Hydrogen transfer reaction on the surface of an oxide catalyst

Article abstract of DOI:10.1021/ja043355h

RuO₂(110) exposes two kinds of active surface species (acidic and basic centers) that govern the interaction of the gas phase in contact with the catalyst's surface. Here we will elucidate the cooperative interplay of these two active surface sites for a simple model reaction, namely the water formation over RuO₂ catalysts when supplying hydrogen and oxygen from the gas phase. The bridging O atoms harvest the hydrogen from the gas phase, while the on-top O atoms pick up those adsorbed hydrogen atoms from the bridging O atoms to form water. This mechanism of hydrogen transfer is mediated by a strong hydrogen bond. Hydrogen transfer is expected to play a vital role for the whole class of catalyzed hydrogenation and dehydrogenation reactions of hydrocarbons over RuO₂. Copyright

Full text of DOI:10.1021/ja043355h

Published on Web 02/18/2005
Hydrogen Transfer Reaction on the Surface of an Oxide Catalyst
Marcus Knapp,† Daniela Crihan,† Ari P. Seitsonen,*,‡ and Herbert Over*,†
Physikalisch Chemisches Institut, Justus-Liebig-UniVersita¨t, Heinrich-Buff-Ring 58, D-35392 Giessen, Germany, and
Physikalisch Chemisches Institut, UniVersita¨t Zu¨rich, Winterthurerstrasse 190, CH-8057 Zu¨rich, Switzerland
Received November 3, 2004; E-mail: Ari.P.Seitsonen@iki.fi; Herbert.Over@phys.chemie.uni-giessen.de
Oxide catalysts are considered to be versatile in heterogeneous
catalysis because of the presence of chemically distinct reactive
centers on their surface, acting as either acidic or basic sites, and
the intimate interplay of these centers in catalyzed reactions.1 For
instance, the model catalyst RuO2(110) exposes two kinds of active
surface species that have shown to govern the interaction of the
gas phase in contact with this surface.2,3 These are the so-called
1f-cus-Ru atoms (1f-cus stands for 1-fold coordinatively unsaturated
sites: acidic sites) and the bridging O atoms Obr (basic sites), both
of which are 1-fold undercoordinated with respect to the Ru and O
atoms in bulk environment (cf. Figure 1).
Most of the molecules studied thus far on the RuO2(110) surface
adsorb from the gas phase initially above the 1f-cus Ru atoms.3
The CO molecule, for instance, adsorbs first on the 1f-cus-Ru site
and then recombines with the bridging O atom to form CO2 above
room temperature.2,4-8 In this oxidation reaction the bridging O
atoms are consumed rather than serve as a reactive center. Quite
in contrast, hydrogen molecules have shown to interact preferen-
tially with the bridging O atoms, forming hydroxyl groups and a
kind of water molecule.9
Here we will elucidate the cooperative interplay of the two active
surface sites on RuO2(110) for a simple model reaction, namely
the water formation over RuO2 catalysts when supplying hydrogen
and oxygen from the gas phase. Oxygen molecules adsorb disso-
ciatively above the 1f-cus Ru atoms, forming on-top O atoms Oot,
while the incoming hydrogen molecules accommodate solely above
the bridging O atoms. Subsequently, the on-top O species picks up
sequentially two hydrogen atoms from the bridging O atomss
hydrogen transfer reactionsthereby forming adsorbed water above
the 1f-cus-Ru atoms.
Figure 1. (Top) Ball and stick model of the RuO2(110) surface under
oxidizing conditions. Oxygen atoms are indicated by large balls, while the
Ru atoms are presented as small balls bridging O ) Obr, on-top O ) Oot,
1-fold undercoordinated Ru site ) 1f-cus Ru. (Bottom) Energy diagram of
the water formation along the reaction path calculated from DFT. Hydrogen
bonds are indicated by dashed lines, and the asterisk marks an unstable
configuration. Without Oot on the surface the binding energy of H2@Obr is
only 0.3 eV. We estimate the error bars in the relative energies to be at
most ca. 0.1 eV.
To study the water reaction on RuO2(110), we exposed the RuO2-
(110) surface to 50 L (1 L ) 1.33 × 10-6 mbar‚s) of D2 at room
temperature and recorded a thermal desorption (TD) spectrum of
D2O using a quadrupole mass spectrometer and a heating rate of
10 K/s. In a second experiment, we exposed the RuO2(110) surface
to 50 L of D2 at room temperature and then saturated the surface
by on-top O. Both D2O TD spectra are shown in Figure 2. Clearly,
the production of D2O is dramatically increased by a factor of 20
when on-top O is present on the surface. These experiments teach
us two lessons: First, D2 adsorbs on the stoichiometric RuO2(110)
surface, barely removing oxygen from the oxide surface, and
second, the presence of on-top O species is mandatory for efficient
water production.
To determine the entrance channel of hydrogen, we reversed the
sequence of the above adsorption experiment. We first saturated
the RuO2(110) surface with on-top O, thereby blocking all the 1f-
cus Ru atoms, and subsequently exposed the surface to 50 L of
D2. If hydrogen enters the surface only via the 1f-cus-Ru atoms,
then blocking these sites by on-top O is to suppress the hydrogen
Figure 2. Temperature-programmed reaction of water D2O by dosing only
50 L of D2 (dotted line) and dosing first 50 L of D2 and then 5 L of O2
(solid line).
adsorption and consequently the water formation. However, the
amount of produced D2O is not affected by the sequence of
adsorption. Therefore D2 adsorbs preferentially through the surface
O atoms rather than via the 1f-cus Ru atoms, challenging recent
DFT-based “predictions”.10
Next we took TD spectra of D2O and O2 (Figure 3), exposing
various D2 doses: in each experimental run we exposed first the
RuO2(110) surface to D2 and saturated subsequently the surface
with on-top O (by dosing 5 L of O2 at room temperature). We
notice that with increasing D2 pre-exposure more D2O is produced
and less on-top O is left on the surface. The integral TD signals of
D2O and O2 (integrated from 300 to 600 K) are complementary in
that the sum of both TD signals is independent of the D2 dose.
From the shape of the TD traces we infer that the O2 desorption
proceeds with second-order kinetics consistent with that of two
neighboring on-top O atoms that recombine to form O2. Quite in
contrast, the water signal follows first-order reaction kinetics with
† Justus-Liebig-Universita¨t.
‡ Universita¨t Zu¨rich.
9
3236
J. AM. CHEM. SOC. 2005, 127, 3236-3237
10.1021/ja043355h CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
determined the energy barrier to be 0.28 eV. This activation barrier
is still quite low compared to the desorption/reaction temperature
of 420 K (cf. Figures 2 and 3). However, the resulting water
molecule adsorbs quite strongly by 0.84 eV over the 1f-cus-Ru
atoms, in agreement with a recent DFT study.10 The strong
adsorption of water is therefore consistent with the desorption
temperature of 420 K and the observed first-order kinetics. The
strong adsorption of water is also supported by a recent X-ray
diffraction study16 and HREELS measurements.17 From our DFT
calculations we infer that hydrogen bonding is important to lower
the activation barriers for the hydrogen transfer reactions, similar
to the well-known Grotthus effect,18 where the fast diffusion of
protons is explained in terms of structure diffusion.
The quintessential point is that the RuO2(110) surface provides
a nice example of a synergy effect in the catalyzed water reaction:
The bridging O atoms harvest the hydrogen from the gas phase,
while the on-top O atoms pick up those adsorbed hydrogen atoms
from the bridging O atoms to form water. The mechanism of
hydrogen transfer is mediated by the strong hydrogen bond. The
hydrogen transfer is expected to play an important role for the whole
class of catalyzed hydrogenation and dehydrogenation reactions of
hydrocarbons over RuO2.19
Figure 3. Predosing various D2 doses (0-500 L) and postdosing 5 L of
O2 to saturate the surface with on-top O. (Top) Desorption of O2 for various
D2 doses. (Inset) Temperature-integrated (300-600 K) O2 TD spectra.
(Bottom) Temperature-programmed D2O reaction for various D2 doses.
(Inset) Temperature-integrated (300-600 K) D2O TPR spectra.
a maximum at 420 K. This reaction order suggests that D2O
desorption is the rate-determining step. At a D2 exposure of 500
L, no O2 leaves the surface below a temperature of 600 K.
The last question we want to settle with mass spectrometry is
whether both surface speciessbridging O and on-top Osare equally
capable in the accommodation of D2. Exposing the stoichiometric
RuO2(110) surface to CO at room-temperature, we replaced all the
bridging O atoms by bridging CO.11 Subsequently, we saturated
the 1f-cus Ru atoms by on-top O atoms (exposure of 5 L of O2 at
200 K) and dosed 500 L of D2. With mass spectrometry we followed
the temperature-dependent production of D2O, CO2, and the O2.
No D2O is detected, and most of the bridging CO molecules
recombine with on-top O to form CO2 identical to the case when
no D2 is post-exposed.8,12 This experiment provides strong evidence
that the on-top O species is hardly able to adsorb D2. This finding
is quite counterintuitive as we would have expected that the on-
top O is much easier to polarize than bridging O atoms and therefore
more prone to adsorbed D2. Again the on-top O species on
RuO2(110) surprises by its inactive behavior.13
With DFT calculations14 we studied the hydrogen transfer
reaction from the bridging O atoms toward the on-top O atoms. In
the first reaction step we started from a configuration where two
hydrogen atoms sit on a single bridging O atom and no hydrogen
atom is adsorbed on the on-top O species. Our DFT calculations
indicate that there is no energy barrier for one hydrogen to shift to
the on-top O site (cf. Figure 1). This means that the first hydrogen
atoms move spontaneously from the bridging to the on-top O
position. The reverse reaction is of course activated by the hydrogen
adsorption energy (i.e., 0.7 eV per hydrogen atom). The final state
of this reaction pathway is characterized by a hydrogen bond
between the bridging O and the H atom now covalently attached
to the on-top O atom.
Acknowledgment. We would like to thank John von Neumann
Institute for Computing for the generous supercomputing time and
the DFG for financial support (SPP 1091).
References
(1) Henrich, V. E.; Cox, P. A. The Surface Science of Metal Oxides;
Cambridge University Press: Cambridge, 1994.
(2) Over, H.; Kim, Y. D.; Seitsonen, A. P.; Wendt, S.; Lundgren, E.; Schmid,
M.; Varga, P.; Morgante, A.; Ertl, G. Science 2000, 287, 1474.
(3) Over, H.; Muhler, M. Prog. Surf. Sci. 2003, 72, 3.
(4) Liu, Z.-P.; Hu, P.; Alavi, A. J. Chem. Phys. 2001, 114, 5957.
(5) Fan, C. Y.; Wang, J.; Jacobi, K.; Ertl, G. J. Chem. Phys. 2001, 114, 10058.
(6) Over, H.; Seitsonen, A. P.; Lundgren, E.; Schmid, M.; Varga, P. J. Am.
Chem. Soc. 2001, 123, 11807.
(7) (a) Seitsonen, A. P.; Over, H. Ruthenium Dioxide, A Versatile Oxidation
Catalyst: First Principles Analysis. In High Performance Computing in
Science and Engineering, Transactions of the First Joint HLRB and
KONWIHR Status and Result Workshop, Munich, Germany, Oct 10-
11, 2002; Wagner, S., Hanke, W., Bode, A., Durst, F., Eds.; Springer-
Verlag: Berlin, 2003; pp 171-187. (b) Ari P. Seitsonen’s Home Page.
http://www.iki.fi/∼apsi/Science/movies/; accessed Aug. 2004.
(8) Wendt, S.; Seitsonen, A. P.; Kim, Y. D.; Knapp, M.; Idriss, H.; Over, H.
Surf. Sci. 2002, 505, 137.
(9) Wang, J.; Fan, C. Y.; Sun, Q.; Reuter, K.; Jacobi, K.; Scheffler, M.; Ertl,
G. Angew. Chem., Int. Ed. 2003, 42, 2151.
(10) Sun, Q.; Reuter, K.; Scheffler, M. Phys. ReV. B 2004, 70, 235402.
(11) Seitsonen, A. P.; Kim, Y. D.; Knapp, M.; Wendt, S.; Over, H. Phys. ReV.
B 2002, 65, 05413.
(12) Knapp, M.; Crihan, D.; Over, H. 2004, unpublished measurements.
(13) Wendt, S.; Knapp, M.; Over, H. J. Am. Chem. Soc. 2004, 126, 1537.
(14) The electronic wave functions were expanded in plane waves up to 36
Ry, and pseudo-potentials were used in the code VASP: (a) Kresse, G.;
Furthmu¨ller, J. Comput. Mater. Sci. 1996, 6, 15. (b) Kresse, G.; Joubert,
D. Phys. ReV. B 1999, 59, 1758. The surface was modeled by five double
layers of RuO2(110), and consecutive slabs were separated by a vacuum
region of 16 Å. The reaction coordinate was defined as the distance
between the reacting H atom and the on-top O atom, and the transition
state of each reaction pathway and the corresponding activation barrier
were searched with a constrained minimization technique.4
The second reaction path modeled was the transfer of the second
hydrogen atom from the bridging O to the on-top O atom.15 In the
initial state of this reaction, both the on-top and the bridging O
atoms carry one hydrogen atom. The minimum energy configuration
is characterized by the OotH group that is inclined toward the next-
row bridging O atom and a second hydrogen bond between the H
atom from the bridging O atom and the on-top O atom. The reaction
coordinate is again the separation of the on-top O species and the
hydrogen atoms sitting on the bridging O atom. DFT calculations
(15) We did not study the reaction path of a second hydrogen atom reacting
from an OotH group with another, neighboring OotH group to form water.
(16) Chu, Y. S.; Lister, T. E.; Cullen, W. G.; You, H.; Nagy, Z. Phys. ReV.
Lett. 2001, 86, 3364.
(17) Lobo, A.; Conrad, H. Surf. Sci. 2003, 523, 279.
(18) Benoit, M.; Marx, D.; Parrinello, M. Nature 1998, 293, 258.
(19) Zhan, B.-Z.; White, M. A.; Sham, T.-K.; Pincock, J. A.; Doucet, R. J.;
Rao, K. V. R.; Robertson, K. N.; Cameron, T. S. J. Am. Chem. Soc. 2003,
125, 2195.
JA043355H
9
J. AM. CHEM. SOC. VOL. 127, NO. 10, 2005 3237