The hydrogenation of acetylene catalyzed by palladium: Hydrogen pressure dependence

Article abstract of DOI:10.1006/jcat.1998.2294

The kinetics of acetylene hydrogenation catalyzed by a clean palladium foil at high pressures are measured and yield an activation energy of 9.6 ± 0.1 kcal/mol when using hydrogen. The rate exhibits a deuterium isotope effect such that the reaction activation energy is 9.0 ± 0.2 kcal/mol for reaction with deuterium. The hydrogen pressure reaction order is 1.04 ± 0.02 at 300 K with an acetylene pressure of 100 Torr and the acetylene order is -0.66 at 300 K and with a hydrogen pressure of 100 Torr. These reaction kinetics closely mimic those of supported model catalysts. In addition, it is found that the rate of benzene formation is accelerated by the addition of hydrogen to the reaction mixture. This is rationalized by proposing that hydrogen enhances the coverage of acetylene under catalytic conditions. This notion can be used to successfully calculate the hydrogen pressure dependence for acetylene hydrogenation as a function of temperature, a value which varies between ～1.05 and 1.3 as the temperature changes from 300 to 380 K. Possible origins for this effect are discussed.

Full text of DOI:10.1006/jcat.1998.2294

Journal of Catalysis 181, 49–56 (1999)
Article ID jcat.1998.2294, available online at http://www.idealibrary.com on
The Hydrogenation of Acetylene Catalyzed by Palladium:
Hydrogen Pressure Dependence
1
H. Molero, B. F. Bartlett, and W. T. Tysoe
Department of Chemistry and Laboratory for Surface Studies, University of Wisconsin-Milwaukee, Milwaukee, Wisconsin 53211
Received June 19, 1998; revised October 5, 1998; accepted October 6, 1998
strated to form by reaction between acetylene and an inter-
The kinetics of acetylene hydrogenation catalyzed by a clean pal- mediate metallocyclic C4H4 species (10, 11). In addition, ad-
ladium foil at high pressures are measured and yield an activation sorbed acetylene has been shown to convert into vinylidene
energy of 9.6 � 0.1 kcal/mol when using hydrogen. The rate exhibits
a deuterium isotope effect such that the reaction activation energy
is 9.0 � 0.2 kcal/mol for reaction with deuterium. The hydrogen
pressure reaction order is 1.04 � 0.02 at 300 K with an acetylene
pressure of 100 Torr and the acetylene order is � 0.66 at 300 K and
with a hydrogen pressure of 100 Torr. These reaction kinetics closely
mimic those of supported model catalysts. In addition, it is found
that the rate of benzene formation is accelerated by the addition of
hydrogen to the reaction mixture. This is rationalized by proposing
on warming to � 300 K (12). It has been demonstrated that
vinylidene is very unreactive so that adsorbed acetylene is
the precursor to benzene formation in ultrahigh vacuum
and also reacts with hydrogen to yield ethylene. Thus, un-
der ultrahigh vacuum conditions, vinylidene neither reacts
with adsorbed hydrogen, nor does it react to form ben-
zene. It has, however, recently been demonstrated that un-
der high-pressure, catalytic conditions, acetylene and vinyli-
that hydrogen enhances the coverage of acetylene under catalytic dene can combine to yield a C4H4 intermediate that finally
conditions. This notion can be used to successfully calculate the leads to benzene formation (13). The saturation coverage
hydrogen pressure dependence for acetylene hydrogenation as a of vinylidene is unity (where coverages are referenced to
function of temperature, a value which varies between � 1.05 and
the number density of palladium substrate atoms) (9) but
1
.3 as the temperature changes from 300 to 380 K. Possible origins
it has been demonstrated in several cases that reactants
can adsorb onto the metal surface even in the presence of
such a carbonaceous layer. Thus ethylene can adsorb on
an ethylidyne-covered surface (14, 15) and CO can adsorb
onto a Mo(100) single crystal surface covered by a thick
for this effect are discussed.
�c 1999 Academic Press
Key Words: acetylene; hydrogenation; benzene formation; kinet-
ics; hydrogen pressure dependence; palladium; model catalyst; iso-
tope effect.
(
� 8 monolayers) carbonaceous layer (16) as well as on a
vinylidene-covered surface (17).
INTRODUCTION
The kinetics of acetylene hydrogenation to ethylene,
Palladium has a unique ability to selectively hydrogenate catalyzed by a model palladium metal catalyst, carried out
alkynes since it is capable of hydrogenating small concen- at high pressures, are described below. This reaction cata-
trations of alkynes to alkenes in the presence of a large ex- lyzed by supported palladium has been investigated pre-
cess of alkene and so, for example, it is used for preparing viously where the activation energy has been shown to be
extremely pure ethylene feedstocks for subsequent poly- � 11.0 kcal/mol (18–20). The reaction kinetics are rather
merization reactions (1–4). The origin of this high selectiv- complicated and contradictory where, according to one re-
ity is not well understood (5–7) but presumably arises either port, the hydrogen pressure dependence is 1.0 at 273 and
from differences between the relative heats of adsorption 303 K but changes to 1.4 at 400 K (18) but, according to an-
of ethylene and acetylene on palladium (4) or their relative other, displays no temperature dependence (19). The acety-
subsequent reactivities with hydrogen. Studies in ultrahigh lene pressure dependence is negative and varies between
vacuum reveal that acetylene adsorbs much more strongly �pr 0e . s5 s 5u ar en d( 2�0 )0 ..67, depending on temperature and hydrogen
on palladium than does ethylene (8, 9). In addition, both
ethylene and acetylene can undergo an array of other sur-
We have therefore measured the acetylene hydrogena-
face reactions. The most striking of these is the formation of tion kinetics for a reaction catalyzed by palladium using a
benzene from acetylene on Pd(111) which has been demon- high-pressure reactor incorporated into an ultrahigh vac-
uum chamber, for two main reasons. The first is to deter-
1
mine whether a palladium foil model catalyst mimics the
activity of the supported catalyst to establish whether it
Author to whom correspondence should be addressed. E-mail: wtt@
csd.uwm.edu.
49
0021-9517/99 $30.00
Copyright �c 1999 by Academic Press
All rights of reproduction in any form reserved.

5
0
MOLERO, BARTLETT, AND TYSOE
provides a suitable model system for the working catalyst. a more sensitive gauge of carbon coverage was to saturate
Despite the extensive surface science investigations that the surface with oxygen and to perform a temperature-
have been undertaken on palladium, this has not yet been programmed desorption experiment. The presence of sur-
established. Second, it is also found that the rate of ben- face carbon is manifest by the desorption of CO. As the sur-
zene formation is accelerated by the addition of hydrogen face becomes depleted of carbon, the CO yield decreases
to the reaction mixture, an effect that has been observed and the yield of oxygen increases correspondingly in inten-
previously using alumina-supported palladium (21). It is sity. The complete absence of carbon is indicated by the
therefore postulated in the following that hydrogen causes desorption of only O2.
an increase in the coverage of acetylene on the surface. This
The acetylene used for these experiments (Linde, Pu-
postulate leads to a successful prediction of the variation in rified grade) was transferred from the cylinder to a glass
hydrogen pressure dependence of the hydrogenation reac- bottle and further purified by repeated bulb-to-bulb distil-
tion with temperature which can be related to the enhance- lations and its cleanliness monitored either mass spectro-
ment in benzene formation rate. Possible origins for this scopically or using the gas-chromatograph. The hydrogen,
effect are also discussed below.
deuterium, and nitrogen (Linde, Ultrahigh Purity grade)
were used without further purification. In all experiments,
the high-pressure reactor was ballasted to 1 atm using ni-
trogen to ensure that the gas circulation rates were identical
for all experiments.
EXPERIMENTAL
The apparatus that was used for these experiments has
been described in detail elsewhere (22, 23). Briefly, how-
ever, it consists of a bakeable, stainless-steel, ultrahigh
vacuum chamber operating at a base pressure of 1 �
RESULTS
�
10
1
0
Torr following bakeout. It is equipped with a qua-
A palladium foil sample catalyzes the hydrogenation of
drupole mass analyzer for residual gas analysis, leak test- acetylene and reaction rates are calculated for low conver-
ing, and temperature-programmed desorption studies.
sions (below � 1% ). As noted above, one of the rationales
The apparatus also incorporates a co-axial, high-pressure for studying acetylene hydrogenation catalyzed by palla-
reactor which can be sealed and filled to 1 atm while main- dium is its unique properties as a highly selective catalysts.
�
10
taining ultrahigh vacuum (� 2 � 10
Torr) in the rest of Other products found in the reaction are C4 compounds
the apparatus. This is attached to a loop and the gas recir- (primarily butene) and benzene and some ethane. Benzene
culated by a pump. The pressure is monitored by means is also formed following acetylene adsorption on Pd(111)
of a capacitance manometer and the gas composition an- in ultrahigh vacuum (8, 9) and this reaction is effectively
alyzed by diverting aliquots of the gas mixture to a gas catalyzed by the single crystal (24). This forms via the rapid
chromatograph equipped with a 1-m-long column packed initial formation of a C4 metallocycle which then reacts with
with Poropak-N and products were detected using a flame- adsorbed acetylene to yield benzene (10, 11). The detection,
ionization detector, where the resulting current is amplified therefore, of butene and benzene is entirely consistent with
by a Keithly picoammeter and collected using a computer this surface chemistry. In addition, a small amount of ethane
for subsequent analysis. The reaction rate is measured di- is detected due to further hydrogenation to ethylene. The
rectly from the product accumulation curve for low (<1% ) selectivity to ethylene S is shown plotted as a function of hy-
2
conversions. The low area of the model catalyst (� 1 cm ) drogen pressure when using 100 Torr of hydrogen at 300 K
means that, even at these low conversions, each site has in Fig. 1. In this case, the selectivity is defined as
catalyzed many reactions.
r(C2H4)
The palladium foil sample is attached to the end of
S = P
,
the sample manipulator and can be resistively heated to
2000 K or cooled to 80 K by thermal contact to a liquid-
r(products)
products
�
nitrogen-filled reservoir. A palladium foil is used as a model where r(C2H4) is the rate of ethylene formation and
hydrogenation catalyst rather than an oriented single crys- r(product) is the rate of formation of reaction products in-
tal since palladium adsorbs large amounts of hydrogen cluding ethylene so that 0 < S < 1. As expected, the selec-
which can potentially destroy the crystallinity of the sample. tivity increases with increasing hydrogen pressure to attain
The sample is cleaned using a standard procedure which a maximum for hydrogen pressure of 200 to 300 Torr and
�
8
consists of heating at 1000 K in � 4 � 10 Torr of oxy- then remains constant or decreases slightly to a relatively
gen and the annealing at 1200 K in vacuo to remove any low value of � 32% . The other products are, as noted above,
remaining oxygen. Since the carbon KLL Auger feature benzene and C4 hydrocarbons. The corresponding data for
is effectively obscured by a strong palladium peak, Auger the selectivity as a function of catalyst temperature when
spectroscopy is not particularly sensitive to the presence of using 100 Torr of both ethylene and hydrogen are displayed
small amounts of carbon on the surface. It was found that in Fig. 2. Here the selectivity increases drastically with

HYDROGENATION OF ACETYLENE
51
FIG. 1. The selectivity to ethylene as a function of hydrogen pressure
when using 100 Torr of ethylene at 300 K for acetylene hydrogenation
catalyzed by palladium.
FIG. 3. Arrhenius plot of ln(rate) versus 1/T for the temperature
dependence of the palladium-catalyzed hydrogenation of acetylene
(
100 Torr) when using hydrogen as a reactant (100 Torr) (᭹). The cor-
responding Arrhenius plot when using deuterium is also shown (᭿).
increasing temperature from � 30% at 300 K to � 94% at
4
70 K.
and the value when using deuterium is 9.0 � 0.1 kcal/mol.
These values of activation energy agree well with those ob-
tained by previous workers (11.0 � 0.5 (18), 10.6 � 0.5 kcal/
mol (19), and 11.0 � 0.5 (20)).
The temperature dependence for acetylene hydrogena-
tion measured both using hydrogen and deuterium is shown
in Fig. 3. This reveals a normal isotope effect, where the ra-
tio of the reaction rates with hydrogen and deuterium is
The hydrogen pressure dependence is shown in Fig. 4 at a
reaction temperature of 300 K using an acetylene pressure
1
.9 � 0.3 at 300 K. This result indicates that hydrogen is in-
volved in the rate-limiting step at the surface. The reaction
activation energy when using hydrogen is 9.6 � 0.2 kcal/mol
FIG. 4. The hydrogen pressure dependence of acetylene hydrogena-
tion catalyzed by palladium plotted as ln(rate) versus ln(P(H2)), where
P(H2) signifies hydrogen pressure for reaction at 300 K using 100 Torr of
acetylene.
FIG. 2. The selectivity to ethylene as a function of temperature when
using 100 Torr of acetylene and 100 Torr of hydrogen for acetylene hydro-
genation catalyzed by palladium.

5
2
MOLERO, BARTLETT, AND TYSOE
of 100 Torr over a wide hydrogen pressure range from 20 to
00 Torr and shows a very linear plot with a reaction order
6
of 1.04 � 0.02. This value is also in good agreement with
that measured in other work (17–19) and the error limits
imply that the reaction kinetics are slightly faster than first
order and the difference from unity is greater than the
experimental error. This will be discussed in greater detail
below.
The corresponding acetylene pressure dependence mea-
sured using a hydrogen pressure of 100 Torr at 300 K, again
over a wide acetylene pressure range from 5 to 300 Torr,
also yields a good straight line corresponding to a reaction
order of � 0.66 � 0.02. This is again within the range of val-
ues measured by others where between � 0.55 and � 0.67
have been measured by Bond (18), depending on temper-
ature and pressure, between � 0.4 and � 0.6 by Yasunobu
et al. (19) and � 0.5 by Grignon-Dumoulin et al. (20).
As noted above, benzene is also found in the reaction
products, consistent with previous observations that palla-
dium is an effective acetylene cyclotrimerization catalyst.
It has been shown previously using high-surface-area cata-
lysts that the benzene formation rate on a freshly prepared
catalysts is very large and diminishes to a low, steady-state
rate with time on-stream (21). Interestingly, at least part of
this initial high activity could be recovered by adding hy-
drogen (21). This effect is illustrated on this model system
in Fig. 5 which displays the rate of benzene formation at
FIG. 6. An Arrhenius plot of ln(�/�) versus 1/T, where (�/�) is
a) measured directly from the variation in benzene formation rate as
a function of hydrogen pressure at various temperatures (᭺) and (b) from
the hydrogenation reaction order in hydrogen (see text) (᭹).
(
well represented by
R(C H ) ∝ � + � P(H )
6
6
2
[1]
so that the rate increases linearly with hydrogen pressure,
where � represents the rate without any hydrogen present
24) and � represents the effect of hydrogen and are illus-
trated in Fig. 5. This equation can be simply rewritten as
3
00 K using 100 Torr of acetylene as a function of hydro-
gen pressure. The rate with no added hydrogen present is
in good agreement with that found elsewhere (24) and the
rate increases linearly with hydrogen pressure and so it is
(
R(C6H6) ∝ (1 + (�/�)P(H2))
[2]
�
2
and the data of Fig. 5 show that (�/�) = 1.2 � 0.1 � 10 at
3
00 K using 100 Torr of acetylene. The value of (�/�) was
also measured as a function of temperature from 300 to
65 K using 100 Torr of acetylene by repeating the exper-
3
iment that resulted in the plot shown in Fig. 5 at several
temperatures. The resulting Arrhenius plot (ln(�/�) versus
1
/T) is displayed in Fig. 6 (᭺). The origin of the other data
plotted on this curve (᭹) will be discussed below.
The corresponding temperature dependence of the reac-
tion order was also measured using an acetylene pressure
of 100 Torr over the temperature range 300 to 365 K and
the results are displayed in Fig. 7, where the reaction or-
der varies with temperature from � 1.05 at 300 K (Fig. 4) to
�
1.25 at 365 K.
In general, an additional potential advantage to studying
catalytic reactions using an isolatable high-pressure reactor
is that the nature of the surface can be analyzed following
reaction. Unfortunately, this was not possible in this case
using the analytical tools available in this vacuum cham-
ber since the carbon KLL Auger feature is effectively ob-
scured by a palladium peak which renders this essentially
FIG. 5. The effect of hydrogen on the rate of benzene formation for a
reaction catalyzed by a palladium foil. The reaction temperature is 300 K
and the acetylene pressure is 100 Torr.

HYDROGENATION OF ACETYLENE
53
ladium foil provides a good model for a supported acety-
lene hydrogenation catalyst. The data of Fig. 1 show that
the selectivity S increases with hydrogen pressure for low
pressure and then reaches a plateau and the data of Fig. 2
show that the selectivity also increases with temperature so
that, at the very highest temperatures, palladium provides
a reasonably selective catalyst.
The other products that are found from the reaction are
C4 hydrocarbons and benzene as well as some ethane. The
amount of ethane formed in the reaction is in fact rather
small, particularly for low conversions suggesting that ethy-
lene rapidly desorbs from the surface immediately after it
is formed and has little chance to hydrogenate further. Cer-
tainly ultrahigh vacuum experiments show that ethylene
adsorbs only weakly on palladium (111) and thermally des-
orbs below 300 K (9). The low conversions used for the ki-
netic measurements means that the chance of subsequent
product reaction is very low. The formation of C4 and C6
species is also in accord with the surface chemistry found in
ultrahigh vacuum since palladium catalyzesthe cyclotrimer-
ization of acetylene to benzene via a C4H4 metallocyclic
FIG. 7. The hydrogen pressure dependence plotted asa function ofre-
action temperature for palladium-catalyzed acetylene hydrogenation us-
ing 100 Torr of acetylene.
useless for monitoring the amount of carbon on the surface. intermediate (10, 11). The gas-phase C4 product presum-
In addition, since palladium can absorb a large amount of ably arises from the hydrogenation of this intermediate
hydrogen into the bulk of the sample, hydrogen tempera- and C4 hydrocarbons are also detected in temperature-
ture-programmed desorption is also not a very sensitive programmed desorption when acetylene is adsorbed on a
measure of the nature of the species present on the surface. hydrogen precovered surface (12). The increase in selectiv-
ity as the hydrogen pressure increases can be explained in
DISCUSSION
terms of this model since increasing the hydrogen pressure
increases the rate of acetylene hydrogenation relative to
Palladium provides an effective hydrogenation catalyst. the rate of dimerization so that less C hydrocarbons are
4
The activation energy for the reaction is 9.6 � 0.2 kcal/mol formed.
when using hydrogen and 9.1 � 0.1 kcal/mol when us-
Note, however, that the benzene formation rate increases
ing deuterium (Fig. 3) and the rates are proportional to with increasing hydrogen pressure (Fig. 5). In fact, the op-
p(H2)1
.04� 0.02
(Fig. 4) and p(C2H2)
� 0.66� 0.02
. Note that all posite would be expected based on the above assertion that
of the rates presented here are calculated as turnover fre- the selectivity increases due to loss of acetylene from the
quencies by assuming that the palladium site density on the surface. The surface chemistryofacetylene hasbeen studied
polycrystalline foil is equal to that on the (111) face of palla- extensively on Pd(111). Two forms of adsorbed acetylene
dium. Since this is the lowest energy face, this is likely to be have been detected, one a flat-lying acetylene which is the
the predominant facet on an annealed foil. In addition, the precursor to benzene formation (11) which can also react
integrated area of the infrared spectrum of CO adsorbed at with adsorbed atomic hydrogen to form ethylene (25). The
8
0 K on the foil and on a Pd(111) single crystal agree within other is a vinylidene species (CH == C == ) that forms when
2
1
5% confirming the validity of this assumption. This allows an acetylene-covered surface is heated to � 300 K (12). This
a reaction rate to be given as
species binds more strongly to the palladium surface and
is unreactive in ultrahigh vacuum. The activation energy
measured for acetylene hydrogenation by temperature-
programmed desorption is 5.4 � 0.2 kcal/mol (25), com-
pared with 9.6 � 0.2 kcal/mol for the catalytic reaction mea-
6
1.04� 0.02
� 0.66� 0.02
p(C2H2)
rate = 5 � 1 � 10 p(H2)
�
exp(� 9600 � 200/RT) reactions/site/s.
These kinetics are typical for transition-metal catalyzed hy- sured here. It has also been demonstrated that acetylene
drogenation reactions which have reaction activation ener- hydrogenation is first order in hydrogen coverage in ul-
gies � 10 kcal/mol, a zero or negative order in hydrocarbon trahigh vacuum (25). This would be expected to lead to
pressure and some positive order between � 1.0 and � 1.5 an approximately zero to half-order hydrogen pressure de-
in hydrogen pressure. These values are also in very good pendence assuming dissociative adsorption, whereas an or-
agreement with kinetic data collected for other supported der close to unity is found experimentally (see above).
palladium catalysts(18–20) and therefore suggest that a pal- Note that recent molecular beam studies of ethylene

5
4
MOLERO, BARTLETT, AND TYSOE
hydrogenation catalyzed by platinum also show a reaction fragment bonded to the surface (36). This simple analysis
1.2
0.8
rate proportional to 2(C2H4) 2(H) on Pt(111) in ac- also predicts that the ratio of the rates for hydrogenation
cord with this view (26). Clearly, then, there are large dis- and deuteration at 300 K should be 2.7 � 0.1 in reason-
paritiesbetween the reaction kineticsmeasured in ultrahigh able agreement with the value measured experimentally
vacuum and at high pressures. One obvious difference be- (1.9 � 0.3, Fig. 3).
tween the nature of the sample in ultrahigh vacuum and
In addition, the reaction rate measured in ultrahigh
under high external pressures is the presence of a strongly vacuum which is proportional to 2(hydrocarbon) �
bound, relatively unreactive hydrocarbon species present 2(hydrogen) (11, 25) is consistent with this pathway which
on the surface in the latter case. Such hydrocarbons are implies therefore that the catalytic reaction rate depends
found on the surface of model catalysts after reaction both on hydrogen pressure as
on platinum (27) and molybdenum (28) and recent sum-
�
�
�
�
p
�
�
frequency generation (SFG) (14, 15) and reflection– ab-
sorption infrared spectroscopy (RAIRS) (29) experiments
have confirmed the presence of such strongly bound hy-
Rate ∝ 1 +
P(H2)
�
P(H ),
[3]
2
drocarbons during reaction at high pressures. It is likely where the first term represents the acetylene coverage de-
that the vinylidene species referred to above (12) are also termined empirically from the benzene formation kinetics
present on the catalyst surface during acetylene hydrogena- above (Eq. [2]), and the second term represents the hy-
tion. It has been suggested, based on SFG data, that ethy- drogen coverage assuming that this adsorbs dissociatively
lene can adsorb on the surface in spite of the presence of on the vinylidene-covered surface. Note that it has been
an ethylidyne layer (14, 15). Line broadening found in the suggested that adsorbed hydrocarbon species contribute to
infrared spectrum of an ethylidyne on Pd(111) is in accord the source of hydrogen in addition to adsorbed hydrogen
with this view (29) and it has recently been shown that CO (37–41). First, there is no thermodynamic driving force for
can adsorb onto molybdenum even when it is covered by this reaction and second catalysts covered by thicker car-
a thick (several monolayers) carbonaceous layer (16) and bonaceous layers such as molybdenum (42) would be ex-
on a vinylidene-covered Pd(111) surface (17). Acetylene pected to be very active, whereas they are not. Equation [3]
must also adsorb on vinylidene-covered palladium, since it predicts that the hydrogen pressure dependence is a func-
is capable of catalyzing acetylene cyclotrimerization at high tion of (�/�) and that reaction orders can vary between 0.5
pressures. Recent results have suggested that the reaction and 1.5. The value of (�/�) measured at 300 K with 100 Torr
proceeds at high pressures by reaction between adsorbed of acetylene can be taken directly from the data of Fig. 5
acetylene and vinylidene to form a C4 precursor to benzene and is shown marked on the theoretic plot of Fig. 8 derived
formation (13). There are precedents for this reaction in the from the function shown in Eq. [3]. This predicts a hydro-
organometallic literature (30–32).
It is also evident from the data shown in Fig. 5 that the agreement with the value measured experimentally under
addition of hydrogen to the reaction mixture causes an exactly identical experimental conditions.
gen pressure dependence of 1.04 � 0.04, in extremely good
increase in the rate at which benzene is formed. This effect
Since � in the data in Eq. (1) is taken to represent the
has ben noted previously (21). Based on arguments made amount of acetylene adsorbed onto a vinylidene-covered
above, it is proposed that this means that the acetylene surface, this will vary with temperature, the temperature
coverage increases with hydrogen pressure. Possible origins variation presumably depending on the heat of adsorption
of this effect will be discussed below. As noted above, of acetylene on this surface. The value of � is also likely to
hydrogen adsorbs dissociatively on the metal surface to be temperature dependent although the interpretation of
form atomic hydrogen which, according to the Horiuti– any temperature dependence that it does have will rely on
Polanyi model (33, 34), are proposed to add sequentially identifying the origin of the variation in benzene formation
to adsorbed acetylene to form ethylene. The observed rate as a function of hydrogen pressure (Fig. 5). This pro-
isotope effect (Fig. 3) is consistent with this pathway. posal is borne out experimentally by the data of Fig. 6 which
A simple semi-classical treatment suggests that rH /rD � reveal that the value of (�/�) measured directly from the
1
E/RT, where 1E is the difference in zero-point energy variation in benzene formation rate with hydrogen pres-
between the deuterated and hydrogenated transition sure indeed varies in an Arrhenius fashion with tempera-
state and where the normal mode leading to reaction is ture (᭺). The data of Figs. 5 and 6 taken together therefore
assumed to be localized on the hydrogen (deuterium) also imply that the hydrogen reaction order should depend
(
35). The activation energy difference (1E) can be mea- on temperature. Previous experiments on supported palla-
sured from Fig. 3 (0.6 � 0.2 kcal/mol) and directly yields dium model catalysts have suggested that it does and this
a vibrational frequency of the transition state of 1400 �
is confirmed by the data of Fig. 7 which reveal that the
00 cm . This value is consistent with reaction via the hydrogen pressure dependence varies between 1.04 and
transfer of an adsorbed hydrogen atom to a hydrocarbon 1.25 as the temperature changes from 300 to 365 K. The
�
1
4

HYDROGENATION OF ACETYLENE
55
A final possibility, in view of the fact that, as noted above,
these strongly bound species can be titrated from the sur-
face at high hydrogen pressures, is that vinylidene is reac-
tively removed from the surface to reveal active metal sites
below thereby leading to an increase in reaction rate where
several acetylenic species hydrogenate to ethylene before
the site is once again blocked by the formation of a vinyli-
dene species. Note that, on molybdenum where the car-
bonaceous layer is much thicker and so can easily be mon-
itored using Auger spectroscopy, a decrease in the amount
of carbon on the surface correlates with a corresponding
increase in the rate of olefin metathesis (43).
CONCLUSIONS
Palladium foils provide good model acetylene hydro-
genation catalysts since they exhibit identical pressure and
temperature dependences as do supported palladium cata-
lysts. It is found that the rate of benzene formation is en-
FIG. 8. Theoretical plot of (�/�) versus the hydrogen pressure de- hanced by the addition of hydrogen to the reaction mix-
pendence for acetylene hydrogenation. Plotted also on this graph is the
ture and this is taken to mean that hydrogen increases the
value of (�/�) taken from Fig. 6, yielding a predicted hydrogen pressure
coverage of acetylene on the surface. This simple model
dependence of 1.04 � 0.04, in good agreement with experiment (Fig. 4).
successfully predicts the variation in hydrogen pressure de-
pendence of acetylene hydrogenation as a function of tem-
corresponding values of (�/�) as a function of temperature
perature. Possible origins for this effect are discussed.
can also be calculated from the data displayed in Fig. 7 using
the theoretical relationship between (�/�) and hydrogen
reaction order plotted in Fig. 8 and shown in Eq. [3]. The
ACKNOWLEDGMENTS
resulting Arrhenius plot of ln(�/�) versus 1/T measured us-
ing hydrogen pressure dependences is also shown in Fig. 6
We gratefullyacknowledge support ofthiswork bythe U.S. Department
of Energy, Division of Chemical Sciences, Office of Basic Energy Sciences,
under Grant DE-FG02-92ER14289. We also acknowledge the donors of
the Petroleum Research Fund, administered by the American Chemical
Society, for partial support of this research.
(
(
᭹). This yields a good straight line so that the values of
�/�) measured directly from the benzene formation rate
and those obtained from the kinetic model described above
are in excellent agreement. The slope of the Arrhenius plot
yields an energy of 4.3 � 0.2 kcal/mol. The significance of
this value remains to be established and, as noted above,
will rely on successfully identifying the processes involved
in increasing the rate of benzene formation as a function of
hydrogen pressure.
REFERENCES
1
2
3
. Bond, G. C., Dowben, D. H., and Mackenzie, N., Trans. Faraday. Soc.
4, 1537 (1958).
. Bond, G. C., and Wells, P. B., 2nd International Congress on Catalysis.
Editions Technis, Paris, 1961.
5
. Bond, G. C., Catalysis by Metals. Academic Press, New York, 1962.
Note that the range of hydrogen pressure dependence
predicted by the model (between 0.5 and 1.5, Fig. 8) is con-
sistent with values measured for a range of acetylene and
ethylene hydrogenation reactions so that these ideas may
be generally applicable to other model systems. There are
several possible explanations for the observed apparent in-
crease in surface acetylene coverage with increasing hy-
drogen pressure. First, adsorbed hydrogen may modify the
surface in some way to increase the reactivity of the acety-
lene that is adsorbed on the surface so that increasing the
hydrogen pressure does not actually increase the acetylene
coverage but merely modifies its activity. Another possible
explanation for the observed increase in rate may be that
adsorbed vinylidene reacts with hydrogen to form some
other surface species, perhaps adsorbed vinyls or ethyli-
dyne, which less effectively block the surface.
4. Phillipson, J. J., Wells, P. B., and Winterbottom, J. M., Trans. Faraday.
Soc. 60, 1847 (1965).
5
6
7
8
. Twigg, C. H., Disc. Faraday. Soc. 57, 1781 (1961).
. Halsey, G. D., J. Chem. Phys. 67, 2038 (1963).
. Halsey, G. D., J. Chem. Phys. 17, 758 (1949).
. Tysoe, W. T., Nyberg, G. L., and Lambert, R. M., J. Chem. Soc. Chem.
Commun., 623 (1983).
9. Tysoe, W. T., Nyberg, G. L., and Lambert, R. M., J. Phys. Chem. 88,
1960 (1984).
1
1
1
1
1
0. Hoffman, H., Zaera, F., Ormerod, R. M., Lambert, R. M., Saldin, J. M.,
Wang, L. P., Bennett, D. W., and Tysoe, W. T., Surf. Sci. 268, 1 (1992).
1. Hoffman, H., Zaera, F., Ormerod, R., Lambert, R. M., Wang, L. P.,
and Tysoe, W. T., Surf. Sci. 232, 259 (1990).
2. Tysoe, W. T., Nyberg, G. L., and Lambert, R. M., Surf. Sci. 135, 128
(
1983).
3. Kaltchev, M., Molero, H., Wu, G., Blumenfeld, A., and Tysoe, W. T.,
in preparation.
4. Cremer, P. S., Su, X., Shen, Y. R., and Somorjai, G. A., J. Phys. Chem.
100, 16302 (1996).

5
6
MOLERO, BARTLETT, AND TYSOE
1
5. Cremer, P. S., Su, X., Shen, Y. R., and Somorjai, G. A., J. Am. Chem.
Soc. 118, 2942 (1996).
30. Beckhaus, R., Angew. Chem. Ed. Engl. 36, 1686 (1997).
31. Beckhaus, R., Oster, J., Ganter, B., and Englert, U., Organometallics
16, 3902 (1997).
32. Beckhaus, R., Sang, T., Wagner, T., and Ganter, B., Organometallics
15, 1176 (1996).
1
6. Kaltchev, M., and Tysoe, W. T., Catal. Lett. 53, 145 (1998).
7. Kaltchev, M., Molero, H., Wu, G., Stacchiola, D., and Tysoe, W. T., in
preparation.
1
1
1
2
8. Bond, G. C., and Wells, P. B., J. Catal. 5, 65 (1965).
9. Yasunobu, I., and Yasumori, I., J. Phys. Chem. 75, 880 (1971).
0. Grignon-Dumoulin, A., and Thonon, C., Rev. Int. Franc. Petrole. 14,
33. Horiuti, J., J. Catalyst 2, 1 (1948).
34. Horiuti, J., and Polanyi, M., Trans. Faraday Soc. 30, 1164 (1930).
35. Laidler, K. J., Chemical Kinetics, p. 434. Harper Collins, New York,
1987.
36. Ho, W., Willis, R. F., and Plummer, E. W., Phys. Rev. Lett. 40, 1463
(1978).
2
14 (1959).
1. Ormerod, R. M., and Lambert, R. M., J. Chem. Soc. Chem. Commun.
0, 1421 (1990).
2
2
2
2
2. Wang, L. P., and Tysoe, W. T., J. Catal. 28, 320 (1991).
3. Huezo, L. A., Kotvis, P. V., Crumer, C., Soto, C., and Tysoe, W. T.,
Appl. Surf. Sci. 78, 113 (1994).
37. Bos, A. N. R., and Westerterp, K. R., Chem. Eng. Proc. 32, 1
(1993).
38. Al-Ammar, A. S., and Webb, G., J. Chem. Soc. Faraday Trans. 1 74,
195 (1978).
39. Al-Ammar, A. S., and Webb, G., J. Chem. Soc. Faraday Trans. 1 75,
1900 (1979).
40. McGowan, W. T., Kemball, C., Whan, D. A., and Scurrell, M. S.,
J. Chem. Soc. Faraday Trans. 1 73, 632 (1977).
41. Margitfalvi, J., Guczi, L., and Weiss, A. H., J. Catal. 72, 185
(1981).
2
2
4. Rucker, T. J., Logan, M. A., Muetterties, E. M., and Somorjai, G. A.,
J. Phys. Chem. 90, 2703 (1986).
5. Ormerod, R. M., Lambert, R. M., Bennett, D. W., and Tysoe, W. T.,
Surf. Sci. 330, 1 (1995).
2
2
2
2
6. Ofner, H., and Zaera, F., J. Phys. Chem. B 101, 396 (1997).
7. Zaera, F., and Somorjai, G. A., J. Am. Chem. Soc. 106, 2288 (1984).
8. Bartlett, B. F., Molero, H., and Tysoe, W. T., J. Catal. 167, 470 (1997).
9. Kaltchev, M., Thompson, A. W., and Tysoe, W. T., Surf. Sci. 391, 145
42. Wang, L., and Tysoe, W. T., J. Catal. 128, 320 (1991).
43. Wu, G., Bartlett, B., and Tysoe, W. T., J. Catal. 173, 172 (1998).
(
1997).