5th Asia-Pacific Conference on Combustion,
The University of Adelaide, Adelaide, Australia
17-20 July 2005
A New Natural Gas Reforming Concept: Chemical Looping Reforming
J. N. Zhu, J. Bromly and D. K. Zhang
Centre for Fuels and Energy
1 Turner Avenue, Technology Park, Bentley, WA 6102, AUSTRALIA
Abstract oxidation reforming that takes the advantage of the exothermic
feature of the partial oxidation reforming to compensate the
A novel natural gas reforming concept, chemical looping endothermic nature of the steam reforming to achieve high
reforming or CLR, has been proposed, which uses metal overall energy efficiency.
oxides as an oxygen carrier and a reforming catalyst for partial
oxidation of the gas. The metal oxides alternate between a Steam reforming:
combustor where they oxidise with air and a reformer where CH4+H2O(g)=CO+3H2 Ho298K =205.9 MJ/kmole (R1)
they are reduced by reforming the natural gas by partial Carbon dioxide reforming:
oxidation. Thus the expensive air separation plant commonly CH4+CO2= 2CO+2H2 Ho298K = 247.1 MJ/kmole (R2)
required in the partial oxidation reforming process can be Partial oxidation reforming:
avoided. This is a substantial advantage which will offer major CH4 + 0.5O2 = CO + 2H2 Ho298K = -35.9 MJ/kmole (R3)
cost reductions and make the natural gas reforming simpler
and more economically feasible. In order to assist in The dominant reforming process is steam reforming.
identification of suitable metal oxides, a thermodynamic Numerous studies can be found in the literature concerning
analysis has been performed on the CLR process with special improvement of the catalysts, operating conditions and heat
attention paid to a number of candidate metal oxides, transfer to achieve better performance. However, the process
including Al2O3, BaO, CaO, CuO, Fe2O3, GeO2, MgO, MoO3, has some inherent drawbacks that mainly relate to its highly
NbO and Pb2O3. It has been found that these metal oxides may endothermic nature. Because of the high process energy
be divided into three classes. The class 1 metal oxides are not requirement, large, but inefficient inter-surface heat
capable of reacting with CH4. However, both classes 2 and 3 exchangers are typically used. This leads to, high capital and
metal oxides may be used for chemical looping reforming. operating costs. To achieve good economy, most of such
The final selection of suitable oxides will be determined by applications are of very large scales. Another disadvantage of
their reaction kinetics, mechanical strength and structural steam reforming is the synthesis gas produced has an
integrity and, of course, their costs inherently high H2:CO ratio of about 3:1 and usually high CO2
content . However, the key advantages of the steam
1 Introduction reforming is that it can provide high quality synthesis gas and
the process is less affected by carbon deposition problems than
Conventional approaches to natural gas utilisation (other than other techniques. Also it does not require oxygen separation.
combustion) require the natural gas to be converted into
synthesis gas, consisting mainly of H2 and CO, which is then Carbon dioxide reforming, which is an even more endothermic
reacted to produce other chemicals, such as methanol, higher process than steam reforming, was proposed as a means to
alcohols and hydrocarbons using typically the Fischer Tropsch reuse carbon dioxide, which is advantageous from the
syntheses and similar processes. Furthermore, future trends in greenhouse gas emission reduction viewpoint . It has been
natural gas utilisation will also require hydrogen to be suggested that it could utilise renewable energy, such as solar
produced from natural gas for use in the more advanced fuel energy, to drive the endothermic reaction. The process is also
cell technologies to generate power with much greater overall rather prone to carbon deposition. Unfortunately unsteady
energy efficiency and less emission. Currently, the dominant supply and scattered availability of such renewable energy
method of converting natural gas to hydrogen is steam sources make this process impractical and they are only in the
reforming followed by water gas shift reaction and CO2 laboratory stage at present. Nevertheless, studies aimed at
removal. However, the costs of reforming process is generally searching for highly selective and less coking catalysts are
high, accounting for 60-70% of the total cost of a gas continuing.
conversion process . Also these technologies are best suited
Partial oxidation reforming is a mildly exothermic process and
to large scale operation. It follows that more effective and
theoretically yields a synthesis-gas with a H2:CO ratio of 2:1,
flexible reforming techniques are the key to effective
which is desirable for most downstream synthesis processes
utilisation of the gas resource. This paper introduces a new
. However, severe carbon deposition on the metal catalysts
concept for reforming natural gas that can potentially simplify
virtually caused abandonment of attempts to develop it for 50
the process and substantially reduce costs.
years . In 1990 Green and co-workers  reported
laboratory results showing that some noble metals could
2 Current approaches to gas reforming catalyse CH4 partial oxidation to the thermodynamic
equilibrium composition of product gases with little or no
There are three basic approaches to reform natural gas, carbon deposition. However, this process, though technically
namely, steam reforming (R1), carbon dioxide reforming (R2) feasible, requires pure oxygen. If air was used, large amounts
and, partial oxidation reforming (R3). Various combinations of nitrogen would be present in the product stream, which is
of the above three are also possible. For instance auto-thermal costly to remove. Air separation plants substantially increases
reforming is a combination of steam reforming and partial the reforming costs and introduce increased complexity.
3 A new concept of natural gas reforming sequestration process can be simplified with reduced
efficiency losses [7, 9]. Figure 1 schematically illustrates this
Due to its mildly exothermic nature, partial oxidation chemical looping combustion concept. Research on CLC has
reforming is the most attractive candidate to produce synthesis been vigorous and will soon enter the industrialisation stage
gas with inherently low energy consumption and thus greater with a pilot plant of CLC is being constructed in South Korea
efficiency. However, the cost associated with nitrogen .
separation, either before or after the reforming, is the major Based on the CLC concept, a new, simple reforming process
difficulty. To overcome this, the novel concept of chemical- may be devised, which has significant potential to overcome
looping combustion (CLC)  may be modified to suit the the inherent deficiencies of the conventional reforming
natural gas reforming. Although the original CLC process was processes. Assuming a suitable medium can be found, which
proposed as a method to enhance the thermal efficiency of not only carries chemically bonded oxygen but also partially
combustion, it was soon realised that it could be an effective oxidise natural gas to produce synthesis gas. Natural gas
means of obtaining nearly pure CO2 from the combustion of reforming can then be achieved without the need to separate
gaseous fuels [7-10], paving a way for CO2 sequestration and nitrogen either from the air feed or the synthesis gas product
storage. The idea of CLC may be summarised as follows: stream. This new natural gas reforming method may be termed
Air and fuel gas (CnH2m) are alternately reacted with an as chemical looping reforming (CLR). The search for suitable
intermediate metal (Me) and the metal oxide (MeO), metals capable of carrying oxygen and catalytically partially
respectively. Firstly, the air reacts with the metal to form the oxidising the natural gas is the key. Certain catalysts capable
metal oxide (R4) in a suitably designed reactor. Once the of driving steam and carbon dioxide reforming may be
oxidation is essentially completed, the fuel gas is introduced to considered as potential candidates. In this scenario, any
reduce the metal oxide and achieve the complete combustion, possible H2O and CO2 produced may be expected to further
generating heat and the products of H2O and CO2 (R5). By react with the natural gas to form further synthesis gas.
continually repeating the above steps, the metal alternates Combinations of two or more metals that can perform the
between the oxidised and reduced states. It follows that there above different tasks separately may also warrant
is virtually no consumption of the metal and its role in the examination. As the metals have to withstand repeated
process is only to transfer chemically bonded oxygen. oxidation and reduction cycles, maintaining a good
However the feed gases, the fuel and air, are sequentially mechanical strength and structural integrity is also a challenge
consumed and the overall reaction is still the same as a normal to be met.
fuel combustion process (R6). The mechanism of the chemical looping reforming may be
The above process, in which the metal is effectively an oxygen considered as follows: Firstly, air is introduced into a reactor,
carrier (OC), can be represented as:- where oxygen in air reacts with the metals or lower metal
oxides to form their oxides or higher metal oxides. Then the
MeOx + 0.5 O2 MeOx+1 (R4) air is turned off and replaced with natural gas, the metal
oxides or higher metal oxides are reduced back to original
(2n+m)MeOx+1+CnH2m (2n+m)MeOx+ mH2O+nCO2 (R5)
metals or lower metal oxides. Synthesis gas is produced in the
CnH2m + (n+0.5m) O2 mH2O+nCO2 (R6) process. However, the complete oxidation to CO2 and H2O, as
seen in the CLC may also occur to some extent. Ideally, in this
case, the CO2, H2O will further react with excess natural gas
N2, O2 CO2, H2 O
supplied for this purpose As noted earlier, it is intended that
the metal oxides should also serve as reforming catalyst (RC),
MeO via steam reforming (R1) and carbon dioxide reforming (R2).
The overall reaction will be the same as the partial oxidation
Reactor Reduction Reactor
Reduction Reactor reforming (R3). Accordingly, the overall thermal behaviour is
Me still exothermic.
By-product Synthesis gas
Air Fuel N2 N2 N2 N2 N2 2CO H2 H2 3H2 2CO 3H2
OC reduction and fuel
gas reforming reactor
Figure 1 A schematic of the chemical-looping combustion H2 CO
OC oxidation rea ctor
concept RC H2 CO
OC RC RC 3H2
By devising a suitable twin reactor system that allows the N2 N2 CH4 H O
metal and its oxide particles to circulate between the two O2 N2 CH4 CO2 2
reactors, each of which is respectively fed with air and fuel N2 OC CH4
gas, there is no possibility that these gases can mix.
Therefore, the first reactor (air-metal oxidation reactor) will
only discharge the excess air and nitrogen, while the other
reactor (fuel-metal oxide reduction reactor) will only deliver O 2 O2 N 2 N2 N2 N 2 N2 CH4 CH4 CH4 CH4
the gaseous reaction products, which will not contain nitrogen Ai r Natural gas
and air but mainly CO2 and H2O. Since water vapour can
Figure 2 Conceptual reformer and schematic reactions in a
easily be condensed, essentially pure CO2 is obtained and is
chemical looping reforming process (double dotted
suitable for sequestration or storage . It is noted that the
dash line – border of reactor; dashed arrow line –
process for concentrating CO2 is achieved only by this novel
directions of oxygen carrier (OC), reforming catalyst
looping concept, concentrating CO2 does not involve
additional energy consumption. Therefore the CO2
The key advantages of the chemical looping reforming are: 0.8
It retains the exothermic feature of partial oxidation GeO2
reforming and therefore high energy efficiency is 0.4 MoO3
It uses air as the oxidant without requiring the costly air
separation plant. 0
500 700 900 1100 1300 1500 1700 1900
Heat transfer directly occurs between gas and OC/RC Temperature K
particles. Therefore, heat transfer is far more efficient and 1
the reformer is expected to be smaller.
0.8 Class 1
The ratio of H2:CO of synthesis gas is near the ideal value Fe2O3
of 2:1. GeO2
As with chemical looping combustion, a chemical looping 0.4
reforming process can also employ a twin-reactor 0.2
arrangement. A schematic of this is illustrated in Figure 2. .
4 Thermodynamic analysis 500 700 900 1100 1300
1500 1700 1900
In order to assess this new reforming process and assist in
identification of appropriate metals, a thermodynamic analysis Normalised CO2 0.8 Class 1
was performed specific to OC and CH4 reactions. The NASA Fe2O3
Glenn Thermodynamic Data  was used. Based on 0.6 GeO2
available thermodynamic data, Al2O3, BaO, CaO, CuO, Fe2O3, 0.4
GeO2, MgO, MoO3, NbO, Pb2O3 were examined. For
comparison purposes, all the results were normalised (that is, 0.2
based on one mole of CH4 in the initial feed. An oxidation
potential ratio is introduced to quantify the reaction
500 700 900 1100 1300 1500 1700 1900
stoichiometry. This is defined as the ratio of the number of Temperature K
moles of initial OC divided by the number of moles of OC 1
required to completely reform one mole of CH4.. An oxidation
potential greater than 1 means that the oxidised OC is in 0.8
excess. It has been found that the major species in the 0.6 Fe2O3
equilibrium state are CH4, CO2, CO, H2O, H2, C(cr) and GeO2
metal/metal oxides. All the other species are negligible in
terms of synthesis gas production. In general the above metal 0.2
oxides may be divided into three classes. The first class (Class
1) includes metal oxides, which are not capable of reacting 500 700 900 1100 1300 1500 1700 1900
with CH4 and they only serve to provide surfaces for CH4 Temperature K
pyrolysis or cracking. For this work, the Class 1 metal oxides 2
are Al2O3, BaO, CaO, MgO and NbO. The second class (Class Class 1
2) are CuO and Pb2O3. Their role is the same as oxygen in 1.6
partial oxidation reforming (that is, they react with the natural 1.2
gas). The last class (Class 3) includes Fe2O3, GeO2, and GeO2
MoO3. These are somewhat different from the other two 0.8
classes, but fall between CH4 pyrolysis and CH4 partial
oxidation. However when temperature is greater than 1000 K,
their behaviour becomes very much the same as that of Class 0
2 metal oxides. Figure 3 shows the calculated results of 500 700 900 1100 1300 1500 1700 1900
stoichiometric equilibrium at atmospheric pressure. Temperature K
Carbon deposition tendency was found to be higher at low
1.6 Class 1
temperatures but minimal at high temperatures. However, our Class 2
previous experimental work on partial oxidation reforming 1.2
, showed no carbon deposition at low temperatures (T < GeO2
1100 K) In contrast when the temperature was greater than 0.8
1250 K carbon formation and deposition was very significant
for oxidation potentials less than 1.0 . As carbon deposition
is a critical issue in gas reforming processes, further 0
experimental verification is necessary. Nevertheless, as this 500 700 900 1100 1300 1500 1700 1900
new reforming process alternately brings the OC/RC into Temperature K
contact with air and natural gas and it should be possible to
burn out any carbon deposits during the air cycle. Therefore, Figure 3 Equilibrium yields under different temperatures
the effect of carbon deposition is expected to be less critical in calculated under the conditions of 1 bar and 1.0
a reactor of this type. oxidation ratio
The effects of oxidation potential and temperature are shown
in Figure 4. At very high temperatures the maximum synthesis A novel chemical looping reforming concept has been
gas yield occurs at an oxidation potential of 1.0. However this proposed, which is expected to possess the following
peak will shift towards right (> 1.0) when the temperature is advantages:
lowered, similar to the situation in the partial oxidation It does not require an air separation oxygen plant ;
reforming . According to the studies of direct partial
oxidation reforming, to achieve equilibrium yield in a short The reformer volume and capital costs can be significantly
residence time is difficult unless an extremely high reduced.
temperature is employed  and this is why reforming
catalysts are required. The process is less affected by carbon deposition.
A thermodynamic analysis has been performed, which
1.0 suggests that Class 1 metals should be excluded as candidates
for chemical looping reforming. However, both Class2 and 3
0.8 metals may be considered as potential candidates. For the
Normalised CO (Class 2)
purpose of CLR, the thermodynamic differences between
these two classes of metals are not significant, particularly at
the temperatures exceeding 1000 K. The final selection of
suitable metals will be determined by their reaction kinetics,
0.4 mechanical strength, structural integrity and costs.
0.2 840 K
0.5 0.7 0.9 1.1 1.3 1.5 1.7
. Dave, N., Forlds, G. A., Comparative Assessment of
Catalytic Partial Oxidation and Steam Reforming for the
production of Methanol from Natural Gas. Ind. Eng.
Figure 4 Equilibrium compositions for different oxidation Chem. Res., 1995. 34: p. 1037-1043.
ratios and temperatures at atmospheric pressure . Zhu, J.N., Zhang, D. K., Bromly, J, A Study of CH4 Partial
Oxidation Reforming with Detailed Mechanisms.
International Journal of Thermal and Fluid Sciences, 2000.
Pressure effects are shown in Figure 5. As expected a high 04: p. 175 - 181.
pressure is not favoured for metal reduction or gas reforming. . Zhu, J.N., Zhang, D. K. A Thermodynamic Study of
This is explained by the Le Chatelier’s principle . In this Carbon Deposition Tendency During CH4 Reforming
work, attention to pressure effects was focused on the Processes. in Chemca 2000. 2000. Perth, WA, Australia.
differences in the metal oxides’ behaviour. For the metals . Zhu, J.N., Zhang, Dong-ke, Bromly, J., Barnes, F., King,
studied, only slight differences in behaviour were found D. K. An Experimental Study of Partial Oxidation
between the oxides and these were only seen at low Reforming of Methane at 1000-1600 K. in World Congress
temperatures. Therefore, it may be concluded that as long as a of Chemical Engineering. 2001. Melbourne, Australia.
metal oxide can react with CH4, i.e., any of the Class 2 and . Zhu, J., Zhang, D., King, K. D., Reforming of CH4 by
Class 3 metal oxides, they are thermodynamically equivalent Partial Oxidation: Thermodynamic and Kinetic Analysis.
for use as an OC. The final selection will obviously be based Fuel, 2001. 80(7): p. 899-905.
on reaction kinetics, mechanical strength and cost. . Tsang, S.C., Claridge, J. B. Green, M. L. H., Recent
Advances in the Conversion of Methane to Synthesis Gas.
Catalysis Today, 1995. 23: p. 3-15.
Marks: (1/3)MoO3+CH4 . Anheden, M. and G. Svedberg, Exergy analysis of
Lines: Class 2 chemical-looping combustion systems. Energy Conversion
and Management, 1998. 39(16-18): p. 1967-1980.
0.6 . Ishida, M. and H. Jin, CO2 recovery in a power plant with
chemical looping combustion. Energy Conversion and
1 bar Management, 1997. 38: p. S187-S192.
10 bar . Lyngfelt, A., B. Leckner, and T. Mattisson, A fluidized-
40 bar bed combustion process with inherent CO2 separation;
0.2 100 bar
application of chemical-looping combustion. Chemical
Engineering Science, 2001. 56(10): p. 3101-3113.
0.0 .Ryu, H., et al. Reduction and oxidation reactivity of
500 700 900 1100 1300 1500 1700 1900 NiO/Bentonite oxygen carrier particles for chemical-
looping combustor. in The fourth Asia-Pacific Conference
on Combustion. 2003. Nanjing China.
Figure 5 Effect of pressure on CO yield at oxidation ratio of 1, .http://cea.grc.nasa.gov/.
lines CH4 reacts with class 2 metal oxides, marks CH4
reacts with MoO3