Kinetics and mechanism of catalytic partial oxidation reactions of alkanes on rhodium surfaces

Article abstract of DOI:10.1021/ja807475g

The kinetics of the partial oxidation of isobutane with molecular oxygen on Rh(111) single-crystal surfaces were studied by using a collimated molecular beam under ultrahigh vacuum conditions. Both hydrogen and water were shown to form as primary products, not after secondary reforming or water-gas shift steps as it has been suggested in the past. The production of carbon monoxide (but not of carbon dioxide) was also detected. Water production reaches its steady-state rate in a slower fashion than the rest of the products, presumably because of the kinetics of formation and consumption of the hydroxo surface intermediate involved. Copyright

Full text of DOI:10.1021/ja807475g

Published on Web 11/01/2008
Kinetics and Mechanism of Catalytic Partial Oxidation Reactions of Alkanes
on Rhodium Surfaces
Jarod N. Wilson, Ryan A. Pedigo, and Francisco Zaera*
Department of Chemistry, UniVersity of California, RiVerside, California 92521
Received September 19, 2008; E-mail: zaera@ucr.edu
Typically, the oxidation of alkanes leads to the production of
carbon dioxide and water, the most thermodynamically favored
products. Nevertheless, it has been shown that, by using a metal
catalysts and relatively high (∼1000 K) temperatures, it may be
possible to control that oxidation to produce a more desirable carbon
1-3
monoxide + hydrogen (syngas) mixture.
In fact, processes with
short contact times between the catalyst and the reaction mixture,
which may be achieved by using high space velocities, short catalyst
beds, and/or new reactor designs, have been shown to enhance
3
-9
hydrogen production.
However, optimization of the selectivity
of these processes has been hampered by the limited understanding
presently available on the kinetics of the reactions involved. Two
main mechanisms have been proposed, a direct partial oxidation
to H
2
and CO and the total combustion to H
2
0-15
O and CO
and the resolution
2
followed
1
by reforming back to the desired syngas,
between the two is still ongoing. The fact that oxidations are fast
exothermic reactions, together with heat and mass transport issues,
has hindered the extraction of elementary step information from
Figure 1. Kinetic data for the conversion of a 2:1 isobutane/oxygen mixture
on a Rh(111) surface at 940 K. Shown are the rates for the uptake of both
oxygen and the alkane as well as for the production of hydrogen, water,
carbon monoxide, and carbon dioxide, both the experimental data (dots)
and the results from kinetic modeling (solid lines). All but CO are produced
2
in this process, even if the evolution of water reaches its final steady-state
rate slowly over time.
15,16
experiments.
Here we report on studies using molecular beams
on single crystals aimed at obtaining direct information on the
kinetics of these surface reactions. Interestingly, both hydrogen and
water (but not carbon dioxide) were found to be products of primary
reactions on rhodium substrates. The selectivity between the two
was found to depend on the conditions used to carry out the reaction.
The experiments reported herein were carried out in an ultrahigh
the rate of which raises more slowly, over a period of ∼20 s. It
should be emphasized that since these experiments were carried
out under vacuum and by using a low-surface-area single-crystal
sample, they are void of any mass or heat transfer problems and
directly reflect the kinetics of the reactions occurring upon a single
collision of the reactants with the surface.
A simple mechanism was used to help interpret the kinetic data
2
involving the rapid dissociative adsorption of both O and the
alkane, the reversible recombination of surface atomic oxygen back
to the gas phase, the rapid recombination of surface carbon with
oxygen to produce CO, and the stepwise formation of H O via a
2
hydroxo surface intermediate. The results from this modeling are
reported as solid lines in Figure 1, with more details provided in
the Supporting Information. This model is quite simplistic and
cannot reproduce all the data, in particular some of the transient
behavior, but does highlight a few of the key features of the alkane
partial oxidation reaction. For one, it explains the sharp transient
2
in the uptake of O as the consequence of a buildup of a steady-
state concentration of atomic oxygen on the surface before reaching
equilibrium with the reverse oxygen recombination. In fact, a similar
1
7-20
vacuum (UHV) apparatus described previously.
A Rh(111)
single crystal was mounted on a manipulator capable of liquid-
nitrogen cooling and resistive heating to any temperature between
1
00 and 1200 K, as set by a homemade temperature controller and
monitored by a chromel-alumel thermocouple spot-welded to the
back side of the crystal. The surface was cleaned before each
+
-7
experiment by sequential Ar sputtering, heating in 1 × 10 Torr
of O , and annealing at 1200 K. The kinetic experiments were
2
carried out by following a variation of the so-called King and Wells
method where the front surface of the crystal is exposed to an
effusive molecular beam of the desired premixed alkane + oxygen
1
8,19
gas generated by a multichannel microcapillary array doser.
The rates of evolution of all reactants and products are followed
versus time by mass spectrometry while maintaining the crystal at
the desired temperature. Conversion of the raw signals to rates is
17,20
carried out by following established procedures.
Liquid Carbonic 99.999%), O (Nellcor Purtain Bernett, 99.5%),
and i-C 10 (Matheson, >99.9%), were all used as provided.
Figure 1 displays typical results from these studies, in this case
for a 2:1 i-C mixture impinging on the Rh(111) surface while
holding its temperature constant at 940 K (total gas flux ) 1.2 ML
The gases, Ar
(
2
4
H
4
H10/O
2
spike is seen in kinetic studies with O
shown). This uptake has already been well documented.
2
beams alone (data not
2
1,22
-1
s ). A number of observations derive from these data, in particular:
1) the production of H , H O, and CO, but not of CO ; (2) a sharp
transient within the first 2-3 s of reaction characterized by an excess
uptake of O and a spike in the production of H ; and (3) a fast
attainment of a steady-state condition by all compounds but water,
Our kinetic mechanism also explains the slow initial production
of water. That is simply the reflection of the kinetics of formation
and subsequent conversion of hydroxo intermediates on the surface.
In this model, the rate of OH(ads) formation from recombination
of oxygen and hydrogen surface atoms defines the ultimate yield
(
2
2
2
2
2
1
5796 9 J. AM. CHEM. SOC. 2008, 130, 15796–15797
10.1021/ja807475g CCC: $40.75  2008 American Chemical Society

C O M M U N I C A T I O N S
In summary, our results demonstrate the primary nature of the
steps responsible for the formation of both H and H O in the partial
2 2
oxidation of alkanes on rhodium surfaces. Here we report on results
for isobutane, but similar qualitative behavior has also been seen
in preliminary experiments with methane, propane, n-butane, and
n-hexane. It was found that although water production stems directly
from the conversion of the alkane with oxygen on the rhodium
surface, it proceeds via the formation of a hydroxo surface
intermediate and can be minimized at higher temperatures. It was
also noted that under no circumstances carbon dioxide production
was detected in these molecular beam experiments. This is due to
the fact that carbon monoxide desorbs rapidly after formation, before
it has time to react with surface oxygen; that behavior has already
Figure 2. Half-life time for water to reach its steady state (t1/2(H
2
O), left),
been shown in molecular beam experiments on clean and oxygen-
fraction of water produced (S(H O)dR(H O)/[R(H )], center), and
2
2
2
O)+R(H
2
19,25
precovered Rh(111) surfaces.
2
CO formation has been seen on
steady-state oxygen surface coverage (Θ(O), right), all estimated from
26
Pt(111) under similar molecular-beam conditions, suggesting that
it could also occur in catalytic alkane oxidation processes with real
catalysts containing that metal. It is hoped that the insight into the
kinetics of partial oxidation reactions reported here will be
incorporated into more complete mechanistic models to settle the
questions remaining on these systems and to be able to design better
processes for the direct production of hydrogen from natural gas
and other hydrocarbon sources.
isothermal kinetic experiments such as that in Figure 1 carried out at different
temperatures and gas compositions. A general increase in t1/2(H
decreases in S(H
2
O) and
2
O) and Θ(O) are seen as the reaction temperature is
increased. The changes induced by variations in the gas composition are
subtler, but a clear proportionality is seen between the Θ(O) and the flux
of O at low temperatures.
2
of water, whereas the rate of reaction of those OH intermediates
with additional hydrogen controls the time it takes for water
production to reach its steady state. In the simulations the rates of
both steps are typically comparable in magnitude, so a significant
Acknowledgment. Funding for this work was provided by the
U.S. National Science Foundation.
(
∼0.1 ML) coverage of OH(ads) builds up on the surface. The role
23,24
Supporting Information Available: Details of the mechanism used
to simulate the kinetic data in Figure 1 and kinetic parameters from
simulations of the runs in Figure 2. This material is available free of
charge via the Internet at http://pubs.acs.org.
of OH surface groups has certainly been discussed in the past,
but our results highlight the fact that they are intermediates in the
network of the primary reactions that take place after adsorption
2
of O and the alkane.
Further details of the kinetics of these alkane partial oxidation
reactions were obtained by carrying out experiments at different
surface temperatures and with various gas compositions and fluxes.
Some of the resulting data, specifically the half-life time for water
References
(1) Prettre, M.; Eichner, C.; Perrin, M. Trans. Faraday Soc. 1946, 42, 335.
(
2) Ashcroft, A. T.; Cheetham, A. K.; Foord, J. S.; Green, M. L. H.; Grey,
C. P.; Murrell, A. J.; Vernon, P. D. F. Nature 1990, 344, 319.
(3) Hickman, D. A.; Schmidt, L. D. Science 1993, 259, 343.
(
(
4) Goralski, C. T.; Schmidt, L. D. Catal. Lett. 1996, 42, 47.
5) Fathi, M.; Hofstad, K. H.; Sperle, T.; Rokstad, O. A.; Holmen, A. Catal.
Today 1998, 42, 205.
6) Williams, K. A.; Leclerc, C. A.; Schmidt, L. D. AIChE J. 2005, 51, 247.
7) Roychoudhury, S.; Castaldi, M.; Lyubovsky, M.; LaPierre, R.; Ahmed, S.
J. Power Sources 2005, 152, 75.
8) Neumann, D.; Veser, G. AIChE J. 2005, 51, 210.
9) Burke, N. R.; Trimm, D. L. Catal. Today 2006, 117, 248.
production to reach its steady state (t1/2(H
water production over hydrogen (S(H O)), and the steady-state
coverage of atomic oxygen on the surface (Θ ), are summarized
in Figure 2. Although Θ , estimated by integration of the initial
spike in the uptake of O , clearly scales with the flux of oxygen in
2
O)), the selectivity for
2
(
(
O
O
(
(
2
the beam, especially at the lower temperatures, the other two
parameters do not exhibit a strong dependence on gas composition
or total flux. The main effects are seen as a function of reaction
temperature. It is seen that the steady-state coverage of oxygen
decreases with temperature, a straight consequence of the fact that
(10) Hickman, D. A.; Schmidt, L. D. AIChE J. 1993, 39, 1164.
(
11) Hu, Y. H.; Ruckenstein, E. AdV. Catal. 2004, 48, 297.
12) Basini, L. Catal. Today 2006, 117, 384.
(
(13) Horn, R.; Williams, K. A.; Degenstein, N. J.; Schmidt, L. D. J. Catal.
2
006, 242, 92.
(
14) York, A. P. E.; Xiao, T. C.; Green, M. L. H.; Claridge, J. B. Catal. ReV.
Sci. Eng. 2007, 49, 511.
(
(
15) Enger, B. C.; Lødeng, R.; Holmen, A. Appl. Catal., A 2008, 346, 1.
16) Bitsch-Larsen, A.; Horn, R.; Schmidt, L. D. Appl. Catal., A 2008, 348,
165.
atomic oxygen recombination to O
dissociative adsorption of O is not. The selectivity for water
production also decreases with increasing temperature, suggesting
that the activation energy for hydrogen recombination to H may
be larger than that for OH(ads) formation. Perhaps more surprising
is the increase in t1/2(H O) with temperature. This points to a more
2
is activated whereas the reverse
2
(
(
17) Liu, J.; Xu, M.; Nordmeyer, T.; Zaera, F. J. Phys. Chem. 1995, 99, 6167.
18) Zaera, F.; Liu, J.; Xu, M. J. Chem. Phys. 1997, 106, 4204.
2
(19) Gopinath, C. S.; Zaera, F. J. Catal. 1999, 186, 387.
(
20) Wilson, J.; Guo, H.; Morales, R.; Podgornov, E.; Lee, I.; Zaera, F. Phys.
Chem. Chem. Phys. 2007, 9, 3830.
2
(21) Peterlinz, K. A.; Sibener, S. J. J. Phys. Chem. 1995, 99, 2817.
22) Ganduglia-Pirovano, M. V.; Reuter, K.; Scheffler, M. Phys. ReV. B 2002,
(
significant contribution of the reverse OH(ads) dissociation step in
the overall kinetics at higher temperatures, an effect not included
in our model. Also missing from our model is an explanation for
the spikes in hydrogen production seen within the first couple of
seconds of the reaction (see above), which may be due to an effect
exerted by coadsorbed oxygen on the rate of hydrogen recombination.
6
5, 2454261.
(23) Mhadeshwar, A. B.; Vlachos, D. G. J. Phys. Chem. B 2005, 109, 16819.
24) Dalle Nogare, D.; Degenstein, N. J.; Horn, R.; Canu, P.; Schmidt, L. D. J.
(
Catal. 2008, 258, 131.
(25) Gopinath, C. S.; Zaera, F. J. Catal. 2001, 200, 270.
26) Liu, J.; Xu, M.; Zaera, F. Catal. Lett. 1996, 37, 9.
(
JA807475G
J. AM. CHEM. SOC. 9 VOL. 130, NO. 47, 2008 15797