixed in the cylinder by its rotation for several
hours and letting to stay overnight. The resulting gas
mixture was analyzed chromatographically.
Special experiments showed that for the reactor at
a shaking rate of 180 cycle/min and a catalyst particle
size of less than 0.044 mm, the hydrogenation reaction
proceeds without significant diffusion limitations.
Kinetic experiments were performed under the folꢀ
lowing conditions: a hydrogen partial pressure of 0.30
to 1.00 atm; a catalyst charge of 0.05–0.2 g/L; and iniꢀ
tial PhA and St concentrations of 0 to 0.64 and
0.42 mol/L, respectively.
Kinetic Simulation Procedure
The kinetic model was developed on the basis of
our own design of a kinetic experiment (Table 1). It is
known that a design of kinetic experiments limited by
the number and conditions of experimental runs can
often be a source of ambiguity of evaluation of kineticꢀ
model parameters, wherein the greater the number of
parameters to be evaluated, the larger should be the
number of different experiments [21, 22]. Therefore,
the effectiveness of the design intended for use should
be assessed by calculating the rank of the correspondꢀ
ing information matrix and identifying the linear
dependence of its columns. These operations make it
possible to establish which model parameters are
determined uniquely and which are found only in the
form of the soꢀcalled “complexes” of the parameters
and to calculate their errors [23–25].
Our experimental design included measuring PhA,
St, and EB concentrations in samples for 4–15 samꢀ
pling times. The initial PhA concentration, the partial
pressure of hydrogen, the amount of the catalyst, and
the temperature were varied in the experiments. The
design also included experiments on styrene hydrogeꢀ
nation in the absence of PhA and the cohydrogenation
of PhA and St. Thus, the experimental design used for
kinetic analysis almost completely eliminated the
ambiguity of estimates of the model parameters,
which could occur in the case of incorrect experimenꢀ
tal design.
The catalyst support was Vulcan XCꢀ72 carbon
available from Cabot, which is a kind of carbon black
having a specific surface area of about 250 m2/g and a
particle size less than 0.044 mm [18]. Prior to syntheꢀ
sis, the support was calcined in an oxygen atmosphere
at a temperature of 450°C for 4 h [19].
Catalyst Preparation
Acetone in an amount of 50 mL was added to 0.5 g
of carbon support. The resulting mixture was thorꢀ
oughly agitated by sonication in a Eumax UD150SHꢀ
6 L ultrasonic cleaner at room temperature for 20 min
to form a uniform suspension. Then, a preliminarily
prepared solution of a specified amount of Pd diaceꢀ
tate in 15 mL of solvent was poured. The resulting
mixture was stirred for 15 min, and then the solvent
was evaporated by gentle heating (50–60°C) to dryꢀ
ness in a rotary evaporator.
Carbonꢀsupported palladium diacetate was
reduced in a stream of molecular hydrogen using the
equipment and conditions described in [20].
Procedures for Catalytic Experiments
A jacketed glass shaker reactor was charged with an
appropriate amount of the catalyst and 10 mL of the
Programs for solving the direct and inverse kinetic
solvent (DMF). Then, the catalyst in the shaking reacꢀ problems and for numerical analysis of the identifiꢀ
tor was treated with hydrogen (temperature, 25 ability of the parameters and calculation of the errors
°C;
hydrogen flow rate, 0.3 L/h; treatment time, 30 min). of the estimates are described in [23]. The mathematꢀ
Thereafter, the corresponding amount of PhA under ical model was a set of autonomous systems of ordiꢀ
the hydrogen stream was introduced and hydrogenaꢀ nary differential equations (rate equations of PhA, St,
tion started at a predetermined temperature (25 or and EB concentrations for the hypothesis under conꢀ
50
°C). The reaction mixture was periodically sampled sideration) with the initial conditions defining the
for GLC analysis. A Kristall 2000M chromatograph number of experiments.
PETROLEUM CHEMISTRY Vol. 55
No. 2
2015

120
BERENBLYUM et al.
Initial concentrations
Table 1. Conditions of kinetic experiments
Catalyst,
mmol Pd/L
Hydrogen
pressure, atm
Run no. Temperature, °C
PhA, mol/L
St, mol/L
EB, mol/L
1
2
25
25
25
25
25
25
25
25
25
25
50
50
50
50
50
0.220
0.433
0.0045
0.635
0.433
0.433
0.433
0.433
0
0
0
0.167
0.178
0.169
0.164
0.0797
0.0428
0.195
0.163
0.155
0.0985
0.163
0.150
0.163
0.179
0.163
1.00
1.00
1.00
1.00
1.00
1.00
0.505
0.309
1.00
1.00
1.00
0.485
1.00
1.00
0.309
0
0
3
0
0
4
0
0
5
0
0
6
0
0
7
0
0
8
0
0
9
0.436
0.396
10
11
12
13
14
15
0.433
0.220
0.433
0.635
0.433
0.433
0.416
0
0
0
0
0
0
0
0
0
0
0
The inverse kinetic problem was solved by the least
squares method. The objective function was a
weighted meanꢀsquare “experiment – calculation”
deviation, where measurement error of the correꢀ
RESULTS AND DISCUSSION
For kinetic studies, we used the nanocatalyst 9.1%
palladium on carbon, which had a metal particle size
of 2.8 0.2 nm according to TEM data and a narrow
size distribution with no large particles [25].
Typical results of the kinetic experiments are shown
in Figs. 1–3. Note that in the absence of the catalyst,
neither PhA nor St undergoes transformations under
the experimental conditions. In the case of catalytic
hydrogenation, no compounds other than PhA, St, or
EB were detected by GLC.
sponding concentration was taken as the weight, and it
1/2
had the following form
F
= {Σ[(Сei–С
ci)/εi]2/
N}
×
100%, where Сe is the experimentally measured conꢀ
centration; Сc is the corresponding value calculated by
the model,
ε
is the measurement error; i is the
sequence number of a measurements through all runs,
measured concentrations, and sampling times; and
is the total number of measurements.
N
Analysis of the data reveals certain characteristic
features. The rate of PhA conversion is almost indeꢀ
pendent of the concentration in the range examined
(Fig. 1), corresponding to zeroꢀorder kinetics. This
fact was reported in other papers as well (see, for
example, [11, 15]), but substantially different values
for the kinetic order in acetylene compounds in the
range of –2 to +1 were also observed [1, 2, 15]. The
rate of the PhA hydrogenation reaction is proportional
to the catalyst charge as has been also shown in [15,
26]. Note that this behavior was observed over the
entire range of the reaction conditions examined in
this study.
While PhA is present in the solution, the buildup of
EB is very slow and has no induction period. After the
almost complete consumption of PhA, the St hydroꢀ
genation and EB formation rate increases dramatiꢀ
cally and has almost zero order with respect to styrene
(Figs. 2, 3). A similar behavior was observed in the case
of hydrogenation of a preliminarily prepared PhA
mixture with St. It is interesting that during the initial
period of the reaction when the concentration of styꢀ
PhA concentration, mol/L
0.7
0.6
4
0.5
0.4
0.3
7
2
0.2
0.1
5
8
6
12
1
0
50
100
Time, min
150
200
Fig. 1. Time dependence of the phenylacetylene (PhA)
concentration (data point symbols and lines refer to experꢀ
iment and calculation, respectively). The experimental run
numbers and conditions are as given in Table 1.
PETROLEUM CHEMISTRY Vol. 55
No. 2
2015

SUPPORTED PALLADIUM NANOMATERIALS AS CATALYSTS
121
St concentration, mol/L
Concentration, mol/L
1.0
0.6
0.5
0.4
0.3
0.2
0.1
0.8
0.6
0.4
0.2
9, EB
10, St
4
10, ЭБ
6
12
9, St
8
5
2
7
1
0
50
100
150
200
250
0
50
100
150
Time, min
200
250
300
Time, min
Fig. 2. Time dependence of the styrene (St) concentration
(data point symbols and lines refer to experiment and calꢀ
culation, respectively). The run numbers and experimental
conditions are as given in Table 1.
Fig. 3. Time dependence of the styrene (St) and ethylbenꢀ
zene (EB) concentration (data point symbols and lines
refer to experiment and calculation, respectively). The
experimental run numbers and conditions are as given in
Table 1. Run 9 is the hydrogenation of St after PhA hydroꢀ
genation, and run 10 is the hydrogenation of a mixture of
St and PhA.
rene is very small, ethylbenzene seems to be produced
immediately from PhA. This conclusion is consistent
with the assumption [1, 3] that alkanes can be formed
in this reaction directly from alkynes without the
release of olefins into the reaction medium.
this behavior in terms of the previously published
models.
One of the most common models is as follows:
molecules of the substrate to be hydrogenated and
molecular hydrogen are activated at different sites of
the same nature (hypothesis I). In this case, one site
adsorbs one hydrogen or substrate molecule. The
resulting adsorption complexes react with each other
[7, 15]:
Our data also show that the PhA and St hydrogenaꢀ
tion rates have the rate order close to unity in hydrogen
pressure and also increase with temperature.
As we suggested previously [1] in accordance with
Xꢀray diffraction data [27], the interaction of Pd parꢀ
ticles with acetylene compounds in the presence of H2
leads to the formation of active sites that contain Pd–
C bonds of the carbide type and are sufficiently stable
under the reaction conditions [28]. The nascent new
sites are less active in the hydrogenation of styrene.
This fact is confirmed by our experiments showing
that styrene is hydrogenated at twice greater rate on
the fresh catalyst before its treatment with PhA and
hydrogen than after such treatment. This treatment
increases the selectivity of the hydrogenation of acetyꢀ
lene compounds, which is important for practical purꢀ
poses. During the PhA hydrogenation experiment, the
reaction rate was not found to vary. This means that
the formation of new sites in this case is faster than
hydrogenation, and the latter proceeds on these newly
formed sites. This assumption is also supported by the
fact that the PhA and St hydrogenation rates are conꢀ
stant in several successive cycles. For other catalysts
and solvents, a stationary catalyst activity can be estabꢀ
lished not immediately, as in our case, but only after
several hydrogenation cycles [15].
Z + Pa = ZPa,
Z + St = ZSt,
Z + H2 = ZH2,
ZPa + ZH2
ZSt + ZH2
→
2Z + St,
→
2Z + Eb,
where Z is the active site, PhA is phenylacetylene, St is
styrene, and Eb is ethylbenzene. Symbol = implies
that equilibrium is established quickly, and
a slow nonequilibrium step.
→
refers to
The treatment of our experimental data in accorꢀ
dance with this scheme showed the model to result in
relations of the current reactant concentrations to the
time, the initial PhA and St concentrations, and the
hydrogen pressure that strongly differ from the
observed relationships. The difference is especially
prominent for the concentration of EB during the iniꢀ
tial period of the reaction (e.g., see the data shown in
Fig. 4). The use of this model gives the “experiment –
calculation” standard deviation of 38%.
Another documented scheme that we have used
includes sites of two types, of which one adsorbs a subꢀ
strate molecule and the other absorbs a hydrogen molꢀ
ecule (hypothesis II):
Our experiments showed that the replacement of
DMF by ethanol does not lead to significant changes
in the kinetic results.
The behavior discussed above does not go beyond
the limits of the facts reported previously by other
authors. Accordingly, it seemed reasonable to describe
Z + Pa = ZPa,
PETROLEUM CHEMISTRY Vol. 55
No. 2
2015

122
EB concentration, mol/L
BERENBLYUM et al.
In this case, the “experiment – calculation” stanꢀ
dard deviation was 11% (Fig. 4).
0.5
0.4
0.3
0.2
0.1
The next model suggests one type of site that can
bind up to two molecules of the reactants, including
hydrogen, nine adsorption complexes of different
compositions altogether. Furthermore, the reaction of
the adsorbed substrate molecule with hydrogen actiꢀ
vated at the same (or another) site and hydrogen from
the gas phase as well (Eley–Rideal mechanism [7]) is
suggested (hypothesis IV). Thus, the “experiment –
calculation” standard deviation decreased to 8.3%
(Fig. 4).
I
V
III
IV
II
0
50
100
150
Time, min
200
250
300
The best results were obtained with the use of the
model in which each active site of the same nature can
adsorb to two molecules of PhA and/or St and to two
hydrogen molecules (Fig. 4, hypothesis V). Note that
this model suggests the mode of the “noncompetitive
adsorption” of hydrogen [16]; i.e., the Н2 adsorption
constants do not depend on the amount and type of
substrate molecules adsorbed on the active site and
vice versa. Involved in the hydrogenation reaction is
only the hydrogen adsorbed on the same site that conꢀ
tains substrate molecules. This description is adeꢀ
quate.
Fig. 4. Time dependence of the ethylbenzene (EB) conꢀ
centration (data points and lines refer to experimental run
8 and calculation, respectively). The run number and
experimental conditions are as given in Table 1. The roman
numbers refer to the hypotheses defined in the text.
Z + St = ZSt,
X + H2 = XH2,
ZPa + XH2
ZSt + XH2
→
Z + X + St,
→
Z + X + Eb,
where X is an active site of the second type. Although
the description is somewhat improved, the error is still
as high as 18% (Fig. 4).
This model will be considered in detail below. It
includes the following steps.
A significant improvement in the results is obtained
by using the schemes that we proposed previously for
various reactions in which the active site is able to bind
more than one reactant molecule [1, 24]. The same
assumption was made for the catalytically active pallaꢀ
dium metal complex clusters in solution [29]. It is
believed that these mechanisms can mostly characterꢀ
ize supported heterogeneous nanocatalysts and soluꢀ
ble clusters because of the specifics of their structure
[14]. Within the frame of this approach, we examined
the following three models that significantly improved
the quality of the description.
Contribution of
step
1
2
Equation of step
Step number
strong
weak
Z + Pa = ZPa
ZPaH2 St + Z
(1)
(11)
(17)
(18)
(2)
→
weak
ZPa(H2)2
ZPa(H2)2
→
St + ZH2
Eb + Z
strong
medium
strong
medium
strong
strong
strong
strong
weak
→
Z + St = ZSt
ZStH2 Eb + Z
ZSt(H2)2 Eb + ZH2
→
(12)
(19)
(3)
→
Z* + H2 = Z*H2
Z + 2Pa = ZPa2
First, we considered the model in which one subꢀ
strate molecule is activated on a site of one type and a
site of the other type can activate up to two hydrogen
molecules (hypothesis III):
(6)
ZPa2H2
→
St + ZPa
St + ZPaH2
Eb + ZPa
(13)
(20)
(21)
(7)
ZPa2(H2)2
ZPa2(H2)2
→
→
Z + Pa = ZPa,
Z + St = ZSt,
strong
medium
strong
medium
strong
strong
strong
weak
Z*+ 2H2 = Z*(H2)2
Z + 2St = ZSt2
(8)
X + H2 = XH2,
X + 2H2 = X(H2)2,
ZSt2H2
ZSt2(H2)2
Z+ Pa + St = ZPaSt
→
Eb + ZSt
(14)
(23)
(9)
→
Eb + ZStH2
ZPa + XH2
→
→
→
Z + X + St,
Z + X + Eb,
Z + XH2 + St,
ZPaStH2
ZPaStH2
→
St + ZSt
(15)
(16)
(25)
ZSt + XH2
→
Eb + ZPa
ZPa + X(H2)2
ZPaSt(H2)2
1
→
St + ZStH2
Z* is the active site with molecules of unsaturated substrates adꢀ
sorbed on it; the number refers to the program by which the modꢀ
el calculations were carried out and is a mere formality.
ZSt + X(H2)2
→
Z + XH2 + Eb,
Z + X + Eb.
2
ZPa + X(H2)2
→
PETROLEUM CHEMISTRY Vol. 55
No. 2
2015

SUPPORTED PALLADIUM NANOMATERIALS AS CATALYSTS
123
Table 2. Equilibrium constants of adsorption steps (25
°
C)*
Step number
Equation
Equilibrium constant
Dimension
Error, logK
1
2
3
6
7
8
9
Z + PhA = ZPhA
Z + St = ZSt
Z* + H2 = Z*H2
Z + 2PhA = ZPhA2
Z* + 2H2 = Z*(H2)2
Z + 2St = ZSt2
9.88
7.00
1.01
8.04
5.80
7.80
8.59
×
×
×
×
×
×
×
108
102
100
109
10–1
102
109
L/mol
L/mol
1/atm
L2/mol2
1/atm2
L2/mol2
L2/mol2
0.38
0.24
Z + PhA + St = ZPhAS
t
Table 3. Equilibrium constants of adsorption steps (50
°
C)*
Step number
Equation
Equilibrium constan
Dimension
Error, logK
1
2
3
6
7
8
9
Z + PhA = ZPhA
Z + St = ZSt
Z* + H2 = Z*H2
Z + 2PhA = ZPhA2
Z* + 2H2 = Z*(H2)2
Z + 2St = ZSt2
1.07
7.00
3.84
1.30
1.80
1.41
2.96
×
×
×
×
×
×
×
108
L/mol
L/mol
1/atm
L2/mol2
1/atm2
L2/mol2
L2/mol2
100
10–1
1010
10–1
102
0.23
0.28
Z + PhA + St = ZPhASt
1010
* The dimensions of the parameters include concentrations of the reactants (mol/L), except for hydrogen, for which partial pressure
(atm) is used. Empty box means that the parameter is determined only in the “complex”.
ecules (steps 18, 21, 26). The standard deviation of the
concentrations calculated by the model from experiꢀ
mentally measured values (total 170 measurements in
Contribution of
step
1
2
Equation of step
Step number
strong
weak
ZPaSt(H2)2
→
Eb + ZSt
(26)
(28)
10 experiments at 25°С and 64 measurements in
ZPaSt(H2)2
1
→
Eb + ZPaH2
5 experiments at 50 ) is 7.8 or 6.7% at 25 or 50°С,
°С
respectively. Tables 2–5 show the values of the equilibꢀ
rium constants and rate constants calculated for this
model.
Z* is the active site with molecules of unsaturated substrates adꢀ
sorbed on it; the number refers to the program by which the modꢀ
el calculations were carried out and is a mere formality.
2
Identifiability analysis showed that the equilibrium
constants of the fast reversible steps of PhA and St
adsorption (steps 1, 2, 6, 8, 9) cannot be uniquely
determined. In this group of parameters, only their
It is evident that this scheme is based on the Langꢀ
muir–Hinshelwood approach suggesting the adsorpꢀ
tion of both reactants and products [7], as well as the
intermediates [24]. The adsorption involves fast
reversible steps 1–3 and 6–9, and the rest steps are the
irreversible reactions of adsorbed substrate molecules
with hydrogen.
This scheme does not involve the unlikely steps in
which two molecules of unsaturated compounds
adsorbed on one site are synchronously hydrogenated
by two hydrogen molecules adsorbed on the same site.
These are kinetically equivalent to steps 20, 23, and
26, included in the scheme, in which only one subꢀ
strate molecule is hydrogenated by one hydrogen molꢀ
ecule. It is noteworthy that limiting the number of
hydrogen molecules adsorbed on one site to one Н2
molecule leads to an inadequate description of the
experimental data, but an increase in their number to
more than two does not improve the description.
ratios, for example, К1/К2, К6/К8, К1/К9, etc., are
uniquely determined because of large values of the
constants. This finding suggests that there are almost
no unoccupied active sites on the surface in the presꢀ
ence of PhA and/or St.
The equilibrium constants for hydrogen adsorption
on active sites (steps 3 and 7) are uniquely determined,
though with a considerable error (Tables 2, 3). Their
value indicates comparable amounts of free sites and
the sites occupied by one or two hydrogen molecules.
The rate constants of the hydrogenation steps were
uniquely determined, except for the two steps at 50°С
(Table 5) because of their relatively low rate.
Based on the above data, including the catalyst
charge (CPd) expressed as the palladium concentration
per unit reaction volume (mol/L), we derived the folꢀ
lowing rate equations:
The formation of ethylbenzene in the initial step of
the reaction according to the scheme is explained in
terms of the simultaneous hydrogenation of the
adsorbed PhA molecule by two bound hydrogen molꢀ
wPha = (– k13K1θI
–
k15K9θI
–
k18K1θII
– k20K6θII – k26K9θII)CPd
/Den,
PETROLEUM CHEMISTRY Vol. 55
No. 2
2015

124
BERENBLYUM et al.
Table 4. Significant rate constants and activation energies of the steps (25
°
C)**
Step number
Equation
EB + Z
Rate constant, min–1
Error, log
k
E
, kJ/mol
a
12
13
14
15
16
18
19
20
23
26
ZStH2
→
5.69
4.28
2.79
1.09
9.24
2.62
5.16
2.80
8.30
3.10
×
×
×
×
×
×
×
×
×
×
101
101
102
102
100
101
101
102
101
101
0.36
0.52
0.53
0.26
0.40
0.19
0.54
0.23
1.04
0.23
124
54.6
~0
ZPhA2H2
→
St + ZPhA
ZSt2H2
→
EB + ZSt
ZPhAStH2
ZPhAStH2
ZPhA(H2)2
→
→
→
St + ZSt
52.0
33.0
148
~0
EB + ZPhA
EB + Z
ZSt(H2)2
→
EB + ZH2
St + ZPaH2
EB + ZStH2
EB + ZSt
ZPhA2(H2)2
→
29.7
~0
ZSt2(H2)2
→
ZPhASt(H2)2
→
30.0
Table 5. Significant rate constants and activation energies of the steps (50
°
C)**
Step number
Equation
EB + Z
Rate constant, min–1
Error, logk
12
13
14
15
16
18
19
20
23
26
ZStH2
ZPhA2H2
ZSt2H2
→
2.76
2.35
2.79
5.52
2.59
2.62
5.16
7.07
8.30
7.90
×
×
×
×
×
×
×
×
×
×
103
102
102
102
101
103
101
102
101
101
0.22
0.28
0.27
0.22
0.23
0.52
–
→
St + ZPhA
→
EB + ZSt
ZPhAStH2
ZPhAStH2
ZPhA(H2)2
→
→
→
St + ZSt
EB + ZPhA
EB + Z
ZSt(H2)2
→
EB + ZH2
St + ZPhAH2
EB + ZStH2
ZPhASt(H2)2 EB + ZSt
ZPhA2(H2)2
→
0.30
–
ZSt2(H2)2
→
→
0.28
** The dash means that the error is too large because of a small contribution of the step.
w
St = (– k12K2θI
+
k13K1θI
k19K2θII
– k23K8θII CPd
EB = ( k12K2θI k14K8θI
k19K2θII k23K8θII
where wi is the rate of buildup of the
K3PH /D is the coverage of the active site by one
–
k14K8θI
k20K6θII
en,
k16K9θI
+
k15K9θI
Recall that the kinetic equations obtained comply
with the Langmuir–Hinshelwood mechanism involvꢀ
ing competitive adsorption of unsaturated substances
and noncompetitive adsorption of hydrogen.
– k16K9θI
–
+
)
/D
The hydrogenation of PhA to St is predominantly
mediated by the active sites containing two adsorbed
PhA molecules, whose proportion in the total amount
of active sites is close to 100% up to a high conversion
of PhA. In addition, the rate constants of transformaꢀ
tion of such adsorption complexes, which also contain
one or two hydrogen molecules, have the highest value
among the other rate constants. Thus, the hydrogenaꢀ
tion of St to EB is slow while the system contains PhA
and the main contribution to the rate of nonselective
direct conversion of PhA into EB is made by the active
site containing one PhA molecule, one St molecule,
and two hydrogen molecules.
w
+
+
+
k18K1θII
en,
+
+
+
k26K9θII
)CPd/
D
ith substance, θI =
2
hydrogen molecule, θII
=
K7P2H /D is the coverage of
2
the active site by two hydrogen molecules,
Den = 1 +
K1CPa
+ K2CSt + + + K9CPaCSt, D = (1 +
K6C2Pa K8C2St
K3PH
+
K7P2H ).
2
2
In the rate equations, the concentrations of liquid
After exhaustion of PhA, as in the case of experiꢀ
reactants are expressed in mol/L and the concentraꢀ ments on hydrogenation of pure styrene, the process
tion of hydrogen, which is in a very small amount in further develops similarly to the above picture for St
the liquid phase, was taken as its pressure in the gas alone: the main contribution to the St conversion rate
phase in accordance with Henry’s law.
is made by active sites containing two adsorbed St
PETROLEUM CHEMISTRY Vol. 55
No. 2
2015

SUPPORTED PALLADIUM NANOMATERIALS AS CATALYSTS
125
molecules, whose concentration and rate constant of
In summary, it can be concluded that in order to
conversion into EB have the highest values. This derive an adequate kinetic model of selective hydrogeꢀ
explains the almost zero rate order for the hydrogenaꢀ nation of phenylacetylene on a palladiumꢀonꢀcarbon
tion of PhA or St in the absence of PhA.
nanocatalyst, an extended experimental design conꢀ
taining a sufficiently large array of experimental data
obtained under various experimental conditions
should be used. Furthermore, the selection of various
reaction mechanisms showed that the best schemes of
the reaction of interest are those differing from the
classical Langmuir–Hinshelwood concepts by the
possibility of adsorbing more than one PhA and St, as
well as hydrogen molecule on one active site. Note that
the descriptions of the kinetics of selective hydrogenaꢀ
tion in terms of the classical scheme as found in the litꢀ
erature were satisfactory only in a limited range of
variation of the experimental conditions [11, 15], so
that reliability and uniqueness of the mathematical
description of the results are questionable.
The results obtained in this study allowed us to forꢀ
mulate some new ideas on the mechanism of hydrogeꢀ
nation of acetylenic and olefinic compounds. They
primarily concern the determining role of active sites
containing two alkyne (or alkene plus alkyne, or two
alkene) molecules, as well as one or two hydrogen
molecules. Selective hydrogenation of phenylacetyꢀ
lene to styrene or nonselective hydrogenation to ethylꢀ
benzene occurs on the active site with one or two H2
molecules, respectively.
The analysis of the calculated parameters of the
model shows that the adsorption constant of the first
PhA molecule is many orders of magnitude above that
1
for St . At the same time, the rate constants for hydroꢀ
genation of adsorbed PhA and St differ significantly
smaller (Tables 2, 3). This implies that the high selecꢀ
tivity of the hydrogenation of PhA to St on the pallaꢀ
dium catalysts has the thermodynamic nature. Note
that the selectivity of PhA hydrogenation to St
decreases with an increase in PhA conversion, temꢀ
perature, and hydrogen pressure or a decrease in the
initial concentration of PhA.
The dependences on the hydrogen pressure are also
interesting. Namely, the adsorption constants for one
or two hydrogen molecules have the same order of
magnitude and are small as compared with those for
PhA and St (Tables 2, 3). As the hydrogen pressure
increases, the proportion of the adsorption complex
with two hydrogen molecules increases, but the proꢀ
portion of sites that are not occupied by hydrogen
remains quite high over the entire range of its partial
pressure. Note that the rate constant of the conversion
of the adsorption complex with two hydrogen moleꢀ
cules into the reaction products is higher than that
of the complex with one molecule of hydrogen
(Tables 4, 5). What is responsible for the nonselective
hydrogenation of PhA to EB at the initial time period
of the reaction (when there is no St yet) is the adsorpꢀ
Furthermore, despite a wide variety of contact palꢀ
ladium catalysts (supported metal and nanocatalysts,
homogenous metal complex systems) and reaction
conditions (gasꢀphase and liquidꢀphase hydrogenaꢀ
tion, the solvent), their high selectivity is associated
with the thermodynamics of the step of adsorption of
acetylene and olefin compounds [1, 30].
tion complex ZPa(H2)2
.
As noted above, the dependence of the hydrogenaꢀ
tion rate upon the partial pressure of hydrogen is close
to the firstꢀorder rate law, but it does not follow this
law. This is obviously due to the following circumꢀ
stances. The proportion of adsorption complexes conꢀ
taining hydrogen increases with the increasing presꢀ
sure, but H2 does not occupy completely all of the
active sites. In our case, the mixing of different depenꢀ
dences of the coverage of sites by one and two hydroꢀ
gen molecules, with the rate of the conversion being
different, leads to the overall dependence that is close
to a linear relation. The hydrogenation rate equation
given in [16] as a function of hydrogen pressure makes
it possible to obtain different dependences of the reacꢀ
tion rate on hydrogen pressure (for example, zero, linꢀ
ear, and quadratic relations or a linear transforming to
a zeroꢀorder relation) in different cases.
REFERENCES
1. I. M. Zhvanetskii, F. D. Klebanova, E. A. Katsman,
and A. S. Berenblyum, Kinet. Katal. 31, 893 (1990).
2. S. D. Jackson and L. A. Shaw, Appl. Catal. A: Gen. 134
,
91 (1996).
3. A. Molnar, A. Sarkany, and M. Varga, J. Mol. Catal. A:
Chem. 173, 185 (2001).
4. New Organic Synthesis Processes, Ed. by S. P. Chernykh
(Mir, Moscow, 1991).
5. S. A. Nikolaev, L. N. Zanaveskin, V. V. Smirnov, et al.,
Usp. Khim. 78, 248 (2009).
6. Siqin Li, Juntao Liu, Fengxia Sun, et al.; Method for
selective hydrogenation of phenylacetylene in the presꢀ
ence of styrene, RU Patent No. 2 505 519 (2009) [in
Russian].
According to our estimates, the activation energies
of steps 13, 15, and 20 of PhA hydrogenation to St and
styrene hydrogenation to EB (steps 12, 14, 16, 19, 23)
lie in the ranges of ~30–55 and 0–124 kJ/mol, respecꢀ
tively (Table 4).
7. O. V. Krylov, Heterogeneous Catalysis (Akademkniga,
Moscow, 2004) [in Russian].
8. J. M. Nadgeri, M. M. Telkar, and C. V. Rode, Catal.
Commun. 9, 441 (2008).
1
9. Jiawei Hu, Zhiming Zhou, Rui Zhang, et al., J. Mol.
Note that the adsorption constants of the second PhA or St molꢀ
ecule is lower compared with the first molecule (Tables 2, 3).
Catal. A: Chem. 381, 61 (2014).
PETROLEUM CHEMISTRY Vol. 55
No. 2
2015

126
BERENBLYUM et al.
10. P. Tribolet and L. KiwiꢀMinsker, Catal. Today 105, 337 21. V. G. Gorskii, Design of Kinetic Experiments (Nauka,
(2005).
Moscow, 1984) [in Russian].
22. V. G. Gorskii, E. A. Katsman, F. D. Klebanova, and
11. N. Semagina, A. Renken, and L. KiwiꢀMinsker, Chem.
A. A. Grigor’ev, Teor. Eksp. Khim. 23, 191 (1987).
Eng. Sci. 62, 5344 (2007).
23. A. S. Berenblyum, H. A. AlꢀWadhaf, E. A. Katsman,
12. H. Jin, S. E. Park, J. M. Lee, and S. K. Ryu, Carbon 34
,
and V. R. Flid, Kinet. Catal. 52, 296 (2011).
429 (1996).
24. A. S. Berenblyum, T. A. Podoplelova, E. A. Katsman,
et al., Kinet. Catal. 53, 595 (2012).
25. A. S. Berenblyum, H. A. AlꢀWadhaf, E. A. Katsman,
et al., Pet. Chem. 54, 105 (2014).
26. R. V. Chaudhari, R. Jaganathan, D. S. Kolhe, et al.,
13. S. DominguezꢀDominguez, A. BerenguerꢀMurcia,
B. K. Pradhan, et al., J. Phys. Chem. C 112, 3827
(2008).
14. A. Yu. Stakheev, I. S. Mashkovskii, G. N. Baeva, and
N. S. Telegina, Russ. J. Gen. Chem. 80, 618 (2010).
15. R. V. Chaudhari, R. Jaganathan, D. S. Kolhe, et al.,
Chem. Eng. Sci. 41, 3073 (1986).
16. Z. M. Shakoor, M. A. Zablouk, and A. A. Shuhaib,
Eng. Tech. J. 72, 1301 (2009).
17. B. A. Wilhite, M. J. McCready, and A. Varma, Ind. Eng.
Appl. Catal. 29, 141 (1987).
27. A. Frankievich, A. Yanko, and D. Stachurski, in Proꢀ
ceeding of the 4th International Symposium on Heterogeꢀ
neous Catalysis (Varna, 1979), part I, p. 135.
28. Wei Tew Min, M. Nachtegaal, M. Janousch, et al.,
Phys. Chem. Chem. Phys. 14, 5761 (2012).
Chem. Res. 41, 3345 (2002).
29. O. N. Temkin, Homogeneous Metal Complex Catalysis:
Kinetic Aspects (Akademkniga, Moscow, 2008) [in Rusꢀ
sian].
18. F. Su, X. Li, L. Lv, and X. S. Zhao, Carbon 44, 801
(2006).
19. E. Antolini, Appl. Catal. B: Environ. 88, 1 (2009).
30. A. S. Berenblyum, Zh. Vses. Khim. Oꢀva im. D.I. Menꢀ
deleeva 32 (1), 82 (1987).
20. A. S. Berenblyum and H. A. AlꢀWadhaf, Vestn. MITHT
5
(5), 43 (2010).
Translated by S. Zatonsky
PETROLEUM CHEMISTRY Vol. 55
No. 2
2015