Permanent alkene selectivity enhancement in copper-catalyzed propyne hydrogenation by temporary CO supply

Article abstract of DOI:10.1016/j.jcat.2010.11.024

Temporary supply of CO during propyne hydrogenation over a copper-based catalyst derived from a Cu-Al hydrotalcite permanently increased the alkene selectivity from 60% to 92% at 100% alkyne conversion (473 K). The superior performance of the CO-modified

Full text of DOI:10.1016/j.jcat.2010.11.024

Journal of Catalysis 278 (2011) 167–172
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Research Note
Permanent alkene selectivity enhancement in copper-catalyzed propyne
hydrogenation by temporary CO supply
a
b
b
a,
⇑
Blaise Bridier , Miguel A.G. Hevia , Nuria López , Javier Pérez-Ramírez
a
Institute for Chemical and Bioengineering, Department of Chemistry and Applied Biosciences, ETH Zurich, HCI E 125, Wolfgang-Pauli-Strasse 10, CH-8093 Zurich, Switzerland
Institute of Chemical Research of Catalonia, Av. Paisos Catalans 16, 43007 Tarragona, Spain
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Temporary supply of CO during propyne hydrogenation over a copper-based catalyst derived from a Cu–
Al hydrotalcite permanently increased the alkene selectivity from 60% to 92% at 100% alkyne conversion
Received 8 August 2010
Revised 28 November 2010
Accepted 30 November 2010
Available online 6 January 2011
(473 K). The superior performance of the CO-modified copper catalyst (higher alkene selectivity and
resistance against deactivation by fouling) was further confirmed at lower temperature (373 K). Our cat-
alytic results suggested that carbon monoxide in the presence of propyne and hydrogen irreversibly
restructures the catalyst surface leading to a smaller copper ensemble that minimizes C–C coupling. Con-
sequently, the enhanced alkene production in the CO-modified catalyst was coupled to a decreased oli-
gomer production and very little propane production. The pretreatment mixture, feed CO concentration,
and type of alkyne are important aspects to attain a selective copper catalyst. The CO effect in alkyne
hydrogenation over copper radically differs from that over palladium or nickel. On the latter two metals,
the increased alkene selectivity in the presence of CO is mostly due to a reduction in the over-hydroge-
nation channel. In addition, the effect is fully reversible, that is, it vanishes on removing CO from the feed.
The permanent modification of the copper ensemble by CO might be advantageous for (chemo) selective
hydrogenations of high acetylenic substrates.
Keywords:
Alkyne hydrogenation
Alkene selectivity
Carbon monoxide
Oligomerization
Copper ensemble
Geometric effect
Ó 2010 Elsevier Inc. All rights reserved.
1
. Introduction
The importance of modifiers has been recognized since catalysts
being poisoned by lead and quinoline in order to minimize side
paths reducing the alkene selectivity.
The partial hydrogenation of alkynes on palladium is widely ap-
plied in the petrochemical industry for the purification of olefin
streams (ethylene, propylene, and butylenes) from steam crackers
[8]. The standard catalyst contains 0.01–0.05 wt.% Pd on low-
surface-area alumina and is modified by a second metal (e.g. Ag
and Au) in order to increase the alkene selectivity [4,5]. Continuous
CO addition is a common way in practice to enhance the alkene
selectivity in Pd-catalyzed front-end and tail-end alkyne hydroge-
nation [4,5,8,9]. The inlet CO concentration varies in the range of
150–5000 ppm depending on the hydrogen content in the stream.
Both electronic and geometric effects account for the role of carbon
monoxide as selectivity modulator in alkyne hydrogenation on Pd
[10,11]. CO is a reversible modifier, that is, its positive effect on the
alkene selectivity vanishes once removed from the feed [4,5].
Recently, Trimm et al. [12] observed increased propene selectivity
were first applied in industrial processes [1,2], and their use in
commercial heterogeneous catalysts is common practice for a wide
range of reactions [3]. A modifier can be generally defined as a
component of the catalyst or an additional substance introduced
in the chemical process that improves the activity, selectivity, or
lifetime of the active substance without having any significant
activity on its own [4]. The effect of modifiers on catalysts may
be textural, structural, electronic, or a combination thereof, and
in most cases, these promoting additives are identified in a purely
empirical way [1]. A detailed understanding of the precise role(s)
of modifiers in the catalytic reaction is often a very complex task.
Modifiers are widely applied in partial hydrogenation of acety-
lenic compounds. Their presence is mandatory to increase the
selectivity and/or lifetime of palladium, which is the preferred ac-
tive metal for this type of reactions [5,6]. For example, the Lindlar
catalyst [7] is a classical for the three-phase selective hydrogena-
tion of acetylenic compounds in the fine chemical industry. A typ-
2
on Ni/SiO when CO was fed during propyne hydrogenation. Disad-
vantageously, CO addition led to increased oligomerization and de-
creased alkyne conversion due to the lowered concentration of
surface hydrogen atoms. Similarly to palladium, the CO effect on
nickel was found to be fully reversible.
3
ical Lindlar catalyst contains 5 wt.% Pd on CaCO and requires
Our recent studies in this field have accomplished an improved
understanding of alkyne hydrogenation over different metals
⇑
Corresponding author. Fax: +41 446 337 120.
E-mail address: jpr@chem.ethz.ch (J. Pérez-Ramírez).
0
021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.11.024

1
68
B. Bridier et al. / Journal of Catalysis 278 (2011) 167–172
(
namely Pd, Au, Ni, and Cu) in order to design more selective cat-
CO (0.5–5 vol.%) to mixture A (denoted as mixture B) for 5 h, and
(iii) switch from mixture B to mixture A or vice versa. In addition,
the reaction with mixture A was carried out over the catalyst pre-
alysts [10,11,13–18]. Copper-based catalysts have been investi-
gated due to their intrinsic high alkene selectivity. However,
their poor ability to split H
2
often leads to extensive oligomeriza-
treated at 473 K for 5 h in CO, H
2
+ CO, C
+ CO (7.5 vol.% H
, and 5 vol.% CO, He balance). The total flow in the various
3
H
4
+ CO, C
3
H
6
+ CO,
tion [16], unless being modified by nickel and iron [17]. To the best
of our knowledge, the influence of CO on copper-catalyzed alkyne
hydrogenation has not been assessed. In this study, a standard
Cu–Al catalyst was namely tested in propyne hydrogenation in
the absence and presence of CO. Temporary supply of carbon mon-
C
C
3
H
H
4
+ H
2
+ CO, and C
3
H
6
+ H
2
2
, 2.5 vol.% C
3
H
4
or
3
6
steps (reduction, pretreatment, and reaction) was kept at
3
À1
42 cm STP min
by balancing the feed mixture with helium.
Fresh catalyst was loaded for each inlet CO concentration and pre-
treatment mixture. Ethyne hydrogenation was studied in the same
3 4 3 6
oxide during C H hydrogenation permanently increased the C H
selectivity from 60% in the unmodified catalyst [16] to >90% in the
CO-modified counterpart (this work). CO irreversibly restructures
the catalyst so that the oligomerization path is minimized. The pre-
treatment conditions and the type of alkyne are crucial aspects for
attaining a selective copper catalyst. The function of CO as selectiv-
ity modulator on Cu is different compared to Pd and Ni and might
open interesting opportunities in (chemo) selective hydrogena-
tions of high acetylenic substrates.
reactor set-up at 523 K using a feed mixture of 2.5 vol.% C
7.5 vol.% H in He.
2 2
H and
2
The concentrations of alkyne, alkene, and alkane at the reactor
outlet were analyzed by a gas chromatograph (Agilent GC6890N)
equipped with a GS-GasPro column and a thermal conductivity
detector. The alkyne conversion was determined as the amount
of reacted alkyne divided by the amount of alkyne introduced in
the reactor. The selectivity to alkene (alkane) was determined as
the amount of alkene (alkane) formed divided by the amount of re-
acted alkyne, and the selectivity to oligomers was obtained as
S(oligomers) = 1 À S(alkene) À S(alkane).
2
. Experimental
Some of the catalysts used in propyne hydrogenation were
characterized ex situ by X-ray diffraction (XRD) and temperature-
programmed oxidation coupled to mass spectrometry (TPO-MS).
XRD was measured in a Bruker AXS D8 Advance diffractometer
equipped with a Cu tube, a Ge(1 1 1) incident beam monochroma-
tor (k = 0.1541 nm), and a Vantec-1 PSD. Data were recorded in the
range of 5–70° 2h with a step size of 0.02° and a counting time of
The preparation and characterization of the copper catalyst
used in this study was detailed elsewhere [16]. Briefly, the precur-
sor was a Cu–Al hydrotalcite with a nominal Cu/Al ratio = 3 pre-
pared by continuous coprecipitation. The clay-like precursor was
fully decomposed by calcination in static air at 873 K for 15 h using
À1
a heating rate of 5 K min , leading to crystalline tenorite (CuO)
and amorphous alumina. The gas-phase hydrogenation of propyne
was studied at ambient pressure and 373–473 K in a MicroActivity
Reference set-up (PID Eng&Tech). The calcined sample (150 mg,
4
s per step. TPO in air was measured in a Thermo TPDRO 1100 unit
coupled to a quadrupole mass spectrometer (Omnistar GSD 320
O1, Pfeiffer Vacuum). The sample (ca. 20 mg) was loaded in the
quartz micro-reactor (11 mm i.d), and the temperature was
sieve fraction 200–400
lm) was loaded in the 12 mm i.d. quartz
micro-reactor, heated in He up to 773 K and reduced in 5 vol.% H
2
À1
3
À1
ramped from 323 to 1173 K at 10 K min in 20 cm STP min of
in He at this temperature for 30 min to attain full copper reduction,
and finally cooled down to the reaction temperature in He. The re-
2 2
air. The atomic mass units m/z = 18 (H O) and m/z = 44 (CO ) were
monitored.
2
À1
duced catalyst presented a total surface area of 68 m g (BET
method, N adsorption at 77 K), an average Cu crystallite size of
0 nm (Scherrer method, X-ray diffraction), and a copper surface
2
2
3. Results
2
À1
2
area of 5.7 m g (pulse N O chemisorption) [16]. The weight loss
of the calcined sample during reduction was estimated at 16%, so
that the amount of reduced catalyst prior to the alkyne hydrogena-
tion tests was ca. 125 mg.
Fig. 1 shows the selectivity to propene, propane, and oligomers
of the copper catalyst in propyne hydrogenation at 473 K alternat-
ing mixtures A and B and other intermediate treatment steps. The
reader is guided through this figure by the encircled codes 1–11.
The propyne conversion in the whole set of experiments at this
temperature was 100%. In the first 5 h (step 1), the catalyst was
The standard catalyst testing protocol includes the following
steps: (i) propyne hydrogenation using 2.5 vol.%
3 4
C H and
7
.5 vol.% H in He (denoted as mixture A) for 5 h, (ii) addition of
2
Fig. 1. Selectivity to propene (circles), propane (diamonds), and oligomers (triangles) versus time in propyne hydrogenation at 473 K over the copper catalyst. The test
involved the alternation of two different feed mixtures (A and B) and other intermediate steps. Mixture A (C /H /He = 2.5/7.5/90) and mixture B (C /CO/H /He = 2.5/5/
.5/85).
3
H
4
2
3
H
4
2
7

B. Bridier et al. / Journal of Catalysis 278 (2011) 167–172
169
3 4 2
tested in mixture A (C H /H /He = 2.5/7.5/90), yielding the antici-
pated propene selectivity of 60% [16]. The selectivity to oligomers
was 40%, and no over-hydrogenated product was observed. Contin-
uous addition of carbon monoxide (C
H
3 4
2
/CO/H /He = 2.5/5/7.5/85,
mixture B) progressively increased the C
H
3 6
selectivity up to 80%,
coupled to a decreased oligomer production (step 2). Upon re-
moval of CO from the feed (step 3), that is, restoring mixture A,
the most striking result was observed. The selectivity to propene
was boosted to 92%, the selectivity to oligomers dropped to 5%,
and little but measurable propane production (3% selectivity)
was observed. Together with the ternary Cu–Ni–Fe catalysts re-
ported elsewhere [17], this is one the highest alkene selectivity val-
ues reported over
a copper catalyst in gas-phase alkyne
hydrogenation. After 5 h, CO was readmitted (step 4) and C
3
H
6
selectivity went quickly back to 80% as in step 2 and oligomers
comprised the only by-product. A new interruption of the CO feed
(
3 6
step 5) rapidly resumed the C H selectivity at >90% as in step 3.
The instantaneous decrease or increase in the propene selectivity
when shifting from mixture A to B or vice versa after step 2 can
be related to the fast CO adsorption/desorption equilibrium on
the catalyst.
Fig. 2. Selectivity to propene (circles) and oligomers (triangles) versus time in
propyne hydrogenation at 473 K over the copper catalyst using different concen-
trations of CO in the feed. Black symbols: C
symbols: C /CO/H /He = 2.5/1/7.5/89, and open symbols: C
0.5/7.5/89.5. After the first 5 h, CO was removed from the feed and the inlet mixture
contained C /H /He = 2.5/7.5/90 (mixture A).
H
3 4
2
/CO/H /He = 2.5/5/7.5/85, gray
3
H
4
2
3
H
4
/CO/H /He = 2.5/
2
In order to demonstrate the permanent effect by CO addition on
the alkene selectivity, the propyne hydrogenation in step 5 was
sustained for 60 h, resulting in a very stable performance. These
results strongly suggest that the presence of CO in the alkyne +
hydrogen mixture during step 2 irreversibly changed the structure
of the active sites, making them far more selective to the alkene.
The state of the copper catalyst upon CO modification was very ro-
bust. The sample used in steps 1–5 was discharged from the quartz
reactor and stored for 10 days in an open vial under ambient con-
ditions (step 6). Afterwards, the sample was retested in mixture A
H
3 4
2
Fig. 3 compares the performance of unmodified (freshly re-
duced) and CO-modified (5 h in mixture B at 473 K) copper cata-
lysts in propyne hydrogenation at 473 K and 373 K. At 473 K, full
propyne conversion was attained over both catalysts with the
(
step 7). In the first minutes, the selectivity to propene was 85%
and increased gradually to almost 90%, approaching the stable
selectivity value of the long run in step 5. The irreversible modifi-
cation by CO on the alkene selectivity in alkyne hydrogenation over
copper radically differs from its reversible CO effect on Pd and Ni
[
4,5,12]. On the latter two metals, the selectivity modulation disap-
pears when CO in the feed is cut off. This implies that, in practice,
continuous CO feeding is required to sustain the higher alkene
selectivity. Advantageously, the copper catalyst only requires a
relatively short pretreatment in propyne, hydrogen, and carbon
monoxide to attain an outstanding selectivity to the double
bond.
Fig. 2 shows that the inlet CO concentration plays an important
role in the alkene selectivity enhancement on copper. Carbon mon-
oxide was fed at levels of 0.5, 1, and 5 vol.% to C
and then, it was discontinued from the feed mixture for another
h. Addition of 0.5 vol.% CO did not influence the selectivity of
3 4 2
H + H during 5 h,
5
the catalyst with respect to the CO-free mixture (60% to propene
and 40% to oligomers, open symbols). When CO was interrupted,
the propene selectivity experienced a slight increase (67%). An in-
creased inlet CO concentration led to a gradual increase in propene
selectivity from 60% to 80% and the coupled decrease in oligomers
selectivity from 40% to 20%. The attainment of the 80% alkene
selectivity level was significantly faster in 5 vol.% CO (solid sym-
bols) than in 1 vol.% CO (gray symbols). This indicates that the
changes occurring in the catalyst are accelerated with the CO con-
centration. However, once the catalyst has been subjected to a suf-
ficient amount of CO (or for sufficient time), the resulting system is
equally selective once carbon monoxide is removed from the feed.
This can be seen in Fig. 2 upon switching from C
3 4 2
H + H + CO to
C H
3 4
+ H . The C selectivity was stable at ca. 90% after modifica-
2
3 6
H
tion by 1 or 5 vol.% CO during 5 h. The propane selectivity was not
plotted in Fig. 2 for the sake of clarity. As shown in Fig. 1, the selec-
tivity to propane over the modified catalyst (after CO removal)
amounted to ca. 3%.
Fig. 3. Conversion of propyne (a) and selectivity to propene (b) and oligomers (c)
versus time in propyne hydrogenation (mixture A) at 473 K and 373 K over
unmodified and CO-modified (5 h in mixture B) copper catalysts.

1
70
B. Bridier et al. / Journal of Catalysis 278 (2011) 167–172
indicated propene selective differences (60% and 92% for the
unmodified and CO-modified samples, respectively). The alkyne
conversion can be decreased by lowering the temperature and/or
by shortening the contact time. It was decided to compare the
unmodified and CO-modified catalysts at a lower temperature to
demonstrate the superiority of the latter system in terms of higher
selectivity and resistance toward deactivation by fouling as a con-
sequence of the presence of smaller copper ensembles. Previous
studies [16,19] have shown that copper catalysts strongly deacti-
vate on propyne hydrogenation below 423 K due to extensive
green oil formation. Accordingly, the propyne conversion on the
unmodified catalyst at 373 K dropped to 3% and the oligomeriza-
tion pathway is increased (selectivity ꢀ87%). The propene selectiv-
ity decreased from 60% to 13%. On the contrary, the CO-modified
catalyst showed stable behavior at a remarkably low temperature.
The propyne conversion and propene selectivity decreased from
copper catalyst at 473 K after 5 h pretreatment in several mixtures
(x-axis) followed by propyne hydrogenation (mixture A) for 5 h.
The propyne conversion in these tests was 100%. Pretreatments
2 3 6 3 6 2
in CO, H + CO, C H + CO, and C H + H + CO did not lead to any
measurable change, that is, the propene selectivity was the same
as on the non-pretreated catalyst (60%). Only mixtures containing
carbon monoxide and propyne (and additionally hydrogen) were
able to induce modification of the catalyst surface. The attainment
of a highly selective catalyst strongly depends on the nature of the
hydrocarbon, that is, the mixture of carbon monoxide, propene,
and hydrogen caused no effect on the selectivity. According to
Fig. 4, pretreatments in C
similar steady-state C
However, the achievement of this selectivity during reaction was
instantaneous when H was included in the pretreatment mixture
3
H
4
+ CO and C
3 4 2
H + H + CO led to very
3
H
6
selectivity after 5 h on stream (ꢀ90%).
2
and took longer (1–2 h) in its absence. We put forward that specific
partially hydrogenated moieties derived from propyne and carbon
monoxide are responsible for the selectivity enhancement on the
copper catalyst.
The X-ray diffraction patterns of the catalyst after reaction in
mixture A (without CO) for 5 h or to mixture B (with CO) for 20 h
were nearly identical (Fig. S1). No shift or broadening of the reflec-
tions of metallic copper was observed, indicating no change in the
bulk structure of the catalyst after alkyne hydrogenation in the ab-
sence or in the presence of CO.
1
00% to 60% and from 92% to 67%, respectively. The turnover fre-
quency, expressed as mol C produced per second and mol Cu
3 6
H
exposed in the fresh catalyst, can be estimated at ca. 2 orders of
magnitude higher in the CO-modified catalyst. The fact that the
propyne conversion remains high at 373 K can point out to the fact
2
that the rate-limiting step for hydrogenation on copper (H split-
ting) is more favorable on the CO-modified catalyst. Thus, the pro-
posed surface restructuring could account for this increased
conversion.
Figs. 1 and 2 concluded that the pretreatment of the copper cat-
alyst in mixtures containing C H , H , and CO raises its selective
3 4 2
We believe that the carbonaceous deposit generated during al-
kyne hydrogenation in the presence of CO might lead to a reduced
size of the copper ensemble. This would explain the attenuated
oligomerization since C–C coupling reactions require a larger cop-
per ensemble than the hydrogenation to propene or propane [16].
Such deposit should have the distinctive feature of incorporating
CO, since the deposit generated during propyne hydrogenation in
the absence of CO did not improve the selective character of the
catalyst. It is known that carbon monoxide can be incorporated
in the oligomers formed during alkyne hydrogenation on palla-
dium catalysts, leading to carboxylic acids or aldehydes [20].
Taking this into account, the CO-modified copper catalyst (after
step 7 in Fig. 1) was calcined to eliminate the carbonaceous
adlayer, which in principle should resume the original perfor-
mance, that is, setting the alkene selectivity back to 60%. The
calcination conditions in step 8 (773 K, 30 h) secured the full
combustion of the deposit. This was confirmed by temperature-
programmed oxidation of a CO-modified copper catalyst (Fig. S2).
The hydrogenation performance of the calcined sample was reeval-
uated in mixture A (step 9). After a 2-h period, in which the reox-
idized catalyst gets reduced in the reaction mixture, the propene
selectivity reached the steady-state value of ca. 80%. Our original
hypothesis was not validated since this value is significantly higher
than the 60% propene selectivity for the fresh catalyst that never
character toward the olefin in a substantial and permanent manner
once carbon monoxide is cut off. We have tentatively suggested
that this modification is caused by the CO-induced restructuring
of the active copper sites. The gradual increase in the propene
selectivity in step 2 is attributed to the progressive modification
of the surface in the presence of CO, which irreversibly turned
the copper catalyst more selective. In contrast, the propene selec-
tivity from step 3 to step 4 immediately reached a stable value
since the catalyst was already activated. This induction period is
again required when the activated catalyst was calcined (steps 8
and 9). A key question to further understand the catalyst modifica-
tion arises: does CO on its own alter the catalyst or does it require
additional components in the mixture? Fig.
4 shows the
steady-state selectivity of propene, propane, and oligomers of the
‘
saw’ CO. The removal of the carbonaceous deposit formed in the
presence of CO did not fully resume the original selectivity values.
Consequently, the deposit is not the only responsible for the
outstanding selectivity enhancement of copper by CO. We cannot
discard an irreversible restructuring of the catalyst surface induced
by the mixture of propyne, hydrogen, and carbon monoxide. Expo-
sure of the catalyst to CO for 5 h (step 10) maintained the C H
3 6
selectivity at 80% (like in steps 2 and 4), and subsequent discontin-
uation of CO in the feed (step 11) fully restored the situation in
steps 3, 5, and 7 with the CO-modified catalyst, that is, S(C
2%, S(oligomers) = 5%, and S(C ) = 3%.
The gas-phase hydrogenation of ethyne was studied over the Cu
3 6
H ) =
9
3 8
H
catalyst to check whether the remarkable selectivity enhancement
observed with propyne applies to other practically relevant alkynes.
Ethyne hydrogenation was studied at 523 K since operation at 473 K
Fig. 4. Influence of pretreatment on the selectivity to propene, propane, and
oligomers in propyne hydrogenation at 473 K over the copper catalyst. The plotted
values were acquired after pretreatment for 5 h followed by reaction in C H /H /
3 4 2
He = 2.5/7.5/90 (mixture A) for another 5 h. The composition of the pretreatment
mixtures is given in Section 2.
(
used for propyne) led to catalyst deactivation due to extensive
oligomerization. In C /H /He = 2.5/7.5/90, the selectivity to
2
H
2
2

B. Bridier et al. / Journal of Catalysis 278 (2011) 167–172
171
Fig. 5. Selectivity to ethene, ethane, and oligomers versus time in ethyne hydrogenation at 523 K over the copper catalyst. Feed mixture: C
2
H
2
/H
2
/He = 2.5/7.5/90. Treatment
2 2 2 3 4 2
C: C H /CO/H /He = 2.5/5/7.5/85, 523 K, 5 h. Treatment B: C H /CO/H /He = 2.5/5/7.5/85, 473 K, 5 h.
ethane was 40% and the remaining 60% comprised oligomers (Fig. 5).
The degree of ethyne conversion was 100%. The lower alkene selec-
tivity achieved with ethyne (40%) compared to propyne (60%) is a
general characteristic of partial hydrogenation catalysts [5,11,15].
Higher alkynes are less prone to oligomerize due to steric hindrance
for Pd and Ni. Accordingly, the electronic effect of carbon monoxide
on copper is regarded as minor. Rather, the alkene selectivity in-
crease upon CO addition should mainly derive from a geometric
(ensemble) effect.
NMR analysis of oligomers on palladium formed in the presence
of carbon monoxide indicated that CO was incorporated in the
hydrocarbon chain, and carboxylic acids or aldehydes were pro-
duced [20]. Borodzi n´ ski [21] pointed out the role of carbonaceous
deposits on the selective character of Pd. It was hypothesized that
the deposits diminish the abundance of large palladium ensembles
and increase alkene selectivity. Similar incorporation of CO in the
oligomers on Cu may generate a stable state of the surface render-
ing higher propene selectivity after the addition of CO and mainly
after its removal. Our catalytic data strongly suggest that carbon
monoxide in the presence of propyne and hydrogen leads to smal-
ler copper ensembles. The ensemble needed for polymerization is
larger than for hydrogenation, which substantiates the increased
propene selectivity on increasing the copper dispersion in the cat-
alysts [16,22]. One can anticipate that a small copper cluster would
be highly selective toward the alkene due to minimized oligomer-
ization. The difficulty is to prepare such clusters and, more impor-
tantly, to stabilize them under reaction conditions. We think that
the use of carbon monoxide enables to synthesize stable catalysts
with an extremely high degree of copper isolation.
in the C–C bond-formingstep. After 5 h in C
2 2 2
H + H , the catalyst was
exposed to C /CO/H /He = 2.5/5/7.5/85 for additional 5 h (treat-
2
H
2
2
ment C in Fig. 5), followed by the removal of CO from the feed. The
selectivity to ethene stayed at 40%, that is, CO does not influence
the selectivity of the catalyst in ethyne hydrogenation. However, if
the catalyst is treated in C
in Fig. 5) and switched back to C
markedly increased to 73%, the oligomer selectivity decreased to
3
4 2
H /CO/H /He = 2.5/5/7.5/85 (treatment B
H
2 2
+ H , the ethene selectivity
2
2
4%, and, similarly to propyne hydrogenation, little ethane (ꢀ3%
selectivity) was produced. In contrast to the case of propene, the
enhanced ethene selectivity was not stable; it decreases with
time-on-stream due to increased oligomerization. This result
further emphasizes the importance of the pretreatment mixture to
attain a selective state of the surface.
4
. Discussion
The increased alkene selectivity in alkyne hydrogenation by
continuous CO feeding over Pd catalysts is widely practiced in
industry for olefin purification in steam cracker cuts [8,9]. This
gas-phase modifier induces electronic and geometric effects
The genesis of small ensemble size from CO addition in the feed
is supported in Fig. 3 as the pathway to oligomers is largely ham-
pered at low temperature on the CO-modified catalyst compared to
its unmodified counterpart that deactivated totally. Nonetheless,
the higher propene selectivity after removal of the carbonaceous
deposit by calcination (step 9 in Fig. 1) highlights that the latter
are not the only responsible for the outstanding selectivity
enhancement of copper. The catalyst might have also experienced
surface restructuring. The production of propane in the active cat-
alyst (steps 3 and 5 in Fig. 1), which is rarely observed for copper-
based systems, also suggests an irreversible change in nature of the
active sites.
[
2
5,10,11,18]. The binding energies of H , alkyne, and alkene on
CO-covered palladium are slightly lowered than that on clean pal-
ladium, thus leading to a lower H coverage and a better thermody-
namic control in partial alkyne hydrogenation. The geometric
effect of CO derives from its preferential adsorption on Pd, reducing
the amount of active sites on the catalyst surface and blocking the
formation of unselective hydride and carbide phases. The effect of
CO in partial alkyne hydrogenation over nickel apparently shares
many features with palladium [12]. Both metals dissociate H
2
easily [18]. CO lowers the hydrogen coverage and blocks the
appearance of the hydride phase, which is strongly linked to the
over-hydrogenation path. The nature of the CO modification on
Fig. 4 reveals that CO individually was not the responsible for
the distinctive modification of the Cu catalyst. The different behav-
ior in the presence of a double or a triple bond unravels the deep
interaction between the nature of beneficial carbonaceous deposit
and the presence of highly unsaturated moieties. Our previous
work [16] highlighted that Cu, in opposition to Ni, produces highly
unsaturated oligomers that contain conjugated double and triple
bonds. The latter carbonaceous moieties were detected only at
423 K and correlated with the fast deactivation of copper at low
2
copper is different. H dissociation on Cu(1 1 1) is marginally endo-
thermic (0.16 eV) and hindered by a sizeable barrier (0.83 eV). Due
to the low H coverage on copper, extensive oligomerization and no
over-hydrogenation occurs (step 1 in Fig. 1). The increased propene
selectivity upon CO feeding on Cu (from 60% to 80%, step 2 in Fig. 1)
cannot be attributed to a lowered H coverage as one would reason

1
72
B. Bridier et al. / Journal of Catalysis 278 (2011) 167–172
temperature [16]. The presence of such highly unsaturated oligo-
mers at 473 K is unlikely but as it was mentioned above, in the
presence of adsorbed CO, H dissociation is hindered and unsatu-
2
rated oligomers causing the copper modification are produced.
Therefore, it is not surprising that this phenomenon appears only
upon Cu-catalyzed hydrogenation as the specific partially hydroge-
nated moieties needed to build up the modified surface result from
the ensemble size responsible for undesired C–C bond formation
reactions. This effect is mainly attributed to the formation of ben-
eficial carbonaceous deposits, although surface restructuring can
also contribute to the result. The irreversible effect of CO on cop-
per-catalyzed alkyne hydrogenation differs from the typically rec-
ognized reversible character of this selectivity moderator on
palladium and nickel catalysts. The remarkable alkene selectivity
increase in propyne hydrogenation over the CO-modified copper
catalysts was not fully attainable in ethyne hydrogenation due to
the marked oligomerization ability of the C2 alkyne. These novel
catalytic results are of significant interest in the hydrogenation
field, although assessing the industrial potential requires catalyst
scale-up and pilot studies. We put forward that the permanent
modification of the copper ensemble by CO might be advantageous
in (chemo) selective hydrogenations of high acetylenic substrates.
2
its poor ability in H splitting.
Besides the degree of insaturation of the hydrocarbon, Fig. 5
shows that the nature of the alkyne is also a decisive factor in
the modification. No increase in alkene selectivity was observed
upon addition and removal of 5 vol.% CO during 5 h in ethyne
hydrogenation. Nevertheless, a Cu-based catalyst which was mod-
ified 5 h in mixture B at 473 K and used in ethyne hydrogenation
showed a markedly increased selectivity to ethene. The same
explanation of the enhanced alkene selectivity in propyne hydro-
genation upon short CO supply (reduction in the ensemble leading
to minimized oligomerization) applies to ethyne hydrogenation.
However, the decreased ethene selectivity during time-on-stream
is attributed to the proneness of ethyne toward oligomerization
Acknowledgment
ETH Zürich is acknowledged for financial support.
3 4 2
that overwhelmed the beneficial effect of the C H + H + CO acti-
Appendix A. Supplementary material
vation. The ensemble required for oligomerization in ethyne
hydrogenation is smaller due to the smaller size of the alkyne sub-
strate, and the barrier of C–C coupling is in general lower for C2
than C3 [11].
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2010.11.024.
At this final point, it must be emphasized that the above inter-
pretations are based on observations derived from detailed cata-
lytic evaluation and some of the conclusions require further
support by characterization studies. Our attempts to identify,
through standard (ex situ) XPS and FTIR techniques, conclusive dif-
ferences between the catalysts used in propyne hydrogenation
with or without CO were futile. A rigorous understanding of the
complex mechanism of CO modification on copper in propyne
and ethyne hydrogenation requires dedicated operando studies.
These are beyond the scope of the manuscript and should be tack-
led in future work using infrared spectroscopy coupled to online
analysis of reaction products.
References
[1] G.J. Hutchings, Catal. Lett. 75 (2001) 1.
[
[
2] A. Mittasch, W. Frankenburg, Adv. Catal. 2 (1950) 81.
3] H.F. Rase, Handbook of Commercial Catalysts, CRC Press, United States,
2000.
[4] A. Borodzi n´ ski, G.C. Bond, Catal. Rev. Sci. Eng. 50 (2008) 379.
[
[
5] Á. Molnár, A. Sárkány, M. Varga, J. Mol. Catal. A: Chem. 173 (2001) 185.
6] G.C. Bond, Metal-Catalysed Reactions of Hydrocarbons, Springer, New York,
2005. pp. 395–434.
[7] H. Lindlar, Helv. Chim. Acta 35 (1952) 446.
[
8] M.L. Derrien, in: L. Cˇ erven y´ (Ed.), Catalytic Hydrogenation, Stud. Surf. Sci. Catal,
Elsevier, Amsterdam, 1986, p. 613.
[
9] A. Borodzi n´ ski, G.C. Bond, Catal. Rev. Sci. Eng. 48 (2006) 91.
[10] N. López, B. Bridier, J. Pérez-Ramírez, J. Phys. Chem. C 112 (2008) 9346.
[
11] M. Garcia-Mota, B. Bridier, J. Pérez-Ramírez, N. López, J. Catal. 273 (2010)
2.
9
5
. Conclusions
[
[
[
12] D.L. Trimm, I.O.Y. Liu, N.W. Cant, Appl. Catal. A 374 (2010) 58.
13] Y. Segura, N. López, J. Pérez-Ramírez, J. Catal. 247 (2007) 383.
14] M. Garcia-Mota, N. Cabello, F. Maseras, A.M. Echavarren, J. Pérez-Ramírez, N.
Lopez, ChemPhysChem 9 (2008) 1624.
We have discovered that temporary addition of CO during pro-
pyne hydrogenation over a conventional copper-based catalyst
permanently increases the alkene selectivity from 60% to 92% at
[
15] S. Abelló, D. Verboekend, B. Bridier, J. Pérez-Ramírez, J. Catal. 259 (2008) 85.
[16] B. Bridier, N. López, J. Pérez-Ramírez, J. Catal. 269 (2010) 80.
[
[
[
17] B. Bridier, J. Pérez-Ramírez, J. Am. Chem. Soc. 132 (2010) 4321.
18] B. Bridier, N. López, J. Pérez-Ramírez, Dalton Trans. 39 (2010) 8412.
19] J.T. Wehrli, D.J. Thomas, M.S. Wainwright, D.L. Trimm, N.W. Cant, Appl. Catal.
66 (1990) 199.
1
00% alkyne conversion (473 K). The superior performance of the
CO-modified copper catalyst was also evidenced at lower temper-
ature (373 K), where the unmodified sample completely deacti-
vated due to extensive oligomerization and fouling. Pretreatment
[
[
20] A. Sárkány, A.H. Weiss, T. Szilágyi, P. Sándor, L. Guczi, Appl. Catal. 12 (1984)
373.
3 4 2
in C H + H + CO mixtures is essential to attain a selective copper
21] A. Borodzi n´ ski, Catal. Lett. 63 (1999) 35.
catalyst. Our catalytic results suggest that CO in reaction mixture
alters the structure of the catalyst surface, effectively reducing
[22] J.T. Wehrli, D.J. Thomas, M.S. Wainwright, D.L. Trimm, N.W. Cant, Appl. Catal.
0 (1991) 253.
7