Oxidation of methane over a Rh/Al2O3 catalyst using microfabricated reactors with integrated heating

Article abstract of DOI:10.1016/j.jcat.2006.04.014

Microfabricated reactors with integrated local heating were fabricated for investigating fast exothermic reactions with independent control of temperature and flow rate. As an example, oxidation of methane over a supported rhodium catalyst was investigate

Full text of DOI:10.1016/j.jcat.2006.04.014

Journal of Catalysis 241 (2006) 74–82
www.elsevier.com/locate/jcat
Oxidation of methane over a Rh/Al2O3 catalyst using microfabricated
reactors with integrated heating
Osnat Younes-Metzler a,1, Johnny Johansen b, Sune Thorsteinsson c, Søren Jensen c, Ole Hansen c,
Ulrich J. Quaade a,∗
a
Danish Research Foundation’s Center for Individual Nanoparticle Functionality (CINF), NanoDTU, Department of Physics, Technical University of Denmark,
2800 Kgs. Lyngby, Denmark
b
Department of Chemical Engineering, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark
c
Danish Research Foundation’s Center for Individual Nanoparticle Functionality (CINF), NanoDTU, MIC – Department of Micro and Nanotechnology, Technical
University of Denmark, 2800 Kgs. Lyngby, Denmark
Received 22 December 2005; revised 7 April 2006; accepted 15 April 2006
Available online 30 May 2006
Abstract
Microfabricated reactors with integrated local heating were fabricated for investigating fast exothermic reactions with independent control of
temperature and flow rate. As an example, oxidation of methane over a supported rhodium catalyst was investigated. The reaction was studied
at different oxygen-to-methane ratios, covering both fuel-lean and fuel-rich compositions with full and partial oxidation. Good agreement with
previously reported results indicates that the microfabricated reactors are useful for studying fast exothermic reactions.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Microreactors; Partial oxidation; Methane
1. Introduction
experimental setup. Together, these advantages make the micro-
fabricated reactors ideal for studies of fast exothermic reactions,
where it is difficult to vary, for example, temperature and flow
rate independently due to the extensive self-heating from the re-
actions and where safety is an important issue if reactants are
not diluted in inert gases.
In a recent paper, we described a high-temperature microre-
actor setup for studying partial oxidation reactions using micro-
fabricated silicon reactors [5]. In the present paper, we further
develop this experimental setup, now including an integrated
heating element, which enables local heating of the catalyst
bed and provides even better temperature control of the reac-
tion zone.
To test the performance of the reactor, both partial and full
oxidation of methane over 0.1 wt% Rh/Al2O3 catalysts under
fuel-rich and fuel-lean conditions are studied in the microfabri-
cated reactor, with emphasis on partial oxidation. This reaction
system is chosen because the mixture of oxygen and methane is
potentially explosive, the reactions are fast and exothermic (de-
veloping a lot of heat), and the system itself is interesting for
production of synthesis gas [6].
Microfabricated reactors [1–4] have numerous advantages
over more traditional reactors. These advantages are direct con-
sequences of the small dimensions of the reactor. For example,
the corresponding large surface-to-volume ratio implies supe-
rior heat dissipation and control of temperature, and the small
channel dimensions ensure a low Reynolds number with well-
defined laminar flow for realistic flow rates. The small size also
enhances quenching of free radicals, which, together with the
good heat dissipation, effectively suppresses gas-phase reac-
tions in flammable or explosive gas mixtures. Even if an ex-
plosion should occur in the reactor, the small volume will make
it harmless. Further, the small reactors respond quickly to vari-
ations in reaction conditions, increasing the throughput in the
*
Corresponding author.
E-mail addresses: osnat@nano.ku.dk (O. Younes-Metzler),
quaade@fysik.dtu.dk (U.J. Quaade).
1
Current address: Nano-Science Center, University of Copenhagen, Univer-
sitetsparken 5, DK-2100 Copenhagen, Denmark.
0021-9517/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.04.014

O. Younes-Metzler et al. / Journal of Catalysis 241 (2006) 74–82
75
The production of synthesis gas (syngas) is a step in the
indirect conversion of methane to liquid fuel. The conversion
of methane to liquid fuel is of great importance in the utiliza-
tion of natural gas resources, and both direct and indirect routes
have been suggested. The indirect routes require conversion of
methane to syngas, which can then be converted to liquid fuel
by a Fischer–Tropsch reaction [7] or methanol synthesis [8,9].
Conversion of methane to syngas through catalytic partial ox-
idation is mildly exothermic (CH4 + (1/2)O2 → 2H2 + CO,
ꢀH = −36 kJ/mol), producing syngas with an H2/CO ra-
tio (ꢀ2) suitable for both the Fischer–Tropsch reaction and
methanol synthesis. The reaction is fast, with full conversion
measured even at millisecond contact times. This causes high
heat generation in the reactor at high flow rates, thus making it
difficult to study the reaction’s dependence on temperature and
flow rate independently. For the partial oxidation reaction, sup-
ported noble metal catalysts, including Rh, Ru, Pd, and Pt [10–
12], as well as supported Ni catalysts, have been used [13,14].
Among these catalysts, rhodium is considered the most reac-
tive, with a high and stable syngas yield [11].
The literature includes numerous experimental studies
[15–20] as well as model development and comparison be-
tween model and experimental data [21–26]. Various reaction
pathways have been suggested [27–30], and the studies have
revealed a good understanding of the process with respect to
ignition behavior, selectivity, influence of contact times and
gas composition, and other factors. In microstructured reactors,
partial oxidation of methane has been performed in a metallic
multichannel reactor made entirely of rhodium [31], and par-
tial oxidation of propane has been performed in multichannel
metallic reactors impregnated with porous Rh-catalysts [32,33].
In the present work, many of the previously obtained re-
sults are reproduced to confirm the function of the microreactor.
The reaction rate is measured as a function of the fuel-to-air
ratio, and the ignition temperature (at which the partial oxida-
tion reaction starts) is identified for each fuel-to-air ratio. For
full oxidation, the apparent activation energy is determined. At
temperatures above the ignition temperature, where partial ox-
idation contributes to the reaction, the gas composition is lim-
ited mostly by equilibrium, and the measured gas compositions
are compared with calculated equilibrium compositions. To in-
vestigate the transition between kinetic limitation and equilib-
rium limitation, measurements at varying contact times are per-
formed.
Fig. 1. Microreactor layout: (a) front, (b) back. The integrated heating element
is indicated by 1 and the integrated temperature sensor is indicated by 2.
This has been demonstrated to be a safe way to study the reac-
tion [5]. The reactor is sealed with a Pyrex glass lid. The heating
element is located on the back of the microreactor (Fig. 1b),
along with access ports for the inlet and outlet channels.
Using an integrated heating element significantly reduces
the system’s response time compared with earlier versions, in
which the heater was in the form of a 525 µm thick piece of
silicon located on top of the Pyrex lid. In the current reactors,
the integrated heating element is a thin-film conductor located
directly underneath the catalyst bed with only 20–50 µm of sil-
icon and 200 nm of SiO2 separating them. The time constant
for heat to propagate through such a layer is on the order of
microseconds. The low mass of the integrated heating element
itself reduces the amount of heat stored in the system and thus
reduces the heating and cooling time constants. Heating and
cooling rates exceeding 100 ◦C/s can be easily obtained. The
heating element comprises a 500 nm thick layer of nickel di-
silicide (NiSi2), which is compatible with temperatures up to
960 ◦C.
2. Experimental
The microreactors are fabricated using silicon micromachin-
ing methods. The heaters are realized by reacting a lift-off struc-
tured nickel layer with an underlying polycrystalline silicon
layer by annealing them at 850 ◦C. Thereafter, the remaining
silicon is removed by KOH wet etching, ensuring lateral elec-
trical insulation at elevated temperatures, at which silicon be-
comes conducting. The finished heater structures are insulated
from the silicon substrate by a previously grown SiO2 layer.
After the heaters are fabricated, the access ports and channel
structures are etched from the back and front of the substrate,
respectively, using deep reactive ion etching. Finally, after cata-
The microreactor used in the present study is similar to that
used in a previous study [5], with the important difference that
the present reactor contains an integrated electrical heating ele-
ment, which significantly reduces the thermal response time of
the system. The microreactor is shown in Fig. 1. The gas chan-
nels, located on the front of the microreactor (Fig. 1a), consist
of two inlet channels, a reaction chamber in which a catalyst
is placed, and one outlet channel. The two inlet channels make
on-chip mixing of oxygen and methane possible, thereby min-
imizing the volume of the potentially explosive gas mixture.

76
O. Younes-Metzler et al. / Journal of Catalysis 241 (2006) 74–82
lyst deposition in the reactor chamber, a Pyrex lid is anodically
bonded to the reactor to seal the channel system.
a residence time of approximately 60 ms. The oxidation of
methane is studied in the temperature range of 300–700 ◦C. The
temperature is increased manually in steps of approximately
15 ◦C by increasing the current in the heating element. After
steady state is reached, the temperature and gas composition are
recorded. Because of the rapid response of the complete system,
steady state is reached in about 1 min, and the complete mea-
surement is carried out in about 20 min.
The catalyst deposition procedure is identical to the sol–gel
technique described previously [5]. First, alumina gel is de-
posited in the reactor channel while the reactor temperature is
maintained at 80 ◦C. When the desired amount of material is
deposited, the reactor is calcined at 400 ◦C for 30 min. The re-
sulting porous alumina (usually 130 m2/g) is then impregnated
with active material—in this case rhodium-acetyleacetonate—
to obtain Rh(0.1 wt%)/Al2O3.
3. Equilibrium calculations
The temperature is measured from the Pyrex side using an IR
sensor focused to a spot size of 1 mm on the reactor. Because
of its small thickness, the Pyrex lid is somewhat transparent to
IR radiation, meaning that the measured temperature is an av-
erage over the lid and catalyst bed temperatures. This gives the
possibility of detecting step changes of the temperature of the
catalyst surface associated with ignition/extinction of the reac-
tion. The IR sensor is calibrated by making simultaneous IR and
thermocouple recordings using a thermocouple inserted into the
reactor channel of a reactor in which one end has been bro-
ken off. The absolute accuracy is estimated to 20 ◦C, and the
sensitivity is 0.2 ◦C. Measurements performed with a thermo-
camera show that the temperature is uniform within 1 ◦C over
>90% of the reaction zone and does not vary more than a few
percentage points along the complete reaction zone. Because of
the local heating of the reaction zone and the efficient heat dis-
sipation, very little heat propagates to other parts of the chip,
resulting in temperatures below 200 ◦C in the inlet and outlet
access port regions over the entire range of working tempera-
tures. This enables the use of Viton O-rings for the seal between
the microreactor and the interfacing, which is a significant ad-
vantage over the previous use of gold O-rings [5]. The quality
of the seal is improved, and the handling is much easier when
using Viton O-rings.
The gas flow through the reactor is controlled by mass flow
controllers, and gas analysis is done with a mass spectrome-
ter. Argon is added to the gas mixture to keep the total flow
through the microreactor constant. Argon is also used as an
internal standard. This is especially important for the partial
oxidation reaction, in which the reaction generates an increase
in the total number of moles. Calibration factors for all gases
are measured separately. With the experimental setup, it is not
possible to measure the water signal reliably, because the max-
imum temperature in the chamber of the mass spectrometer is
limited to around 70 ◦C. At this temperature, water condenses
at the chamber surfaces, causing an error in the measured water
concentration; therefore, this is not used.
Once the ignition temperature is reached, all of the oxygen
is consumed in the reactions. At this point, the reaction changes
from being kinetically limited to a state where it can be limited
by both equilibrium and kinetics. Assuming that the reaction
approaches equilibrium, changes in concentrations with tem-
perature correspond to an altered equilibrium composition. To
test this, the equilibrium composition of the reaction gas has
been calculated as function of temperature and compared with
the measured concentrations.
The equilibrium calculations have been done by the method
developed by Michelsen [35], which minimizes the total Gibbs
free energy of a multicomponent mulitiphase system at con-
stant pressure and temperature. This method has been modified
so that it includes fugacity coefficients to account for nonideal
behavior. The thermodynamic data needed for the calculations
are obtained from published data sources [36,37].
4. Results and discussion
Fig. 2 shows a typical methane oxidation experiment to-
gether with the calculated equilibrium composition. The ex-
periment is performed under fuel-rich conditions (Φ = 0.7).
At low temperatures, the methane is converted mainly to CO2
and H2O. As the temperature of the reactor reaches a critical
value (in this case 590 ◦C), the H2 and CO signals suddenly
appear, and the CO2 signal decreases. This critical tempera-
ture is known as the ignition temperature of the reaction [34].
This demonstrates that at low temperatures, only the full oxida-
tion reaction occurs, whereas at high temperatures, both the full
and partial oxidation reactions occur. The ignition of the partial
oxidation reaction is also seen as a sharp increase in the con-
version of methane and the total consumption of oxygen. With
a further temperature increases, the conversion of methane and
the concentrations of the partial oxidation products CO and H2
increase, and the concentration of CO2 decreases. Similar be-
havior has been observed in an earlier study [19].
The oxidation of methane is studied under both fuel-rich
and fuel-lean conditions. Instead of using the conventional stoi-
chiometric ratio, φ = %CH4/%O2, we use the modified stoi-
chiometric ratio, Φ = φ/(φ + 1) [34]. The modified ratio has
the advantage of being limited to the range 0–1 and symmetric
around the stoichiometric point of 0.5. Throughout this paper,
the term “stoichiometric ratio” refers to this modified stoichio-
metric ratio.
Because all of the oxygen has reacted at temperatures above
the ignition temperature, the partial oxidation reaction might
be limited by the thermodynamic equilibrium, rather than by
surface kinetics or mass transport. Increasing the temperature
further after ignition then shifts the equilibrium concentrations,
resulting in increased methane conversion. The calculated equi-
librium concentrations of the different gas components in the
outlet gas are represented by the solid lines in Fig. 2. The cal-
culations fit well with the measured concentrations and show
that in this case, the reaction proceeds close to equilibrium.
The catalyst used in all experiments is 0.1 wt% Rh/Al2O3.
The total gas flow in all experiments is 4.5 mL/min, giving

O. Younes-Metzler et al. / Journal of Catalysis 241 (2006) 74–82
77
Fig. 2. Outlet molar flow of the different gas species as a function of temperature. Experimental data are displayed as symbols and solid lines show calculated
equilibrium data.
Fig. 3 shows the conversion of methane as a function of tem-
perature for different stoichiometric ratios. It is clearly seen that
the conversion of methane increases with increasing temper-
ature for all methane-to-oxygen ratios. Under fuel-lean con-
ditions, the full oxidation reaction is the main reaction, and
no ignition is seen in the studied temperature range, except
when Φ = 0.45. Under fuel-lean conditions, the conversion of
methane depends only slightly on Φ and reaches about 35%
conversion at around 625 ◦C.
Under fuel-rich conditions below the ignition temperature,
the methane conversion depends only weakly on Φ, as in the
fuel-lean case. But above the ignition temperature, methane
conversion is strongly dependent on Φ, as was also observed
previously [17]. This finding is expected, because the reaction
is limited by oxygen, and less oxygen in the inlet gas will result
in a lower conversion rate of methane. To illustrate this, Fig. 4
shows the methane conversion as a function of Φ for four dif-
ferent temperatures along with the corresponding equilibrium
calculations. The measured data and calculated curves closely
agree and demonstrate an increased equilibrium conversion of
methane with increasing temperature and decreasing stoichio-
metric ratio Φ.
to-oxygen ratio. Similar observations have been reported by
Veser et al. [34], who explained this decrease in ignition tem-
perature as a result of competitive adsorption of methane and
oxygen. At low temperatures, the surface of the catalyst is al-
most completely covered with oxygen. As the temperature of
the catalyst is increased, desorption of oxygen starts, thereby
creating free sites for methane to adsorb and allowing surface
reaction to take place. As the partial pressure of oxygen de-
creases with increasing Φ, the critical CH4/O2 surface ratio
needed for surface ignition can be reached at lower tempera-
tures. A simplified model describing competitive adsorption of
hydrocarbon species and oxygen has been analyzed, giving the
following relation between the stoichiometry and temperature
at ignition [34]:
Φ = φ/(1 + φ), φ = aT 2 exp(Eapp/RT ),
where T is the ignition temperature, Eapp is the apparent acti-
vation energy, R is the gas constant, and a is a proportional-
ity factor. Fig. 5 plots this expression for a = 4 × 10−11 K−2
and Eapp = 80 kJ/mol and shows that it fits the data well. For
comparison, a value of Eapp ≈ 110 kJ/mol has been found for
methane oxidation on platinum [34].
Another feature illustrated in Fig. 3 is the variation of igni-
tion temperature with Φ. Fig. 5 provides a plot of the ignition
temperature as function of Φ. The ignition temperatures are in
the range of 450–650 ◦C and decrease with increasing methane-
Arrhenius-type plots are created for different fuel-to-oxygen
ratios based on the total conversion of methane as a function of
temperature. Under fuel-lean conditions, a single linear regime
is observed for each Φ. Under fuel-rich conditions, two linear

78
O. Younes-Metzler et al. / Journal of Catalysis 241 (2006) 74–82
Fig. 3. The measured conversion of methane as a function of the temperature for different stoichiometric ratios Φ.
Fig. 4. The conversion of methane as a function of the stoichiometric ratio Φ at different temperatures. Experimental data are displayed as symbols and solid curves
show calculated equilibrium data.
regions are observed, as shown in Fig. 6: one region at tem-
peratures below ignition, corresponding to the full oxidation
reaction, and a second region at temperatures above ignition,
corresponding to the partial oxidation reaction.
In the case of full oxidation, the reaction is kinetically lim-
ited, and the slope can be interpreted as apparent activation
energy. Apparent activation energies determined for the differ-
ent methane-to-oxygen ratios are shown in Fig. 7A. Although

O. Younes-Metzler et al. / Journal of Catalysis 241 (2006) 74–82
79
Fig. 5. The ignition temperature as a function of the stoichiometric ratio Φ. Symbols are experimental data and the solid curve is a model adapted from [34].
Fig. 6. Arrhenius-type plot for the stoichiometric ratio Φ = 0.7. Two linear regions are observed. At temperatures below ignition the slope reflects an apparent
activation energy E . Above ignition the slope determines the temperature coefficient c in the temperature dependent equilibrium.
A
Φ
the data are scattered, the apparent activation energy seems to
decrease with increasing Φ, from EA = 65 kJ/mol at Φ = 0.2
to EA = 45 kJ/mol at Φ = 0.85. Lower oxygen content in the
inlet gas lowers the activation energy for full oxidation.
In the case of partial oxidation, the conversion of oxygen is
almost complete, and the process might become limited by ther-
modynamic equilibrium rather than by surface kinetics or by a
combination of the two factors. Therefore, this linear part of the
Arrhenius plot cannot be interpreted as activation energy, but
is related to the overall temperature-dependent equilibrium of
the different reactions occurring at temperatures above ignition.
To test whether this linear part can be assigned to equilibrium
alone, a comparison is made to the calculated equilibrium; the
appearance of straight lines for the methane conversion in an
Arrhenius-like plot in the equilibrium-limited temperature re-
gion above ignition indicates that the equilibrium concentration
of methane changes with temperature as exp(−cΦ/RT ), where
cΦ is a temperature coefficient. Fig. 7B shows the measured
temperature coefficients as a function of Φ. In the temperature
range investigated, the calculated equilibrium concentration of
methane can also be fitted to an Arrhenius-type expression to
get a prediction for the temperature coefficient. This prediction,
shown in Fig. 7B, fits well with the measured values. Both the
predicted and measured values show significant changes in the

80
O. Younes-Metzler et al. / Journal of Catalysis 241 (2006) 74–82
Fig. 7. (A) The apparent activation energy for the full oxidation reaction as a function of the stoichiometric ratio Φ. (B) The equilibrium temperature coefficient c
as a function of Φ.
Φ
Fig. 8. The outlet H /CO ratio as a function of the temperature for different values of the stoichiometric ratio Φ. Experimental data are displayed as symbols and
2
solid curves show calculated equilibrium data.
equilibrium around Φ = 0.5. Again, this indicates that equilib-
rium is approached in the reactor.
Fig. 8 shows the H2/CO ratio, which is important when using
the product gas for further synthesis, as a function of tempera-
ture for the different Φ’s. After ignition (lowest temperature
plotted for each Φ), the ratio first increases and then approaches
2 in a seemingly asymptotic way. The higher the Φ, the lower
the temperature required for obtaining ratios close to 2. The

O. Younes-Metzler et al. / Journal of Catalysis 241 (2006) 74–82
81
◦
Fig. 9. Outlet composition as function of total flow rate for Φ = 0.8 and a temperature of 570 C. Data from water are calculated from the mass balance.
curves represent the predicted H2/CO ratio as determined by
the equilibrium calculations. At lower temperatures, the pre-
dicted equilibrium H2/CO ratio deviates significantly from the
measured ratio, indicating that equilibrium is not reached at the
lower temperatures close to the ignition temperature. To ad-
dress this further, the flow was varied for Φ = 0.8 at a constant
temperature of 570 ◦C, which is close to but above ignition.
As shown in Fig. 9, the conversion of CH4 and the selec-
tivity toward H2 and CO increase with space velocity up to
2.5 × 104 h−1, then flatten or even decrease slightly. At low
temperatures, the calculated equilibrium is shifted toward pro-
duction of H2O and CO2, with very limited production of H2
and CO. This is clearly shown in Fig. 9 at low flow rates. At
higher flow rates, the composition moves away from equilib-
rium, the conversion of CH4 increases, and production of CO
and H2 is greatly increased. This is in qualitative agreement
with a previous model [21] that predicts that at relatively low
space velocities, the conversion of CH4 and selectivity toward
H2 will increase with space velocity. This increase, caused by
the kinetics of the reaction, illustrates the benefits of using
rhodium catalysts for efficient syngas production at low tem-
peratures.
The IR sensor was first focused close to the inlet of the reactor,
and subsequently the experiment was repeated with the sensor
focused on the middle of the reactor. In the middle of the re-
actor, a small temperature increase (<1 ◦C) was observed at
ignition, but close to the inlet, no abrupt temperature change
was observed. This is in contrast to previous results for a larger
reactor [16], in which significant heating of the catalyst surface
was observed. The difference is probably due to the improved
thermal properties of the microfabricated reactor, in which heat
was applied externally to maintain a high temperature, com-
pared with the earlier study [16] in which heat came from the
chemical reaction, which is self-sustained. In all of the mea-
surements, the heat generated chemically inside the reactor is
<0.1 W, which should be compared with the 40 W applied to
maintain the high temperature. This demonstrates that the heat
is effectively dissipated and that flow rate and temperature can
be varied independently.
5. Conclusion
Microfabricated reactors with integrated local heating of the
reaction zone have been fabricated and investigated thermally.
As an example, the partial oxidation of methane over rhodium
was investigated. Our results agree well with earlier studies and
demonstrate that flow rate and reactor temperature can be var-
ied independently. At low flow rate or high temperature, the
reaction in the microfabricated reactor approached equilibrium,
Finally, the temperature during ignition and extinction was
closely monitored to detect any possible change in temperature
caused by heat developed by the chemical reaction. Ignition is
reached by increasing the temperature at a slow, steady rate,
and extinction is accomplished by shutting off the oxygen flow.

82
O. Younes-Metzler et al. / Journal of Catalysis 241 (2006) 74–82
whereas at high flow rate or low temperature, the reaction was
kinetically limited with high selectivity toward hydrogen, as
reported previously. Our findings demonstrate that the micro-
fabricated reactor is convenient and reliable for studying fast
exothermic reactions.
[14] V.R. Choudhary, A.M. Rajput, B. Prabhakar, J. Catal. 139 (1993) 326.
[15] A.T. Ashcroft, A.K. Chetham, J.S. Foord, M.L.H. Green, C.P. Grey, A.J.
Murrell, P.D.F. Vernon, Nature 344 (1990) 319.
[16] L. Basini, A. Guarinoni, A. Aragno, J. Catal. 190 (2000) 284.
[17] D.A. Hickman, E.A. Haupfear, L.D. Schmidt, Catal. Lett. 17 (1993)
223.
[18] P.M. Witt, L.D. Schmidt, J. Catal. 163 (1996) 465.
[19] J.-D. Grunwaldt, L. Basini, B.S. Clausen, J. Catal. 200 (2001) 321.
[20] K.L. Hohn, L.D. Schmidt, Appl. Catal. A 211 (2001) 53.
[21] M. Bizzi, G. Saracco, R. Schwiedernoch, O. Deutschmann, AIChE J. 50
(2004) 1289.
[22] A.B. Mhadeshwar, D.G. Vlachos, J. Chem. Phys. B 109 (2005) 16819.
[23] D.G. Vlachos, L.D. Schmidt, R. Aris, AIChE J. 40 (1994) 1018.
[24] C. Appel, J. Mantzaras, R. Schaeren, R. Bombach, A. Inauen, N. Tylli,
M. Wolf, T. Griffin, D. Winkler, R. Carroni, Proc. Combust. Inst. 30 (2005)
2509.
[25] O. Deutschmann, F. Behrendt, J. Warnatz, Catal. Today 46 (1998) 155.
[26] R. Schwiedernoch, S. Tischer, C. Correa, O. Deutschmann, Chem. Eng.
Sci. 58 (2003) 633.
[27] M. Prettre, C. Eichner, M. Perrin, J. Chem. Soc., Faraday Trans. 43 (1946)
335.
Acknowledgments
The authors thank the STVF for financial support. CINF is
sponsored by The Danish National Research Foundation. S.J. is
supported by the Danish Research Council for Technology and
Production Sciences.
References
[1] K.F. Jensen, Chem. Eng. Sci. 56 (2001) 293.
[2] G. Kolb, V. Hessel, Chem. Eng. J. 98 (2004) 1.
[3] K. Jähnisch, V. Hessel, H. Löwe, M. Baerns, Angew. Chem. Int. Ed. 43
(2004) 406.
[28] D.A. Hickman, L.D. Schmidt, J. Catal. 138 (1992) 267.
[29] O.V. Buyevskaya, D. Wolf, M. Baerns, Catal. Lett. 27 (1994) 131.
[30] H.Y. Wang, E. Ruckenstein, Catal. Lett. 59 (1999) 121.
[31] M. Fichtner, J. Mayer, D. Wolf, K. Schubert, Ind. Eng. Chem. Res. 40
(2001) 3475.
[32] I. Aartun, T. Gjervan, H. Venvik, O. Görke, P. Pfeifer, M. Fathi, A. Hol-
men, K. Schubert, Chem. Eng. J. 101 (2004) 93.
[33] I. Aartun, H.J. Venvik, A. Holmen, P. Pfeifer, O. Görke, K. Schubert,
Catal. Today 110 (2005) 98.
[4] L. Kiwi-Minsker, A. Renken, Catal. Today 110 (2005) 2.
[5] O. Younes-Metzler, J. Svagin, S. Jensen, C.H. Christensen, O. Hansen,
U.J. Quaade, Appl. Catal. A: General 284 (2005) 5.
[6] D.A. Hickman, L.D. Schmidt, Science 259 (1993) 343.
[7] G.P. Van der Laan, A.A.C.M. Beenachers, Catal. Rev. Sci. Eng. 41 (1999)
255.
[8] F.J. Keil, Microporous Mesoporous Mater. 29 (1999) 49.
[9] P. Reubroycharoen, T. Yamagami, T. Vitidsant, Y. Yoneyama, M. Ito,
N. Tsubaki, Energy Fuels 17 (2003) 817.
[34] G. Veser, M. Ziauddin, L.D. Schmidt, Catal. Today 47 (1999) 219.
[35] M.L. Michelsen, Fluid Phase Equilib. 53 (1989) 73.
[36] I. Barin, Thermodynamic Data of Pure Substances, VCH, Weinheim,
1989.
[37] O. Knacke, O. Kubaschewski, K. Hesselmann, Thermodynamic Properties
of Inorganic Substances, second ed., Springer, Berlin, 1991.
[10] P.M. Witt, L.D. Schmidt, J. Catal. 163 (1996) 465.
[11] P.M. Torniainen, X. Chu, L.D. Schmidt, J. Catal. 146 (1994) 1.
[12] M. Lyubovsky, L.L. Smith, M. Castaldi, H. Karim, B. Nentwick, S. Ete-
mad, R. LaPierre, W.C. Pfefferle, Catal. Today 83 (2003) 71.
[13] D. Dissanayake, M.P. Rosynek, J.H. Lunsford, J. Phys. Chem. 97 (1993)
3644.