Energy Conversion and Management 47 (2006) 560569

www.elsevier.com/locate/enconman

Technologies for direct production of exible H2/CO

synthesis gas
Xueping Song, Zhancheng Guo *

Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100080, PR China

Received 26 October 2004; accepted 18 May 2005

Available online 14 July 2005

Abstract

The use of synthesis gas oers the opportunity to furnish a broad range of environmentally clean fuels and
high value chemicals. However, synthesis gas manufacturing systems based on natural gas are capital inten-
sive, and hence, there is great interest in technologies for cost eective synthesis gas production. Direct pro-
duction of synthesis gas with exible H2/CO ratio, which is in agreement with the stoichiometric ratios
required by major synthesis gas based petrochemicals, can decrease the capital investment as well as the oper-
ating cost. Although CO2 reforming and catalytic partial oxidation can directly produce desirable H2/CO
synthesis gas, they are complicated and continued studies are necessary. In fact, direct production of exible
H2/CO synthesis gas can be obtained by optimizing the process schemes based on steam reforming and auto-
thermal reforming as well as partial oxidation. This paper reviews the state of the art of the technologies.
2005 Elsevier Ltd. All rights reserved.

Keywords: Natural gas; Synthesis gas; Steam reforming; Autothermal reforming; Partial oxidation

1. Introduction

Synthesis gas, a mixture of hydrogen and carbon monoxide usually containing carbon dioxide
as additives, is a very good intermediate product for further syntheses. Depending on the reaction

*
Corresponding author. Tel.: +86 010 62579402; fax: +86 010 62561822.
E-mail address: [email protected] (Z. Guo).

0196-8904/$ - see front matter 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enconman.2005.05.012
 X. Song, Z. Guo / Energy Conversion and Management 47 (2006) 560569 561

conditions and the catalyst used, dierent chemicals may be produced on a large industrial scale
[15]. The conversion of synthesis gas to liquid fuels, which are sulphur free and have high cetane
or octane numbers, via FischerTropsch (FT) synthesis has been interesting ever since 1923.
Methanol, one of the most industrially important chemicals, together with high alcohols can be
simultaneously produced from synthesis gas via many dierent catalysts. High value chemicals
such as aldehydes or alcohols containing one more -CHO- or -CH2OH- group can be obtained
through the hydroformylation (oxo) reaction, which consists of direct reactions of olen and syn-
thesis gas. Chemicals containing the acetyl group (CH3CO, often designated as Ac), which are
produced through the carbonylation process, have experienced steady growth for 30 years and
will increase dramatically over the next 10 years [6].
The required properties of the synthesis gas are dierent for the dierent syntheses [79]. In the
synthesis of methanol, CO2 and CO are both reactants because the water gas shift reaction is
active in addition to the synthesis. Therefore, a module M = (H2
CO2)/(CO + CO2) that is
derived from simple stoichiometric calculations, should be close to 2.0. In contrast, for FT syn-
theses, the desirable synthesis gas composition is best characterized by a H2/CO ratio of about 2.0
because the water gas shift reaction is not active and the CO2 remains inert, whereas the optimal
H2/CO ratio should be 1.0 for the oxo-synthesis process. In general, the requirement of the H2/CO
ratio is in a range of about 2.0 to 1.0, and sometimes, even pure CO is required for the carbon-
ylation process.
The generation of synthesis gas may, from a point of view of feedstock used, be categorized in
two dierent types of processes. One is based on natural gas or high hydrocarbons and the other is
coal. However, natural gas and high hydrocarbons will remain the major feedstock for the man-
ufacture of synthesis gas due to its lower investment compared with a coal based plant [10]. With
decline of the oil resource, natural gas seems to be the most suitable for synthesis gas production.
However, economically viable processes based on natural gas for manufacturing synthesis gas typ-
ically produce a gas that is far too rich in hydrogen to meet the stoichiometric ratios required by
major synthesis gas based petrochemicals [6]. Thus, H2/CO adjustment units, such as the PSA
(pressure swing adsorption) unit or the reverse shift of hydrogen and CO2 unit, are necessary prior
to downstream synthesis (Fig. 1).
Synthesis gas manufacture may be responsible for ca. 60% of the investments of large scale gas
conversion plants based on natural gas [8]. Therefore, researches for reducing operating costs and
conserving energy have been the focus of synthesis gas production. Direct production of exible
H2/CO synthesis gas can avoid the H2/CO adjustment units, as shown in Fig. 1 (dotted line), and
thus, energy conservation and cost reduction is achieved. Hence, there is a strong incentive to nd
solutions for providing exible H2/CO ratios.

F-T synthesis
Synthesis H2/CO Methone
Natural gas
gas adjustment Oxo-synthesis
Carbonylation
Flexible H2/CO

Fig. 1. Synthesis gas route from natural gas to chemicals.

562 X. Song, Z. Guo / Energy Conversion and Management 47 (2006) 560569

Catalytic partial oxidation (CPO) (Eq. (1)) and CO2 reforming (Eq. (2)) can directly produce
desirable H2/CO ratio synthesis gas. CPO has been subjected to intensied research eorts in
recent years due to its ameless, mild exothermic reaction [1114]. However, the CPO reaction
is complicated, and a comprehensive understanding is presently lacking [15]. Researches on
CO2 reforming [1619] are also attractive because it converts two greenhouse gases into valuable
synthesis gas, but synthesis gas from CO2 reforming is considered to be rarely feasible [8,20].
In fact, direct production of exible H2/CO ratio synthesis gas can be obtained by optimization
of the existing processes based on steam reforming and autothermal reforming as well as partial
oxidation. In this paper, the technologies for direct production of exible H2/CO ratio synthesis
gas are discussed.
CH4 1=2O2 CO 2H2 DH 0298 K 38 kJ=mol 1
CH4 CO2 2H2 2CO DH 0298 K
247 kJ=mol 2

2. Technologies for exible H2/CO

2.1. Steam reforming

Steam reforming is a well established process for the manufacture of synthesis gas [10,2123].
The steam reforming reaction (Eq. (3)) is an endothermic reaction that takes place in a high alloy
tube reformer loaded with a nickel based catalyst, which is placed inside a furnace equipped with
burners in the side or top of the furnace. The process is typically operated with excess steam (H2O/
CH4 mole ratio 2:14:1) at temperatures of about 750 to 900 C depending on the use of the gas.
The exit raw synthesis gas composition is a mixture of hydrogen, CO, CO2 and unreacted meth-
ane, which is near the equilibrium of the steam reforming reaction and water gas shift (WGS)
reaction (Eq. (4)). A typical outlet gas composition is shown in Table 1, which corresponds to
a H2/CO ratio of about 4.9 [6].

CH4 H2 O 3H2 CO DH 0298 K
206 kJ=mol 3

CO H2 O H2 CO2 DH 0298 K 41 kJ=mol 4

Table 1
Typical reformer furnace outlet gas compositions
Furnace outlet temperature, C 890
Furnace outlet pressure, bar 24
S/C ratio 3.0
Component, vol.%
Hydrogen 51.0
Carbon monoxide 10.4
Carbon dioxide 5.0
Methane 2.0
Water vapor 31.6
 X. Song, Z. Guo / Energy Conversion and Management 47 (2006) 560569 563

Synthesis gas obtained from steam reforming is usually hydrogen rich (Table 1). An increase
in temperature, decrease in pressure and reduction of the steam to carbon (S/C) ratio will
lower the hydrogen to CO ratio. However, operating conditions can be adjusted only to reduce
the H2/CO ratio to a small extent [6]. Recycling CO2 or addition of CO2 in the feedstock is an
alternative to adjust the H2/CO ratio of the synthesis gas to a large extent. However, a problem asso-
ciated with CO2 addition in the feedstock is carbon formation by methane decomposition (Eq. (5))
and the Bouduard reaction (Eq. (6)), which causes catalyst deactivation as well as metal dusting cor-
rosion [24].
CH4 C 2H2 DH 0298 K
74.9 kJ=mol 5
2CO C CO2 DH 0298 K 172.4 kJ=mol 6
The relations between the H2/CO ratios of the exit gas and the thermodynamic carbon limit,
which is calculated by the atomic ratio O/C and H/C in the feed stream and the temperature
and pressure at the reformer exit, is shown in Fig. 2. The H2/CO ratios correspond to equilibrium
at the reformer outlet conditions of 900 C and 5 bar. For the O/C and H/C ratios on the left side
of the carbon limit curve, there is a thermodynamic potential for carbon formation in the equil-
ibrated gas [2527].

Fig. 2. Equilibrium chart of thermodynamic carbon limit.

564 X. Song, Z. Guo / Energy Conversion and Management 47 (2006) 560569

It can be seen that exible H2/CO ratio synthesis gas may be obtained by changing the oper-
ating conditionsthe H2O/CH4 ratio and the CO2/CH4 ratio. However, selection of the process
parameters is limited by the potential for carbon formation. It is obvious that exible H2/CO
ratios as well as more economic conditions are obtained when operated below the values indicated
by the curve, but there is the thermodynamic potential for the formation of carbon.
SPARG (sulfur passivated reforming) [2529] may be a solution to carry out the process over
noble metal catalysts without carbon formation on the left side of the carbon limit curve. The
SPARG principle allows carbon free operation below the carbon limit curve by ensemble con-
trol, which means that the sites for carbon formation are blocked while sucient sites for the
reforming reactions are maintained. This eect is obtained by adding sulfur to the process stream.
Usually, hydrogen sulde is chosen because it is a stable sulfur compound at the conditions for the
tubular reformer and because the catalyst activity can be easily controlled by increasing or low-
ering the sulfur content in the feed stream.
A pilot plant situated at the facilities of Haldor Topsoe Inc., Bayport, Texas, was developed to
test the eects of various process parameters on carbon formation and H2/CO ratios. SPARG
tests were performed in a full size monotube reformer, shown in Fig. 3, and a highly active catalyst
R-67 developed by Topsoe was used. The process conditions and the H2/CO ratios of the exit gas
are summarized in Table 2.
The pilot tests showed that the R-67 catalyst has sucient activity for an almost complete con-
version of the natural gas. Direct production of exible H2/CO ratio synthesis gas on the left side
of the carbon limit, as shown in Fig. 2, was obtained without carbon formation. A mild CO2 addi-
tion in the feed steam will produce synthesis gas of H2/CO = 1.8. However, the H2/CO ratio can
be further decreased by increasing the CO2 addition to the feed stream. For instance, operation at
CO2/CH4 = 2.43 and H2O/CH4 = 0.7 can result in a H2/CO ratio close to 0.6, but this means
more unconverted CO2 and a high recycle ratio.

CO2 (H2S)

CO2 import

Steam

Natural gas

Synthesis
CO2 Wash gas

Combustion air

Fig. 3. SPARG process ow diagram.

 X. Song, Z. Guo / Energy Conversion and Management 47 (2006) 560569 565

Table 2
Results from SPARG monotube demonstration test
Test T, C P, atm. H2O/C CO2/C CH4 leakage, dry mol.% H2/CO, mol./mol. Duration, h
A [25] 930 5.7 0.35 1.55 0.88 0.89 500
B [25] 630 14.5 0.60 1.66 1.7 0.92 160
C [25] 945 14.5 0.72 2.43 0.6 0.7 400
D [26] 880 6.2 0.93 0.65 2.27 1.89 500

2.2. Autothermal reforming

Autothermal reforming (ATR) is a stand alone process, which combines partial oxidation and
steam reforming in a single reactor. The ATR process was rst developed in the late 1950s by
Topsoe mainly for industrial synthesis gas production in ammonia and methanol plants [30].
However, new developments and optimizations have been made in the 1990s to produce cost eec-
tive synthesis gas for dierent employments [22,3133].
The ATR reactor, the heart of the whole process as shown in Fig. 4, is a refractory lined cylin-
drical vessel swaged to a smaller diameter at the top to provide a combustion zone. A specially
designed oxygen burner is installed in this section. The bottom, a larger diameter section, contains
the catalyst [6,31].
The natural gas feedstock and oxygen are sent to the reactor as two separate streams. Natural
gas enters at the top of and ows down through an annular space between the outer and inner tubes
of the oxygen burner. Oxygen and a small amount of steam ows through the inner tube and exits
through ports in the bottom of the tube [6]. At the combustion chamber, the two streams are grad-
ually mixed and combusted in a sub-stoichiometric ame, which is assumed mixed-is-burnt and
simplied, by Eq. (7). The hot gases ow through the catalyst bed in which the synthesis gas is fur-
ther equilibrated. Because the temperature leaving the combustion zone is approximately 1200 to

Oxygen/air
(or enriched air)
Feedstock
and steam

Combustion
chamber

Catalyst
bed

Synthesis
gas

Fig. 4. Diagram of an ATR reactor.

566 X. Song, Z. Guo / Energy Conversion and Management 47 (2006) 560569

1250 C, a high thermal stability catalyst and carrier system are required. A nickel catalyst sup-
ported by a magnesiumalumina spinel carrier has shown high stability and activity for ATR reac-
tors [32]. The temperature leaving the catalyst bed is 870 to 955 C, and the produced synthesis gas
is completely soot free [6,32].
CH4 3=2O2 CO 2H2 O DH 0298 K 519 kJ=mol 7
At the reactor outlet, the composition of the product gas will be determined by the thermody-
namic equilibrium of the steam-methane reforming and water gas shift reactions (Eqs. (3), (4)) at
the exit pressure and temperature, which are determined through the adiabatic heat balance based
on the composition and ow of the feed, steam and oxygen added to the reactor. Thermodynamic
calculation shows that for a process without CO2 feed, the H2/CO ratio under traditional oper-
ating conditions is in a range of 3.3 to 4.0. Direct production of a gas with H2/CO ratio about
2 is achieved only at very low S/C ratio and high exit temperature [32,33].
However, recycled or imported CO2 can reduce the H2/CO ratio to a large extent as shown in
Fig. 5 [32]. It can be seen that moderate CO2 addition, e.g. CO2/C = 0.0 to 0.5, has the largest
eect on the synthesis gas ratio. Total CO2 recycle will give a CO2/C ratio of 0.3 to 0.5 resulting
in a H2/CO = 1.7 to 2.0. Considerable excess CO2 must be available as an import if the H2/CO
ratio is to decrease below 1.0. In summary, the composition of the raw synthesis gas produced
by the ATR reactor can be adjusted within certain ranges by selection of the process conditions
and by applying recycling of CO2.
A pilot plant was built to test and demonstrate the ATR technology. The plant operated on
natural gas at a capacity corresponding to 100 Nm3/hr producing synthesis gas from 250 to
300 Nm3/hr. The oxygen source was pure oxygen, and the typical natural gas composition is

Fig. 5. Thermodynamic calculation results of ATR with CO2 import or recycle Pressure = 25 bar; Equilibrium
temperature = 1000 C; Natural gas = 100% CH4.
 X. Song, Z. Guo / Energy Conversion and Management 47 (2006) 560569 567

Table 3
Natural gas compositions for pilot plant tests
Component N2 CO2 CH4 C2H6 C3H8 C4H10 C5H12 R
Mol., % 0.55 1.10 95.10 2.45 0.50 0.20 0.10 100

Table 4
ATR pilot plant tests
Test A B C D E F
Feed ratios
H2O/C, mol/mola 1.43 0.59 0.86 0.68 0.67 0.58
CO2/C, mol/mol 0.0 0.0 0.0 0.47 0.75 0.98
O2/C, mol/molb 0.60 0.62 0.60 0.67 0.70 0.73
Product gas
Temperature, C 950 1065 1000 1018 1030 1028
Pressure, bar 24.5 24.5 24.5 24.5 24.5 24.5
H2/CO, mol/mol 2.97 2.25 2.47 1.49 1.19 0.99
CH4 leakage, dry mol% 1.00 0.42 0.90 0.32 0.15 0.11
a
Mole per mole of carbon atoms in natural gas feed.
b
The O2/C ratio is 510% higher than for typical industrial plants due to heat loss compensation.

shown in Table 3. The process conditions used in the pilot demonstration tests and the H2/CO
ratios of the product gas are listed in Table 4. No soot was formed when operated at these process
conditions in tests lasting between one week and one month [32].
Demonstrations A, B and C showed that when operated without CO2 addition or CO2 recycle,
the variation of the H2/CO ratio of the raw synthesis gas is only in a small range of about 2.23.0.
However, the H2/CO ratio can be decreased and varied in a range of 1.51.0 by CO2 addition in
the feedstock as listed in demonstrations D, E and F. It can be shown that the ATR process is
suitable for direct production of exible H2/CO ratio synthesis gas.

2.3. Partial oxidation

This process is the non-catalytic reaction of natural gas and oxygen at a high temperature and
pressure to produce synthesis gas. A process ow diagram for a partial oxidation (POX) unit is
shown in Fig. 6 [7]. The partial oxidation reactor is a refractory lined pressure vessel that contains
a specially designed burner in the top. Feedstocks enter the reactor through the burner and ow
down through the reactor. The raw synthesis gas exits near the reactors bottom [6,34].
In this process, oxygen, which is usually 4050% higher than the required amount, and natural
gas is preheated, mixed thoroughly and ignited. The principal reaction can be described as Eq. (1).
However, having no catalyst, the peak temperature has to be high, up to 12001500 C, to obtain
complete conversion. Thus, some complete combustion to CO2 and H2O followed by endothermic
reactions like steam reforming and CO2 reforming take place. The outlet temperature is in the
order of 10001100 C, and the gas at this stage is near thermodynamic equilibrium [6,7,35].
Some soot is always formed and must be removed in a separate soot scrubber system down-
stream of the partial oxidation reactor. The gas composition cannot be adjusted by addition of
568 X. Song, Z. Guo / Energy Conversion and Management 47 (2006) 560569

Natural gas
H.P. Steam

Waste heat
exchanger
Reactor

Synthesis gas

Oxygen
Desulphurizer
Steam

Fig. 6. Schematic diagram of partial oxidation process.

Table 5
Typical H2/CO ratio of synthesis gas by POX from various feeds
Feed type Natural 64 API 9.6 API 4.3 API H-Oil 0 API Petr.
gas naphtha fuel oil vacresid btms. asphalt coke
C/H2 ratio, wt/wt 3.22 5.17 8.82 8.68 9.49 9.50 24.8
Specic oxygen consumption, 255 248 254 265 292 292 355
Nm3(O2)/MNm3(H2+CO)
H2/CO ratio, mol/mol 1.75 1.13 0.95 0.92 0.81 0.84 0.4

steam because the addition of steam to the POX process increases soot formation. Therefore, the
non-partial oxidation process typically results in a product gas within a narrow range of H2/
CO = 1.71.8 [6,8] (Table 5).
However, the most important advantage of partial oxidation is that it can process virtually any
hydrocarbon feedstock, from natural gas to petroleum residue and petroleum coke, even solid
feeds such as coal or even metallurgical coke. The hydrogen to CO ratio primarily depends on
the carbon to hydrogen ratio of the feedstock. Table 4 shows typical H2/CO ratios of product
gas for a range of hydrocarbon feedstocks. However, the capital cost is high for post treatment
of raw synthesis gas to remove soot and acid gases [6].

3. Conclusions

More synthesis gas will be needed to satisfy the demand for synfuels and high value chemicals
in the near future. A key parameter is the cost of the manufacture of the synthesis gas. Direct pro-
duction of exible H2/CO ratio synthesis gas near the stoichiometric ratios required by synthesis
gas based major petrochemical plants seems to be cost eective. Although catalytic partial oxida-
tion and CO2 reforming can directly produce desirable H2/CO ratio synthesis gas, however, opti-
mizing the existing processes such as steam reforming and autothermal reforming as well as
partial oxidation provides an alternative.
 X. Song, Z. Guo / Energy Conversion and Management 47 (2006) 560569 569

Acknowledgements

This work was nancially supported by both the National Science Fund for Distinguished
Young Scholars (50225415) and the National High Technology Research and Development Pro-
gram of China (2002AA529090).

References

[1] Stelmachowski Marek, Nowicki Lech. Appl Energy 2003;74:85.

[2] Wender Irving. Fuel Process Technol 1996;48:189.
[3] Zaman Jasimuz. Fuel Process Technol 1999;58:61.
[4] Alex Mills G. Fuel 1994;73:1243.
[5] Sheldon Roger A. Chemicals from synthesis gas. Dordrecht, Holland: D. Reidel Publishing Company; 1983.
[6] Gunardson HH, Abrardo JM. Hydrocarbon Process 1999;78:87.
[7] Solbakken A. Stud Surf Sci Catal 1991;61:447.
[8] Aasberg-Petersen K, Bak Hansen J-H, Christensen TS, Dybkjaer I, Seier Christensen P, Stud Nielsen C, et al.
Appl Catal A 2001;221:379.
[9] Rostrup-Nielsen JR. Catal Today 2002;71:243.
[10] Rostrup-Nielsen JR. Catal Today 1993;18:305.
[11] Hickman DA, Schmidt LD. J Catal 1992;138:267.
[12] Mleczko L, Wurzel T. Chem Eng J 1997;66:193.
[13] Fathi M, Heitnes Hofstad K, Sperle T, Rokstad OA, Holmen A. Catal Today 1998;42:205.
[14] Basini L, Aasberg-Petersen K, Guarinomi A, stberg M. Catal Today 2001;64:9.
[15] Joensen Finn, Rostrup-Nielsen JR. J Power Sources 2002;105:195.
[16] Rostrup-Nielsen JR, Hansen J-HB. J Catal 1993;144:38.
[17] Zhang Zhaolong, Verykios Xenophon E, MacDonald Susan M, Arossman Stanley. J Phys Chem 1996;100:744.
[18] Choudhary VR, Uphade BS, Mamman AS. Appl Catal A 1998;168:33.
[19] Goldwasser MR, Rivas ME, Pietri E, Perez-Zurita MJ, Cubeiro ML, Gingembre L, et al. Appl Catal A
2003;255:45.
[20] Rostrup-Nielsen JR. Stud Surf Sci Catal 1994;81:25.
[21] Rostrup-Nielsen JR. Catalytic steam reforming. In: Anderson JR, Boudart M, editors. Catalysis, science and
technology, vol. 5. Berlin: Springer; 1983. p. 1110.
[22] Dybkjax Ib. Fuel Process Technol 1995;42:85.
[23] Reyes Sebastian C, Sinfelt John H, Feeley Jennifer S. Ind Eng Chem Res 2003;42:1588.
[24] Hohmann FW. Hydrocarbon Process 1996;75:71.
[25] Udengaard NR, Bak Hansen J-H, Hanson DC, Stal JA. Oil Gas J 1992;90:62.
[26] Dibbern HC, Olesen P, Rostrup-Nielsen JR, Tottrup PB, Udengaard NR. Hydrocarbon Process 1986;65:71.
[27] Rostrup-Nielsen JR. Stud Surf Sci Catal 1988;36:73.
[28] Rostrup-Nielsen JR. J Catal 1984;85:31.
[29] Rostrup-Nielsen JR. Stud Surf Sci Catal 1991;68:85.
[30] Haldor Topsoe A/S. Hydrocarbon Process 1988;67:77.
[31] Ernst WS, Venables SC, Christensen PS, Berthelsen AC. Hydrocarbon Process 2000;79:100-C.
[32] Christensen TS, Primdahl II. Hydrocarbon Process 1994;73:39.
[33] Christensen TS, Christensen PS, Dybkjaer I, Bak Hansen J-H, Primdahl II. Stud Surf Sci Catal 1998;119:883.
[34] Texaco Development Corp. Hydrocarbon Process 1988;67:78.
[35] Tindall BM, Crews MA. Hydrocarbon Process 1995;74:75.