Furan hydrogenation over Pt(111) and Pt(100) single-crystal surfaces and Pt nanoparticles from 1 to 7 nm: A kinetic and sum frequency generation vibrational spectroscopy study

Article abstract of DOI:10.1021/ja105800z

Sum frequency generation surface vibrational spectroscopy and kinetic measurements using gas chromatography have been used to systematically study the adsorption and hydrogenation of furan over Pt(111) and Pt(100) single-crystal surfaces and size-controlled 1.0-nm, 3.5-nm and 7.0-nm Pt nanoparticles at Torr pressures (10 Torr of furan, 100 Torr of H₂) to form dihydrofuran, tetrahydrofuran, and the ring-cracking products butanol and propylene. As determined by SFG, the furan ring lies parallel to all Pt surfaces studied under hydrogenation conditions. Upright THF and the oxametallacycle intermediate are observed over the nanoparticle catalysts under reaction conditions. Butoxy increases in surface concentration over Pt(111) with increasing temperature in agreement with selectivity trends.

Full text of DOI:10.1021/ja105800z

Published on Web 08/26/2010
Furan Hydrogenation over Pt(111) and Pt(100) Single-Crystal
Surfaces and Pt Nanoparticles from 1 to 7 nm: A Kinetic and
Sum Frequency Generation Vibrational Spectroscopy Study
Christopher J. Kliewer,†,‡ Cesar Aliaga,†,‡ Marco Bieri,†,‡ Wenyu Huang,†,‡
Chia-Kuang Tsung,†,‡ Jennifer B. Wood,†,‡ Kyriakos Komvopoulos,§ and
Gabor A. Somorjai*,†,‡
Departments of Chemistry and Mechanical Engineering, UniVersity of California, Berkeley,
California 94720, and Materials Sciences DiVision, Lawrence Berkeley National Laboratory,
Berkeley, California 94720
Received July 1, 2010; E-mail: somorjai@socrates.berkeley.edu
Abstract: Sum frequency generation surface vibrational spectroscopy and kinetic measurements using
gas chromatography have been used to systematically study the adsorption and hydrogenation of furan
over Pt(111) and Pt(100) single-crystal surfaces and size-controlled 1.0-nm, 3.5-nm and 7.0-nm Pt
nanoparticles at Torr pressures (10 Torr of furan, 100 Torr of H2) to form dihydrofuran, tetrahydrofuran,
and the ring-cracking products butanol and propylene. As determined by SFG, the furan ring lies parallel
to all Pt surfaces studied under hydrogenation conditions. Upright THF and the oxametallacycle intermediate
are observed over the nanoparticle catalysts under reaction conditions. Butoxy increases in surface
concentration over Pt(111) with increasing temperature in agreement with selectivity trends.
catalysis research for many years,8 and it provides the reference
state needed to probe and compare more complex nanoparticle
surfaces which are more closely related to what is used
industrially. The technique of sum frequency generation vibra-
tional spectroscopy, a surface specific vibrational spectroscopy,
has proven to be a powerful surface analysis technique over
single crystal catalysts to reveal both adsorption modes and
reaction intermediates in situ from UHV to atmospheric reactant
pressures.6,9-15 SFG-VS has now been applied to the study of
size-controlled nanoparticle systems,16,17 which was not possible
until recently due to the fact that the capping or stabilizing agent
used in the synthesis of the nanoparticles interfered with the
surface spectra of adsorbed reaction intermediates. However,
Aliaga et al.16 solved this problem by applying a UV-ozone
cleaning treatment in the presence of oxygen to combust capping
1. Introduction
Catalytic reactions involving aromatic cyclic and heterocyclic
molecules are important for the chemical industry for both fuel
reforming and environmental concerns.1 The adsorption of six-
membered molecules, e.g. benzene and pyridine, has been
extensively studied on various metal surfaces.2-7 Five-
membered ring systems, such as pyrrole, furan, and thiophene
have received comparatively much less attention. This may be
due in part to the more complex behavior of five-membered
rings upon adsorption to catalytic surfaces, making them less
attractive for fundamental surface science study. A thorough
understanding of the adsorption and reactions of these five-
membered aromatic molecules not only is important fundamen-
tally as model systems for studying heterogeneous catalysis but
also is directly important to the petrochemical industry as it
applies to the hydrodenitrogenation, hydrodesulfurization, and
hydrodeoxygenation processes.
(8) Somorjai, G. Introduction to Surface Science and Catalysis; Wiley:
New York, 1994.
(9) Cremer, P. S.; McIntyre, B. J.; Salmeron, M.; Shen, Y. R.; Somorjai,
G. A. Catal. Lett. 1995, 34, 11.
(10) Cremer, P. S.; Su, X. C.; Shen, Y. R.; Somorjai, G. J. Chem. Soc.,
Faraday Trans. 1996, 92, 4717.
The use of single crystals to study surface chemical reactions
over metal catalysts has been a mainstay of surface science and
(11) Cremer, P. S.; Su, X. C.; Shen, Y. R.; Somorjai, G. A. J. Am. Chem.
Soc. 1996, 118, 2942.
(12) Kliewer, C. J.; Bieri, M.; Somorjai, G. A. J. Phys. Chem. C 2008,
112, 11373.
† Department of Chemistry, University of California.
‡ Lawrence Berkeley National Laboratory.
(13) McCrea, K.; Parker, J. S.; Chen, P. L.; Somorjai, G. Surf. Sci. 2001,
494, 238.
§ Department of Mechanical Engineering, University of California.
(1) Cooper, B. H.; Donnis, B. B. L. Appl. Catal., A 1996, 137, 203.
(2) Netzer, F. P.; Bertel, E.; Matthew, J. A. D. Surf. Sci. 1980, 92, 43.
(3) Horsley, J. A.; Stohr, J.; Hitchcock, A. P.; Newbury, D. C.; Johnson,
A. L.; Sette, F. J. Chem. Phys. 1985, 83, 6099.
(14) Buck, M.; Himmelhaus, M. J. Vac. Sci. Technol., A 2001, 19, 2717.
(15) Rupprechter, G.; Bruce, C. G.; Helmut, K. Sum Frequency Generation
and Polarization-Modulation Infrared Reflection Absorption Spectros-
copy of Functioning Model Catalysts from Ultrahigh Vacuum to
Ambient Pressure. In AdVances in Catalysis; Academic Press: 2007;
Vol. 51, p 133.
(4) Ogletree, D. F.; Vanhove, M. A.; Somorjai, G. A. Surf. Sci. 1987,
183, 1.
(5) Weiss, P. S.; Eigler, D. M. Phys. ReV. Lett. 1993, 71, 3139.
(6) Bratlie, K. M.; Kliewer, C. J.; Somorjai, G. A. J. Phys. Chem. B 2006,
110, 17925.
(16) Aliaga, C.; Park, J. Y.; Yamada, Y.; Lee, H. S.; Tsung, C. K.; Yang,
P. D.; Somorjai, G. A. J. Phys. Chem. C 2009, 113, 6150.
(17) Dellwig, T.; Rupprechter, G.; Unterhalt, H.; Freund, H. J. Phys. ReV.
Lett. 2000, 85, 776.
(7) Kliewer, C.; Somorjai, G. Catal. Lett. 2010, 137, 118.
9
13088 J. AM. CHEM. SOC. 2010, 132, 13088–13095
10.1021/ja105800z  2010 American Chemical Society

Furan Hydrogenation over Pt Catalysts
A R T I C L E S
Scheme 1. Reaction for the Hydrogenation of Furan
butanol once desorbed, but the conversion of DHF or THF to
butanol was not observed.
In this study the hydrogenation of furan to form the observed
products DHF, THF, and the ring-cracking products butanol and
propylene (Scheme 1) was carried out over Pt(111) and Pt(100)
single crystals to help elucidate the effect catalyst structure has
on the reaction mechanism and selectivity, and these results are
compared to the reaction carried out over 1.0-nm, 3.5-nm and
7.0-nm Pt nanoparticles to help determine the effect of catalyst
size. Further, sum frequency generation vibrational spectroscopy
was carried out in situ to elucidate the surface species over all
of the model catalysts in the temperature range of 23 to 140 °C
under 10 Torr of furan and 100 Torr of hydrogen.
hydrocarbons from the surface of Pt nanoparticles deposited on
a fused silica prism. In this study, this technique was applied
so that the results over the single crystals could be compared
to 1-7-nm platinum nanoparticles.
Furan (C4H4O) is a five-membered aromatic ring system
(Scheme 1) in which one of the two lone pairs of electrons of
the oxygen is delocalized over the π-system of the ring. The
adsorption of furan on various single crystal surfaces has been
studied. Sexton et al.18 investigated the adsorption of furan on
a Cu(100) surface using EELS and TDS determining that furan
adsorbed to the surface at 85 K with a monolayer of furan
molecules π-bonded and lying parallel to the surface, while a
second “bilayer” builds on top of this which is tilted with respect
to the surface. Gui et al.19 studied the adsorption of the
5-membered heterocycles, including furan, on Pt(111) electrodes
using EELS ex-situ after immersion in a solution of furan at a
given pH. In this case the furan underwent hydrolysis on the
Pt(111) surface. While the conditions are quite different than
gas phase hydrogenation, it is interesting to note the ring-
cracking over Pt(111). The adsorption structure of furan on
Pd(111) was investigated by Knight et al.20 using NEXAFS and
scanned-energy mode photoelectron differaction (PhD), with the
conclusion being that the furan molecule is adsorbed close to
parallel (within 10°) to the Pd(111) surface. This is in agreement
with STM images taken by Loui and Chiang21 which demon-
strate furan adsorbed nearly parallel to the Pd(111) surface.
Solomon and Madix22 determined both furan and dihydrofuran
(DHF) adsorb to Ag(110) at a 22 (( 7)° angle using NEXAFS.
Limited work has been performed over Pt surfaces as to the
adsorption and reaction of furan. Hlavathy et al.23 studied furan
and tetrahydrofuran (THF) adsorption to a Pt foil using work
function changes and Auger electron spectroscopy. Their
determination was that furan adsorbs to a Pt foil via a surface
π-complex, which would necessitate a near-parallel adsorption
scheme. They determine further that THF is adsorbed to Pt via
a C-Pt σ-bond and that the addition of hydrogen may increase
C-O dissociation (bond scission) within the reactants.
2. Experimental Section
2.1. Materials. Prior to use, furan (>99%, Sigma-Aldrich), 2,3-
dihydrofuran (99%, Aldrich), tetrahydrofuran (>99%,Sigma-Ald-
rich) and 1-butanol (99.8%,Sigma-Aldrich) were subjected to
several freeze-pump-thaw cycles and the purities were checked
by means of gas chromatography.
2.2. The High-Pressure/Ultrahigh-Vacuum System. All single
crystal experiments reported here were carried out in a high-
pressure/ultrahigh-vacuum (HP/UHV) system. The UHV chamber
is operated at a base pressure of 2 × 10-10 Torr and is isolated
from the HP cell by a gate valve. The UHV system is equipped
with an Auger electron spectrometer (AES), a quadrupole mass
spectrometer (Stanford Research Systems) and an ion bombardment
gun (Eurovac). The HP cell consists of two CaF2 conflat windows
that allow transmission of infrared (IR), visible (VIS) and sum
frequency radiation for sum frequency generation (SFG) experi-
ments. The product gases in the HP cell are constantly mixed via
a recirculation pump, and kinetic data is acquired by periodically
sampling the reaction mixture and analyzing the relative gas phase
composition in a flame ionization detector (FID) of a gas chro-
matograph (Hewlett-Packard HP 5890 on a 5% Carbowax 20 M
packed column).
2.3. Single Crystal Sample Preparation. Prior to each experi-
ment, the Pt(111) and Pt(100) crystal surfaces were cleaned in the
UHV chamber by Ar+ (1 keV for Pt) sputtering for 20 min at about
3 × 10-5 Torr of Ar. After sputtering, the crystals were heated to
1103 K in the presence of O2 of 5 × 10-7 Torr and annealed at the
same temperature for 2 min. The cleanliness of the crystal surfaces
was verified by AES and the crystallographic structure verified with
low energy electron diffraction (LEED). The samples were then
transferred into the HP cell for SFG and kinetic studies.
2.4. Nanoparticle Preparation. The synthesis and characteriza-
tion of the nanoparticles used in this study has been previously
reported and is not the focus of this work16,25-28 Further, the
stability of the nanoparticles against agglomeration during the
UV-ozone cleaning treatment has been previously reported.16
2.5. 7-nm Pt Cubic and 3.5-nm Nanocrystal Synthesis. A total
of 0.05 mmol of Pt ion (NH4)2Pt(IV)Cl6, 0.75 mmol of tetra-
methylammonium bromide, and 1.00 mmol of poly(vinylpyrroli-
done) (in terms of the repeating unit; MW 29,000) were dissolved
into 10 mL of ethylene glycol in a 25 mL round-bottom flask at
room temperature. The mixed solution was heated to 180 °C in an
oil bath at 60 °C/min. For 3.5 nm Pt spherical nanocrystal synthesis,
a total of 0.05 mmol of Pt ion (NH4)2Pt(II)Cl4), 0.75 mmol of
The catalytic study of furan hydrogenation was first carried
out in 1949 by Smith and Fuzek24 over Adams platinum
(platinum dioxide reduced in hydrogen). It was observed that
furan hydrogenation yielded one product, butanol. By using
kinetic arguments, it was argued that over Adams platinum the
reaction mechanism proceeded along one of two pathways. The
furan ring could hydrogenate to yield dihydrofuran and tetrahy-
drofuran sequentially, or the furan ring could break to form
(18) Sexton, B. A. Surf. Sci. 1985, 163, 99.
(25) Huang, W.; Kuhn, J. N.; Tsung, C. K.; Zhang, Y.; Habas, S. E.; Yang,
P.; Somorjai, G. A. Nano Lett. 2008, 8, 2027.
(19) Gui, J. Y.; Stern, D. A.; Lu, F.; Hubbard, A. T. J. Electroanal. Chem.
1991, 305, 37.
(26) Kuhn, J. N.; Huang, W. Y.; Tsung, C. K.; Zhang, Y. W.; Somorjai,
G. A. J. Am. Chem. Soc. 2008, 130, 14026.
(20) Knight, M. J.; Allegretti, F.; Kro¨ger, E. A.; Polcik, M.; Lamont,
C. L. A.; Woodruff, D. P. Surf. Sci. 2008, 602, 2524.
(21) Loui, A.; Chiang, S. Appl. Surf. Sci. 2004, 237 (1-4), 555.
(22) Solomon, J. L.; Madix, R. J.; Stohr, J. J. Chem. Phys. 1991, 94, 4012.
(23) Hlavathy, Z.; Tetenyi, P. Surf. Sci. 2007, 601, 2026.
(24) Smith, H. A.; Fuzek, J. F. J. Am. Chem. Soc. 1949, 71, 415.
(27) Lee, H.; Habas, S. E.; Kweskin, S.; Butcher, D.; Somorjai, G. A.;
Yang, P. D. Angew. Chem., Int. Ed. 2006, 45, 7824.
(28) Zhang, Y. W.; Grass, M. E.; Kuhn, J. N.; Tao, F.; Habas, S. E.; Huang,
W. Y.; Yang, P. D.; Somorjai, G. A. J. Am. Chem. Soc. 2008, 130,
5868.
9
J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010 13089

A R T I C L E S
Kliewer et al.
tetramethylammonium bromide, and 1.00 mmol of poly(vinylpyr-
rolidone) (in terms of the repeating unit; M.W. 29,000) were
dissolved into 10 mL of ethylene glycol in a 25 mL round-bottom
flask at room temperature. The mixed solution was heated to 160
°C in an oil bath at 60 °C/min. The solutions were held at these
respective temperatures for 20 min under argon protection and
magnetic stirring, resulting in a dark brown solution. After the
solution was cooled to room temperature, acetone (90 mL) was
then added to form a cloudy black suspension, which was separated
by centrifugation at 3,000 rpm for 10 min. The black product was
collected by discarding the colorless supernatant. The products were
further washed three times by precipitation/dissolution (redispersed
in 20 mL of ethanol and then precipitated by adding 80 mL of
hexanes). The nanocrystals were then redispersed in 10 mL of
ethanol for characterization and catalyst preparation.
2.6. Synthesis of Dendrimer Templated NPs. Dendrimer
encapsulated NPs were synthesized following a previously published
method25 with minor modifications. Generation 4 dendrimers
(G4OH) were purchased from Dendritech Inc. (Midland, MI) as
10.2% (mass) methanol solution and used without further purifica-
tion. A dendrimer stock solution (250 µM) was prepared by adding
the dendrimer methanol solution to water. The dendrimer stock
solution was mixed with 40 molar equiv of an aqueous solution of
0.01 M K2PtCl4 in a 20 mL vial. The vial was purged with Ar for
30 min, tightly sealed with a septum, and allowed 66 h for
complexation. Then, a 20-fold excess of freshly prepared 0.5 M
NaBH4 (stored at 0 °C before use) was injected into the vial
dropwise under vigorous stirring. The reaction solution was stirred
for an additional 8 h. The reaction solution (10 mL) was purified
by dialysis against 2 L of deionized water in cellulose dialysis sacks
with a molecular weight cutoff of 12,000 (Sigma-Aldrich, Inc., St.
Louis, MO). Dialysis occurred over 24 h with the water being
changed 4 times. The average number of metal atoms per
nanoparticle is controlled by the Pt ion to dendrimer ratio. Since
the particles were too small to successfully image with TEM, sizes
were calculated from the average number of metal ions added per
dendrimer, and this method has been verified using X-ray absorp-
tion.29 In this case approximately 40 Pt atoms were added to the
dendrimer (Pt-40), creating a 1.0-nm Pt nanoparticle.
The 1.0-nm Pt nanoparticles were then drop cast onto a fused
silica prism for catalytic and SFG-VS study. The 3.5-nm and 7-nm
nanoparticles were deposited onto the fused silica prism using the
Langmuir-Blodgett technique as described previously.16 The
particles were all UV-ozone treated on the fused silica prism16 to
remove the organic capping layer. The removal of the capping layer
was verified by observing the disappearance of the C-H stretching
vibrations in the SFG-VS spectrum.
2.7. Sum Frequency Generation Vibrational Spectroscopy.
For SFG measurements, an active/passive mode-locked Nd:YAG
laser (Leopard D-20, Continuum) with a pulse width of 20 ps and
a repetition rate of 20 Hz was used. The fundamental output at
1064 nm was sent through an optical parametric generation/
amplification (OPG/OPA) stage where a tunable IR (2300s4000
cm-1) and a second harmonic VIS (532 nm) beam were created.
The IR (150 µJ) and visible (200 µJ) beams were spatially and
temporally overlapped on the crystal surface at angles of
incidence of 55° and 60°, respectively, with respect to the surface
normal for the single crystal studies. In the case of the
nanoparticle studies, the beams were directed into a fused silica
prism on top of which the nanoparticles had been deposited.
The generated SFG beam was collected and sent through a
motorized monochromator equipped with a photomultiplier tube
to detect the SFG signal intensity. The signal-to-noise ratio was
further increased by using a gated integrator while the IR beam was
scanned through the spectral region of interest. The SFG process is
enhanced when the IR beam comes into resonance with a vibrational
Figure 1. Temperature dependence of the selectivity during the hydrogena-
tion of furan over Pt(111) using 10 Torr of furan and 100 Torr of hydrogen
as measured by GC. The products are dihydrofuran (]), tetrahydrofuran
(0), and butanol (4). There is a complete selectivity reversal from 90 to
150 °C from tetrahydrofuran to the ring-cracking product butanol.
mode of a molecule adsorbed at the surface, giving rise to a vibrational
spectrum of adsorbed species. More information on the HP/UHV
system and SFG measurement can be found elsewhere.30-35
3. Results and Discussion
3.1. Furan Hydrogenation over Pt(111): Kinetic and SFG-
VS Results. Furan hydrogenation was carried out over Pt(111)
with 10 Torr of furan, 100 Torr of hydrogen, and 550 Torr of
argon in the temperature range of 22 to 170 °C. Figure 1 displays
the initial reaction product distribution as a function of tem-
perature as measured by GC while the total conversion is less
than 5%. Dihydrofuran, tetrahydrofuran and butanol were all
detectable products. The catalytic activity of the high-pressure
cell and sample holder was measured at all experimental
temperatures used and found to be negligible. As can be seen,
up to 90 °C the dominant product is THF, the saturated ring
product. As the temperature is raised however, we see a
complete reversal of the selectivity and the dominant product
(∼90%) becomes the ring-cracking product butanol. DHF is a
minor product (<4%) at all temperatures. It is interesting to note
that Smith and Fuzek24 report only butanol as the reaction
product over an Adams platinum catalyst while all three products
are observed over Pt(111).
Figure 2 shows the SFG-VS spectrum of 10 Torr of furan
over Pt(111) in the absence of excess hydrogen. Four major
vibrational stretches are observed. The strongest two stretches
are clearly seen at 2845 cm-1 and 2925 cm-1 corresponding to
the CH2 symmetric and asymmetric stretch of the THF ring
respectively. It is possible that DHF on the surface could
contribute to the CH2 stretches as well. A third stretch is seen
at 3145 cm-1 which we can attribute to the aromatic C-H
(30) Bratlie, K. M.; Flores, L. D.; Somorjai, G. A. Surf. Sci. 2005, 599,
93.
(31) Kung, K. Y.; Chen, P.; Wei, F.; Rupprechter, G.; Shen, Y. R.;
Somorjai, G. A. ReV. Sci. Instrum. 2001, 72, 1806.
(32) Yang, M. C.; Tang, D. C.; Somorjai, G. A. ReV. Sci. Instrum. 2003,
74, 4554.
(33) Shen, Y. R. The Principles of Nonlinear Optics; John Wiley & Sons:
Hoboken, NJ, 2003.
(29) Alexeev, O. S.; Siani, A.; Lafaye, G.; Williams, C. T.; Ploehn, H. J.;
(34) Shen, Y. R. Annu. ReV. Phys. Chem. 1989, 40, 327.
(35) Shen, Y. R. Nature 1989, 337, 519.
Amiridis, M. D. J. Phys. Chem. B 2006, 110, 24903.
9
13090 J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010

Furan Hydrogenation over Pt Catalysts
A R T I C L E S
Figure 2. SFG-VS spectrum of 10 Torr of furan over Pt(111) in the absence of excess hydrogen taken at 23 °C. Stretches are seen for the C-H aromatic
stretch of furan, the C-H vinylic stretch of DHF, and the CH2 symmetric and asymmetric stretches of the THF ring. In the molecule representations, oxygen
is red, carbon is gray, and hydrogen is white.
stretch of the furan ring. A thorough description of the
vibrational modes of furan was done by Klots et al.,36 and the
gas phase aromatic C-H stretch of furan arises at 3160 cm-1
.
Thus, a 15 cm-1 red-shift of this mode takes place upon the
adsorption of furan to Pt(111). Lastly, a C-H stretch is observed
at 3040 cm-1 which we assign to the vinylic CdC-H stretch
of the DHF ring. Clearly, some hydrogenation is taking place
on the surface species prior to the addition of excess hydrogen
to the reaction mixture, although no detectable turnover was
measured by GC. This could be due either to the small
background pressure of hydrogen in the high-pressure cell or
to intermolecular hydrogen transfer.
Figure 3 displays the temperature dependent spectra of 10
Torr of furan in the presence of 100 Torr of hydrogen over
Pt(111). The spectra are characterized by three peaks at 2860
cm-1, 2925 cm-1, and 2960 cm-1. The peak at 2860 cm-1 can
be attributed to the CH2(s) stretch of the saturated ring, THF.
The peak at 2960 cm-1 is very clearly indicative of the CH3(a)
stretch of the ring-cracking product, butanol. The peak at 2925
cm-1 red-shifts to 2915 cm-1 upon heating and may be attributed
to a CH2(a) stretch. The CH2(s) stretch seen at 2860 cm-1 is
blue-shifted slightly from standard hydrocarbon CH2(s) stretches,
due to the electron withdrawing effects of the O atom in the
molecule. Further, the dominance of the CH2(s) stretch relative
to the CH2(a) stretch at 23 °C is indicative of a “standing up”
THF species with respect to the metal surface.37 No C-H
aromatic stretch is seen once hydrogen is added to the reaction
mixture, not even at room temperature when the gas phase is
still >99.9% furan. This indicates that the furan ring is lying
parallel to the Pt(111) surface upon the coadsorption of furan
and hydrogen.
The three peaks undergo a general trend made clear by Figure
4 upon heating the crystal. In Figure 4 the spectrum of 10 Torr
of furan and 100 Torr of hydrogen at 107 °C has been subtracted
from the spectrum at 23° to highlight the trend observed as the
crystal is heated to 107 °C. First, a very strong negative peak,
or “loss peak”, at 2860 cm-1 is seen. This occurs concomitantly
Figure 3. SFG-VS spectra of 10 Torr of furan and 100 Torr of hydrogen
over Pt(111) from 23 to 147 °C.
with a gain in intensity at 2915 cm-1 and 2965 cm-1. The gain
observed in the peak at 2965 cm-1 can be attributed to an
increasing surface concentration of butoxy, the ring-cracking
product, with increasing temperature. Clearly, the desorption
of the products is a rate limiting step for this reaction as the
surface builds up an appreciable concentration. SFG-VS is not
a suitable technique to quantify the relative surface coverages.
(36) Klots, T. D.; Chirico, R. D.; Steele, W. V. Spectrochim. Acta, Part A
1994, 50, 765.
(37) Yang, M.; Somorjai, G. A. J. Am. Chem. Soc. 2004, 126, 7698.
9
J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010 13091

A R T I C L E S
Kliewer et al.
Figure 6. Two possible intermediates on Pt(111) during furan hydrogena-
tion at 107 °C and higher. On the left is a parallel THF molecule, and on
the right is an oxametallacycle.
prepared by thermally treating 2-iodoethanol, and resulted in a
39
single band at 2922 cm-1
.
However, in the case of furan
Figure 4. Difference spectrum of 10 Torr of furan and 100 Torr of
hydrogen over Pt(111) at 107 °C as compared with 10 Torr of furan and
100 Torr of hydrogen at 23 °C to clarify the trends observed as the
temperature is raised. A “loss” peak is observed at 2865 cm-1 upon heating
corresponding to the CH2 (s) stretch, while two positive peaks are seen at
hydrogenation over Pt(111), one would expect to see a nearly
equivalent peak intensity associated with the CH2(s) stretch of
the carbon connected to the oxygen (Figure 6), which is not
observed as by 107-127 °C there is significantly more intensity
at the 2915 cm-1 peak than at the 2865 cm-1 peak. Thus, the
oxametallacycle intermediate cannot fully explain the spectrum.
In the case of the orientational change in the THF ring as the
interpretation, the 2915 cm-1 peak would be attributed to the
CH2(a) stretch off of the THF ring. This is further evidenced in
that the stretch red-shifts by 10 cm-1 upon heating, indicating
more interaction with the metal surface. As the ring goes from
more perpendicular to the surface to more parallel to the surface,
we would expect the methylene symmetric stretch to vanish
and the methylene asymmetric stretch to become dominant due
to the metal surface selection rule. This is in agreement with
the spectra.
Finally, as discussed above, the third interpretation is simply
an increasing presence of gauche (GTT) butoxy bound to the
surface as we raise the temperature. The presence of the
asymmetric peak at 2965 cm-1 informs us that this is occurring
on the surface to some extent.
Thus there are several conclusions from the interpretation of
the SFG-VS results during furan hydrogenation on Pt(111). First,
furan lies parallel to the Pt(111) surface during hydrogenation.
Second, at 23 °C in the presence of 10 Torr of furan and 100
Torr of hydrogen we have predominantly an “upright” THF
molecule on the surface. Third, butoxy is bound to the Pt(111)
surface in the gauche (GTT) conformation growing in with
temperature bound though the oxygen atom. Last, there may
exist on the surface either an increasingly “parallel” lying THF
species as the temperature is raised or an oxametallacycle,
neither of which can be ruled out or completely verified by the
acquired spectra in the available spectral range.
2915 cm-1 and 2965 cm-1
.
Figure 5. SFG-VS reference spectrum of 1 Torr of butanol over Pt(111)
at 23 °C. The dominant CH3(s) peak at 2875 cm-1 indicates that butanol
adsorbs upright on the surface.
The presence of the methyl asymmetric stretch and the absence
of the methyl symmetric stretch, which would appear at 2875
cm-1, indicate that the butanol molecule is adsorbed to the
surface with the chain parallel to the surface. Further, there is
no evidence of an O-H bond anywhere in the 3200-4000 cm-1
range which leads us to conclude that butanol is adsorbed to
the surface through an O-Pt bond, forming a butoxy species.
Figure 5 displays the SFG-VS spectrum of 1 Torr of butanol
over Pt(111) as a reference. As can be seen, without the presence
of coadsorbates the spectrum for butanol on Pt(111) is domi-
nated by the methyl symmetric stretch at 2875 cm-1 indicating
that the molecular axis of the methyl must be aligned nearly
perpendicular to the metal surface, and the methyl asymmetric
stretch is very weak. No resonance in the O-H region was
observable. This is in contrast to what is observed for the butoxy
product on the Pt surface during the course of furan hydrogena-
tion, for which the asymmetric methyl stretch is the visible one,
indicating a different adsorption mode in the presence of
coadsorbates such as furan and hydrogen on Pt(111).
3.2. Furan Hydrogenation over Pt(100): Kinetics and SFG-
VS. Figure 7 displays the steady state selectivity for furan
hydrogenation over Pt(100) as a function of temperature for
comparison to Pt(111) so as to help elucidate the effect that
catalyst structure plays in this reaction. As can be seen, at 90
°C a nearly 50/50 product distribution is obtained between
tetrahydrofuran and butanol. When compared to Pt(111) at 90
°C, it is evident that there is significantly more ring-cracking
on the Pt(100) surface at lower temperatures than on the (111)
face. As the temperature is raised to 160 °C, the selectivity
becomes 86% for the cracking product butanol. This is the same
general trend observed over Pt(111) in that butanol becomes
The strong increase of the stretch observed at 2915 cm-1
,
which occurs concomitantly as the 2865 cm-1 vanishes, has
three possible interpretations. It may be due to the THF ring
adsorption going from an “upright” geometry to a more “flat
lying” geometry, due to the formation of an oxametallacycle
intermediate as the ring cracks, or due to a parallel lying butoxy
species on the surface with a gauche rotation (Figure 3). In the
case of the oxametallacycle, the 2915 cm-1 stretch would be
interpreted as the CH2(s) stretch off of the C which is bound to
the Pt surface (Figure 6). This type of shift in the symmetric
stretch has been observed before with RAIRS38 upon the
formation of a C6 metallacycle on Pt(111) and appeared around
∼2910 cm-1. Over Ag(110), a 2 carbon oxametallacycle was
(38) Ilharco, L. M.; Garcia, A. R.; Hargreaves, E. C.; Chesters, M. A. Surf.
Sci. 2000, 459, 115.
(39) Jones, G. S.; Mavrikakis, M.; Barteau, M. A.; Vohs, J. M. J. Am.
Chem. Soc. 1998, 120, 3196.
9
13092 J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010

Furan Hydrogenation over Pt Catalysts
A R T I C L E S
Figure 7. Steady state product distribution (selectivity) as a function of
temperature on Pt(100) under 10 Torr of furan and 100 Torr of hydrogen. The
products are dihydrofuran (]), tetrahydrofuran (0), and butanol (4).
Figure 9. SFG-VS spectra of 10 Torr of furan and 100 Torr of hydrogen
over Pt(100). The structures indicated are upright THF, upright butoxy,
and parallel furan molecules.
implies that upon the adsorption of furan to Pt(100) at room
temperature there exist vertically oriented furan, THF, DHF,
and butoxy on the surface.
Figure 9 displays the spectra of 10 Torr of furan and 100
Torr of hydrogen over the Pt(100) surface as a function of
temperature. Similarly to Pt(111) the aromatic and vinylic
resonances disappear upon the addition of 100 Torr of hydrogen.
This is an indication that under reactive conditions the furan
ring lies parallel to the surface. The absence of the DHF peak
may be due to a parallel orientation, but more likely due to this
part of the reaction having a fast reaction rate constant forward
to THF, i.e. the double bond is quickly hydrogenated. The
methylene and methyl symmetric stretching vibrations are seen
very close together at 2850 cm-1 and 2860 cm-1 at 23 °C,
appearing as one wide peak with a doublet at the top. As the
reaction mixture is heated to 127 °C however, the peaks resolve
a little more and are seen at 2845 cm-1 and 2870 cm-1
respectively. By 147 °C nearly all resonant signal is lost with
just a small peak seen at 2870 cm-1 for the methyl symmetric
stretch, in agreement with an increased surface coverage of
vertical butoxy. The methylene asymmetric stretch is seen at
2930 cm-1 while the methyl asymmetric stretch is not visible.
These spectra indicate that both THF and butoxy are present
on the surface in upright geometries (Figure 9) and as the surface
is heated to 147 °C the adsorbates become less ordered
generating less SFG-VS signal. The lack of a major peak around
∼2910 cm-1 rules out the presence of an oxametallacycle as a
major surface species during furan hydrogenation over Pt(100).
This does not eliminate the oxametallacycle as a transient
intermediate between the ring and ring-cracked structures of
furan and butanol, but if present it must nevertheless be short-
lived enough as to not build up any appreciable surface
concentration.
Figure 8. SFG-VS spectrum of 10 Torr of furan over Pt(100) at 23 °C.
No excess hydrogen was added. Four resonances are seen. At 3145 cm-1
is the C-H aromatic stretch indicating a furan ring which is not parallel to
the surface. At 3040 cm-1 is observed a vinylic CdC-H stretch of the
partially saturated DHF ring. At 2915 cm-1 and 2850 cm-1 the methylene
asymmetric and symmetric stretches are seen respectively. A narrow peak
at 2870 cm-1 is seen which would correspond to the methyl symmetric
stretch of the ring cracking product butanol.
the dominant product at high temperatures. The production of
the partially saturated ring, which may be a desirable intermedi-
ate for synthetic applications, is very low as in the case over
Pt(111). DHF is 8.9% of the product at 90 °C over Pt(100),
and it is 0.9% of the product at 90 °C over Pt(111). Although
it is a minor product, Pt(100) generates 10 times more DHF
than Pt(111) at this temperature.
Figure 8 displays the SFG-VS spectrum of 10 Torr of furan
over Pt(100) in the absence of excess hydrogen. At 3145 cm-1
a strong C-H aromatic stretch is seen for the furan ring, which
is quite different from the very weak C-H aromatic stretch seen
on Pt(111) under the same conditions. This indicates that the
furan ring is significantly tilted upward with respect to the
Pt(100) surface. A vinylic CdC-H stretch is observed at 3040
cm-1 corresponding to the vinylic portion of the partially
saturated ring of dihydrofuran. The methylene symmetric and
asymmetric stretches are observed at 2845 cm-1 and 2915 cm-1
respectively with the symmetric stretch significantly more
dominant. The methyl symmetric and asymmetric stretches are
also observed at 2870 cm-1 and 2965 cm-1 respectively, also
exhibiting much stronger symmetric intensity. This spectrum
3.3. Furan Hydrogenation over Size-Controlled Pt Nano-
particles from 1.0 to 7.0 nm. To elucidate the effect that catalyst
size has on this reaction, the hydrogenation of furan was also
carried out over 1-nm, 3.5-nm and 7-nm Pt nanoparticles.
Figure 10 displays the reaction selectivity as a function of
temperature during furan hydrogenation carried out over the
three Pt nanoparticle sizes which had been UV-ozone cleaned
9
J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010 13093

A R T I C L E S
Kliewer et al.
Figure 11. Temperature-dependent SFG-VS spectra acquired during furan
hydrogenation (10 Torr of furan and 100 Torr of H2) over Pt-40
nanoparticles.
39% for the formation of DHF. Butanol is always present in
the product distribution, with a similar selectivity as that seen
for the single crystals at lower temperatures, but as the
temperature is raised the selectivity for butanol decreases as
the formation of propylene becomes dominant over the three
nanoparticle sizes studied. Propylene was not detected during
furan hydrogenation over Pt(111) or Pt(100) and is unique to
the nanoparticle catalysts. The production of propylene also
indicates the formation of CO over the nanoparticle catalysts,
which may act as a surface poison to the overall TOF over these
catalysts. The GC column combination used to separate the larger
products was not adequate to accurately characterize the CO
production. The three nanoparticle sizes studied exhibited similar
selectivity trends with temperature, as is seen in Figure 10.
Figure 11 displays the SFG-VS spectra of 10 Torr of furan
and 100 Torr of hydrogen over a drop-cast film of the 1.0-nm
Pt nanoparticles deposited onto a fused silica prism. Similarly
to the single crystal results, no aromatic C-H stretch is observed
indicating that the adsorbed furan is parallel to the nanoparticle
surface. Only two vibrational resonances are observed. The first
stretch, corresponding to the methylene symmetric mode of THF
on the surface, is observed at 2865 cm-1 at room temperature
and red-shifts to 2850 cm-1 by 120 °C. Again, the dominant
symmetric stretch in combination with a very weak asymmetric
stretch indicates an upright THF structure on the nanoparticle
surface. The second mode appears at 2905 cm-1, which
corresponds to a highly perturbed CH2(a) stretching mode
consistent with the oxametallacycle intermediate as depicted in
Figure 11.
Figure 10. Product selectivities during furan hydrogenation on (a) 1.0-
nm, (b) 3.5-nm, and (c) 7-nm Pt nanoparticles (10 Torr of furan, 100 Torr
of hydrogen). The particles had been UV-ozone treated to remove the
organic capping layer. The products are propylene (O), dihydrofuran (]),
tetrahydrofuran (0), and butanol (4).
until no C-H modes were present in the SFG-VS spectrum
prior to reaction. A significant difference is seen in the product
selectivities as compared with the single crystals. The partially
saturated ring, DHF, was always a minor product over the single
crystals, with Pt(100) having the highest selectivity of ∼9% at
90 °C. At 80 °C, the Pt-40 nanoparticles have a selectivity of
Although propylene is a significant reaction product over the
Pt nanoparticle catalysts, it is not observed as a surface species
during reaction. Propylidyne is characterized9 by a strong
resonance at 2965 cm-1, not observed in these spectra. The
propyl intermediate exhibits two stretches below 2850 cm-1
,
9
13094 J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010

Furan Hydrogenation over Pt Catalysts
A R T I C L E S
Figure 12. Temperature dependent SFG-VS spectra acquired during furan hydrogenation (10 Torr of furan and 100 Torr of H2) over (a) 7-nm Pt nanoparticles
and (b) 3.5-nm Pt nanoparticles.
and di-σ propylene adsorbed to the Pt surface gives a stretching
mode at 2920 cm-1. A broad peak centered at 3000 cm-1 is
seen for π-bound propylene. As none of these characteristic
features are seen, propylene is not adsorbed to the 1.0-nm Pt
nanoparticle surface during furan hydrogenation indicating that
desorption of propylene is rapid following molecular cracking
on the surface.
Figure 12 displays the temperature-dependent SFG-VS spectra
of furan hydrogenation over (a) 7.0-nm and (b) 3.5-nm Pt
nanoparticles. The same two vibrational resonances are seen as
over the 1.0-nm Pt nanoparticles, and the only difference is a
slightly differing amount of red-shifting for the CH2(a) mode.
This is in agreement with the macroscopic selectivities observed
for the furan hydrogenation reaction over the three nanoparticle
sizes which exhibited very similar trends, much different than
the results over Pt(111) and Pt(100). While Pt(111) and Pt(100)
demonstrate both differing surface species as detected by SFG-
VS and differing selectivities as detected by GC as a function
of temperature, the nanoparticle catalysts all exhibit both similar
surface-bound reaction intermediates and reaction selectivity
trends when compared among the nanoparticle sizes.
as the product selectivities as a function of catalyst temperature.
These results are then compared to those obtained over
UV-ozone cleaned Pt nanoparticles from 1.0 to 7.0 nm in size.
The SFG-VS data indicates that on Pt(111) under furan
hydrogenation conditions the planar furan ring lies parallel to
the metal surface. Butanol is bound to the surface through the
O atom, adsorbed in a gauche (GTT) orientation as the butoxy
intermediate. The surface concentration of butoxy increases with
increasing temperature over Pt(111), and THF is bound to the
surface in an upright binding geometry. This was in agreement
with the selectivity observed with GC over Pt(111) as the
dominant product went from THF to butanol as the temperature
was raised. Over Pt(100) the SFG-VS spectra also indicate a
parallel furan adsorption under hydrogenation conditions. Both
butoxy and THF are seen bound to the Pt(100) surface in upright
geometries. The 1-7-nm Pt nanoparticles had upright THF and
the oxametallacycle intermediate bound to the surface during
reaction observed with SFG, and the nanoparticles exhibited a
tendency toward further molecular cracking to form propylene
in the gas phase observed by GC.
Acknowledgment. This work was supported by the Director,
Office of Energy Research, Office of Basic Energy Sciences, and
Materials Sciences Division of the U.S. Department of Energy under
Contract DE-AC02-05CH11231. M.B. thanks the Swiss National
Science Foundation (SNF) for financial support.
4. Conclusions
In this study, sum frequency generation vibrational spectros-
copy and gas chromatography have been applied to systemati-
cally study the surface species present on the Pt(111) and Pt(100)
single crystal surfaces in situ during furan hydrogenation as well
JA105800Z
9
J. AM. CHEM. SOC. VOL. 132, NO. 37, 2010 13095