Catalysis Letters Vol. 99, Nos.

34, February 2005 ( 2005) 113

DOI: 10.1007/s10562-005-2103-y

Catalytic partial oxidation of methane to syngas at elevated pressures

Maxim Lyubovsky, Subir Roychoudhury and Rene LaPierre
Precision Combustion, Inc., 410 Sackett Point Road, North Haven, CT 06473

Received 27 May 2004; accepted 21 September 2004

In this work we demonstrate catalytic partial oxidation of methane into syngas at pressures up to 0.8 MPa, power densities up
to 15 MW/L and selectivity greater than 85%. The product composition proles indicate high initial selectivity to CO and low
initial selectivity to H2 , suggesting direct partial Oxidation of methane primarily into CO and water, while most hydrogen is
produced in the consecutive steam reforming of methane.
KEY WORDS: catalytic Partial Oxidation; natural Gas; syngas; microlith.

1. Introduction Catalytic partial oxidation (CPOX) of methane is a

relatively inexpensive alternative to syngas generation.
Converting natural gas into a mixture of H2 and CO,
In the CPOX process, generally represented by equation
known as synthesis gas or syngas, is an important
(2), methane is reacted with oxygen over a catalyst bed
intermediate step in many existing and emerging energy
to yield syngas with H2 to CO ratio near 2. Examples of
conversion technologies (gure 1). In recent years Gas-
recent publications in this area are those of D.A.
to-Liquids (GTL) process is becoming a mature
Hickman and L.D. Schmidt who demonstrated that
technology attracting multi-billion dollar investments.
near comp1ete conversion of methane to mostly
For example, in October of 2003 Shell and Qatar
hydrogen and carbon monoxide could be achieved at
Petroleum signed an agreement for construction of the
reaction times as short as 1 ms, promising dramatic
largest to-date GTL plant with a production capacity of
reduction in a reactor size and complexity, as compared
140,000 barrels per day, in which Shell will invest some
to existing syngas production technologies [9,10].
5 billion [1]. Construction of several smaller GTL plants
having 70,00075,000 barrel per day production capa- CH4 H2 O ! CO 3H2 DH 206 kJ/mole 1
city has also been announced [2]. Preparation of
synthesis gas is the most capital-intensive part of this 1
size GTL plant, accounting for 5075% of the capital CH4 O2 ! CO 2H2 DH
38 kJ/mole 2
2
cost [3]. In addition, reforming natural gas or liquid
hydrocarbons into syngas is the initial step in all fuel To avoid the energy losses associated with compression
reforming systems being developed to power the of hot, high hydrogen content, gases the CPOX reactor
hydrogen economy of the future [4]. In hydrogen should be operated at pressures in the range of 0.5 to
Production systems syngas generation is generally 4 Mpa compatible with downstream processes (gure 1).
followed by a water gas shift (WGS) and a CO clean Surprisingly, despite much interest over the last decade
up reactors or a hydrogen separation device to yield fuel and extensive development of this technology by the
cell quality hydrogen [5,6]. In combustion systems
operating on natural gas, such as power turbines and
utility burners, preliminary conversion of methane into
syngas can help in stabilizing lower temperature ames,
reducing NOx emissions [7].
Currently most syngas is produced by steam reform-
ing (SR). In this endothermic reaction, generally
represented by equation (1), methane is contacted with
steam over a heated catalyst at high pressures and
temperatures to produce a high hydrogen content
syngas. Although, SR is a mature technology, it requires
large heat exchange reactors, demanding large initial
investment [8]. Furthermore, stringent heat balance Figure 1. Routes for natural gas utilization opened or facilitated by
requirements in the SR process make scaling these conversion of natural gas into syngas in a high power density CPOX
reactors to smaller sizes extremely difcult. rector.

1011-372X/05/02000113/0 2005 Springer ScienceBusiness Media, Inc.

114 M. Lyubovsky et al./ Oxidation of methane to syngas

major petrochemical companies, such as Shell, Exxon- 6.3 cm3 amounting to the total amount of precious
Mobil, ConocoPhillips, only very limited work on high- metal loading of 0.25 g.
pressure CPOX of methane has been reported in Prior to the start up the catalyst was pre-heated to
scientic literature [11,12,13]. Furthermore, in these 250  C by passing the ow of hot air through the
works testing at very small scale [11], augmentation with reactor. Ignition of the reaction occurred as soon as the
steam and CO2 [12] or exotic catalyst structures, such as methane/air mixture was introduced and the reactor
honeycombs made of fused Rh foils [13], were used, achieved steady state operation in less than 15 s
making the results only indirectly applicable to practical (gure 2).
applications of the CPOX process.
In this work, we demonstrate robust, steady-state,
operation of a methane CPOX reactor at pressures up 3. Results and discussion
to 0.8 MPa with near equilibrium conversion of
In all experiments reported here the ow rates of
methane to above 90% and process selectivity about
methane and air were increased proportionally to the
90%. A series of discrete, metallic screen substrates
operating Pressure, to maintain constant linear ow
(trademarked by Precision Combustion, Inc. as
velocity and residence time. In the experiment at
Microlith ) [14] coated with a high surface area
P 0:8 Mpa, methane was fed into the reactor at a
ceramic washcoat and impregnated with Rh were used
rate of 160 standard liters per minute (SLPM) (0.11
to construct a catalytic reactor. A characteristic feature
mole. CH4 per second), which is equivalent to 88 kWt of
of these substrates is the short length of each element,
thermal power input. Normalized to the reactor volume
which prevents development of a boundary layer over
this equals 14 MWt /L power density in the tested
the catalytic surface greatly enhancing heat and mass
CPOX reactor.
transfer coefcients for the system [15,16]. These
At this power density a 1 L CPOX reactor applied to
properties of the Microlith based catalytic systems
the GTL process can produce enough syngas to make
make them perfectly suited for short contact time
100 barrels Per day of synthetic fuel (assuming 50%
oxidation reactions. In recent years such catalytic
overall conversion efciency of CH4 into (CH2 )n and
systems have been applied to catalytic combustion
0.8 g/cm3 density of the produced Synthetic oil).
[17], automotive exhaust aftertreatment [18] and fuel
Alternatively, for a fuel cell power system having
reforming [19]. Microlith based catalysts are relatively
35% overall efciency this amount of syngas may be
inexpensive, manufactured via an automated contin-
converted into hydrogen to fuel 5 MWe PEM fuel cell.
uous production process and allow a high degree of
For an estimated 50 kWe fuel cell powered mid-sized
exibility in reactor design, making them readily
car this is equivalent to the amount of hydrogen
adaptable for applications on an industrial scale.
consumed by a eet of 100 cars. In gas turbine
applications a 1 L CPOX reactor can produce enough
hydrogen (without additional WGS reactor) to provide
2. Experimental 10% augmentation to natura1 gas fuel for a 100 MW
gas turbine (assuming 35% turbine efciency) for ame
The reactor used in this work was built by stacking
stability and reduced NOx emissions.
Microlith screens with catalyst loading of 5.5 wt%.
Gas composition and temperature proles measured
Gas sample probes made of 1/16 OD high temperature
along the length of the catalyst bed (gure 3)
alloy tubing and 0.02 K-type thermocouples were
inserted between the screens at several locations along
the length of the reactor. Gas samples extracted by the
sample probes were passed through a chiller to remove
water vapor and analyzed by an on-line gas chromato-
graph. Concentrations of H2 , N2 , O2 , CH4 , CO, CO2 ,
C2 H4 and C2 H6 species were measured. Nitrogen was
assumed to be internal in the reaction and was used as
an internal standard. Water concentration in the sample
was calculated using hydrogen balance. Overall mass
balance of all species was within 10%.
The total length of the reactor was 6 cm (including
the length of the partitions, required to install the
instrumentation). The length of the catalyst itself was
2 cm, and this length is used below in the reactor prole
Plots (gure 3) and in the estimates of the reactor
volume and power density. The gas ow diameter was Figure 2. Light off of the Microlith based COPX of methane reactor.
20 mm, such that the volume of the catalyst bed was P 0:6 Mpa.
 M. Lyubovsky et al./Oxidation of methane to syngas 115

Under all tested operating pressures essentially all of the

oxygen was consumed by 4 mm into the catalyst bed
(corresponding residence time 0.5 msec), at which
point, methane conversion was between 55 and 75%. It
was also found that at constant inlet ow velocity the
reaction rate linearly increases with pressure between 0.2
and 0.8 MPa. For multi-step processes, such as partial
oxidation, rst order dependence of the kinetic rate on
pressure is unlikely. Scaling of the reaction rate with
pressure, therefore, suggests that the CPOX process rate
is limited by the transport of oxygenthe limiting
reactantacross the boundary layer to the surface of
the catalyst.
Surprisingly, much more water than CO2 was formed
at the front of the reactor. A large peak in the H2 O
prole is consistent with the indirect pathway for the
CPOX process, in which complete oxidation of methane
on the front of the reactor is followed by reforming
reactions [20]. On the other hand, low and near constant
concentration of CO2 along the reactor, is consistent
with direct partial oxidation pathway in which partial
oxidation products directly form on the front of the
reactor [10,21,22]. Linear extrapolation of product
selectivity vs. conversion plots to zero methane conver-
sion (gure 4) suggests initial partial oxidation reaction
as represented by equation (4). Therefore, CO is directly
produced with high selectivity on the oxidation step at
the onset of the CPOX process. However, selectivity for
direct production of H2 is low, such that high H2
selectivity of the CPOX process is mainly due to the
secondary steam reforming reaction.

CH4 1:34O2 ! 0.82CO 0.18CO2 0:5H2 1:5H2 O

DH
449 kJ/mole 4

A detailed model for the CPOX process based on

Figure 3. Species concentration and temperature proles along the literature data for the kinetic parameters of multiple
length of the catalyst bed at 0.2, 0.4 and 0.8 Mpa and O:C = 1.2. surface elementary reactions, proposed by L.D. Schmidt
Concentration of N2 in the sample was used to convert species et al. [23,24,25] did not predict the dierence in the
concentrations into the ow rate units (mmole/sec) based on the
known rate of air supply to the reactor inlet and assuming that
initial selectivity for CO and H2 observed in our
nitrogen remains inert under the reaction conditions. Values at x 0 experiments. This discrepancy can be attributed to the
position are made of the inlet methane and air ow as measured by the two assumptions made in the Schmidt model: (1) of
mass ow meters. reversible adsorption of O2 onto the surface of the
catalyst and (2) of high activation energy for the surface
demonstrate that in spite of very short contact time, the reaction between the adsorbed O* and H* species on the
CPOX process does not occur in a single reaction, but Rh catalyst, eectively eliminating the surface hydrogen
consists of sequential oxidation and reforming steps. oxidation pathway.
While oxygen to methane ratio in the feed stream is At low coverages of all surface species characteristic
O2 :CH4 = 0.6, oxygen is consume at a rate up to 1.34 of a high temperatures CPOX process and with the
times higher than methane at the front of the reactor. oxygen adsorption rate being limited by the rate of mass
This stoichiometry suggests that simultaneous complete transfer, adsorption of both methane and oxygen on the
and partial oxidation reactions occur at the onset of the surface of the catalyst is most likely irreversible.
CPOX process. (For the partial oxidation reaction (2) Therefore, all molecules dissociated onto the surface
the O2 :CH4 consumption ratio equals 0.5, while for are consumed in the oxidation reaction, Equation (4)
complete oxidation, equation (3), it equals 2). indicates that the ratio between partial and complete
oxidation products is determined by the relative rates of
CH4 2O2 ! CO2 2H2 O DH
802 kJ/mole 3 methane and oxygen dissociation onto the surface. The
116 M. Lyubovsky et al./ Oxidation of methane to syngas

rate of CO and H oxidation observed in the experi-

ment can be explained by the difference in the desorp-
tion mechanism for these species. At low surface
coverages rst order CO desorption is much faster
than the second order H desorption. This leads to
higher steady state surface coverage of H species,
making them more susceptible for reaction with O to
produce OH and water.
Before oxygen is consumed from the gas phase on the
front of the reactor homogeneous oxidation reactions
may take place. In these non-selective free radical
reactions methane, hydrogen and carbon monoxide
can be oxidized to CO2 and H2 O decreasing selectivity
of the CPOX process and increasing the peak temper-
Figure 4. Selectivity to complete and partial oxidation products as a ature. Good t by a single line of all data points
function of methane conversion. Calculated based on the species collected in the experiments with a 4-fold difference in
concentrations shown in Figure 3 for all tested conditions.
SH2 H2 =H2 H2 O; SH2 O H2 O=H2 H2 O; S(CO) CO= the operating pressure (gure 4) demonstrates that in
CO CO2 ; S(CO2 CO2 =CO CO2 : the tested pressure interval the gas phase reactions were
insignicant. Excellent transport properties of the
Microlith substrate ensure high rate of the mass transfer
mass transfer limited rate of oxygen adsorption has only limited surface reactions and free radical quenching,
weak dependence on temperature, while the rate of such that high selectivity to partial oxidation products
methane adsorption increases exponentially with tem- can be achieved even at high operating pressure. In fact,
perature. Therefore, selectivity towards partial oxida- it was found that peak temperature decreased slightly
tion products on the oxidation step of the CPOX when the process pressure was increased from 0.2 to
process increases exponentially with the increase in the 0.4 and to 0.8 MPa (gure 5).
catalyst temperature. This was also observed in our Under all tested conditions methane conversion
previous work on methane oxidation under fuel rich increased linearly with increasing O:C ratio (gure 5).
conditions over the catalysts with externally controlled The decrease in the peak temperature with pressure
surface temperature [26]. allowed compensation for decreasing methane conver-
For the intermediate surface speciesCO and H sion by operating the reactor at higher O:C ratio, such
two competing pathways exist-to desorb as molecular that conversion above 90% could be reached even at
CO and H2 forming partial oxidation products; or to be P 0:8 MPa. This also suggests the likelihood of
further oxidized to CO2 and H2 O. The difference in the successful operation of the CPOX process at even

Figure 5. Dependence of methane conversion and peak TC reading on the amount of air added to the reactor. Methane ow 40 SLPM at
0.2 MPa, 80 SLPM at 0.4 Mpa and 160 SLPM at 0.8 MPa constant throughout each test. Air ow is increased stepwise to increase O:C ratio.
Dotted lines show the thermodynamic prediction for methane conversion in adiabatic process under these inlet conditions.
 M. Lyubovsky et al./Oxidation of methane to syngas 117

higher pressures. (Note that in our experiments the [3] K. Aasberg-Petersen, T.S. Christensen, C. Stub Nielsen and
pressure was limited to 0.8 MPa by the design of the test I. Dybkjaer, Fuel Process. Technol. 83 (2003) 253.
[4] D.L. Trinm and Z.I. Onsan, Catal. Rev. 43 (2001) 31.
rig, but not by the observed performance of the
[5] S. Ahmed and M. Krumpelt, Int. J. Hydrogen Energ. 26 (2001)
catalyst). Although gure 5 shows that methane con- 291.
version did not reach the equilibrium value over the [6] D.G. Loer, K. Taylor and D. Mason, J. Power Sources 117
length of the reactor, we believe that minimizing heat (2003) 8491.
losses would allow bringing the conversion closer to the [7] M.B. Colket III, A.S. Kesten, J.J. Sangiovanni, M.F. Zabielski,
D.R. Pandy and D.J. Seery, US Patent #5,235, 804 (1993) 12.
equilibrium values. After oxygen consumption at the
[8] K. Aasberg-Petersen, J.-H. Bak Hansen, T.S. Christensen,
front of the reactor, methane conversion proceeds I. Dybkjder, P. Seier Christensen, C. Stub Nielsen, S.E.L. Winter
through a slower SR reaction. Therefore, increasing Madsen and J.R. Rostrup-Nielsen, Appl. Catal. A Gen. 221
the length of the catalyst bed and adding steam to the (2001) 379.
reactor inlet may also help bring the nal conversion [9] D.A. Hickman and L.D. Schmidt, J. Catal. 138 (1992) 267.
[10] D.A. Hickman and L.D. Schumidt, Science 259 (1993) 343.
closer to the equilibrium point.
[11] A.G. Dietz and L.D. Schmidt, Catal. Lett. 33 (1995) 15.
In addition to smaller size, control over a CPOX [12] L. Basini, K. Aasbergni Petersen, A. Guarinoni and M. [sberg,
reactor is simpler than over a SR reactor. In the SR Catal. Today 64 (2001) 9.
process assuring balance between heat generation on the [13] M. Fichtner, I. Mayer, D. Wolf and A. Schubert, Ind. Eng. Chem.
combustion side and heat consumption on the reforming Res. 40 (2001) 3475.
[14] Microlith Catalytic Reaction System. U.S. Patent 5,051, 241.
side along the entire length of the reactor requires
Sept. 24, 1991.
intensive control over the ow rates and ow distribu- [15] S. Roychoudhury, J. Bianchi, G. Muench and W.C. Pfeerle,
tion. In the CPOX process, performance is essentially SAE 971023, (SAE Intl, Warrendale, PA, 1997).
determined by the methane to oxygen ratio in the inlet [16] M. Lyubovsky, H. Karim, P. Menacherry, S. Boorse,
mixture, such that only this parameter should be R. LaPierre, W.C. Pfeede and S. Roychoudhury, Catal.
Today 83 (2003) 183.
controlled. Furthermore, at the extremely high power
[17] G. Kraemer, T. Strickland, W.C Pfeerle and J. Ritter in: Proc.,
density for the CPOX process (in excess of 14 MW/L) EC-Vol. 5, Proceedings of the International Joint Power Generation
small, simple and inexpensive bench scale reactors can Conference, (eds) A. Gupta, A. Sanyal and J. Veilleux (ASME
be directly used to produce syngas for industrial scale International, Book No. G01072, 1997).
applications. This will promote utilization of natural gas [18] Fast Lighto Non-electrically and Electrically Heated Microlith
Catalytic Converter. S. Roychoudhury, D. Hixon, W. Pfeerle
in a broad range of efcient and environmentally
and R.E. Gibbs, et al. SAE 940467, (SAE Intl, Warrendale, PA,
friendly energy generation processes. 1994).
[19] M. Castaldi, M. Lyubovsky, R. LaPierre, W.C. Pfeerle and S.
Roychoudhury, SAE, 200301-1366.
Acknowledgments [20] O.V. Buyevskaya, K. Walter, D. Wolf and M. Baerns, Catal. Lett.
38 (1996) 81.
The authors are grateful to Mr Andy Heeps and tie
[21] D. Wang, O. Dewaele, A.M. De Groote and G.F. Froment,
technical support group at PCI for help in designing and J. Catal. 159 (1996) 418.
building the apparatus for the experiments described in [22] M. Fathi, F. Monnet, Y. Schuurman, A. Holmben and
this paper. C. Mirodatos, J. Catal. 190 (2000) 439.
[23] D.A. Hickman and L.D. Schmidt, AIChE J. 39 (1993) 1164.
[24] O. Deutschmann and L.D. Schmidt, AIChE J. 44 (1998) 2465.
[25] C.T. Goralski Jr., R.P. OConnor and L.D. Schmidt, Chem. Eng.
References
Sci. 55 (2000) 1357.
[1] Gas-to-Liquid comes of age; Shell in the Middle East magazine, [26] M. Lyubovsky, L. Smith, M. Castaldi, H. Karim, B. Nentwick,
#24, Jan. 2004. www.shell-me.com/english/jan04/cover-story.htm S. Etemad, R. LaPierre and W.C. Pfeerle, Catal. Today 83
[2] S.L. Weeden, Oil Gas J. 99 (March 12, 2001) 58. (2003) 71.