Selective hydrogenation of the C=O bond in acrolein through the architecture of bimetallic surface structures

Article abstract of DOI:10.1021/ja070264m

In the current study we have performed experimental studies and density functional theory (DFT) modeling to investigate the selective hydrogenation of the C=O bond in acrolein on two bimetallic surface structures, the subsurface Pt-Ni-Pt(111) and surface Ni-Pt-Pt(111). We have observed for the first time the production of the desirable unsaturated alcohol (2-propenol) on Pt-Ni-Pt(111) under ultra-high vacuum conditions. Furthermore, our DFT modeling revealed a general trend in the binding energy and bonding configuration of acrolein with the surface d-band center of Pt-Ni-Pt(111), Ni-Pt-Pt(111), and Pt(111), suggesting the possibility of using the value of the surface d-band center as a parameter to predict other bimetallic surfaces for the selective hydrogenation of acrolein.

Full text of DOI:10.1021/ja070264m

Published on Web 05/10/2007
Selective Hydrogenation of the CdO Bond in Acrolein
through the Architecture of Bimetallic Surface Structures
Luis E. Murillo,† Amit M. Goda,‡ and Jingguang G. Chen*,‡
Contribution from the Departments of Materials Science and Engineering and of Chemical
Engineering, Center for Catalytic Science and Technology, UniVersity of Delaware,
Newark, Delaware 19716
Received January 12, 2007; E-mail: jgchen@udel.edu
Abstract: In the current study we have performed experimental studies and density functional theory (DFT)
modeling to investigate the selective hydrogenation of the CdO bond in acrolein on two bimetallic surface
structures, the subsurface Pt-Ni-Pt(111) and surface Ni-Pt-Pt(111). We have observed for the first time
the production of the desirable unsaturated alcohol (2-propenol) on Pt-Ni-Pt(111) under ultra-high vacuum
conditions. Furthermore, our DFT modeling revealed a general trend in the binding energy and bonding
configuration of acrolein with the surface d-band center of Pt-Ni-Pt(111), Ni-Pt-Pt(111), and Pt(111),
suggesting the possibility of using the value of the surface d-band center as a parameter to predict other
bimetallic surfaces for the selective hydrogenation of acrolein.
of acrolein on Pt(111).6,7 These studies suggested that the
absence of gas-phase 2-propenol was likely due to its strong
Introduction
In this paper we report the selective hydrogenation of acrolein
(CH2dCHCHdO) to produce 2-propenol (CH2dCHCH2OH),
which represents for the first time the hydrogenation of the Cd
O bond in acrolein on bimetallic surfaces under ultra-high
vacuum (UHV) conditions. The selective hydrogenation of R,â-
unsaturated aldehydes to unsaturated alcohols is of growing
interest for the production of fine chemicals and pharmaceutical
precursors.1,2 In addition, a fundamental understanding of the
competitive hydrogenation mechanisms of the CdC and CdO
bonds in unsaturated aldehydes offers the opportunity to identify,
and potentially design, catalyst formulations and structures for
the selective hydrogenation of organic molecules containing
multiple functional groups. Experimental efforts for the hydro-
genation of unsaturated aldehydes have been primarily focused
on supported catalysts of group VIII transition metals, in
particular platinum-based bimetallic catalysts.1-3 Due to the
polycrystalline nature of the supported catalysts, a fundamental
understanding of the hydrogenation pathways is lacking at
present. Zaera et al. have attempted to determine the reaction
pathways of acrolein, the smallest R,â-unsaturated aldehyde,
on the single crystal surface of Pt(111) under UHV conditions.4,5
Although novel decomposition pathways were observed on Pt-
(111), the hydrogenation of either the CdO or CdC bond was
not detected.5 More recently, density functional theory (DFT)
modeling has been performed on the adsorption and reaction
bonding on the Pt(111) surface.
In principle, hydrogenation pathways can be selectively
modified by the formation of bimetallic surfaces on Pt(111), in
particular by controlling the structures of the bimetallic
surfaces.8-12 For instance, depositing one monolayer (ML) of
Ni on Pt(111) can lead to the formation of either the Ni-Pt-
Pt(111) surface structure, with Ni atoms residing on the topmost
surface, or the Pt-Ni-Pt(111) subsurface structure, with most
Ni atoms residing between the first and second layers of Pt, as
described previously.11 These structures have been previously
confirmed using low-energy ion scattering (LEIS) and grazing-
angle X-ray photoelectron spectroscopy (XPS).11 These two
bimetallic surfaces have very unique electronic and chemical
properties as compared to Pt(111). For example, results from
DFT modeling revealed that the surface d-band center of the
subsurface Pt-Ni-Pt(111) was further away from the Fermi
level as compared to those of both Pt(111) and Ni-Pt-Pt(111)
surfaces,13,14 leading to weaker binding energies of atomic
hydrogen and alkenes on Pt-Ni-Pt(111). The weaker binding
energies in turn led to a novel low-temperature hydrogenation
pathway of alkenes on Pt-Ni-Pt(111), as confirmed experi-
(6) Loffreda, D.; Delbecq, F.; Fabienne, V.; Sautet, P. Angew. Chem., Int. Ed.
2005, 44, 5279.
(7) Loffreda, D.; Delbecq, F.; Fabienne, V.; Sautet, P. J. Am. Chem. Soc. 2006,
128, 1316.
(8) Beccat, P.; Bertolini, J. C.; Gauthier, Y.; Massardier, J.; Ruiz, P. J. Catal.
1990, 126, 451.
† Department of Materials Science and Engineering.
‡ Department of Chemical Engineering.
(9) Jerdev, D. I.; Olivas, A.; Koel, B. E. J. Catal. 2002, 205, 278.
(10) Hwu, H. H.; Eng, J., Jr.; Chen, J. G. J. Am. Chem. Soc. 2002, 124, 702.
(11) Kitchin, J. R.; Khan, N. A.; Barteau, M. A.; Chen, J. G.; Yakshinksiy, B.;
Madey, T. E. Surf. Sci. 2003, 544, 295.
(1) Gallezot, P.; Richard, D. Catal. ReV.sSci. Eng. 1998, 40, 81 and references
therein.
(2) Claus, P. Top. Catal. 1998, 5, 51 and references therein.
(3) Maki-Arvela, P.; Hajek, J.; Salmi, T.; Murzin, D. Y. Appl. Catal., A 2005,
292, 1 and references therein.
(4) de Jesus, J. C.; Zaera, F. Surf. Sci. 1999, 430, 99.
(5) de Jesus, J. C.; Zaera, F. J. Mol. Catal. A 1999, 138, 237.
(12) Olivas, A.; Jerdev, D. I.; Koel, B. E. J. Catal. 2004, 222, 285.
(13) Skoplyak, O.; Barteau, M. A.; Chen, J. G. J. Phys. Chem. B 2006, 110,
1686.
(14) Kitchin, J. R.; Nørskov, J. K.; Barteau, M. A.; Chen, J. G. J. Chem. Phys.
2004, 120, 10240.
9
10.1021/ja070264m CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 7101-7105
7101

A R T I C L E S
Murillo et al.
mentally.10 In addition, studies on other bimetallic systems, such
as Fe8 and Sn15 on Pt(111), also demonstrated novel hydrogena-
tion pathways of â-substituted acrolein, including crotonalde-
hyde and 3-methylcrotonaldehyde. It was observed that the
hydrogenation activity was enhanced by the formation of
bimetallic surfaces.
Sham equations were solved using a plane-wave basis set with a cutoff
energy of 396 eV. The PW91 functional was used to describe the
exchange correlation term. The core electrons and the nuclei of the
atoms were described by the Vanderbilt ultrasoft pseudopotential.19,20
Electronic energies were calculated using a 3 × 3 × 1 k-point grid
mesh. The coadsorption of acrolein and atomic hydrogen was studied
using 3 × 3 supercells, with nine metal atoms in each layer and with
two hydrogen atoms and one acrolein molecule adsorbed per unit
cell. Four metal layers were used, with the top two layers allowed to
relax in each case. A vacuum region equivalent in thickness to six
metal layers was used to separate the slabs to avoid any electronic
interactions between them. Calculations for gas-phase acrolein and
adsorbate-metal systems were carried out spin-unpolarized. The value
of the surface d-band center was calculated as the first moment of the
projected d-band density of states on the surface atoms with reference
to the Fermi level.
In the current study we have performed experimental studies
and DFT modeling to explore whether the unique chemical
properties of the Pt-Ni-Pt(111) subsurface structure, such as
the weaker binding energies of atomic hydrogen and alkenes,
would lead to the selectiVe hydrogenation of the CdO bond in
acrolein. The formation of the desirable hydrogenation product,
2-propenol, was detected from the Pt-Ni-Pt(111) subsurface
structure using temperature programmed desorption (TPD).
Parallel surface vibrational studies provided evidence of the
interaction of the CdO bond of acrolein with Pt-Ni-Pt(111),
most likely responsible for the selective hydrogenation of the
CdO bond on this surface. Furthermore, DFT modeling
confirmed that one of the adsorption configurations of acrolein
occurred via the interaction of the CdO bond with Pt-Ni-Pt-
(111), which was not favored on either Ni-Pt-Pt(111) or Pt-
(111). The DFT results also showed a correlation between the
adsorption energy and bonding configuration of acrolein with
the surface d-band center, which demonstrated the possibility
to potentially predict, on the basis of the value of the d-band
center, other desirable bimetallic structures for the selective
hydrogenation of the CdO bond in acrolein.
Results and Discussion
Figure 1 compares the TPD results following the hydrogena-
tion of 0.5 langmuir (1 langmuir ) 1 × 10-6 Torr‚s) of acrolein
on Pt-Ni-Pt(111), Ni-Pt-Pt(111), and Pt(111), after the
coadsorption of acrolein and atomic hydrogen on these surfaces
at 100 K. Figure 1 displays the TPD spectra of mass 31 amu,
a cracking fragment for both CH2dCHCH2OH (2-propenol) and
CH3CH2CH2OH (1-propanol), masses 58 and 57 amu, cracking
fragments for both CH2dCHCH2OH and CH3CH2CHdO (pro-
panal), and mass 60 amu, a cracking fragment of CH3CH2CH2-
OH. The production of either CH2dCHCH2OH or CH3CH2-
CH2OH is not detected from Pt(111) and Ni-Pt-Pt(111), as
indicated by the absence of any desorption peaks in the 31 and
60 amu TPD spectra from the two surfaces. In contrast, a
relatively intense desorption peak is observed at 260 K from
the Pt-Ni-Pt(111) subsurface structure and is attributed to the
formation of primarily 2-propenol, on the basis of the peak
intensity ratio in the 31 and 60 amu spectra.
Experimental and Modeling Methods
The surface science experiments were performed in a UHV chamber
equipped with capabilities for TPD and high-resolution electron energy
loss spectroscopy (HREELS). The bimetallic surfaces were prepared
by depositing Ni onto a Pt(111) single crystal (Metal Crystals and
Oxides, Ltd., Cambridge, U.K.). The Ni-Pt-Pt(111) surface structure
and the Pt-Ni-Pt(111) subsurface structure were obtained by
depositing one monolayer of Ni at 300 and 600 K, respectively.11
TPD experiments were performed at a heating rate of 3 K/s. Absolute
TPD yields were estimated using the procedure developed by Ko et
al.,16 correcting for mass spectrometer sensitivity factors and using
experimental fragmentation patterns. The integrated area of the CO
desorption spectrum was used as a base for the absolute yield
calculation. The absolute yield of CO at saturation coverage was
calculated experimentally by comparing it with the expected saturation
coverage on Pt(111) (0.68 ML);17 the yields of all the desorption species
were calculated by using the corresponding sensitivity factors relative
to CO.
The HREEL spectra were collected with a primary beam energy of
6 eV. In the case of on-specular measurements, the angles of incidence
and reflection were 60° with respect to the surface normal. For the
off-specular measurements, a reflection angle of 50° with respect to
the surface normal was used. Count rates in the elastic peak were in
the range of 3 × 104 to 3 × 105 counts per second (cps), and the spectral
resolution was between 30 and 40 cm-1 full width at half-maximum
(fwhm). The sample was annealed to the specified temperature with a
linear rate of 3 K/s, held for 5 s, and then cooled to ∼100 K for data
collection.
The formation of propanal, from the hydrogenation of the
CdC bond in acrolein, from the 57 and 58 amu TPD spectra,
shows a relatively weak peak from the Pt(111) surface at 192
K. The detection of desorption peaks at 215, 254, and 302 K in
the 57 and 58 amu spectra on the Ni-Pt-Pt(111) surface,
coupled with the absence of desorption peaks in the 31 amu
TPD, suggests the formation of propanal. In addition, the peak
area ratios of the individual peaks between masses 57 and 58
amu are similar to the ones found for the intensity ratios of the
same masses in the mass spectrum of propanal (area58/area57
≈
3). On the Pt-Ni-Pt(111) surface, the peaks at 184 and 215 K
in the 57 and 58 amu spectra are attributed to the desorption of
propanal, while the peak at 260 K is from the contribution from
both propanal and 2-propenol on the basis of the comparison
with the 31 amu spectrum.
The yields for the three gas-phase hydrogenation products,
2-propenol, propanal, and 1-propanol, are shown in Table 1.
The yields were quantified from the TPD peak areas from the
three surfaces. The empirical procedure developed by Ko
et al.16 for the correction of mass spectrometer sensitivity
factors of these yield calculations is used. Peak deconvolution
is performed when the desorption feature is due to the
Self-consistent periodic slab calculations were carried out on the
basis of the gradient-corrected DFT method. All calculations were
performed using periodic density functional theory as implemented in
the code VASP (Vienna ab initio Simulation Package).18 The Kohn-
(18) (a) Kresse, G.; Hafner, J. Phys. ReV. B 1993, 47, 558. (b) Kresse, G.;
Furthmuller, J. Comput. Mat. Sci. 1996, 6, 15. (c) Kresse, G.; Furthmuller,
J. Phys. ReV. B 1996, 54, 11169.
(15) Jerdev, D. I.; Koel, B. E. Surf. Sci. 2002, 513, L391.
(16) Ko, E. I.; Benziger, J. B.; Madix, R. J. J. Catal. 1980, 62, 264.
(17) Ertl, G.; Neumann, M.; Streit, K. M. Surf. Sci. 1977, 64, 393.
(19) Vanderbilt, D. Phys. ReV. B 1990, 41, 7892.
(20) Kresse, G.; Hafner, J. J. Phys.: Condens. Matter 1994, 6, 8245.
9
7102 J. AM. CHEM. SOC. VOL. 129, NO. 22, 2007

Selective Hydrogenation of the CdO Bond in Acrolein
A R T I C L E S
Figure 1. TPD spectra following the reaction of acrolein with coadsorbed atomic hydrogen at 100 K on different surfaces: (a) 2-propenol (31 amu), (b, c)
propanal and 2-propenol (58 and 57 amu, respectively), and (d) 1-propanol (60 amu).
Table 1. TPD Yields (Molecules per Surface Pt Atom) of
Hydrogenation Products after the Reaction of 0.5 langmuir of
are the calculated gas-phase values of acrolein using Gaussian
Acrolein with Preadsorbed Atomic Hydrogen on Different Surfaces
frequencies listed at the top of the figure. These frequencies
03.21 At 100 K (not shown), the frequencies of adsorbed acrolein
propanal
2-propenol
1-propanol
on H/Pt-Ni-Pt(111) are very similar to the corresponding gas-
phase values, suggesting weakly bonded acrolein on this
bimetallic surface. After annealing to 200 K, the ν(CdO) mode
at ∼1700 cm-1 disappears from the subsurface Pt-Ni-Pt(111),
indicating either the formation of di-σ-bonded acrolein through
the C-O moiety or the onset of the hydrogenation of the CdO
bond. However, the absence of the ν(O-H) mode, even in the
off-specular spectrum, suggests that the CdO hydrogenation
does not occur on the surface at 200 K. The remaining ν(Cd
C) mode at ∼1569 cm-1 is characteristic of the ν(CdC)
stretching mode in the gas-phase spectrum of 2-propenol. The
corresponding off-specular spectrum shows vibrational features
similar to those described in the on-specular mode. In contrast,
the nature of the surface intermediate on H/Ni-Pt-Pt(111) is
significantly different. For example, the ν(CdO) mode remains
on the H/Ni-Pt-Pt(111) surface at 200 K, indicating that the
CdO bond of acrolein does not interact strongly with the
surface, which should be most likely responsible for the lack
of CdO hydrogenation on Ni-Pt-Pt(111), as confirmed in the
TPD measurements.
surface
(CH3CH2CH
d
O)
(CH2dCH2CH2OH)
(CH3CH2CH2OH)
H/Pt(111)
H/Ni-Pt-Pt(111)
H/Pt-Ni-Pt(111)
0.004
0.025
0.12
0
0
0
0
0.025
0.002
contribution of the cracking patterns of the mixed products, such
as the desorption peaks at 260 K on the H/Pt-Ni-Pt(111)
surface. Overall, the TPD results indicate that the hydrogenation
of the CdO bond does not occur on either Pt(111) or Ni-Pt-
Pt(111). In contrast, the subsurface Pt-Ni-Pt(111) is active
for the hydrogenation of acrolein to produce both CH2dCHCH2-
OH and CH3CH2CHdO.
To further understand the onset of the novel CdO hydroge-
nation pathway on the Pt-Ni-Pt(111) surface, the bonding
configurations of acrolein have been investigated using HREELS,
which is an ideal technique to differentiate between the various
bonding configurations of acrolein. For example, Figure 2 shows
HREEL spectra following the coadsorption of acrolein and
atomic hydrogen on the two bimetallic structures, subsurface
Pt-Ni-Pt(111) and surface Ni-Pt-Pt(111), after the annealing
to 200 K, which is below the desorption temperature of the
hydrogenation products. The physisorbed spectrum of acrolein,
after multilayer adsorption on Pt(111), is also included at the
bottom of Figure 2 for comparison with the vibrational
To correlate the TPD results in Figure 1 and Table 1 and the
vibrational fingerprints observed in HREELS with the adsorption
(21) Frisch, M. J.; et al. Gaussian 03, reVision B.04; Gaussian, Inc.: Pittsburgh,
PA, 2003.
9
J. AM. CHEM. SOC. VOL. 129, NO. 22, 2007 7103

A R T I C L E S
Murillo et al.
Figure 2. On and off-specular HREEL spectra of the coadsorption of
acrolein and atomic hydrogen on H/Pt-Ni-Pt(111) and H/Ni-Pt-Pt(111)
surfaces after annealing to 200 K. Physisorbed acrolein at 100 K (bottom
spectrum) and the calculated gas-phase vibrational features are included
for comparison.
configurations of acrolein, we have performed DFT modeling
to determine the binding energies (BEs) of different adsorption
configurations of acrolein on H/Pt-Ni-Pt(111), H/Ni-Pt-Pt-
(111), and H/Pt(111) surfaces. Figure 3a displays the three
binding configurations of acrolein on the Pt-Ni-Pt(111) surface
through di-σ-C-C, di-σ-C-O, and η4(C,C,C,O). Figure 3b
shows the self-consistent periodic DFT calculations for atomic
hydrogen and acrolein bonded in the di-σ-C-C, di-σ-C-O, and
η4(C,C,C,O) configurations. The DFT results show that the BE
values of acrolein are in the order H/Pt-Ni-Pt(111) < H/Pt-
(111) < H/Ni-Pt-Pt(111), which follows a trend similar to
that of the BE values of atomic hydrogen on these surfaces. In
addition, on all three surfaces the di-σ-bonding through the C-C
bond of acrolein is stronger than through the C-O bond. In
fact, on the H/Pt(111) and H/Ni-Pt-Pt(111) surfaces, the
binding through the di-σ-C-O configuration is so weak that it
converts to the flat η4(C,C,C,O) configuration during the course
of the DFT calculations. Hence, on H/Pt(111) and H/Ni-Pt-
Pt(111) surfaces acrolein tends to be adsorbed through either
the di-σ-C-C or the η4(C,C,C,O) mode. Last, the η4(C,C,C,O)
configuration shows a binding energy similar to that of the di-
σ-C-C configuration on all the surfaces. By comparing the
trends in the TPD results in Table 1 and DFT modeling in Figure
3b, it appears that the higher hydrogenation activity on H/Pt-
Ni-Pt(111) is related to the weaker bonding of both acrolein
and atomic hydrogen, which is consistent with our previous
studies of correlating the presence of weakly bonded alkenes
and weakly bonded atomic hydrogen with enhanced hydrogena-
Figure 3. (a) Side view and top view of configurations of adsorbed acrolein
through di-σ-C-C, di-σ-C-O, and η4(C,C,C,O) on the Pt-Ni-Pt(111)
surface. (b) DFT calculations of acrolein and hydrogen binding energies
vs the surface d-band center for H/Pt-Ni-Pt(111), H/Pt(111), and H/Ni-
Pt-Pt(111) surfaces.
tion activity of alkenes.10 The onset of the selective hydrogena-
tion pathway on H/Pt-Ni-Pt(111) is most likely due to the
similar BE values of the three adsorption configurations of
acrolein. The similar binding energies would lead to the
adsorption of acrolein through both the CdO and CdC moieties,
making it possible to hydrogenate both the CdO and CdC
bonds, as summarized in the hydrogenation activities in Table
1. On the other hand, as predicted by DFT, the adsorption of
acrolein on H/Pt(111) and H/Ni-Pt-Pt(111) surfaces does not
occur through the di-σ-C-O configuration, hence making it
difficult to hydrogenate the CdO bond, as confirmed by the
absence of the 2-propenol product in the TPD measurements.
It should be pointed out that DFT modeling is often
inadequate in describing weakly adsorbed species, such as
acrolein on H/Pt-Ni-Pt(111), as indicated by the weak binding
energies. Other types of modeling approaches, such as those
that can more accurately describe dispersive van der Waals
interactions, should be performed to quantitatively differentiate
the three adsorption configurations on the H/Pt-Ni-Pt(111)
surface. A more accurate prediction of the hydrogenation
selectivity should also include the calculations of the activation
barriers for the different reaction pathways. However, these
calculations are computationally much more expensive than the
9
7104 J. AM. CHEM. SOC. VOL. 129, NO. 22, 2007

Selective Hydrogenation of the CdO Bond in Acrolein
A R T I C L E S
DFT modeling of binding energies, such as those shown in
Figure 3. Overall, the comparison of the trend in the binding
energies in Figure 3 suggests that the onset of the selective
hydrogenation pathway of the CdO bond on Pt-Ni-Pt(111)
is most likely related to the following two factors: (1) the
weaker binding energies of acrolein and atomic hydrogen, which
should favor the hydrogenation and subsequent desorption of
the reaction products from the surface, and (2) the smaller
difference between the binding energies of the three bonding
configurations, which should enhance the possibility of bonding
through the CdO moiety and therefore the subsequent hydro-
genation of the CdO bond. Assuming that both factors are
related to the selective hydrogenation of the CdO bond, and
assuming that the correlation of binding energies with the surface
d-band center continues in Figure 3b for other surfaces, one
can predict that bimetallic surfaces with the d-band center further
away from the Fermi level might have higher selectivity than
Pt-Ni-Pt(111). From our previous DFT calculations,14,22 the
surface d-band center of the Pt-3d-Pt(111) subsurface structure
shifts away from the Fermi level as the subsurface 3d metal
moves toward the left-hand side of the periodic table. If the
trend in Figure 3b continues for other Pt-3d-Pt(111) surfaces
as the subsurface 3d metal moves from Ni to Ti, the binding
through the CdO bond of acrolein could potentially become
preferred, making those Pt-3d-Pt(111) surfaces more selective
toward the CdO bond hydrogenation. However, it is also
important to keep in mind that, as the surface d-band center
moves further away from the Fermi level, the bonding of
acrolein might become too weak for the subsequent hydrogena-
tion to occur. Experimental and DFT studies are under way to
extend the correlation in Table 1 and Figure 3b by investigating
the binding energies and hydrogenation selectivity of acrolein
on the subsurface structures of other Pt-3d-Pt(111) bimetallic
surfaces.
Conclusions
Our combined experimental and theoretical investigations
demonstrate that the formation of bimetallic surfaces leads to
the onset of the selective hydrogenation of the CdO bond in
acrolein. More importantly, our results reveal that the structure
of the bimetallic surface plays a critical role, with the subsurface
Pt-Ni-Pt(111) structure being much more active and selective
than the surface Ni-Pt-Pt(111) structure for the hydrogenation
of the CdO bond in acrolein. The different behavior of the two
bimetallic structures is explained with DFT modeling in terms
of the differences in binding energies and bonding configurations
of acrolein, which appear to be related to the surface d-band
center of these surfaces. Such correlation in turn demonstrates
the possibility to identify other bimetallic formulations and
structures, on the basis of the value of the surface d-band center,
that can potentially be more selective than Pt-Ni-Pt(111) for
the preferential adsorption of the CdO moiety and consequently
the hydrogenation of the CdO bond in acrolein and other R,â-
unsaturated aldehydes.
Acknowledgment. We thank the Department of Energy,
Office of Basic Energy Sciences (Grant No. DE-FG02-
00ER15104), for funding. We also thank Prof. Neurock of the
University of Virginia for discussion and for providing CPU
time.
Supporting Information Available: Tables and a figure
showing the absolute energies of hydrogen and acrolein in
hartrees, optimized geometries (as Cartesian coordinates) of all
DFT-calculated structures, and a top view of the adsorption
structure of acrolein and H in a periodic slab containing four
(3 × 3) unit cells to illustrate the relative positions of acrolein
and H. This material is available free of charge via the Internet
at http://pubs.acs.org.
(22) Menning, C. A.; Hwu, H. H.; Chen, J. G. J. Phys. Chem. B 2006, 110,
15471.
JA070264M
9
J. AM. CHEM. SOC. VOL. 129, NO. 22, 2007 7105