Selective hydrogenation of acetylene in excess ethylene using Ag- and Au-Pd/SiO2 bimetallic catalysts prepared by electroless deposition

Article abstract of DOI:10.1016/j.apcata.2013.10.024

Selective hydrogenation of acetylene in excess ethylene has been investigated for a series of Ag- and Au-Pd/SiO₂ bimetallic catalysts having incremental surface coverages of Ag and Au on the Pd surface. The catalysts were prepared by electroless deposition and characterized by atomic absorption spectroscopy and selective chemisorption. Both Ag- and Au-Pd/SiO ₂ catalysts showed increased selectivity of acetylene conversion to form ethylene, rather than ethane, at high coverages. The catalyst performance results suggest that at high coverages of Ag or Au on Pd, that result in small ensembles of Pd sites, acetylene is adsorbed as a π-bonded species that favors hydrogenation to ethylene. At low coverages, where ensemble sizes of contiguous Pd surface sites are much larger, acetylene is strongly adsorbed as a multi-σ-bonded species which preferentially forms ethane, lowering the selectivity to ethylene. Both kinetic analyses and the calculated turnover frequencies of acetylene conversion and ethane formation were consistent with the above explanation. The similar performance trends for Ag- and Au-Pd/SiO ₂ suggest that the bimetallic effect for these catalysts is geometric and not electronic in nature.

Full text of DOI:10.1016/j.apcata.2013.10.024

Applied Catalysis A: General 469 (2014) 419–426
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Selective hydrogenation of acetylene in excess ethylene using Ag- and
Au–Pd/SiO2 bimetallic catalysts prepared by electroless deposition
Yunya Zhang, Weijian Diao, Christopher T. Williams, John R. Monnier∗
Department of Chemical Engineering, University of South Carolina, 2C02 Swearingen Engineering Center, 301 Main Street, Columbia, SC 29208, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 8 July 2013
Received in revised form 8 October 2013
Accepted 14 October 2013
Available online 19 October 2013
Selective hydrogenation of acetylene in excess ethylene has been investigated for a series of Ag- and
Au–Pd/SiO2 bimetallic catalysts having incremental surface coverages of Ag and Au on the Pd surface. The
catalysts were prepared by electroless deposition and characterized by atomic absorption spectroscopy
and selective chemisorption. Both Ag- and Au–Pd/SiO2 catalysts showed increased selectivity of acetylene
conversion to form ethylene, rather than ethane, at high coverages. The catalyst performance results
suggest that at high coverages of Ag or Au on Pd, that result in small ensembles of Pd sites, acetylene is
adsorbed as a ␲-bonded species that favors hydrogenation to ethylene. At low coverages, where ensemble
sizes of contiguous Pd surface sites are much larger, acetylene is strongly adsorbed as a multi-␴-bonded
species which preferentially forms ethane, lowering the selectivity to ethylene. Both kinetic analyses
and the calculated turnover frequencies of acetylene conversion and ethane formation were consistent
with the above explanation. The similar performance trends for Ag- and Au–Pd/SiO2 suggest that the
bimetallic effect for these catalysts is geometric and not electronic in nature.
Keywords:
Hydrogenation
Acetylene
Bimetallic catalysts
Ag–Pd
Au–Pd
Electroless deposition
Kinetics
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
that the adsorption mode of multi-␴ bonded ethylidyne can lead to
the formation of ethane as it undergoes sequential hydrogenation
Ethylene is used as a starting point for many chemical inter-
mediates in the petrochemical industry. The yearly demand for
ethylene is over 150 million tons with an annual growth rate of 3.5%
[1]. It is predominantly produced through steam cracking of higher
hydrocarbons (ethane, propane, butane, naphtha, and gas oil). Dur-
ing the cracking process, a small amount of acetylene is produced
as a side product. However, acetylene must be removed since it
acts as a poison for ethylene polymerization catalysts at even ppm
concentrations (>5 ppm) [2]. Thus, the selective hydrogenation of
acetylene to ethylene is an important process for the purification
of ethylene. Conventional, low weight loading Pd catalysts are used
for this selective reaction in high concentration ethylene streams.
Selectivity of acetylene to ethylene is a key objective, since over
hydrogenation of acetylene to ethane results in a decrease in yield
and must be recycled back to the ethane cracking unit.
steps: ethylidyne → ethylidene → ethyl → ethane [6,7]. The forma-
tion of ethylidyne requires Pd ensembles that have three adjacent
adsorption sites on the catalyst surface. Therefore, it is reasonable
to assume the selectivity of acetylene hydrogenation can be con-
trolled by changing the ensemble sizes of catalytic surface atoms to
influence the forms of the adsorbed species on the catalyst surface.
In more recent work, Neurock [8] has used a combination of Monte
Carlo and DFT computations to study hydrogenation of acetylene
and ethylene mixtures over Pd(1 1 1) and a Pd50%Ag50% structure on
the Pd(1 1 1) surface and concluded that while strongly-adsorbed,
three-fold ethylidyne species are adsorbed on Pd ensembles on
the (1 1 1) surface, they are largely unreactive and lead to catalyst
fouling. His calculations suggest that a three-fold, adsorbed vinyl
species (C=CH2) is adsorbed on the Pd surface which is reactive for
the formation of both ethylene and some of the ethane. Addition of
Ag to the Pd surface lowers the overall rate of hydrogenation but
shifts product selectivity to ethylene. What is common between
three-fold, adsorbed ethylidyne and adsorbed vinyl species is that
large Pd ensembles favor strongly adsorbed, multi-␴-intermediates
that favor the formation of ethane.
To understand the factors that influence the selectivity in acety-
lene hydrogenation, it is important to consider the adsorption
modes of acetylene on the catalyst surface. Fig. 1 shows identi-
fied adsorption modes of acetylene on a Pd surface, some in the
presence of pre-adsorbed H [3–6]. Both associative and dissocia-
tive acetylene adsorption can occur. Margitfalvi et al. [7] showed
Thus, over the years, bimetallic Pd catalysts have been
developed for the selective hydrogenation of acetylene. Various
additives, such as Ag, Ni, Cu, Au, Pb, Tl, Cr and K, have been reported
to improve the performance of Pd catalysts, especially in achiev-
ing high selectivity for ethylene production [9–12]. However, the
∗
Corresponding author. Tel.: +1 803 777 6813; fax: +1 803 777 8265.
E-mail address: monnier@cec.sc.edu (J.R. Monnier).
0926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.10.024

420
Y. Zhang et al. / Applied Catalysis A: General 469 (2014) 419–426
Fig. 1. Adsorption modes of acetylene on Pd surface [3–6].
bimetallic effects of the above additives are not yet well understood,
possibly because the conventional methods of bimetallic catalyst
preparation (co-impregnation and successive impregnation) result
in both monometallic and bimetallic particles with varying com-
positions which make it difficult to determine the position of the
two metallic components. The bimetallic interaction occurs only
when the two metallic components form bimetallic surface com-
positions instead of separate particles. An alternative method to
prepare bimetallic catalysts with controlled and proximal contact
between the two metals is electroless deposition (ED), which can
be applied to a wide range of metals. ED is an industrially feasible
method which has been widely studied in our laboratory to syn-
thesize many different bimetallic catalyst compositions [13–20].
Since the ED process is kinetically controlled, the final composition
of a particular bimetallic catalyst can be controlled to give rather
precise combinations of the two metallic components.
In this paper, catalyst evaluation and kinetic data are presented
for the selective hydrogenation of acetylene in excess ethylene over
Ag–Pd/SiO2 and Au–Pd/SiO2 catalysts prepared by the ED method.
The goals of this paper are not only to discuss the improvements
in catalysts selectivity of acetylene to ethylene during acetylene
hydrogenation, but also to illustrate the roles of Ag and Au additives
on the adsorption modes of acetylene on Pd.
conducted at room temperature, and the baths were placed in dis-
posable polyethylene beakers to prevent cross-contamination. The
experiments were carried out while stirring on a magnetic stir-
ring plate to minimize external mass transfer limitations. Dropwise
addition of a concentrated solution of sodium hydroxide (NaOH,
pellets, J.T. Baker) was used to maintain pH of the bath solu-
tion at ∼9.0 0.5 during the entire experiment; a pH probe was
immersed in the ED bath throughout the deposition sequence.
Baths were prepared by adding all the required components except
the monometallic Pd/SiO2 catalyst. Addition of the monometal-
lic Pd/SiO2 catalyst initiated the deposition. Liquid aliquots (1 mL)
were periodically taken and filtered using 4.5 ␮m/mesh syringe fil-
ters to remove catalyst particles at different time intervals. These
liquid samples were diluted by adding 5 vol% nitric acid (HNO3,
68–70%, BDH) and analyzed by atomic absorption spectroscopy
(AAS, PerkinElmer AAnalyst 400) to monitor the concentrations of
Ag+ and Au+ remaining in the bath. After 4 h of deposition time,
the solution was filtered and the catalyst was washed thoroughly
with 2 L of 18.2 Mꢀ-cm de-ionized water (Thermo Scientific Barn-
stead Nanopure Ultrapure Water System) to remove all soluble
ligands and salts. The filtrates were then dried in vacuo at room tem-
perature and stored at ambient conditions. Thus, complete series
of Ag–Pd/SiO2 and Au–Pd/SiO2 bimetallic catalysts were prepared
with varying weight loadings of Ag and Au metals on the Pd sur-
face.
2. Experimental
2.1. Catalyst preparation
2.2. Catalyst characterization
A master batch of 1.85 wt% Pd/SiO2 (sieved to 60/80 mesh
particles), supplied by BASF Catalysts, was used as the base
monometallic Pd catalyst. The surface area of the silica was mea-
sured by the BET method to be 100 m2/g. The dispersion of Pd
was determined by hydrogen titration of oxygen pre-covered Pd
sites using a Micromeritics AutoChem 2920 Automated Analyzer.
The dispersion of 8.6% corresponded to calculated Pd particles
of 13.2 nm diameter and 9.0 × 1018 surface Pd sites/g cat. The
somewhat large Pd particles were by intent; the same base cat-
alyst had been used in a previous study where specific activities
were higher for larger Pd particles. Controlled loadings of Ag and
Au on the Pd surface were added using the electroless deposi-
tion method described earlier [19,20]. Briefly, potassium silver
cyanide (KAg(CN)2, 99% purity, Technic, Inc.) and potassium gold
cyanide (KAu(CN)2, 99% purity, Sigma-Aldrich) were used as the
metal sources, and hydrazine (N2H4, 35 wt% aqueous solution,
Sigma-Aldrich) was used as the reducing agent. Initial concen-
trations of the metal salts were selected based on the desired,
theoretical coverage of Ag or Au on Pd, assuming monolayer depo-
sition and 1:1 surface stoichiometry of Ag or Au deposited on Pd.
The molar ratio of the reducing agent to metal salt was initially
selected at 10:1. For higher concentrations of the metal salt a 20:1
ratio was used to compensate for normal time-dependent, ther-
mal decomposition of the reducing agent in the ED bath. Typically,
the volume of the electroless bath was 50–100 mL for each gram
of the monometallic catalyst. The deposition of Ag and Au was
Catalyst compositions were confirmed using atomic absorption
spectroscopy. Samples were digested in aqua regia (3:1 of HCl to
HNO3 by volume) at 120 ◦C for 4 h and then diluted within the AAS
detection limits. Calibration curves were determined in advance
using standards with known concentrations.
The concentrations of Pd surface sites for the base Pd/SiO2 cata-
lyst as well as the bimetallic Ag–Pd/SiO2 and Au–Pd/SiO2 catalysts
were determined by chemisorption using hydrogen pulse titra-
tion of oxygen pre-covered Pd sites. Chemisorption was performed
using a Micromeritics Autochem II 2920 automated chemisorption
analyzer. All samples were reduced at 200 ◦C in 20% H2/balance
argon for 2 h and then purged with pure argon for 1 h, before cool-
ing to 40 ◦C in argon. Next, 10% O2/balance helium was flowed
over the catalyst for 30 min to form Pd-O surface species, and
then purged with pure argon for 30 min to remove residual gas
phase oxygen and weakly adsorbed oxygen on the support. Finally,
10% H2/balance argon was dosed until all surface Pd-O species
reacted with hydrogen to form water and Pd–H surface species.
For each Pd atom, 1.5 H2 molecules are consumed. Hydrogen con-
sumption was quantitatively measured using a high sensitivity
thermal conductivity detector (TCD). Since Ag and Au are inac-
tive for hydrogen–oxygen titration, the concentration of Pd surface
sites that were covered by Ag or Au metals can be determined by
subtracting the Pd surface site concentration of the bimetallic cata-
lysts from the total number of surface Pd sites for the monometallic
Pd/SiO2 catalyst.

Y. Zhang et al. / Applied Catalysis A: General 469 (2014) 419–426
421
Table 1
2.3. Catalyst evaluation
Ag–Pd/SiO2 and Au–Pd/SiO2 bimetallic catalysts that were evaluated for acetylene
hydrogenation. Base catalyst is 1.85 wt% Pd/SiO2. Theoretical coverage in mono-
layers (ML) represents monodisperse coverage of Ag or Au on Pd surface, assuming
1:1 adsorption stoichiometry.
The monometallic Pd/SiO2 and the series of bimetallic
Ag–Pd/SiO2 and Au–Pd/SiO2 catalysts were evaluated for selective
hydrogenation of acetylene in the presence of ethylene. Catalysts
were evaluated in a single pass, 0.19 in ID, packed bed, tubu-
lar reactor (316 stainless steel). The reactor was encased in a 1.0
in OD, jacketed shell with liquid inlet and exit ports at the bot-
tom and top of the shell, respectively, which was connected to
an ethylene glycol/H2O recirculation bath to maintain isothermal
behavior at 65 ◦C for this highly exothermic reaction. The reactor
was wrapped with heating tape outside the jacket to heat the reac-
tor to 200 ◦C when the catalyst was pretreated in situ. The reactor
was typically loaded with 0.005 g of catalyst diluted by 0.030 g of
the silica support (20/40 mesh) to form the catalyst bed that was
supported on glass wool in the middle of the reactor. A thermo-
couple was inserted into the catalyst bed to accurately monitor the
reaction temperature and to ensure isothermal behavior. All cat-
alysts were pretreated in situ at 200 ◦C in flowing 20% H2/balance
He for 2 h before being cooled to 65 ◦C for evaluation. All gas flows
were maintained by mass flow controllers; the reaction feed stream
for catalyst screening consisted of 1% C2H2, 20% C2H4, 5% H2, bal-
ance He at a total flow rate of 50 SCCM to roughly approximate
the tail-end feed of an ethylene cracker. To ensure kinetic mea-
surements were free from external mass transfer limitations, a
standard loading (0.005 g) of the most active catalyst, 1.85 wt%
Pd/SiO2, was evaluated over the flow range 50–250 SCCM, corre-
sponding to GHSV values 6.0 × 105 to 3.0 × 106 h−1. In all cases
the TOF for C2H2 conversion varied between 4.2 and 4.6 s−1 indi-
cating no external mass transfer limitations over this flow range.
Acetylene at 1% concentration was added from a pre-mixed cylin-
der of 10% C2H2/He by making the proper dilution. Both reaction
feed and products were evaluated using an automated, on-line
Hewlett-Packard 5890 Series II gas chromatograph using flame ion-
ization detection. The feed and product analyses were typically
made every 1.5 h over the full length of the run. In addition to C2H2,
C2H4, and C2H6, a number of C4 hydrocarbons (n-butane, 1-butene,
cis-2-butene, trans-2-butene, and 1,3-butadiene) were identified
and quantitatively analyzed. These products can be considered as
precursors to the “green oil” or C2 oligomers commonly observed
during industrial operation. All C4 products were combined to give
a total product and rates of formation that were normalized to C2
feeds, i.e., the molar quantities of C4 products were multiplied by 2.
All catalysts exhibited high, initial activities and underwent varying
degrees of deactivation for the first several hours on line. Therefore,
all the reaction data reported here were based on stable catalyst
performance after 15 to 20 h on line to eliminate transient behavior.
The kinetic measurements for acetylene hydrogenation were
studied using the same reactor system. In order to maintain dif-
ferential conversion conditions, flow rates were varied between
50 and 200 SCCM and the reaction temperature was lowered from
65 to 50 ◦C. For determination of C2H2 reaction orders, partial
pressures were varied between 0.005 and 0.015 atm while holding
PH constant at 0.05 atm, and for calculation of H2 dependencies,
pa2rtial pressures ranged between 0.02 and 0.10 atm and PC
Ag (wt%)
Ag (mol%)
Theoretical
Actual
coverage (ML)
coverage (ꢁAg)
0.19
0.25
1.07
1.20
1.55
2.05
2.10
0.0018
0.0023
0.0099
0.0111
0.0144
0.0190
0.0195
1.16
1.52
6.52
7.31
9.44
0.44
0.53
0.73
0.83
0.85
0.91
0.99
12.49
12.79
Au (wt%)
Au (mol%)
Theoretical
Actual
coverage (ML)
coverage (ꢁAg)
0.28
0.58
2.19
3.50
4.40
5.01
0.0014
0.0029
0.0111
0.0178
0.0223
0.0254
0.93
1.93
7.31
11.67
14.68
16.71
0.44
0.54
0.69
0.84
0.91
0.96
were evaluated for the hydrogenation of acetylene. Even though
catalysts with coverages of Ag and Au lower than those shown
in Table 1 were prepared, they were not evaluated since there
were no apparent reaction effects at the lower Group IB coverages.
To achieve the highest coverages of Ag and Au on Pd, autocat-
alytic deposition of Ag on Ag and Au on Au also occurred, which
is observed clearly in Fig. 2. Transition from catalytic deposition
to autocatalytic deposition occurs at approximately 0.4 monolayer
of Ag or Au metal deposited on Pd. Below this level of deposition,
catalytic deposition predominates. Above this point, both catalytic
and autocatalytic depositions occur, so the actual coverage devi-
ates substantially from the theoretical coverage, where theoretical
coverage in monolayers (ML) is defined as monodisperse coverage
of Ag or Au on Pd surface sites, assuming 1:1 adsorption stoichiom-
etry. In order to obtain the highest fractional coverages of Ag and
Au on Pd, it was necessary to deposit considerably more than one
theoretical monolayer amount of Ag and Au. Thus, complete cover-
age can still be attained after substantial autocatalytic deposition.
In all cases, autocatalytic deposition increased as the concentration
of the second metal increased on the surface of Pd, as expected for
kinetically-controlled deposition processes. Since both Ag and Au
were inactive for hydrogenation (2 wt% Ag/SiO2 and 2 wt% Au/SiO2
showed no activity for either C2H2 or C2H4 hydrogenation), the for-
mation of multilayer aggregates on the Pd surface should have no
significant effects on catalyst performance.
H
2
was kept constant at 0.01 atm. The activation energies at differen2-
tial conversion conditions were determined between 40 and 80 ◦C
using the standard feed composition of 1% C2H2, 20% C2H4, 5% H2,
balance He.
3. Results and discussion
Fig. 2. Actual coverage of Ag or Au on Pd/SiO2 as a function of moles of Ag or Au
deposited. Black line is the theoretical coverage of Ag or Au metal on Pd surface at a
1:1 deposition stoichiometry. Dashed lines show coverage based on chemisorption
of Pd sites.
A series of Ag–Pd/SiO2 and Au–Pd/SiO2 bimetallic catalysts were
prepared with increasing coverage of Ag and Au on Pd. Table 1
shows a summary of the Ag–Pd/SiO2 and Au–Pd/SiO2 catalysts that

422
Y. Zhang et al. / Applied Catalysis A: General 469 (2014) 419–426
Fig. 3. (A) Conversion of acetylene and selectivity of acetylene to ethylene and (B)
selectivity of acetylene to ethane and C4s as a function of Ag coverage on Pd/SiO2.
Reaction conditions: 65 ◦C and feed composition of 1% C2H2, 5% H2, 20% C2H4, and
balance He at GHSV = 6.0 × 105. Representative error bars for all data are shown for
samples having ꢁAg = 0.92.
Fig. 4. (A) Conversion of acetylene and selectivity of acetylene to ethylene and (B)
selectivity of acetylene to ethane and C4s as a function of Au coverage on Pd/SiO2.
excess as a feed component, making gas chromatographic analysis
inaccurate.
The Ag–Pd/SiO2 and Au–Pd/SiO2 catalysts that are listed in
Table 1 were evaluated for selective hydrogenation of acety-
lene in the presence of ethylene; the results are summarized in
Figs. 3 and 4, respectively. Conversion of C2H2 and selectivity of
C2H2 to C2H4, C2H6 and C4s were calculated from the following
relationships:
Fig. 3A shows conversion of acetylene and selectivity of hydro-
genation of acetylene to form ethylene as a function of Ag coverage
on Pd/SiO2. The base Pd/SiO2 catalyst exhibited C2H2 conversion of
∼60%, which remained essentially constant over the range of ꢁAg
from 0 to 0.7. Similarly, over this same range, the selectivity to C2H4
remained at approximately 50%. The conversion decreased sharply
at ꢁAg > 0.7, while the selectivity for C2H2 conversion increased
notably, rising to 90% at ꢁAg > 0.9; at these high Ag coverages,
C2H2 reacted
Conversion of C2H2
=
C2H2(in)
C2H2 formed
C2H2 reacted + C2H6 formed + C4s formed × 2
Selectivity of C2H2 to C2H4
Selectivity of C2H2 to C2H6
Selectivity of C2H2 to C4s =
=
=
C2H6 formed
C2H2 reacted + C2H6 formed + C4s formed × 2
C4s formed × 2
C2H2 reacted + C2H6 formed + C4s formed × 2
C2H2 reacted = C2H2(in) − C2H2(out)
C2H6 formed = C2H6(out) − C2H6(in)
C4s formed = n − C4H10(out) + t − 2 − C4H8(out) + 1 − C4H8(out) + c − 2 − C4H8(out)
+1,3 − C4H6(out)
where (in) and (out) represent the feed stream and the product
stream, respectively. This method of calculation cannot accurately
be used for C2H4 hydrogenation, since C2H4 is a reaction product
from selective C2H2 hydrogenation, and is also present in great
the remaining Pd ensembles are necessarily very small, changing
the modes of acetylene adsorption. The results for the Au-Pd/SiO2
series of catalysts in Fig. 4 show similar behavior. Conversion

Y. Zhang et al. / Applied Catalysis A: General 469 (2014) 419–426
423
and selectivity remain essentially constant up to Au coverages of
0.7. At higher Au coverages, the C2H2 conversion decreases and
the selectivity towards C2H2 conversion increases similarly to the
Ag–Pd/SiO2 series. Somewhat surprisingly, for both the Ag–Pd and
Au–Pd series of catalysts the selectivities to C4 products are rela-
tively constant at 8–10% based on C2H2 conversion. The C4 products
are derived from C2H2, and not C2H4, since no C4 products were
observed when C2H2 was not present in the feed. The sites that
give rise to C4 formation do not appear to be affected by the pres-
ence of either Ag or Au on the Pd surface, suggesting that ensemble
effects are not responsible for oligomerization of C2H2. Others [3]
have speculated that influences of the support may be a factor in the
formation of C4 products, such as those that exist at the Pd-support
interface. These sites would not be affected by the presence of either
Au or Ag on the Pd surface.
Somorjai [21] has reported that ethylidyne (Fig. 1) was formed
rapidly upon exposure of C2H4 to a Pt(1 1 1) surface at ∼300 K;
similarly, adsorption of C2H2 in the presence of pre-adsorbed H
also formed ethylidyne at 350 K. This mode of strong adsorption
required a minimum of three contiguous Pt atoms, preferably on
(1 1 1) surfaces according to Somorjai. Hydrogenation of such inter-
mediates resulted primarily in the formation of ethane. Since Pd is
similar to Pt in catalytic reactivity trends, adsorption of C2H2 and
C2H4 on the Pd(1 1 1) surface should exhibit the same pattern as
the adsorption of C2H2 and C2H4 on the Pt(1 1 1) surface. In fact,
Medlin [22] has calculated that C2H2 is preferentially adsorbed
over the three-fold hollow sites on the (1 1 1) surfaces of Pt, Pd,
and Rh, confirming similar adsorption modes for these metals.
However, Medlin concluded that the adsorbed acetylene species
reactive toward selective hydrogenation was not the highly coordi-
nated species which was bound very strongly to the metal surface.
Instead, the high-coordination sites likely become irreversibly
blocked by strongly adsorbed acetylene leaving exposed only the
low coordination sites for reversible C2H2 adsorption and subse-
quent hydrogenation. The computational results of Neurock [8]
likewise support the results of Medlin [22] and Somorjai [21] that
adorbed ethylidyne is catalytically inactive and/or non-selectively
converted to ethylene. Finally, Norskov [23,24] has also used com-
putational methods and found that the heats of adsorption of
acetylene and ethylene were the main factors that determined
activity and selectivity for hydrogenation on Pd surfaces. Because
acetylene and ethylene adsorption energies scaled with the carbon-
surface bond energies, ethylene which adsorbed as a di-␴-species
was adsorbed at approximately 1/2 the strength of C2H2 (tetra-␴-
species). To be both active for C2H2 hydrogenation and selective for
C2H4 formation, a Pd catalyst must have sites that bind acetylene
strongly enough to undergo hydrogenation, but weakly enough to
permit desorption as C2H4. This condition is best achieved, not on
planar (1 1 1) surfaces, but on sites where multi- ␴-bonded species
are not formed. Pd surfaces with subsurface carbon or partially cov-
ered with other metals such as Ag provided the proper balance of
activity and selectivity.
principal mode of ethylene adsorption was altered from ethylidyne
to a ␲-bonded species.
Because the coverage of Ag and Au on the Pd surface can be
systematically changed by electroless deposition, it is possible to
observe the transition from multi-␴ to ␲-adsorbed species. At low
coverages of Ag on Pd, there is an abundance of contiguous Pd sur-
face sites where the strongly adsorbed acetylene and ethylene are
adsorbed as multi-␴-adsorbed species. The formation of ethlidyne
is favored when there are three adjacent adsorption sites on the
catalyst surface, or if 1,1,2,2-tetra-␴-adsorbed acetylene is the pre-
ferred mode of adsorption, four adjacent Pd surface sites would be
required. These species favor hydrogenation of acetylene to form
ethane, rather than ethylene. In addition, the excess ethylene in the
feed stream (at typical ethane cracking conditions) can also adsorb
as ethylidene, or 1,2-di-␴-ethylene and be hydrogenated to ethane.
However, at high coverages of Ag most of the Pd surface is covered
by Ag, leaving only small Pd ensembles, or even single atom sites of
Pd on the surface. In this case, acetylene is more weakly adsorbed
as a ␲-bonded species requiring potentially only a single Pd surface
site, which favors formation of ethylene.
In earlier work from our group [20] FTIR of adsorbed CO on Pd
surfaces as a function of Ag coverage was used to identify the pos-
itions of Ag deposition; Ag deposition was favored on the Pd(1 1 1)
surface up to ∼ ꢁAg = 0.4; FTIR peaks for three-fold adsorption of CO
that is favored on (1 1 1) surfaces was strongly suppressed over this
range of Ag coverage. Deposition of Ag from ∼ ꢁAg = 0.4–0.75 then
occurred primarily on (1 0 0) sites as evidenced by decreasing peak
intensities for two-fold adsorption of CO as Ag coverage increased.
Only when Ag coverages were above ꢁAg = 0.8 was Ag deposited on
the corner and edge sites of Pd, where peak intensities of linearly-
bound CO decreased with higher Ag coverages. Interestingly, for
Au deposition on the same base Pd/SiO2 catalyst, FTIR of CO indi-
cated that Au was deposited in a rather non-discriminate manner
on all surface facets and sites of the Pd surface [19]. However,
since the Pd particles in this study were quite large (average par-
ticle diameter of 13.2 nm), the Pd surface was composed primarily
of (1 1 1) facets with lesser amounts of (1 0 0) and even smaller
amounts of coordinatively-unsaturated corner and edge sites [27].
Thus, even if Au deposition was random while Ag was directed
primarily on (1 1 1) surfaces, the positions of IB deposition would
necessarily be the same to give similar behavior for C2H2 hydro-
genation. Thus, both Ag–Pd and Au–Pd catalysts should have similar
IB coverage effects for selective C2H2 hydrogenation. The exist-
ence of only linear CO adsorption at high Ag and Au coverages
suggests that only isolated Pd sites or very small Pd ensembles
were responsible for acetylene hydrogenation. Li and Shen [26]
observed similar CO adsorption results on Pd/SiO2, Pd–Ag(1:1)/SiO2
and Pd–Ag(1:4)/SiO2. Both bridged and linear adsorbed CO were
seen on the Pd/SiO2 catalyst, but on the surface of Pd–Ag/SiO2 cat-
alysts linearly adsorbed CO was the main species.
Because the concentration of Pd surface sites decreases with Ag
and Au coverage, it is necessary to determine specific activities of
these catalysts in terms of turn over frequency (TOF), or the activity
per exposed Pd site. Fig. 5 shows the TOF values for C2H2 conversion
and C2H6 formation for the Ag–Pd and Au–Pd series of catalyst. The
TOF for C2H2 conversion and C2H6 formation were calculated as
follows:
Finally, Freund [25] has also studied the adsorption of ethyl-
ene and co-adsorption of ethylene and H2 on alumina-supported
Pd catalysts and concluded that the strength of C2H4 adsorp-
tion increased gradually with increasing particle size, which led
to a redistribution of ␲-bonded and di-␴ bonded ethylene, the
latter preferentially adsorbed on the large particles. Similarly,
Li and Shen [26] investigated ethylene adsorption on Pd/SiO2,
Pd–Ag(1:1)/SiO2 and Pd–Ag(1:4)/SiO2 catalysts at room temper-
ature and found that tri-␴-bonded ethylidyne, di-␴ bonded surface
species, and a ␲-bonded species were formed on the Pd/SiO2 cat-
alyst and among the three adsorption modes ethylidyne was the
predominant species. The ␲-bonded and di-␴ bonded species was
formed on Pd–Ag(1:1)/SiO2 while only ␲-bonded was found on
Pd–Ag(1:4)/SiO2. Their work showed that with addition of Ag the
number of C2H2 molecules reacted
TOF of C2H2 conversion =
number of Pd surface sites × time (s)
number of C2H6 molecules formed
TOF of C2H6 formation =
number of Pd surface sites × time (s)
where the number of Pd surface sites was determined from
chemisorption results for each sample. The TOF values for C2H6
formation were relatively constant between 0 and 0.9 Ag and

424
Y. Zhang et al. / Applied Catalysis A: General 469 (2014) 419–426
Fig. 6. The C2H2 pressure dependencies for Pd/SiO2 and Ag–Pd/SiO2 (ꢁAg = 0.92) at
50 ◦C and H2 pressure of 0.05 atm.
electronic properties of Ag and Au should be different enough to
cause measureable differences in catalyst performance.
The kinetics of acetylene hydrogenation in excess ethylene
over the monometallic Pd/SiO2 and bimetallic Ag–Pd/SiO2 catalysts
were then investigated to verify the above hypotheses. The optimal
Ag–Pd/SiO2 catalyst with 0.92 coverage of Ag on Pd was used for
the kinetic study. Fig. 3A shows that the C2H2 conversion was ∼60%
for 5 mg Pd/SiO2 and ∼35% for 5 mg Ag–Pd/SiO2 with 0.92 Ag cov-
erage at 65 ◦C and a total flow rate of 50 SCCM. In order to obtain
differential conversion of C2H2, reaction temperature was lowered
to 50 ◦C and flow rates were changed as needed between 50 and
200 SCCM.
Fig. 5. Effect of (A) Ag and (B) Au coverage on Pd for TOF of acetylene hydrogenation.
The C2H2 pressure dependencies for Pd/SiO2 and Ag–Pd/SiO2
are shown in Fig. 6 at a reaction temperature of 50 ◦C and con-
stant H2 pressure of 0.05 atm (5%) over a C2H2 pressure range from
0.005 (0.5%) to 0.015 atm (1.5%). The plots give good straight line fits
with reaction orders of −0.67 0.06 for Pd/SiO2 and −0.20 0.04
for Ag–Pd/SiO2. The −0.67 reaction order for Pd/SiO2 is in good
agreement with the values of −0.55 to −0.67 observed by Bond [29]
and −0.66 by Tysoe [30]. Further, in the more recent work of Neu-
rock [8] calculations gave C2H2 reaction orders of −0.4 and −0.21
for Pd(1 1 1) and Ag–Pd(1 1 1), respectively, showing less kinetic
inhibition by C2H2 on the Ag-modified Pd(1 1 1) surface. The nega-
tive C2H2 pressure dependencies for both Pd/SiO2 and Ag–Pd/SiO2
catalysts confirm that the C2H2 is strongly adsorbed on the Pd sur-
face. The more negative reaction order for Pd/SiO2 than Ag–Pd/SiO2
implies that C2H2 is more strongly adsorbed on the former catalyst,
consistent with the earlier hypothesis that C2H2 is more strongly
adsorbed as ethylidyne at low coverage and weakly adsorbed as a
␲-bonded species at higher coverages of Ag.
The H2 pressure dependences for these two catalysts were mea-
sured at 50 ◦C and constant C2H2 pressure of 0.01 atm (1%) over a
H2 pressure range from 0.02 (2.0%) to 0.10 (10%) atm. The plots
shown in Fig. 7 also give good fits for reaction orders of 1.17 0.11
for Pd/SiO2 and 1.18 0.10 for Ag–Pd/SiO2, which are also in agree-
ment with the values 1.0–1.4 measured by Bond [29] and 1.04 by
Tysoe [30]. Neurock’s [8] results also showed first order dependen-
cies in H2 pressure for both Pd(1 1 1) and Ag-Pd(1 1 1) surfaces. The
clear first order H2 pressure dependencies suggest that hydrogen
is weakly adsorbed on the Pd surface of both catalysts relative to
the hydrocarbon species.
Au coverages. Conversely, the TOF values for C2H2 conversion
increased markedly at high Group IB coverages, consistent with
the assumptions proposed earlier for conversion and selectivity
requirements. At low coverages, much of the acetylene is strongly
adsorbed as multi-␴ species, which requires more Pd surface sites
than for ␲-bonded species at high coverage. Thus, one explanation
for the lower TOF of C2H2 conversion at low Ag coverage is there
are fewer adsorbed C2H2 molecules per Pd surface site. Hydro-
genation of these strongly bound ethylidyne species leads to the
preferable formation of C2H6 while hydrogenation of ␲-complex
species favors the formation of ethylene.
Further, Somorjai [28] has shown that during ethylene hydro-
genation the relative hydrogenation rates of the species adsorbed
on the Pt(1 1 1) surface are as follows: ␲-bonded ethylene ꢀ di-
ꢂ-bonded ethylene ꢀ ethylidyne and vinylidene. It would be
expected that hydrogenation of ethylidyne also occurs more slowly
than ␲-bonded acetylene on Pd(1 1 1) surfaces during acetylene
hydrogenation. The shift to ␲-adsorbed C2H2 also increases the
rate of C2H4 formation per Pd site and simultaneously decreases
the rate of C2H6 formation, giving the behavior seen in Fig. 5.
The TOF values for acetylene conversion and C2H6 formation
for Au–Pd/SiO2 are also quite similar to the corresponding values
for Ag–Pd/SiO2. Although there are measurable differences in the
TOF values, 22 and 12 s−1 for Ag–Pd and Au–Pd, respectively, the
difference is not significant enough to argue for different types
of bimetallic effects. Rather, the similar performances for Ag–Pd
and Au–Pd catalysts suggest the bimetallic effect for these two
series catalysts is geometric and not electronic in nature, since the
The temperature dependencies for acetylene conversion using
these two catalysts were measured by changing the temperature

Y. Zhang et al. / Applied Catalysis A: General 469 (2014) 419–426
425
For Ag − Pd/SiO2 (ꢁAg on Pd = 0.92),
rate = k2e− B,900 600 [p(C2H2)]−0.20 0.04[p(H2)]1.18
0.10
RT
4. Conclusions
Electroless deposition has been shown to be a very good method
for preparation of Ag–Pd/SiO2 and Au–Pd/SiO2 bimetallic cata-
lysts with selective and controlled coverages of Ag and Au on Pd.
The resulting families of catalysts make it more straightforward
to draw correlations between catalyst performance and bimetallic
surface compositions. Selectivity of acetylene to ethylene and TOF
for acetylene conversion are enhanced at high coverages of Ag or Au
on Pd, implying that the transition of adsorption mode of acetylene
onPdsurfacefromstronglyadsorbedethylidyneonlargePdensem-
bles to weakly adsorbed, ␲-bonded species on small Pd ensembles,
is responsible for selective conversion of acetylene to ethylene. This
explanation of adsorption mode is further confirmed by the kinetics
of acetylene hydrogenation. The presence of optimal coverages of
either Ag or Au on Pd lower the reaction order in C2H2 from −0.67
to −0.20 for Pd and Ag–Pd, respectively. Both catalysts exhibit first
order kinetics for H2 pressures, indicating the surface is near fully
covered by adsorbed C2H2. Likewise, the apparent activation ener-
gies for C2H2 conversion decrease from 12.1 to 8.9 kcal/mol at high
coverages of Ag on Pd. The combination of lower activation energy
and less negative reaction order in C2H2 pressure explains why the
TOF values for optimum compositions of Ag–Pd are much higher
than for Pd catalysts. The similar performance trends for Ag- and
Au–Pd/SiO2 indicate that the bimetallic effect for these catalysts is
likely geometric, and not electronic, in nature.
Fig. 7. The H2 pressure dependencies for Pd/SiO2 and Ag-Pd/SiO2 (ꢁAg = 0.92) at 50 ◦
and C2H2 pressure of 0.01 atm.
C
from 40 to 80 ◦C at C2H2 and H2 partial pressures of 0.01 and
0.05 atm, respectively. The apparent activation energy for Pd/SiO2
shown in Fig. 8 is 12.1 0.8 kcal/mol in good agreement with
11.0 kcal/mol by Bond [29] and 10.6 kcal/mol by Inoue et al. [31].
The activation energy for Ag–Pd/SiO2 is substantially lower at
8.9 0.6 kcal/mol, which is consistent with less strongly adsorbed
C2H2 for Ag–Pd/SiO2 at the generally accepted rate determining
step of the surface reaction between adsorbed C2H2 and H [8].
Again both activation energy values agree quite well with the trend
seen by Neurock [8], specifically, 7.7–10.5 kcal/mol for Pd(1 1 1) and
6.2–7.4 kcal/mol for Ag–Pd(1 1 1). The combination of lower acti-
vation energy and less negative reaction order in C2H2 pressure
explains why the TOF values for optimum compositions of Ag–Pd
are much higher than for Pd catalysts. The reaction rates for C2H2
conversion over Pd/SiO2 and Ag–Pd/SiO2 can be stated as follows,
where R has the standard value of 1.986 cal/◦/K-mol:
Acknowledgments
The authors would like to thank BASF Catalysts LLC for supplying
the Pd/SiO2 catalyst and UOP LLC for funding some of this work. In
addition, we gratefully acknowledge NSF Grant No. CBET 0854339
for support of Yunya Zhang.
For Pd/SiO2,
References
rate = k1e− 12,10D 800 [p(C2H2)]−0.67 0.06[p(H2)]1.17
0.11
RT
[1] W.R. True, Oil Gas J. 111 (2013) 90–94.
[2] H. Zea, K. Lester, A.K. Datye, E. Rightor, R. Gulotty, W. Waterman, Appl. Catal.,
A 282 (2005) 237–245.
[3] A.N.R. Bos, K.R. Westerterp, Chem. Eng. Process. 32 (1993) 1–7.
[4] C. Kemball, D.A. Dowden, G. Webb, Catalysis 2 (1979) 145–175.
[5] G.C. Bond, Catalysis by Metals, Academic Press, London, 1962.
[6] J. Wood, M.J. Alldrick, J.M. Winterbottom, E.H. Stitt, S. Bailey, Catal. Today 128
(2007) 52–62.
[7] J. Margitfalvi, L. Guczi, A.H. Weiss, React. Kinet. Catal. Lett. 15 (1980) 475–479.
[8] D. Mei, M. Neurock, C.M. Smith, J. Catal. 268 (2009) 181–195.
[9] D.C. Huang, K.H. Chang, W.F. Pong, P.K. Tseng, K.J. Hung, W.F. Huang, Catal. Lett.
53 (1998) 155–159.
[10] P. Miegge, J.L. Rousset, B. Tardy, J. Massardier, J.C. Bertolini, J. Catal. 149 (1994)
404–413.
[11] S. Leviness, V. Nair, A.H. Weiss, J. Mol. Catal. 25 (1984) 131–140.
[12] V.H. Sandoval, C.E. Gigola, Appl. Catal., A 148 (1996) 81–96.
[13] M.T. Schaal, A.Y. Metcalf, J.H. Montoya, J.P. Wilkinson, C.C. Stork, C.T. Williams,
J.R. Monnier, Catal. Today 123 (2007) 142–150.
[14] M.T. Schaal, A.C. Pickerell, C.T. Williams, J.R. Monnier, J. Catal. 254 (2008)
131–143.
[15] K.D. Beard, D. Borrelli, A.M. Cramer, D. Blom, J.W. Van Zee, J.R. Monnier, ACS
Nano 3 (2009) 2841–2853.
[16] K.D. Beard, J.W. Van Zee, J.R. Monnier, Appl. Catal., B 88 (2009) 185–193.
[17] M.T. Schaal, M.P. Hyman, M. Rangan, S. Ma, C.T. Williams, J.R. Monnier, J.W.
Medlin, Surf. Sci. 603 (2009) 690–696.
[18] M.T. Schaal, J. Rebelli, H.M. McKerrow, C.T. Williams, J.R. Monnier, Appl. Catal.,
A 382 (2010) 49–57.
[19] J. Rebelli, M. Detwiler, S. Ma, C.T. Williams, J.R. Monnier, J. Catal. 270 (2010)
224–233.
[20] J. Rebelli, A.A. Rodriguez, S. Ma, C.T. Williams, J.R. Monnier, Catal. Today 160
(2011) 170–178.
Fig. 8. Arrhenius plots for the temperature dependencies of C2H2 conversion for (A)
Pd/SiO2 and (B) Ag–Pd/SiO2 (ꢁAg = 0.92) at C2H2 pressure of 0.01 atm and H2 pressure
of 0.05 atm.

426
Y. Zhang et al. / Applied Catalysis A: General 469 (2014) 419–426
[21] L.L. Kesmodel, L.H. Dubois, G.A. Somorjai, J. Chem. Phys. 70 (1979) 2180–2188.
[22] J.W. Medlin, M.D. Allendorf, J. Phys. Chem. B 107 (2003) 217–223.
[23] F. Studt, F. Abild-Pedersen, T. Bligaard, R.Z. Sørensen, C.H. Christensen, J.K.
Nørskov, Angew. Chem. Int. Ed. 47 (2008) 9299–9302.
[26] M. Li, J. Shen, Mater. Chem. Phys. 68 (2001) 204–209.
[27] R. Van Hardeveld, F. Hartog, Surf. Sci. 15 (1969) 189–230.
[28] P.S. Cremer, X. Su, Y.R. Shen, G.A. Somorjai, J. Am. Chem. Soc. 118 (1996)
2942–2949.
[24] F. Studt, F. Abild-Pedersen, T. Bligaard, R.Z. Sørensen, C.H. Christensen, J.K.
Nørskov, Science 320 (2008) 1320–1322.
[25] S. Shaikhutdinov, M. Heemeier, M. Bäumer, T. Lear, D. Lennon, R.J. Oldman, S.D.
Jackson, H.-J. Freund, J. Catal. 200 (2001) 330–339.
[29] G.C. Bond, P.B. Wells, J. Catal. 5 (1965) 65–73.
[30] H. Molero, B.F. Bartlett, W.T. Tysoe, J. Catal. 181 (1999) 49–56.
[31] Y. Inoue, I. Yasumori, J. Phys. Chem. 75 (1971) 880–887.