Partial oxidation of alcohols to produce hydrogen and chemicals in millisecond-contact time reactors

Article abstract of DOI:10.1016/j.jcat.2005.07.015

We compare the autothermal partial oxidation of 1-propanol and 2-propanol with methanol and ethanol on Rh with several additives in short-contact time reactors. All alcohols could produce H₂ at 70-90% selectivity, and Rh-Ce was superior to Rh i

Full text of DOI:10.1016/j.jcat.2005.07.015

Journal of Catalysis 235 (2005) 18–27
www.elsevier.com/locate/jcat
Partial oxidation of alcohols to produce hydrogen and chemicals
in millisecond-contact time reactors
Edvard C. Wanat, Balram Suman, Lanny D. Schmidt ∗
Department of Chemical Engineering and Materials Science, University of Minnesota, 421 Washington Ave SE, Minneapolis, MN 55455, USA
Received 12 May 2005; revised 12 July 2005; accepted 13 July 2005
Available online 18 August 2005
Abstract
We compare the autothermal partial oxidation of 1-propanol and 2-propanol with methanol and ethanol on Rh with several additives in
short-contact time reactors. All alcohols could produce H at 70–90% selectivity, and Rh–Ce was superior to Rh in producing H . Methanol
2
2
◦
produced high conversion and high selectivity to H even at high C/O where temperatures fell to <600 C. 2-Propanol gave lower conversions
2
and less H and CO than the other alcohols, but produced the most chemicals. Above C/O = 1.5, ∼70% of 2-propanol was converted into
2
acetone or propylene. Up to 20% propylene was formed at C/O = 1.5. In contrast, 1-propanol gave <8% propylene and <15% propanal at
any C/O and produced more ethylene than propylene. Much more oxygenates and olefins were formed on Rh than on Rh–Ce. These results
show that different alcohols have very different selectivity in catalytic partial oxidation at short contact times even at high temperatures.
Rapid adsorption of alcohols as alkoxy species leads to complete dissociation to H and CO. Our results suggest that acetone and olefins
2
likely were produced primarily by homogeneous reactions after all O had been consumed in the catalyst. Although alkanes do not form
2
significant oxygenates by partial oxidation at short contact times, alcohols can be made to produce predominantly oxygenates through suitable
adjustments of C/O and catalyst.
2005 Elsevier Inc. All rights reserved.
Keywords: Partial oxidation; Methanol; Ethanol; 1-Propanol; 2-Propanol; Rh and Rh–Ce catalysts
1. Introduction
uct streams containing high (85%) hydrogen selectivity at
conversions exceeding 95%.
The present work examines the products and mechanism
of partial oxidation of alcohols by examining the partial oxi-
dation of the C3 alcohols 1-propanol, which, like ethanol and
methanol is a primary alcohol, and 2-propanol, a secondary
alcohol.
Biorefineries produce a myriad of products, including
hydrogen for fuel cells and chemical feedstocks from re-
newable feedstocks. One possible approach to producing
hydrogen and chemicals from biomass is the catalytic par-
tial oxidation of biomass-derived liquids. Unlike fossil fuels,
biomass-derived liquids contain many oxygenates that affect
the performance of the reforming catalyst.
1.1. Methanol reforming
Previous research in our laboratory and other laborato-
ries has focused on the autothermal reforming of ethanol
and methanol [1,2]. Using contact times in the milliseconds
(<10 ms), both methanol and ethanol have produced prod-
Partial oxidation is an autothermal process that produces
syngas from methanol [1,3]. Noble metal catalysts yield high
activity and selectivity for syngas products. Using Pd/ZnO,
a hydrogen selectivity of 96% and conversion of 70% were
reported [3]. Pt and Rh catalysts on α-alumina monoliths
gave lower hydrogen selectivities but higher conversions.
Typical hydrogen selectivities were 65–75% with conver-
sions >90% at lean C/O [1]. Very low (<1%) methane
*
Corresponding author. Fax: +1-612-626-7246.
E-mail address: schmi001@tc.umn.edu (L.D. Schmidt).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2005.07.015
2005 Elsevier Inc. All rights reserved.

E.C. Wanat et al. / Journal of Catalysis 235 (2005) 18–27
19
selectivity was observed, which is highly desired because
methane consumes hydrogen and lowers the efficiency of re-
forming.
an automotive fuel injector at the top of the quartz reactor.
Heating tape wrapped around the quartz reactor above the
catalyst provided heat to vaporize the fuel. The upstream
temperature was maintained at 130–160 ◦C. Lightoff was
achieved by flowing fuel and air over the catalyst and apply-
ing an external burner until the catalyst temperature reached
∼800 ◦C. Once steady-state operation was obtained, no heat
was applied to the reactor except to vaporize the fuel. No
homogeneous ignition (flames) of the fuel before the cata-
lyst was observed at conditions and temperatures shown.
1.2. Ethanol reforming
The partial oxidation of ethanol produced 95% conver-
sion and 85% hydrogen selectivity in an autothermal re-
former using Rh–Ce [2]. When combined with Pt–Ce water–
gas shift catalyst in a staged reactor, the hydrogen selectiv-
ity (based on hydrogen from ethanol) reached 130% with a
25 mol% ethanol and 75 mol% water mixture. This indicates
that hydrogen was being formed from both ethanol and wa-
ter in the autothermal reactor.
2.3. Data acquisition and analysis
Most data were acquired at a total inlet flow rate, air
plus fuel, of 4 standard liters per minute (slpm), although
flow rates of 2 and 6 slpm were also examined. In all cases
the data at 2 slpm showed lower conversions and lower se-
lectivities to syngas because of lower reactor temperatures.
At 6 slpm, high temperatures prevented the acquisition of
data at low C/O due to the possibility of catalyst loss due
to evaporation. The residence time of the gases in the cat-
alytic monolith at 4 slpm was 10 5 ms and the GHSV
was ∼1 × 105 h−1. Data are presented as a function of C/O.
This ratio was determined by dividing the number of carbon
atoms in the fuel by the number of oxygen atoms from the
air flowing into the reactor. Only oxygen atoms from the air
were included, because only those atoms could oxidize the
fuel. This was different from previous work [2] with ethanol
in which the internal oxygen was included in the C/O cal-
culation. Temperatures reported are regarded as accurate to
20 ◦C.
1.3. 1-Propanol and 2-propanol reforming
There have been few reports of the reforming of 1-pro-
panol and 2-propanol [4]. The steam reforming of 2-pro-
panol was investigated on Rh with various supports includ-
ing CeO2, Al2O3, SiO2, ZrO2, MgO, and TiO2. Conversions
>90% were obtained at temperatures of 400 ◦C at low space
velocities. Rh–Ce produced the highest yields of syngas,
which was suggested to be due to the oxygen storage ca-
pacity of ceria. Al2O3 and SiO2 also produced high yields
of syngas. ZrO2, TiO2, and MgO were less active to syngas,
instead producing more acetone and propylene.
2. Experimental
2.1. Catalyst preparation
Gases were analyzed using gas chromatography by in-
jecting samples with a gas syringe. All data points represent
the average of three samples on a single catalyst. The data
for Rh and Rh–Ce was repeated on a second catalyst. All re-
sults were reproducible within 5%. Selectivities reported
were based on carbon atoms except for H2, which was based
on hydrogen atoms. No deactivation of catalyst activity was
noted for time on stream of ∼30 h with repeated shutdown
and restart.
The catalyst coating procedure was as described previ-
ously [5]. The monoliths used were 80 or 45 pores per lin-
ear inch (ppi), made from α-alumina, 17 mm in diameter
and 10 mm long, with a void fraction of ∼0.8. Rh from
Rh(NO3)3 solution was loaded to 5 wt% for Rh without ad-
ditives. Monoliths with additives contained 2.5 wt% Rh and
2.5 wt% additive.
The additives used included ceria, cobalt, and ruthenium,
all of which have been reported to promote syngas forma-
tion [2,6,7]. The catalyst and additive were deposited on the
monolith concurrently. Ceria was prepared from Ce(NO3)3 ·
6H2O, ruthenium from RuCl3 · xH2O, and cobalt from
CoCl2 · 6H2O. After the monolith was loaded with catalyst,
it was heated in air at 600 ◦C for 6 h to decompose the salts.
Pt was tested in a manner similar to Rh. However, Pt
never achieved light-off or steady-state operation with either
of the C3 alcohols. This observation was consistent with re-
sults from previous experiments with ethanol [2].
3. Results
3.1. Comparison of alcohols on Rh–Ce
Fig. 1 displays the conversion (a), temperature (b), hydro-
gen selectivity (c), and carbon monoxide selectivity (d) for
methanol, ethanol, 1-propanol, and 2-propanol on Rh–Ce.
Fig. 1a shows that methanol gave nearly complete conver-
sion to products except at very rich C/O, where the tem-
perature fell below 500 ◦C. The other primary alcohols both
gave >85% conversion over the range of C/O tested. Fi-
nally, for all C/O, the secondary alcohol, 2-propanol, gave
the lowest conversion. Fig. 1b shows that the reactor tem-
perature increased with increasing alcohol chain length. The
2-propanol operated at a slightly higher temperature than the
2.2. Gas delivery and startup
Compressed air from cylinders and metered by mass flow
controllers was used to provide oxygen for the partial ox-
idation reaction. The liquid fuels were delivered through

20
E.C. Wanat et al. / Journal of Catalysis 235 (2005) 18–27
Fig. 1. Reactions of alcohols on Rh–Ce on a ceramic foam monolith at millisecond contact time. The conversion (a), reactor temperature (b), hydrogen
selectivity (c), and carbon monoxide selectivity (d) are shown as a function of C/O ratio.
1-propanol. Figs. 1c and d display the hydrogen and carbon
monoxide selectivities, respectively. Syngas was the favored
product at low C/O for all the alcohols. The hydrogen and
carbon monoxide selectivity indicates that with increasing
carbon length, the alcohol produced less syngas and more of
the other chemicals.
of 1-propanol and 2-propanol. At all C/O, C2 and C3 prod-
ucts of 2-propanol were favored over those of 1-propanol.
3.2. Role of catalyst in the partial oxidation of C3 alcohols
Fig. 3 shows the conversion (a), reactor temperature (b),
hydrogen selectivity (c), and carbon monoxide selectivity
(d) for 1-propanol. The conversion of 1-propanol was not
strongly dependent on the additive used; nearly all conver-
sions at a given C/O were within ∼5% of one another.
Fig. 3b shows the reactor temperature for the different cat-
alyst/additive combinations. At lean C/O, there was wide
variability in the temperature, with Rh and Rh–Ru oper-
ating at high temperatures and Rh–Ce significantly lower
(∼200 ◦C). This was due to more combustion on Rh and
Rh–Ru at low C/O. However, as the C/O increased, the
temperatures became closer, as less oxygen was available
for combustion. Figs. 3c and d show the selectivities for hy-
drogen and carbon monoxide. The best syngas catalyst was
Rh–Ce; the carbon monoxide selectivity was ∼10% higher
for Rh–Ce, and hydrogen selectivity was even higher.
Fig. 2 shows the methane selectivity (a), ethylene selec-
tivity (b), propylene selectivity (c), and sum of the selec-
tivities observed for all C2 and C3 products for 1-propanol
and 2-propanol (d). Fig. 2a shows that the methane selectiv-
ity remained <5% for all of the alcohols except ethanol, for
which it was as high as 15% for high C/O. Fig. 2b shows the
ethylene selectivity for ethanol, 1-propanol, and 2-propanol.
Ethylene selectivities for ethanol and 2-propanol were <2%
for all C/O, and 1-propanol showed high ethylene selectivity
∼20% at high C/O. Fig. 2c displays the propylene selectiv-
ity for 1-propanol and 2-propanol, showing that 2-propanol
produced the most propylene with a maximum selectivity
of 20% on Rh–Ce at high C/O and that propylene selectiv-
ity went through a maximum at a C/O of 1.5. At lower C/O,
excess oxygen prevented the formation of propylene, and the
syngas selectivity increased. At high C/O, the selectivity to
acetone increased rapidly, which decreased the selectivity to
propylene. Fig. 2d shows the sum of the C2 and C3 products
Fig. 4 shows a more detailed examination of Rh and
Rh–Ce for 1-propanol. The selectivities for carbon monox-
ide, propylene, ethylene, and propanal are shown. Rh–Ce
produced more carbon monoxide than Rh, presumably be-

E.C. Wanat et al. / Journal of Catalysis 235 (2005) 18–27
21
Fig. 2. Reactions of alcohols on Rh–Ce on a ceramic foam monolith at millisecond contact time. The methane selectivity (a), ethylene selectivity (b), propylene
selectivity (c), and sum of the C and C product selectivities from 1- and 2-propanol (d) are shown as a function of C/O ratio.
2
3
Fig. 3. Comparison of Rh, Rh–Ce, Rh–Ru, and Rh–Co for 1-propanol at millisecond contact time. The conversion (a), reactor temperature (b), hydrogen
selectivity (c), and carbon monoxide selectivity (d) are shown as a function of C/O ratio.

22
E.C. Wanat et al. / Journal of Catalysis 235 (2005) 18–27
Fig. 4. Comparison of Rh and Rh–Ce for 1-propanol at millisecond contact time. The carbon monoxide selectivity (a), propylene selectivity (b), ethylene
selectivity (c), and propanal selectivity (d) are shown as a function of C/O ratio.
cause Ce stores oxygen and makes it available for reac-
tion via a redox reaction. Rh produced more ethylene and
propylene. Both catalysts produced approximately the same
amount of propanal.
this was due to the oxygen storage capability of ceria. Rh
produced ∼5% more propylene and ∼10% more acetone.
The propylene selectivity goes through a peak at C/O =
1.5–1.7 due to high syngas selectivity at low C/O and high
acetone selectivity at high C/O.
Fig. 5 shows the conversion (a), reactor temperature (b),
hydrogen selectivity (c), and carbon monoxide selectivity (d)
for 2-propanol. The conversion of 2-propanol did not vary
much on Rh–Ru, Rh–Ce, and Rh but was ∼10% lower on
Rh–Co. The conversions for all catalysts were lower than
those for 1-propanol. Fig. 5b shows the reactor tempera-
ture for the different catalyst/additive combinations. At lean
C/O, there was wide variability in the temperature, with
Rh and Rh–Ru operating at high temperatures and Rh–Ce
and Rh–Co operating at significantly lower temperatures.
Figs. 5c and d show the selectivities to hydrogen and carbon
monoxide for the four different catalysts. Rh–Ce and Rh–Co
were the best hydrogen catalysts at lean C/O, whereas all of
the catalysts but Rh produced approximately the same car-
bon monoxide selectivity.
3.3. Pore size and preheat
Fig. 7 shows the effect of pore size on the partial oxi-
dation of 1-propanol and 2-propanol. The carbon monoxide
selectivity (a) and propanal selectivity (b) are shown for 1-
propanol, and carbon monoxide selectivity (c) and acetone
selectivity (d) are shown for 2-propanol. The 80-ppi mono-
lith had more surface area and active catalytic sites available
for reaction. The carbon monoxide and hydrogen selectivi-
ties for both 1-propanol and 2-propanol were higher for the
80-ppi monolith. This indicates that higher-surface area cat-
alysts yield more syngas products. The olefin and oxygenate
selectivities were higher for the 45-ppi monolith, however.
This suggests that more complex chemical products were
produced by homogeneous reactions.
Fig. 6 shows a more detailed examination of Rh and
Rh–Ce for 2-propanol. Selectivities for carbon monoxide,
propylene, and acetone are shown. Rh–Ce produced more
carbon monoxide at low C/O than Rh. As with 1-propanol,
Fig. 8 shows the temperature (a), carbon monoxide selec-
tivity (b), and propanal selectivity (c) for two preheat tem-
peratures, 160 and 260 ◦C. The catalyst was Rh–Ce, and the

E.C. Wanat et al. / Journal of Catalysis 235 (2005) 18–27
23
Fig. 5. Comparison of Rh, Rh–Ce, Rh–Ru, and Rh–Co for 2-propanol at millisecond contact time. The conversion (a), reactor temperature (b), hydrogen
selectivity (c), and carbon monoxide selectivity (d) are shown as a function of C/O ratio.
Fig. 6. Comparison of Rh and Rh–Ce for 2-propanol at millisecond contact time. The carbon monoxide selectivity (a), propylene selectivity (b), and acetone
selectivity (c) are shown as a function of C/O ratio.

24
E.C. Wanat et al. / Journal of Catalysis 235 (2005) 18–27
Fig. 7. Comparison of catalyst pore size on reaction products. The catalyst used was Rh–Ce for 80 ppi monoliths and 45 ppi monoliths. The carbon monoxide
selectivity (a) and propanal selectivity (b) are shown for 1-propanol as a function of C/O ratio. The carbon monoxide selectivity (c) and acetone selectivity (d)
are shown for 2-propanol as a function of C/O ratio.
Fig. 8. Comparing the effect of preheat on reaction products. The fuel was 1-propanol and the catalyst used was Rh–Ce. Two preheat temperatures were tested:
◦
160 and 260 C. The temperature (a), carbon monoxide selectivity (b), and propanal selectivity (c) are shown as a function of C/O ratio.

E.C. Wanat et al. / Journal of Catalysis 235 (2005) 18–27
25
>800 ◦C, 1-propanol conversions were >90%. However, the
conversion fell considerably <800 ◦C and dropped to 30%
at ∼650 ◦C. The primary products from pyrolysis are eth-
ylene, propylene, methane, and carbon monoxide (from the
CO present in 1-propanol). Both ethylene and methane are
relatively constant, with selectivities between 20 and 35%
throughout the temperature range tested. However, propy-
lene and CO vary widely. Propylene is favored at low tem-
peratures, with ∼45% selectivity at 650 ◦C but only ∼5%
selectivity at 950 ◦C. This decrease is offset by an increase
in CO selectivity from ∼0 to 20% at high temperatures.
Hydrogen selectivity behaved very much the same as CO
selectivity, reaching 25% at high temperatures and falling to
<5% at low temperatures.
One surprising result was the absence of propanal. Even
at high temperatures, propanal selectivity remained <1%.
High selectivity of acetaldehyde from ethanol was observed
when air was fed to the reactor and in the modeling of the
pyrolysis of ethanol [8]. This suggests that aldehydes are
formed with oxygen-assisted reactions.
Fig. 9. The conversion and selectivities of 1-propanol in a tube with a non-
catalytic monolith foam. The figure shows the conversion, carbon monox-
ide, ethylene, propylene, and methane selectivities as a function of tem-
perature in a heated tube with no catalyst. The experiment simulated the
homogeneous pyrolysis of 1-propanol. For these experiments the heated
zone of the tube was ∼30 cm, so the residence time was ∼200 ms.
fuel was 1-propanol. Increasing the preheat temperature led
to higher reactor temperatures, which in turn favored more
homogeneous reactions, shown by the higher propanal selec-
tivity. Conversely, less heterogeneous reactions occurred at
higher temperatures, shown by the lower carbon monoxide
selectivity.
3.5. Acetone
Acetone was flowed over Rh and Rh–Ce monoliths at
4 slpm because of the high selectivity to acetone (>50%)
observed from 2-propanol. Acetone gave low conversion at
virtually all C/O levels. At C/O = 1.5, the conversion was
∼70%; at high C/O, it fell to <40%. The products were vir-
tually all syngas and methane; no C2 or C3 products were
observed. The lack of olefins and oxygenated products is be-
lieved to be due to the absence of any weak α-C–H bonds.
This absence impedes homogeneous chemistry, the mecha-
nism for the production of olefins and oxygenates.
3.4. Noncatalytic monolith
A noncatalytic monolith test was performed to study the
role of homogeneous chemistry in the partial oxidation of the
C3 alcohols. An uncoated, α-Al2O3 monolith was placed in-
side a quartz tube, and 1-propanol and nitrogen were flowed
through the tube at a flow rate of 4 slpm. A furnace was
used to regulate the temperature. The heated zone in the
furnace was ∼30 cm long, and residence time of the gases
was 150–300 ms. Temperatures tested ranged from 600 to
950 ◦C. These conditions simulated the downstream portion
of the catalyst where no oxygen is present and only homoge-
neous reactions producing olefins and oxygenates can occur.
The homogeneous chemistry of ethanol has been exam-
ined in a noncatalytic monolith experiment [8] and com-
pared with the results from a homogeneous combustion and
pyrolysis model [9]. Four parameters affecting the homo-
geneous chemistry of ethanol decomposition were modeled.
First, homogeneous chemistry with oxygen and without oxy-
gen was tested to examine the effect of oxygen on reaction
chemistry. Second, two temperatures (600 and 900 ◦C) were
modeled to examine temperature effects. Third, residence
times of 200 and 600 ms were tested, because the homo-
geneous kinetics showed no reaction at 600 ◦C at 200 ms.
Finally, for experiments with oxygen, C/O was varied be-
tween 0.7 and 1.3.
4. Discussion
A major result from these experiments is that selectivity
varies considerably among the alcohols. In contrast, all alka-
nes can be made to produce primarily syngas at low C/O
and olefins at higher C/O. No oxygenates (much <1%) are
ever observed with alkanes, except for the single gauze re-
actor [10], where surface reactions produce heat that drives
homogeneous combustion reactions in an empty tube.
Our general model of these processes in short-contact
time reactors is that high-velocity premixed gases at low
temperature force surface reactions early in the hot cata-
lyst (typically within the first millimeter), and all oxygen is
consumed within at least a few millimeters of the catalyst en-
trance. Surface reactions of all fuels appear to form mostly
C1 products (CO, CO2, and CH4) and H2 and H2O.
After all of the O2 is consumed, homogeneous reactions
appear to dominate. For C/O <1, most fuel reacts in the
oxidation zone, and C1 products form exclusively. Homoge-
neous reactions occur primarily downstream in the channels
of the monolith and after the monolith. Detailed modeling of
Fig. 9 shows the results of an experiment in which ni-
trogen and 1-propanol were flowed over the reactor with
only an α-Al2O3 monolith with no catalyst. At temperatures

26
E.C. Wanat et al. / Journal of Catalysis 235 (2005) 18–27
these reaction systems [11] shows that the Rh catalyst in the
oxidation zone contains primarily oxygen atoms, whereas
surface carbon dominates downstream in the catalyst. Car-
bon covered surfaces are probably inert toward any surface
reactions, so that in this zone the effect of the monolith sur-
face appears to be primarily to maintain high and uniform
temperatures by solid heat conduction.
Temperature is a critical parameter in controlling rates
of surface and homogeneous reactions, which in turn con-
trol selectivity. Measured temperatures are at the exit of
the catalyst, and these and other experiments have shown
that maximum temperatures are 100–200 ◦C higher than at
the exit. The highest temperatures occur near the entrance
of the monolith, where most surface reactions occur. Fol-
lowing the monolith, rapid cooling of the product gases
rapidly quenches all reactions and freezes the product mix-
ture formed within the monolith.
We observe more ethylene than propylene, which could oc-
cur through a surface reaction,
C3H7O(s) → C2H4 + CH3O(s)
or
C3H7O(s) → C3H6 + OH(s),
although scission of the C–O bond to make propylene should
be more difficult than scission of the C–C bond to make eth-
ylene.
We know of no surface science studies of 2-propanol,
but steric limitations should make it less reactive than
1-propanol (which would lead to more homogeneous chem-
istry), and there are no obvious surface reaction channels
that would yield olefins. The α-C–H bond in absorbed iso-
propoxy may be weaker than other bonds, which could lead
to acetone by a surface reaction,
4.1. Surface reactions of alcohols
C3H7O(s) → CH3COCH3 + H(s).
Alcohol adsorption and decomposition on many noble
metal surfaces [12,13], including Rh [12], has been exam-
ined extensively on well-defined single-crystal surfaces. For
example, the decomposition of ethanol and acetaldehyde
have been studied on Rh(111) [12]. It appears that ethanol
adsorbs, forming an ethoxy species. The next step is the
formation of a bridged oxametallacycle, which readily un-
dergoes C–C bond scission. The resulting species quickly
break down to adsorbed C, H, and O atoms, which recom-
bine to form syngas. However, acetaldehyde adsorbs on two
adjacent Rh sites, forming an η2-acetaldehyde species that
undergoes C–C bond scission to form carbon monoxide and
methyl, which can then form methane [12].
We suggest that surface reactions of all alcohols should lead
to mostly C1 products, and that species larger than C1 prod-
ucts are probably formed by homogeneous reactions.
4.2. Homogeneous reactions
The combustion chemistry of ethanol [9] and metha-
nol [14] has been studied extensively, and detailed reaction
mechanisms of ethanol oxidation involving 57 species and
more than 370 reactions are available [9]. No detailed mech-
anisms appear to be available for the combustion of C3 alco-
hols.
The experiments in Fig. 9 using a tube without a catalyst
show that nearly complete conversion of 1-propanol can be
obtained above 800 ◦C, even in the absence of O2, although
it should be noted that in a tube furnace the reactants are
heated for nearly the entire length of the furnace, so that the
residence time in this experiment is ∼200 ms, compared to
10 ms in the catalytic monolith, where the gases are cool
before they reach the catalyst and reaction is complete.
Fig. 9 shows that the major products from homogeneous
reactions are C2H4, C3H6, CH4, and CO, with less propylene
and more dissociative products at high temperatures. The H2
selectivity was 25%, and CO was ∼20% at the highest tem-
peratures.
We note that at our surface temperatures (800–1000 ◦C),
noble metals are predicted to be essentially clean even at
atmospheric pressure. All experiments agree that alcohols
adsorb initially through the lone pair of electrons on the O
atom, and that above 200 ◦C, rapid dissociation occurs to
form the alkoxy species
ROH → ROH(s) → RO(s) + H(s).
The adsorbed alkoxy is stable under UHV conditions to
∼300 ◦C [12], where it is observed to decompose completely
to adsorbed C, H, and O atoms and CO. This is in agreement
with the present results showing that CO and H2 dominate,
with CO2 and H2O as minor products at low C/O.
Selectivities were predicted using the detailed model for
ethanol and for ethanol–O2 mixtures as functions of time
and temperature. They show that low C/O, high tempera-
tures, and high residence times favor syngas formation. For
example, at 900 ◦C and 200 ms, the CO selectivity falls from
60% at C/O = 0.7 to <40% at C/O = 1.3. Ethylene forma-
tion accounts for the decrease in CO selectivity as ethylene
selectivity increases from ∼10% at low C/O to 35% at rich
C/O. This is in contrast to 600 ◦C, which showed very little
conversion (<1%) even at long residence times of 600 ms.
Although dehydration of alcohols can yield olefins, the
alkoxy species do not have obvious dehydration channels
available. Thus other pathways probably would be needed
to produce olefins on the surface.
From methanol, there are no reaction channels of methoxy
that yield higher species, except perhaps dimerization.
Ethoxy could dissociate with addition of adsorbed H to
ethane (<0.5% observed) or removal of H to form eth-
ylene (<2% observed). 1-propanol forms propoxy, which
could dissociate to form propane, propylene, or ethylene.

E.C. Wanat et al. / Journal of Catalysis 235 (2005) 18–27
27
4.3. Effects of catalyst and additives
tems often can be tuned to produce a single dominant prod-
uct, such as acetone from 2-propanol. Short-contact time
reactors can be tuned to exhibit high selectivities to specific
products, even at very high temperatures. The patterns of re-
action can be adjusted strongly by changing catalyst, feed,
and flow conditions, and at short contact times these product
distributions are frozen to yield simple, and perhaps valu-
able, product distributions.
The results of Fig. 3 clearly show that Rh–Ce is the best
catalyst for H2 and CO while suppressing higher products,
and that Rh and Rh–Co produce much less H2 and CO while
enhancing higher products. Similar trends are observed with
all alcohols, although methanol produces no higher com-
pounds and ethanol very little.
Conversions on all catalysts are high and approximately
the same, although temperatures vary widely, with Rh being
200 ◦C hotter than Rh–Ce, which was the coldest. Tempera-
tures correlate with selectivities because the catalyst produc-
ing most CO2 and H2O will, of course, run hotter. This can,
in turn, generate more products of homogeneous chemistry.
There are large differences between Rh and Rh–Ce for
higher products. For example, with 1-propanol at C/O =
1.4, the C2H4 is 3% on Rh–Ce and 17% on Rh and the C3H6
is 3% on Rh–Ce and 15% on Rh. Acetaldehyde formation
from 1-propanol appears to be essentially the same on Rh
and Rh–Ce.
Acknowledgments
This research was partially supported by grants from the
Minnesota Corn Growers Association and the Initiative for
Renewable Energy and the Environment at the University of
Minnesota.
References
[1] B.E. Traxel, K.L. Hohn, Appl. Catal. A 244 (2003) 129.
[2] G.A. Deluga, J.R. Salge, L.D. Schmidt, X.E. Verykios, Science 303
(2004) 993.
[3] M.L. Cubiero, J.L.G. Fierro, J. Catal. 179 (1998) 150.
[4] T. Mizuno, Y. Matsumura, T. Nakajima, S. Mishima, Int. J. Hydrogen
Energy 28 (2003) 1393.
[5] K. Venkataraman, J.M. Redenius, L.D. Schmidt, Chem. Eng. Sci. 57
(2002) 2335.
[6] S. Cavallaro, N. Mondello, S. Freni, J. Power Sources 102 (2001) 198.
[7] D.K. Liguras, K. Goundani, X.E. Verykios, Int. J. Hydrogen Energy 29
(2004) 419.
[8] J.R. Salge, G.A. Deluga, L.D. Schmidt, J. Catal. 235 (2005) 69.
[9] N.M. Marinov, Int. J. Chem. Kinet. 31 (1999) 183.
[10] R.P. O’Connor, L.D. Schmidt, Stud. Surf. Sci. Catal. 113 (2001) 289.
[11] O. Deutschman, L.D. Schmidt, AIChE J. 44 (1998) 2465.
[12] C.J. Houtman, M.A. Barteau, J. Catal. 130 (1991) 528.
[13] N. Kruse, G.K. Chuah, G. Abend, D.L. Cocke, J.H. Block, Surf.
Sci. 189/190 (1987) 832.
5. Conclusions
The selectivity of catalytic partial oxidation of alcohols
varies strongly with size and structure of the alcohol and
with the catalyst used. We examined only a few catalysts and
reaction parameters in the present study, because our inter-
est was in exploring the mechanisms rather than optimizing
for a particular product. Methanol and ethanol produce es-
sentially only H2 and CO, because they react rapidly on the
surface, and the alkoxy species probably has no surface reac-
tion channels that do not lead to total decomposition on no-
ble metal surfaces. Propanol and (probably) higher alcohols
have many surface and homogeneous reaction channels that
can lead to large amounts of larger products, and these sys-
[14] P.H. Cribb, J.E. Dove, S. Yamazaki, Combust. Flame 88 (1992) 186.