Oxygenates and olefins from alkanes in a single-gauze reactor at short contact times

Article abstract of DOI:10.1006/jcat.1999.2637

Partial oxidation of linear C₁-C₅ alkanes in single-gauze reactors has been examined in the fuel-rich regime at contact times as low as 100 μs. We show that, whereas methane and ethane produce mostly CO and ethylene, respectively, propane, butane, and pentane give high selectivity to oxygenates and olefins. Partial oxidation of propane over a single gauze gives a total selectivity to olefins of nearly 60% with a 2/1 C₃H₆/C₆H₄ ratio and minimal oxygenates. Butane oxidation gives a significant amount of oxygenated products, mainly acetaldehyde and formaldehyde with a more open gauze, giving better selectivities. Pentane oxidation gives the highest oxygenate selectivities (up to 60%), the main products being acetaldehyde and propionaldehyde. We argue that this is a coupled heterogeneously initiated homogeneous reaction system, with total combustion primarily catalyzed by the Pt surface and the oxygenates and olefins formed subsequently by gas-phase reactions.

Full text of DOI:10.1006/jcat.1999.2637

Journal of Catalysis 187, 400–409 (1999)
Article ID jcat.1999.2637, available online at http://www.idealibrary.com on
Oxygenates and Olefins from Alkanes in a Single-Gauze Reactor
at Short Contact Times1
D. I. Iordanoglou, A. S. Bodke, and L. D. Schmidt2
Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, Minnesota, 55455
Received March 16, 1999; revised June 17, 1999; accepted June 25, 1999
with a significant portion of the feed being converted to
methane. Huff et al. have shown that selective production
of ethylene can be achieved upon feeding ethane and oxy-
gen in a Pt-coated monolith reactor (7). They also showed
that using propane as a feed in the above reactor can give
1/1 C2H4/C3H6 ratio (8), a significant improvement since
propylene-rich product streams are particularly desirable.
Recent results using Pt–Sn catalysts with H2 addition in the
feed give olefin selectivities higher than 85% (15). In all of
these situations, less than 1% total selectivity to oxygenated
products was observed.
Oxygenates such as low molecular weight aldehydes are
widely used in industrial organic synthesis. A lot of re-
search is generated on aldehyde production from alcohol
oxidation (23) or acid reduction (24). Of great interest are
methods on direct oxygenate production from alkane feeds
(25, 26). Goetsch et al. (1, 2) showed that a mixture of n-
butane and oxygen fed in a single-gauze reactor gave at
least 40% selectivity of oxygenated products with contact
times in the order of milliseconds.
Partial oxidation of linear C1–C5 alkanes in single-gauze reac-
tors has been examined in the fuel-rich regime at contact times as
low as 100 s. We show that, whereas methane and ethane pro-
duce mostly CO and ethylene, respectively, propane, butane, and
pentane give high selectivity to oxygenates and olefins. Partial ox-
idation of propane over a single gauze gives a total selectivity to
olefins of nearly 60% with a 2/1 C3H6/C2H4 ratio and minimal oxy-
genates. Butane oxidation gives a significant amount of oxygenated
products, mainly acetaldehyde and formaldehyde with a more open
gauze, giving better selectivities. Pentane oxidation gives the high-
est oxygenate selectivities (up to 60%), the main products being
acetaldehyde and propionaldehyde. We argue that this is a cou-
pled heterogeneously initiated homogeneous reaction system, with
total combustion primarily catalyzed by the Pt surface and the oxy-
genates and olefins formed subsequently by gas-phase reactions.
c
1999 Academic Press
1. INTRODUCTION
A reactor consisting of a single layer of woven metal
gauze is a novel catalytic system for alkane oxidation with
several unique characteristics (1–3). Autothermal partial
oxidation of hydrocarbons over single gauzes combines
surface reactions with gas-phase chemistry under fast mix-
ing and quenching conditions at very short contact times.
This process forms mostly nonequilibrium products which
are currently made by multiple-stage, energy-intensive pro-
cesses. In this paper, we extend previous work on single-
gauze reactors to a range of linear hydrocarbon feeds and
examine its potential for production of chemicals.
Olefins such as ethylene and propylene are important
chemicals, whose manufacture attracts the attention of
many researchers. Typical production methods, which are
still producing large pieces of literature, are steam crack-
ing (16), catalytic cracking (17–19), and oxidative dehy-
drogenation (20–22). For many thermal methods (4) C2H4
and C3H6 production is simultaneous and reported carbon
selectivity ratios range from 60/1 (C2H4/C3H6) to 1.5/1 (5),
2. EXPERIMENTAL
We fed C1–C5 linear hydrocarbons premixed with oxygen
and nitrogen to single-gauze reactorsand ran autothermally
at short contact times, investigating conditions for poten-
tially high selectivities to specific products. Gauzes were
99.9% pure Pt–10% Rh, with a wire diameter of 90 m and
either 40 or 80 mesh (wires per inch), corresponding to 640-
or 320- m wire spacing.
The reactor, catalysts, and procedure used in these ex-
periments have been described previously (3). We used a
syringe pump with 500-ml capacity and variable flow rates
up to 420 ml/h to feed the liquid hydrocarbon through a
coil pipe in a fluidized bed heater at a pressure of 800 kPa
(9). The gasified fuel was then expanded to reactor operat-
ing pressure. Before mixing with the oxygen and nitrogen,
the hydrocarbon vapors passed through a superheater to
eliminate condensation.
1 Supported by NSF under Grant CTS96-2902 and by DuPont.
2 To whom correspondence should be addressed.
400
0021-9517/99 $30.00
c
Copyright
1999 by Academic Press
All rights of reproduction in any form reserved.

OXYGENATES AND OLEFINS FROM ALKANES
401
3. RESULTS
3.1. Methane and Ethane
3.2. Propane
The possibility of selectively producing propylene from
propane was examined. As mentioned earlier (1, 2) and
seen in this work, oxygenates are a small by-product in
this process. Experiments were done using both 80- and
40-mesh gauzes, at a total flow rate of 2.5 slpm and 20%
nitrogen dilution, varying the fuel-to-oxygen ratio. Results
are shown in Fig. 2.
Methane and oxygen were fed over a 80-mesh gauze at
a flow rate of 2.5 slpm and 20% nitrogen dilution. Typi-
cal results are shown in Figs. 1a and 1b. Over a range of
CH4/O2 ratios, methane conversion varied between 21 and
15% . Despite the large gauze transparency, oxygen con-
version was always above 90% . This is surprising since no
oxygen is expected to be consumed in the gas phase un-
der these operating conditions. This means that catalytic
consumption rates of oxygen are very high. CO selectivity
was relatively high, between 65 and 80% . However, total
oxidation dominates H-atom consumption, resulting in H2
selectivities <10% .
Similarly, a mixture of ethane and oxygen was fed over
a 80-mesh gauze and the results are shown in Figs. 1c and
1d. Between C2H6/O2 ratios of 1.7 and 2.3, the ethane con-
version changed from 85 to 55% , while oxygen was almost
completely converted. Ethylene selectivity was between 60
and 65% , and other products included CO, CO2, and CH4.
These results are comparable to those obtained by Huff
et al. (7) over Pt-coated monoliths at a residence time of
about 5 ms.
80-mesh gauze. Three operating regimes can be identi-
fied: with C3/O2 varying from 1 to 1.7, propane conversion
exhibited highest values close to 90% while oxygen was
completely converted. The dominant olefin was ethylene
and CO selectivity was higher than that of CO2. For C3/O2
varying from 1.7 to 2.8, the dominant olefin changed from
ethylene to propylene. Fuel conversion dropped from 55 to
20% but oxygen was still completely converted. At a ratio
of 2.8, the propylene-to-ethylene selectivity ratio was 2/1
and propylene selectivity was 37% . Finally, for C3/O2 > 2.8,
oxygen breakthrough was observed and CO2 selectivity
increased dramatically. The reactor was practically extin-
guished, mostly producing combustion products.
40-mesh gauze. Results were similar to those of the 80-
mesh gauze, but in this case, oxygen conversion was always
FIG. 1. Fuel and oxygen conversions, temperatures, and product selectivities for a single 80-mesh Pt–10% Rh gauze reactor over a range of
fuel/oxygen feed ratios. Experimental results for the methane feed are shown in Figs. 1a and 1b, and for the ethane feed are shown in Figs. 1c and 1d.

402
IORDANOGLOU, BODKE, AND SCHMIDT
FIG. 2. Fuel and oxygen conversions, temperature, and product selectivities for n-propane feed over a single Pt–10% Rh gauze reactor for a range
of fuel/oxygen feed ratios. The results for the 80-mesh gauze are shown in Figs. 2a–2c, and for 40-mesh gauze the results are shown in Figs. 2d–2f.
incomplete. Propylene vs ethylene selectivity ratios were tivity of 40% . Olefin selectivity was around 30% and the
higher on the 80-mesh gauze.
combined CO + CO2 selectivity did not exceed 30% . For a
ratio above 3.0, considerable reduction of conversions and
olefin and oxygenate selectivity occurred, with n-butane
3.3. n-Butane
n-Butane was the lightest fuel to yield significant oxy- and oxygen conversions dropping to 5 and 60% , respec-
/O2
genated products(1, 2). It wasproposed and later confirmed tively. This change appears to be reversible with n-C4
feed variation. Negligible oxygenates and olefins were pro-
duced, main products being CO and CO2.
A 40-mesh gauze gave higher oxygenate selectivities and
butane conversions with leaner operation. For a small range
of n-C4/O2 feed ratios, two stable steady states were ob-
served and reproduced on several gauzes. In the first steady
(3) that oxygenates are produced by gas-phase chemistry
only. The results using n-butane feed have been published
earlier and we refer the reader to the original publication
for more details (3). Here, we mention only the most im-
portant features.
Over an 80-mesh gauze and for an n-C4/O2 ratio between
2.3 and 3.0, the main products were oxygenates, with a selec- state, oxygen and fuel conversions were high, whereas in the

OXYGENATES AND OLEFINS FROM ALKANES
403
FIG. 3. Fuel and oxygen conversions, temperature, and product selectivities for n-pentane feed over a single 40-mesh Pt–10% Rh gauze reactor for
a range of fuel/oxygen feed ratios. Reactor operating pressure is 120 kPa.
second they were much lower. Oxygenates were formed in pressure varied from 120 to 195 kPa. Results are shown in
one steady state only, and in the other CO and CO2 were Fig. 4. Elevated pressures increase fuel conversion signif-
the main products. An important observation was that CO2 icantly as expected. The conversion almost doubled upon
yields remained constant in both steady states (3).
increasing the pressure from 120 to 195 kPa. Oxygenate se-
lectivity decreased and olefins were the major products at
higher pressures. In these experiments, two main regimes
were identified: for pressures below 160 kPa, oxygenate
and total COx selectivity dropped and olefin selectivity in-
creased upon increasing the pressure. For pressure above
160 kPa, conversions and selectivities did not vary signifi-
cantly. Exit gas temperatures increased from 480 to 540 C.
We also ran experiments at a constant pressure of
191 kPa, varying the n-C5/O2 ratio, and the results are plot-
ted in Fig. 5. A much higher conversion was obtained, ex-
ceeding 60% . The temperature was 100 C higher than
that in low-pressure experiments. Up to an n-C5/O2 ratio of
2.8, olefins were the major products. Above this ratio, oxy-
genate selectivity was higher. Maximum oxygenate selectiv-
ity approached 50% , with acetaldehyde and propionalde-
hyde being the most important products.
3.4. n-Pentane
Since n-butane oxidation gave higher selectivities to oxy-
genated products using a 40-mesh gauze, we ran pentane ex-
periments with this mesh size only. Typical results are shown
in Fig. 3. Over the entire n-C5/O2 range, oxygen conversion
wascomplete. n-Pentane conversion varied from 40% in the
fuel-lean side to 20% in the fuel-rich side. Exit gas temper-
atures were between 440 and 280 C. Olefins were the dom-
inant products in the fuel-lean regime. For n-C5/O2 > 1.7,
oxygenate selectivity was higher than that of olefins, varying
from 35 to 58% . The maximum selectivity was about 10%
higher than that seen with n-butane. Acetaldehyde was still
the major oxygenated product with 27% selectivityand pro-
pionaldehyde was next with 18% selectivity. The selectivity
ratio of these products varied from 1/1 to 3/2 from the fuel-
lean to richer side. Other minor oxygenated products were
also observed, including formaldehyde, methanol, acetic
acid, acrolein, and ethanol, with all selectivities below 5% .
CO + CO2 selectivity never exceeded 20% .
3.5. Gauze Activation
Activation of catalyst gauzes with time is very important
(10, 11). Fresh gauze has a smooth surface and, after several
Higher pressures. We ran pentane oxidation experi- hours of operation, the surface restructures dramatically
ments at pressures up to 195 kPa. Flow rate, preheat, ni- (1, 10). The morphology of a used gauze indicates that
trogen dilution, and n-C5/O2 were kept constant, and the the activation mechanism is similar to that of ammonia

404
IORDANOGLOU, BODKE, AND SCHMIDT
FIG. 4. Fuel and oxygen conversions, temperature, and product selectivities for n-pentane feed over a single 40-mesh Pt–10% Rh gauze reactor for
a range of reactor operating pressures.
oxidation and HCN production (10). Conversions and se- served by SEM. The activation period can vary from a few
lectivities improve significantly during the activation period hours to several days, depending on the fuel used and the re-
before reaching a steady state. After reaching steady state, action temperature. Attention was paid to avoid gauze de-
the gauze surface shows no further restructuring, as ob- terioration. Repeated high-temperature experiments with
FIG. 5. Fuel and oxygen conversions, temperature, and product selectivities for n-pentane feed over a single 40-mesh Pt–10% Rh gauze reactor for
a range of fuel/oxygen feed ratios. The reactor operating pressure is 195 kPa.

OXYGENATES AND OLEFINS FROM ALKANES
405
extreme composition variations sometimes resulted in constant at 1.6 and nitrogen dilution was 20% . In the first
gauze disintegration, especially during partial oxidation of 5 h, the fuel conversion was below 10% , oxygen conver-
C1–C3, where the catalyst temperatures can be very high. sion was below 30% , and CO + CO2 selectivity was above
All the results reported in this work are from experiments 80% . It is very interesting that selectivities to all products
where gauze deterioration was not an issue.
reached a steady value after about 8–9 h of operation, but
With methane feed, the gauze activated in only a few it took at least 3–4 more h for the conversions to stabilize.
hours and exhibited the highest exit gas temperatures. This was also observed with other fuels. For example, both
Gauze activation with higher alkanes took much longer be- butane and pentane oxidation gave a steady selectivity to
cause temperatures were lower. Before being able to obtain oxygenates much before the fuel and oxygen conversions
consistent and reproducible data with n-pentane, several became stable.
days were required. With all fuel feeds, we observed lower
fuel and oxygen conversions and higher CO and CO2 selec-
tivities during initial runs.
4. DISCUSSION
For a more systematic study of gauze activation, experi-
ments were done with propane using a 40-mesh gauze and
the results are shown in Fig. 6. The C3/O2 ratio was kept
4.1. Oxygenate and Olefin Formation
CH4 yields mostly CO and H2O and C2–C5 linear alkanes
give 60–80% selectivities to olefins and oxygenates. How-
ever, oxygenates are produced in significant quantities only
when there are at least four carbon atoms in the backbone
of the hydrocarbon feed. C2H6 and C3H8 give mostly olefins.
Methane and ethane. Results for methane oxidation
show high CO selectivity ( 80% ) and very low H2 selectiv-
ity ( 10% ). For production of syngas, Rh-coated monoliths
give much higher yields (6). Results for ethane oxidation
show C2H4 selectivities between 60 and 65% , comparable
to those obtained on Pt-coated monoliths. Since residence
times in gauze reactors are at least 50 times shorter than
those over monoliths, these results show that Pt-catalyzed
oxidative dehydrogenation of ethane is an extremely fast
process and only a minimum amount of the noble metal is
required for the reaction to go to completion.
Propylene from propane. Using propane as the fuel,
we observe significant amounts of propylene. In this re-
action ethylene is an important by-product. Recent results
of propane oxidation over monoliths have shown that the
C3H6/C2H4 ratio can be as high as 1/1 (8). We find that this
product ratio can be further improved and can be as high
as 2/1 using 80-mesh gauze.
Oxygenates from n-C4 and n-C5. With butane as the
feed, we observe significant amounts of oxygenated prod-
ucts, primarily acetaldehyde and formaldehyde. A single
gauze is the only configuration to give significant selectivi-
ties to these nonequilibrium products during fast autother-
mal operation. A more open gauze gives higher selectiv-
ities to oxygenates. The gas and surface reactions can be
decoupled over the more open gauze, resulting in multiple
steady states (3).
With pentane feed, a 100 C preheat was applied in all
runs to avoid condensation. Maximum oxygenate selectiv-
ityapproached 60% . Onlythe 40-mesh gauze wasused since
the more open gauze is favorable for oxygenate produc-
tion. The main oxygenated products from pentane are C2
and C3 aldehydes rather than C1 and C2 aldehydes from
butane. Oxygen conversion is complete throughout the
FIG. 6. Fuel and oxygen conversions, temperature, and product selec-
tivities for n-propane feed over a single 40-mesh Pt–10% Rh gauze reactor
during activation. The fuel/oxygen feed ratio is 1.6.

406
IORDANOGLOU, BODKE, AND SCHMIDT
n-C5/O2 feed range, in contrast to butane experiments. This
makes pentane more attractive since the presence of oxy-
gen in the exit stream makes product handling dangerous.
In n-C4 experiments, the oxygenate selectivities rapidly
dropped above C/O 4.7. With n-C5, we ran up to C/O 8.7
without seeing any decrease in oxygenates. This can prob-
ably be attributed to the increased gas-phase reactivity of
pentane compared to that of butane. In addition, n-C5 runs
much colder than n-C4 and lower temperatures promote
higher oxygenates production because their decomposition
is minimized. The exit gas temperature is 50 C lower than
that in n-C4 experiments at the rich limit.
Pressure effects. In Fig. 4, it is seen that higher pressure
has a significant effect only up to 160 kPa and above this
there are not manychangesin the chemistry. CO production
may be a key factor in this situation in that CO is the only
major species whose selectivity remains constant between
120 and 195 kPa. This means that its total yield is rising, thus
consuming oxygen atoms which could otherwise go toward
oxygenate production.
Varying the n-C5/O2 ratio at higher pressure (191 kPa) re-
sults in significantly higher fuel conversion, especially in the
fuel leaner regime. The trade-off is a decrease in oxygenate
selectivity, presumably because of the higher temperatures
observed.
In general, the fact that pressure strongly influences ex-
perimental results confirms the importance of homoge-
neous chemistry. However, it appears that we cannot easily
increase fuel conversion without decreasing oxygenate se-
lectivity in an insulated gauze reactor. For a commercial
process heat removal from the system should probably be
FIG. 7. Fuel conversions and oxygenate and olefin selectivities vs C/O
feed ratio, for n-propane, n-butane, and n-pentane fuel feeds. The reactor
implemented to avoid production of other chemicals such is a 40-mesh single Pt–10% Rh gauze.
as olefins and COx at the expense of oxygenates.
olefins have slopes and values which are quite different than
those of butane or pentane.
Comparison of fuels. In Fig. 7, conversions and major
product selectivities of C3H8, C4H10, and C5H12 are com-
pared, for experiments with a 40-mesh gauze.
A reaction path analysis using gas-phase chemistry
(Fig. 8) shows that the intramolecular rearrangement of
peroxy radicals is the primary mechanism for oxygenates
production for butane and presumably for pentane. The
rates of similar reactions for propane were much lower, re-
sulting in minimal oxygenates production.
It is seen that butane and pentane exhibit similar behav-
ior in that conversions and selectivities of both fuels have
almost identical slopes. The pentane conversion is higher
than butane by 5–10% for C/O between 3.5 and 4.5. As
a consequence, the corresponding temperatures are about
50 C higher for n-C5, and this results in lower oxygenates
and higher olefin selectivities than with n-C4. Butane oxy-
genate production maximizes at C/O 4.5 and then the ho-
mogeneous chemistry extinguishes. Pentane, however, con-
tinues without gas-phase chemistry extinction. Conversions
of C4H10 and C5H12 show that the two fuels behave in the
same manner with respect to conversions and production
of olefins and oxygenates. This was also confirmed by re-
sults shown in Fig. 9, where both fuels show experimental
behavior with % N2 feed variation which was theoretically
predicted by the same homogeneous chemistry mechanism.
In contrast, propane behaves much differently in that ex-
perimental conversion and selectivities to oxygenates and
4.2. Modeling of Propane Oxidation
The selective production of propylene rather than ethy-
lene from propane feed is of economic importance. The
major problem encountered in most existing processes is
C–C cracking, which leads to ethylene and CH4 (5, 8, 12).
It was found that the optimum selectivity ratio of C3/C2
olefins can be attained over the 80-mesh gauze. At a C3/O2
ratio 2.8 and 20% nitrogen dilution, oxygen conversion
was complete, the fuel conversion was 20% , C3H6 selectiv-
ity was 37% , and the C2H4 selectivity was 18% at an exit
gas temperature of 630 C (Fig. 2). Selectivity to methane
was 6% .

OXYGENATES AND OLEFINS FROM ALKANES
407
issues delay the advancement of such modeling efforts: De-
tailed surface mechanisms for hydrocarbon oxidation on Pt
and Rh are still under development. Current research ef-
forts focus on CH4 surface reactions. Single-gauze systems
have economic potential for C3 and higher oxidation sys-
tems. Until surface reaction mechanisms for these higher
hydrocarbons are established, any detailed modeling at-
tempts will be subject to skepticism. Gas-phase reaction
mechanisms are also under development, even though cur-
rent models can adequately simulate reactions of heavier
hydrocarbons. Finally, even if reliable mechanisms existed
for both surface and gas chemistry, the computational re-
quirements for a 3-D detailed calculation are tremendous.
We are extensively working in this direction and have re-
cently performed 2-D simulations of methane partial ox-
idation in Pt- and Rh-coated monoliths with surface and
gas-phase chemistry (27) and 2-D simulations of methane
partial oxidation over Pt gauzes with surface chemistry
only (28).
To obtain a measure of how well existing chemistry
models agree with our experimental results, we performed
simple zero-dimensional (0-D) calculations, assuming only
homogeneous reactions and plug flow at a specified tem-
perature profile, using a detailed homogeneous chemistry
mechanism developed by Warnatz and co-workers (14,
29, 30, 31). The temperature profiles used were based on
2-D calculations which simulated the flow of cold gas over
hot wires (no chemistry). This mechanism incorporates
240 species including various aliphatic and aromatic com-
pounds and 2000 chemical reactions. Computer experi-
ments were performed at a C3/O2 ratio of 2.8, 20% nitro-
gen dilution, and a final temperature of 900 K, using the
following temperature profiles: (1) Constant temperature
of 900 K; (2) initial temperature of 1200 K and a step change
FIG. 8. Reaction path analysis for the homogeneous production of
oxygenates from n-butane.
Equilibrium. At a C3/O2 feed ratio of 2.8 and 20% ni-
trogen dilution, we calculated the equilibrium composition
at 630 C. As shown in Table 1, we found that at equilib-
rium both oxygen and propane conversions are complete,
while C3H6 and C2H4 selectivities are only 6 and 12% , re-
spectively. Equilibrium favors methane and CO formation,
giving 56% CH4 selectivity and 24% CO selectivity. Thus,
we conclude that the experimental products are far from
equilibrium and contain greater amounts of olefins than
those predicted by thermodynamics.
TABLE 1
Comparison between Gauze Reactor Experiments, Homogene-
ous Modeling Calculations, and Equilibrium, of Propane Conver-
sions and Main Product Selectivities for a C3/O2 Feed Ratio of 2.8
(Modeling Calculations Included Heterogeneous CO2 Production
Prior to Gas-Phase Chemistry)
Fuel/Oxygen Ratio = 2.8
Modeling. Single-gauze systems require complicated
models to accurately predict and elucidate all the phenom-
ena that take place during reactor operation. A rigorous
modeling approach would involve three-dimensional (3-D)
simulations (rather than two-dimensional (2-D) where ad-
ditional assumptions should apply) of flow past woven
Pt/Rh wires, involving detailed catalytic and homogeneous
chemistry. An investigation of potential chemical interac-
tions between the surface and gas phase is important. Sharp
temperature gradients and mixing after the wires can dras-
tically affect final product distribution. However, several
Experiments, 900 K
Pt–10% Rh gauze
Model
900 K, 34.5 ms
Equilibrium
900 K
% C3H8 conversion
% O2 conversion
% Selectivities
(C atom basis)
CH4
C2H4
C3H6
CO
CO2
18.0
98.0
24.7
99.9
99.9
100.0
6
18
37
5
3.9
18.3
36.6
1.1
56.2
12
6
23.8
<0.01
28
19.8

408
IORDANOGLOU, BODKE, AND SCHMIDT
to 900 K after 10 s; (3) initial temperature of 1200 K and
a step change to 900 K after 100 s. We found only a mi-
nor influence of the initial temperature profile. In all three
cases, the species concentration profiles were calculated
as a function of reaction time. The propylene and ethy-
lene selectivities were almost identical to the experimental
values, as was the calculated oxygen conversion and CO
selectivity. An important difference was that the model pre-
dicted higher fuel conversion than that seen in the exper-
iment. The selectivity to CO2 was also different between
calculations and experiments, with simulations predicting
very small amounts. As reported earlier (3) this is expected
since CO2 is believed to be formed primarily on the catalyst
surface.
Homogeneous modeling with heterogeneous combustion.
To test this hypothesis, we ran simulations where we mim-
icked the catalytic chemistry, assuming that initially a frac-
tion of the C3H8 and O2 instantly reacts on the surface to
produce CO2 and H2O. The total amount of CO2 produced
was taken to be equal to the experimental value. Then, the
resulting mixture was allowed to react homogeneously at
900 K up to a residence time where the C3H6/C2H4 ratio was
equal to 2. This combined heterogeneous–homogeneous
calculation gives a much better agreement with the experi-
mental results than the purely homogeneous one (Table 1).
The only relatively large difference with the experiment is
the residence time. In the experiment, contact times are of
the order of 1 ms, while simulations predict a reaction time
of around 35 ms. However, this discrepancy is not as big
as that noticed previously with n-butane calculations (3).
We suggest that certain radicals are produced catalytically
and speed up the homogeneous ignition. This suggestion
was reasonably validated when we seeded radicals like O,
OH, H, and n-C3H7 and calculated residence times of 19 ms.
However, this caused a slight decrease in C3H6 and C2H4
selectivities.
Overall, the following mechanism is proposed: The feed
enters the reactor and part of it reacts catalytically on the
gauze surface to produce primarily total oxidation prod-
ucts (CO2 and H2O), releasing a significant amount of heat.
The remaining part of the feed reacts homogeneously at
these high reaction temperatures to yield the final products.
A similar mechanism was suggested for the production of
olefins and oxygenates from the C4 feed (3).
FIG. 9. Product selectivities vs feed N2 dilution for n-butane and n-
pentane. Figure 9a shows results for n-butane partial oxidation over a
single 80-mesh Pt–10% Rh gauze and Figs. 9b and 9c show results for n-
pentane partial oxidation over a single 40-mesh Pt–10% Rh gauze reactor,
for two different fuel/oxygen feed ratios.
4.3. Effect of Nitrogen Dilution
The effect of N2 dilution on oxygenate production from
n-butane is shown in Fig. 9a. We recorded a small decrease fuel feed (calculations with n-pentane were not possible
in oxygenate selectivity for dilutions less than 15% . For because the mechanism does not include n-C5H12). When
values higher than 15% , a steady but gradual decrease in oxygenate selectivity vs N2 dilution was calculated, three
the oxygenate selectivity was seen.
types of behavior were recorded: (1) monotonic decrease
We speculated earlier that this should be an effect of of SOX with increasingly negative slope; (2) increase of SOX
the homogeneous chemistry (3). We performed computer toward a maximum and then decrease; and (3) monotonic
simulations with the same model as before, using n-butane increase of SOX
.

OXYGENATES AND OLEFINS FROM ALKANES
409
The predicted behavior (1) was experimentally observed
REFERENCES
in n-butane oxidation and is shown in Fig. 9a. Assuming
that the chemistry-producing oxygenates is similar for all
higher hydrocarbons, it should also be possible to observe
behaviors like (2) and (3).
To verify this, experiments were performed with n-C5 fuel
with varying dilutions. The results are shown in Figs. 9b
and 9c, and it is seen that the experimental trends do fol-
low those predicted in the simulations. This suggests that
n-C5 oxidation has similar chemistry for oxygenate forma-
tion as that of n-C4 and that N2 dilution can be used as an
important optimization parameter to maximize oxygenate
selectivities.
1. Goetsch, D. A., and Schmidt, L. D., Science 271, 1560 (1996).
2. Goetsch, D. A., Witt, P. M., and Schmidt, L. D., Heterogeneous
Hydrocarbon Oxidation 125 (1996).
3. Iordanoglou, D. I., and Schmidt, L. D., J. Catal. 176, 503 (1998).
4. Sanfilippo, D., 7th Venezuelan Joint Venture Meeting, 1996.
5. Kirk-Othmer, “Encyclopedia of Chemical Technology.” Wiley, New
York, 1993.
6. Hickman, D. A., and Schmidt, L. D., Science 259, 343 (1993).
7. Huff, M., and Schmidt, L. D., J. Phys. Chem. 97, 11815 (1993).
8. Huff, M., and Schmidt, L. D., J. Catal. 149, 127 (1994).
9. Dietz, A., Carlson, A., and Schmidt, L. D., J. Catal. 176, 459 (1998).
10. Satterfield, C. N., “Heterogeneous Catalysis in Industrial Practice.”
McGraw-Hill Inc., New York, 1991.
11. Twigg, M. V., “Catalyst Handbook.” Manson Publishing, Ltd., London,
1996.
5. CONCLUSIONS
12. Cooney, D. O., and Zhenpeng, X., Fuel Sci. Technol. Int. 14, 1315
(1996).
13. Roine, A., “Ver. 1.12, Outokumpu Research Oy, Pori, Finland, 1993.”
14. Warnatz, J., private communication, 1998.
15. Bodke, A. S., Olschki, D., Schmidt, L. D., Bharadwaj, S. S., Siddall, J.,
and Ranzi, E., Science 285, 712 (1999).
16. Chan, K. Y. G., Inal, F., and Senkan, S., Ind. Eng. Chem. Res. 37, 901
(1998).
17. Zhang, W., and Smirniotis, P. G., J. Catal. 182, 400 (1999).
18. Zhao, Y. X., and Wojciechowski, B. W., J. Catal. 163, 365 (1996).
19. Zhao, Y. X., and Wojciechowski, B. W., J. Catal. 163, 374 (1996).
20. Flick, D., and Huff, M. C., J. Catal. 178, 315 (1998).
21. Khodakov, A., Yang, J., Su, S., Iglesia, E., and Bell, A. T., J. Catal. 177,
343 (1998).
22. Blasco, T., Galli, A., Nieto, J. M. L., and Trifiro, F., J. Catal. 169, 203
(1997).
23. Gomez, M. F., Arrua, L. A., and Abello, M. C., Ind. Eng. Chem. Res.
36, 3468 (1997).
A single Pt–10% Rh gauze reactor can be used for rapid
production of chemicals such as olefins and oxygenates. Ex-
periments and modeling from this and previous work sug-
gest that a gauze reactor combines surface and gas-phase
chemistry by oxidizing a part of the fuel on the catalyst sur-
face to produce primarily CO2 and release a lot of heat. This
energy can be transferred to the remaining colder portion of
the feed to ignite a homogeneous reaction sequence, which
subsequently produces olefins and oxygenates. Simultane-
ously, fast quenching due to mixing lowers the exit temper-
ature and prevents the decomposition of oxygenates.
Our reactor could process 6.5 kg/day of pentane and
produce oxygenated species with carbon yields up to 1 kg/
day, from a 15-mm gauze. If that was replaced with a 1-m
diameter gauze reactor, the production would rise up to
4 t/day. It is obvious that, after optimization and recycling,
a plant using this technology could produce large amounts
of oxygenates and olefins with one relatively cheap reactor,
avoiding expensive multistage processes and exotic cata-
lysts.
24. Pestman, R., Koster, R. M., Pieterse, J. A. Z., and Ponec, V., J. Catal.
168, 255 (1997).
25. Parmaliana, A., and Arena, F., J. Catal. 167, 57 (1997).
26. Yang, C., Xu, N., and Shi, J., Ind. Eng. Chem. Res. 37, 2601 (1998).
27. Deutschmann, O., and Schmidt, L. D., in “Twenty-Seventh Symposium
(International) on Combustion,” Boulder, CO, 1998.
28. Iordanoglou, D. I., “Rapid Oxidation ofLight Alkanesin Single Gauze
Reactors.” Ph.D. thesis, University of Minnesota, Minneapolis, MN,
1998.
29. Nehse, M., “Automatische Erstellung von detaillierten Reaktion-
smechanismen zur Modellierung der Selbstzuendung und lam-
inarer Vormischflammen von gasfoermigen Kohlenwassertsoff-
Mischungen.” Ph.D. thesis, Heidelberg University, 1999.
30. Warnatz, J., Maas, U., and Dibble, R. W., “Combustion, Physical and
Chemical Fundamentals, Modeling and Simulation, Experiments, Pol-
lutant Formation.” Springer, New York, 1996.
Experiments are in progress to study the oxidation of
higher hydrocarbons, such as hexane and cyclohexane.
ACKNOWLEDGMENTS
The authors would like to thank Dr. Olaf Deutschmann and Dr.
Jurgen Warnatz of Heidelberg University for providing the mechanism
and computer code for homogeneous calculations and valuable assistance. 31. Deutschmann, O., private communication, 1999.