Energy Fuels 2010, 24, 25892595

Published on Web 03/11/2010

: DOI:10.1021/ef9014058

Hydrogen Production via Steam Reforming of Methane with Simultaneous CO2

Capture over CaO-Ca12Al14O33
Christina S. Martavaltzi,, Eleftheria P. Pampaka, Emmanuela S. Korkakaki, and
Angeliki A. Lemonidou*,,

Department of Chemical Engineering, Aristotle University of Thessaloniki, P.O. Box 1517, University Campus,
GR-54124 Thessaloniki, Greece, and Chemical Process Engineering Research Institute CPERI/CERTH, P.O. Box 361,
sixth km Charilaou-Thermi road, 570 01 Thessaloniki, Greece

Received November 19, 2009. Revised Manuscript Received March 1, 2010

Two processes to produce hydrogen via steam reforming are compared in a laboratory fixed bed reactor:
conventional steam reforming with a nickel-based catalyst and sorbent-enhanced steam reforming with a
mixture of a nickel-based catalyst and CaO-Ca12Al14O33 sorbent. A preliminary thermodynamic study
was performed to determine the operating window of the sorption enhanced reforming at atmospheric
pressure. Operating temperatures less than 650 C are required to take advantage of the favorable
thermodynamics due to the in situ capture of CO2, while maximum gain in methane conversion between
the two process options is attained for a steam to methane molar ratio of around 3. Under these conditions,
CaO-Ca12Al14O33 (85:15 wt) synthesized using calcium acetate as a CaO precursor acted as an effective
CO2 sorbent, reducing by 67% the amount of total CO2 that would have been emitted in conventional
steam reforming. Due to CO2 capture, the methane steam reforming equilibrium shifted to higher
hydrogen purities. Hydrogen concentrations higher than 92%, on a dry basis, were achieved at 650 C,
1 atm of pressure, and a steam to methane molar ratio of 3.4, as opposed to 77% produced under the same
conditions via the conventional steam reforming. The experimental reactor was operated batchwise and
with cyclic alternation between reforming/sorption conditions and higher-temperature conditions to
regenerate the sorbent. The performance of the material is deemed adequate as a moderate loss (ca. 15%)
in sorption capacity after 13 consecutive cycles was observed.

High-temperature operation of the SR reactor, catalyst

deactivation due to coking, use of high-temperature metallurgy for the reactor construction, and complex design of the
multicolumn pressure swing absorption (PSA) system raise
significantly the cost of H2 production.4-6 Therefore, the
development of new concepts for H2 production via SR, with
reduced capital cost and CO2 emissions, is extremely desirable.
The continuous but complex multistep SR process can be
replaced by a much simpler single-step process which employs
a bed packed with an admixture of catalyst and sorbent for the
selective removal of CO2. The latter is known as sorptionenhanced reforming (SER) in which the highly exothermic
carbonation reaction of the sorbent is included in the reaction
scheme. In the case of the use of calcium oxide as a carbon
dioxide sorbent, the main reactions are the following:
CH4 H2 O T CO 3H2 H 298K 206:2kJ=mol 1

Introduction
The continuously increasing standard of living in developing and developed nations results in higher energy demands.
On the other hand, the conventional energy supply via
combustion of fossil fuels and subsequent emission of large
volumes of carbon dioxide has led to one of the most serious
global environmental problems. Therefore, there is a growing
awareness that energy must be produced with significantly
lower greenhouse gas emissions. Hydrogen is often referred to
as a potential pollution-free energy carrier, but its advantages
are unlikely to be realized, unless efficient means are found to
produce hydrogen with reduced CO2 emissions. Biomass,
which is considered a CO2 neutral energy source, can be used
to produce hydrogen via different thermochemical routes
(gasification, fast pyrolysis followed by steam reforming of
the bio-oil produced, etc.).1-3 However, high costs, along with
the several technical problems that need to be solved, currently
prohibit the industrial application of such biomass-based
routes. It is very likely that the production of H2 via steam
reforming (SR) of natural gas will continue to be the dominant
technology for the next few decades, despite the appreciable
amount of CO2 released during the operation of such units.

CaO CO2 T CaCO3 H 298K -178kJ=mol

The concept of sorption-enhanced steam reforming (SER)

is based on Le Chateliers principle, according to which the
conversion of reactants to products and the rate of the
forward reaction in an equilibrium controlled reaction can

*Corresponding author phone: 30 2310 996273; fax: 30 2310

996184; e-mail: alemonidou@cheng.auth.gr.
(1) Huber, G. W.; Dumesic, J. A. Today 2006, 111, 119132.
(2) Vagia, E. Ch.; Lemonidou, A. A. Appl. Catal. A: General 2008,
351, 111121.
(3) Kechagiopoulos, P. N.; Voutetakis, S. S.; Lemonidou, A. A.;
Vasalos, I. A. Catal. Today 2007, 127, 246255.
r 2010 American Chemical Society

CO H2 O T CO2 H2 H 298K -41:2kJ=mol

(4) Johnsen, K.; Ryu, H. J.; Grace, J. R.; Lim, C. J. Chem. Eng. Sci.
2006, 61, 11911198.
(5) Wang, Y.; Rodrigues, A. E. Fuel 2005, 84, 17781789.
(6) Lee, D. K.; Baek, H.; Yoon, W. L. Chem. Eng. Sci. 2004, 59,
931942.

2589

pubs.acs.org/EF

Energy Fuels 2010, 24, 25892595

: DOI:10.1021/ef9014058

Martavaltzi et al.

be increased by selectively removing some of the reaction

products from the reaction zone. As the sorbent is effectively
consumed in reaction 3, the process is inherently dynamic in
operation, requiring a regeneration step. In addition, the
sorbent must maintain its activity through many cycles for
the process to be economically viable.
Sorption enhanced reforming is a hybrid configuration
which not only improves methane conversion and H2 selectivity but also decreases the complexity of the unit by enabling
the in situ separation of CO2. Moreover, the reaction enhancement may enable a lower operation temperature, which in turn
may alleviate the problems associated with catalyst sintering,
high process energy requirements, and poor energy integration
within the plant environment.7,8
Studies on sorption-enhanced hydrogen production using
calcium oxide have been published by several authors.
Balasubramanian et al.9 used a mechanical mixture of a
calcium-based CO2 acceptor and a commercial steam reforming catalyst, producing >95% H2 at 650 C and 15 atm of
pressure using a ratio of solid mixture to methane equal to
0.45 g min cm-3 and a steam to methane molar ratio equal to
4. Johnsen et al.4 used dolomite and a nickel based catalyst to
produce 98% H2 at 600 C and a 0.82 g min cm-3 ratio of solid
mixture over methane. Yi and Harrison10 used 24 g of
uncalcined dolomite admixed with 1.5 g of commercial catalyst to produce 92% H2 at the low temperature of 460 C
under a methane flow of 150 cm3/min and a steam to methane
ratio equal to 3. Li et al.11 used CaO-Ca12Al14O33 (75:25 wt)
with a standard Ni-based reforming catalyst in a continuous
400 min sorption-enhanced hydrogen production test involving two parallel fixed-bed reactors operated in a cyclic
manner. Reaction conditions were 630 C, 1 atm of pressure,
and a steam to methane ratio equal to 5. The product gas
contained about 95% H2 and 2-4% CO2 throughout the test.
The importance but also the main drawbacks of the utilization of CaO-based sorbents for CO2 capture are pointed
out in the recently published review of Florin and Harris.12

Several research groups have been working on improving

CaO-based sorbents.13-32 Lu et al.15 developed nanostructured CaO with a high surface area, thus, with high CO2
uptake efficiency. Calcium oxides are vulnerable to high
temperatures in the carbonator and regenerator, therefore
exhibiting quick activity loss. Abanades and Alvarez16 reported that the absorption capacity for Ca-based sorbents
decays as a function of the number of calcination-carbonation cycles. In another study, Abanades17 collected experimental data from different authors and concluded that the
initial highest carbonation capacity of CaO equal to 14 mol/kg
decreases to 3.78 mol/kg after 20 cycles and keeps decreasing.
Li et al.18,19 used Ca12Al14O33 as a binder to improve the
stability of CaO. Tests in a thermogravimetric instrument
showed 41% adsorption capture after 50 carbonationcalcination cycles. Manovic and Anthony20,21 also worked
on the same direction using calcium aluminate cements to
support CaO for high temperature applications. Our group
further improved the performance of the CaO-Ca12Al14O33
system, using CaO derived from calcined calcium acetate.22
This new material, with low tortuosity in its pore system,
resulted in decreased resistance of the CO2 accessibility to
the active sites. Parametric analysis of the synthesis of the
material revealed that the optimum weight ratio of CaO to
Ca12Al14O33 was 85:15.23 The weight increase achieved was
10 mol of CO2/kg of sorbent in the first 5 cycles with a moderate
loss in capacity (15% after 45 sorption-desorption cycles).
This paper reports steam reforming experimental results
from a laboratory-scale fixed bed reactor containing a
commercial nickel-based reforming catalyst and those of a
mixture of the same catalyst and the new CO2 sorbent
CaO-Ca12Al14O33 (85:15 wt) derived from calcium acetate.
The primary objective was to determine the potential of the
utilization of the new material in sorption enhanced reforming
at near 1 bar of pressure and 650 C. In order to demonstrate
the above, 13 consecutive reforming-regeneration cycles
were conducted without separating the catalyst from the
sorbent between cycles. The selection of the operating conditions was based on the results of preliminary thermodynamic
analysis.

(7) Ding, Y.; Alpay, E. Chem. Eng. Sci. 2000, 55, 34613474.
(8) Harrison, D. P. Ind. Eng. Chem. Res. 2008, 47, 64866501.
(9) Balasubramanian, B.; Ortiz, A. L.; Kaytakoglu, S.; Harrison,
D. P. Chem. Eng. Sci. 1999, 54, 35433552.
(10) Yi, K. B.; Harrison, D. P. Ind. Eng. Chem. Res. 2005, 44, 1665
1669.
(11) Li, Z.-S.; Cai, N.-S.; Yang, J.-B. Ind. Eng. Chem. Res. 2006, 45,
87888793.
(12) Florin, N. H.; Harris, A. T. Energy Fuels 2008, 22, 27342742.
(13) Lu, H.; Smirniotis, P. G. Ind. Eng. Chem. Res. 2009, 48, 5454
5459.
(14) Wang, M.; Lee, C.-G. Energy Convers. Manage. 2009, 50, 636
638.
(15) Lu, H.; Smirniotis, P. G.; Ernst, F. O.; Pratsinis, S. E. Chem. Eng.
Sci. 2009, 64, 19361943.
(16) Abanades, J. C.; Alvarez, D. Energy Fuels 2003, 17, 308315.
(17) Abanades, J. C. Chem. Eng. J. 2002, 90, 303306.
(18) Li, Z.-S.; Cai, N.-S.; Huang, Y.-Y.; Han, H.-J. Energy Fuels 2005,
19, 14471452.
(19) Li, Z.-S.; Cai, N.-S.; Huang, Y.-Y.; Han, H.-J. Ind. Eng. Chem.
Res. 2006, 45, 19111917.
(20) Manovic, V.; Anthony, E. J. Environ. Sci. Technol. 2009, 43,
71177122.
(21) Manovic, V.; Anthony, E. J. Energy Fuels 2009, 23, 47974804.
(22) Martavaltzi, C. S.; Lemonidou, A. A. Microporous Mesoporous
Mater. 2008, 110, 119127.
(23) Martavaltzi, C. S.; Lemonidou, A. A. Ind. Eng. Chem. Res. 2008,
47, 95379543.
(24) Gupta, H.; Fan, L.-S. Ind. Eng. Chem. Res. 2002, 41, 40354042.
(25) Kuramoto, K.; Fujimoto, S.; Morita, A.; Shibano, S.; Suzuki, Y.;
Hatano, H.; Shi-Ying, L.; Harada, M.; Takarada, T. Ind. Eng. Chem.
Res. 2003, 42, 975981.
(26) Salvador, C.; Lu, D.; Anthony, E. J.; Abanades, J. C. Chem. Eng.
J. 2003, 96, 187195.

Experimental Section
Chemicals, Sorbent, Catalyst, and Gases. Calcium acetate
(Panreac, 99%) was used as CaO precursor, while aluminum
nitrate hydrate (Riedel-de Haen, 98%) was used for the synthesis of the nanocomposite oxide CaO-Ca12Al14O33. A commercial nickel based catalyst (18% NiO) from S
ud-Chemie was
employed as the reforming catalyst. Nitrogen (99.999%) was
used as purge gas during the calcination period and as dilution
gas during the carbonation period at TGA. Methane (99.995%),
helium (99.999%), and distilled water were used as feedstock
in the laboratory-scale sorption enhanced reforming experiments.
Material Preparation. The precursor material (Ca(CH3COO)2)
was calcined (850 C/1 h) in order to decompose to pure CaO in a
box type furnace under an air atmosphere. For the synthesis of
(27) Aihara, M; Nagai, T.; Matushita, J.; Negishi, Y.; Ohya, H. App.
Energy 2001, 69, 225238.
(28) Manovic, V.; Anthony, E. J. Ind. Eng. Chem. Res. 2009, 48, 8906
8912.
(29) Li, A.-S.; Cai, N.-S. Energy Fuels 2007, 21, 29092918.
(30) Hughes, R. W.; Lu, D.; Anthony, E. J.; Wu, Y. Ind. Eng. Chem.
Res. 2004, 43, 55295539.
(31) Manovic, V.; Anthony, E. J. Energy Fuels 2008, 22, 18511857.
(32) Wu, S. F.; Li, Q. H.; Kim, J. N.; Yi, K. B. Ind. Eng. Chem. Res.
2008, 47, 180184.

2590

Energy Fuels 2010, 24, 25892595

: DOI:10.1021/ef9014058

Martavaltzi et al.
Table 1. Experimental Conditions

reforming period
regeneration period
a

temp., C

CH4, cm /min

H2O/CH4 molar ratio

He, cm3/min

solid mass, g

S/Ca

650
850

11
0

3.4
0

25
100

4.5
4.5

2
2

S/C: sorbent/catalyst (weight ratio).

inlet flow of methane and helium was 36 cm3/min with a steam to

methane molar ratio equal to 3.4. The reforming reaction and
CO2 sorption (reforming period) were conducted at 650 C with
a duration controlled by the sorption capacity. The sorbent
regeneration was performed in a 100% He flow at 850 C.11,33
The operating conditions for cyclic stepwise sorption-enhanced
steam methane reforming and sorbent regeneration reactions
over CaO-Ca12Al14O33 and a commercial catalyst are reported
in Table 1.

CaO-Ca12Al14O33, aluminum nitrate enneahydrate (Al(NO3)3 3

9H2O) and calcium oxide derived from calcination of calcium
acetate were added in appropriate amounts to distilled water to
achieve a weight ratio of CaO/Ca12Al14O33 equal to 85:15. The
suspension was stirred for 3 h at 70 C and then dried at 120 C
overnight. To decompose the nitrates, the sample was calcined at
500 C for 3 h in synthetic air. The sample was further treated by
adding distilled water to break the aggregates formed at 500 C
and generate the hexagonally plated structure with uniform
Ca/Al distribution.23 The obtained kneaded paste was dried at
120 C overnight and then calcined at 900 C for 1.5 h to form the
nanocomposite synthetic oxide.
Sorbent Characterization. The crystalline structure of the
sorbent was characterized by X-ray diffraction (XRD) on a
Siemens D500 diffractometer using Cu Ka radiation ( =
1.5406 A). Specific surface area and pore size distribution were
measured on a Quantachrome Autosorb Automated Gas Sorption System. The measurements were performed using nitrogen
physisorption and desorption isotherms at -196 C. The sorbents were degassed at 250 C overnight in the degassing port
before the actual measurements.
Material Testing. An SDT Q600 (TA Instrument) thermal
gravimetric analysis (TGA) instrument was employed for the
carbonation and calcination experiments. A small quantity of
the material placed in an aluminum sample cup was initially
heated at 850 C, in the presence of 100 cm3/min pure N2, for
10 min to remove possible humidity and CO2 absorbed. The
sample was then heated to the predetermined carbonation temperature (690 C) under the flow of the inert gas. The selected CO2
partial pressure was 0.15 atm in order to simulate the typical CO2
volumetric concentration at the reformer reactor exit which does
not surpass 15%. The sorption duration was 30 min. Desorption
at 850 C for 5 min under a 100 cm3/min pure N2 flow followed.
The sorption enhanced steam reforming experiments were
performed at atmospheric pressure in a laboratory unit
equipped with a mass flow controlled system for the incoming
gases, a fixed bed quartz reactor, and an online gas chromatograph. An HPLC pump (Gilson) was used for the admission of
the distilled water to the reactor through a preheater. The fixed
bed quartz reactor was equipped with a coaxial thermocouple
for monitoring the temperature in the middle of the solid bed.
Minor variations in the reactor temperature were observed
during the reforming/sorption period. This is because the heat
requirements for the endothermic reforming were almost balanced by the heat generated from the exothermic carbonation
under the conditions studied. The external diameter of the
reactor tube in the reaction zone was 18 mm. The reactor was
heated electrically by a tubular furnace, with three independently controlled temperature zones. The hot gases exiting the
reactor were cooled to condense the unreacted steam. The gas
phase products were analyzed with an online gas chromatograph (Agilent Technologies, 7890A) equipped with TCD. The
CO2 concentration in the reactor exit was also monitored by a
CO2 analyzer (Horiba, VIA 510). An admixture of a commercial
catalyst and CaO-Ca12Al14O33, derived from calcium acetate,
was used for sorption enhanced steam reforming experiments.
The catalyst was crushed and sieved to 100-180 m. Before
testing, the material was reduced in a H2/He flow (30% in H2) at
850 C for 1 h. This treatment at high temperature was also
sufficient for ensuring the desorption of CO2 and water possibly
sorbed on CaO-Ca12Al14O33 from the atmosphere. Approximately 1.5 g of the catalyst and 3 g of the sorbent (bed length:
6 cm), thoroughly mixed, were added to the reactor. The total

Results and Discussion

Thermodynamic Analysis. Thermodynamics clearly portray the potential advantages of combining reaction and
separation for hydrogen production. A preliminary thermodynamic analysis was performed by calculating the equilibrium compositions which minimize the Gibbs free energy.
The Aspen Plus 12.1 software was used for the calculations.
This code requires specification of the systemat least the
reactorfor the reaction calculations. The RGibbs reactor
was selected for the calculations using the Peng-Robinson
property method. The Aspen Plus code requires also the
definition of the possible components in the equilibrium.
Hydrogen, carbon dioxide, carbon monoxide, water,
methane, calcium oxide, calcium hydroxide, and calcium
carbonate were considered as equilibrium components.
Carbon was also included in the products pool, but calculations showed that its concentration in the equilibrium stream
was negligible. For the simplest hydrocarbon, methane, and
with CaO as sorbent, the main reactions scheme can be
described with reactions 1-3.
For the analysis, we assumed that the system of the three
reactions is in equilibrium. Equilibrium compositions of
methane steam reforming were calculated at atmospheric
pressure using a steam to fuel a molar ratio of 1 to 5. The
analysis was restricted to the 400-900 C range. Below this
range, the kinetics limit the reaction of methane steam
reforming, while at T > 900 C, sintering effects of the
materials dominate.
Figure 1 shows the effect of the sorbent to methane molar
ratio on the equilibrium conversion of methane in the case of
conventional steam reforming (SR) and sorption enhanced
steam reforming (SER) as a function of temperature. In both
SR and SER, temperature rise favors the endothermic
reforming reaction, leading to higher methane conversions.
An increase of sorbent has a positive effect on methane
conversion, which attains a maximum value at a sorbent to
methane molar ratio equal to 1. It should be noted that the
above outcome does not take into account the CaO utilization perspective. The carbonation conversion of most of
the CaO-based sorbents is limited due to diffusion related
problems and decreases with the number of cycles, parameters that cannot be taken into account in a thermodynamic analysis. Therefore, for a specified sorbent and
process, the determination of the optimum CaO to methane
ratio can only be evaluated with a techno-economic study.
(33) Ortiz, A. L.; Harrison, D. P. Ind. Eng. Chem. Res. 2001, 40, 5102
5109.

2591

Energy Fuels 2010, 24, 25892595

: DOI:10.1021/ef9014058

Martavaltzi et al.

Figure 1. Effect of temperature and sorbent to methane molar ratio

on the equilibrium conversion of methane in the case of SR and SER
(molar H2O/CH4: 3, 1 atm).

Figure 2. Effect of temperature and steam to methane molar ratio

on the equilibrium conversion of methane in the case of SR and SER
(molar CaO/CH4: 1, 1 atm).

MacKenzie et al.34 reported that the economics of flue gas

CO2 capture using the calcium cycle are dominated by the
cost of the utilized sorbent. The CaO/C ratio and the CaO
deactivation rate are the key variables to improve the
viability of the system. Reduction of these rates, possibly
by development of new calcium-based sorbents, can lead to
lower feedstock and operation costs as less material will be
cycled and consequently wasted.
Another interesting point of Figure 1 is that, at temperatures lower than 430 C and for a sorbent to methane ratio
greater than 1, the consumption of steam for the formation
of Ca(OH)2 limits methane conversion. For temperatures
higher than 430 C and a sorbent to methane ratio equal to 1,
more than 90% of the methane is converted, while for the
same temperature range and in the absence of CaO, the
conversion progressively increases from 30% up to 90%. In
any CaO to CH4 ratio, for a steam to methane ratio equal to
3, under 1 atm of pressure, and temperatures above 680 C,
the CO2 capture reaction is no longer effective and the
equilibrium methane concentration is the same for both
processes.
The steam to methane molar ratio strongly affects
methane conversion, for both SR and SER, as illustrated
in Figure 2. The higher the ratio, the higher is the equilibrium
methane conversion at a specific temperature. This effect is
stronger in SER because higher steam concentrations lead to
higher equilibrium CO2 partial pressures, enhancing sorption and sequentially methane conversion. However, for
each temperature, there is a maximum steam to methane
molar ratio, equal to 3, above which its effect on SER
methane conversion is minimal. On the other hand, the effect
of this ratio in SR methane conversion, in each temperature
up to 600 C, is substantial for all the investigated steam to
methane molar ratios. In accordance with the above, Figure 3
demonstrates that the maximum advantage of SER over SR,
with respect to methane conversion, can be attained with
steam to methane molar ratios around 3 for temperatures up
to 600 C. It is well-known that the excess of steam not only
favors the thermodynamics of the steam reforming reaction
but also inhibits coke formation. However, higher steam

Figure 3. Variation in the CH4 conversion difference between SR

and SER as a function of temperature and steam to methane molar
ratio (molar CaO/CH4: 1, 1 atm).

utilization induces a critical efficiency penalty due to higher

energy demands for steam generation.35 It is of high importance that in SER, for the various steam to methane molar
ratios examined and for a broad temperature range, methane
conversion remains almost constant because of the reverse
effect of the exothermic sorption and water gas shift (WGS)
reactions, as opposed to the endothermic reforming reaction
(Figure 2). It is also remarkable that a steam to methane ratio
equal to 1 is inadequate to cover the demands of the governing reforming reaction at temperatures higher than 600 C.
As a consequence, methane conversion drops at lower values
than in SR.
Figure 4 compares the dry equilibrium composition at the
reactor exit which results from reacting methane and steam
in the presence (SER) and absence of CaO (SR), as a function
of temperature. The formed mixtures consist of hydrogen,
carbon monoxide, carbon dioxide, and unreacted methane.
(35) Bartholomew, C. H.; Farrauto, R. J. Fundamentals of Industrial
Catalytic Processes; John Wiley & Sons, Inc.: Hoboken, NJ, 2006;
pp 268-270.

(34) MacKenzie, A.; Granatstein, D. L.; Anthony, E. J.; Abanades,

J. C. Energy Fuels 2007, 21, 920926.

2592

Energy Fuels 2010, 24, 25892595

: DOI:10.1021/ef9014058

Martavaltzi et al.

Figure 4. Effect of temperature on equilibrium reactor outlet

composition of SR and SER.
Figure 5. X-ray diffractogram of the new material CaO-Ca12Al14O33.

In the conventional reforming process, the carbon monoxide

content increases linearly with the temperature, as its formation depends on the endothermic steam reforming reaction. On the other hand, carbon dioxide concentration,
which is determined by the exothermic water gas shift
reaction, reaches a maximum of almost 15% near 500 C.
Hydrogen concentration depends on both reactions but is
governed by the endothermic reforming reaction. As a result,
it reaches a maximum of 77% at 650 C and remains constant
at higher temperatures. In the presence of the CO2 sorbent,
the equilibrium conversion of methane remains very high
over the entire temperature range. Hydrogen purity is
very high (98%) in the temperature range 450 to 550 C
and decreases at higher temperatures due to the exothermic
nature of the carbonation reaction, which gradually diminishes
for temperatures higher than 600 C, finally ceasing at 710 C.
It is thus possible to produce high purity hydrogen at relatively
low temperatures applying the SER process. Both carbon
oxides are negligible for temperatures 400-600 C due to the
strong effect of the carbonation reaction of the CO2 sorbent.
For higher temperatures, up to 710 C, their concentration
increases, following afterward the course of conventional
steam reforming.
The present thermodynamic analysis reveals that for a
steam to methane molar ratio equal to 3 the maximum
advantage of SER over SR can be attained. Moreover, a
CaO to methane ratio equal to 1 is the threshold to obtain
maximum methane conversion. The study also demonstrates
that the optimum temperature for SER lies in the range
450-600 C. However, as previously mentioned, the selection of the temperature depends not only on thermodynamics but also on kinetics. The reforming kinetics are
favored at high temperatures. In addition to that, the risk of
carbon formation increases at low temperatures. Therefore,
a temperature of 650 C was selected for the SER experiments in this study. On the basis of the results of the SER
thermodynamic analysis, under these conditions, more
than 95% of methane can be potentially converted to
produce high purity hydrogen (almost 93%), while according to the equilibrium of SR the hydrogen concentration
cannot surpass 77%.
Sorbent Characterization. A polycrystalline nanocomposite oxide of Ca and Al, with a surface area and pore volume
of 11.68 m2/g and 0.098 cm3/g, respectively, was successfully

synthesized. Pore volume was calculated from the adsorption branch of the nitrogen isotherm.
The XRD pattern of the new sorbent material is presented
in Figure 5. All characteristic peaks of CaO (2 = 32.2,
37.35, 53.85, 64.15, 67.3) and Ca12Al14O33 (2 = 33.41,
41.21, 55.22, 57.52) were clearly detected. The absence
of any other Ca-Al mixed phases or hydrated mixed structures proved that the calcination conditions applied in
the synthesis process (900 C, 1.5 h) and the weight ratio of
the metal precursors ensure the formation of the desired
CaO-Ca12Al14O33.
Sorption Capacity. The CaO-Ca12Al14O33 (75:25 wt)
sample proved to be an effective CO2 sorbent with adequate
cyclic stability.18-23 The stability of CaO in repeated sorption-desorption cycles was attributed to the uniform dispersion of tolerant at high temperature Ca12Al14O33 among
CaO particles, which provides a stable framework inhibiting
sintering of active CaO sites. A previous study of our group
demonstrated that utilizing calcium acetate as a CaO precursor could lead to 65% improvement in the CO2 uptake
compared to CaO-Ca12Al14O33 derived from other precursors.22 Moreover, a CaO to Ca12Al14O33 weight ratio equal
to 85:15 proved superior to the 75:25 ratio.
The sorption capacity of the improved CaO-Ca12Al14O33
(85:15 wt) sorbent was tested in the TGA at 690 C in 45
repetitive cycles of sorption-desorption using a flow of CO2
in He (15% CO2). The variation in the weight of the sample is
illustrated in Figure 6. The material initially captured almost
10 mol of CO2 per kg, with a moderate loss in capacity (15%)
after the fifth cycle. This loss can be characterized as
negligible compared to other CaO sorbents which exhibit a
far more drastic decrease of their fixing ability, as reviewed
by Abanades.17 It is important to note that, in the present
study, in the TGA experiments, the desorption took place at
850 C in the presence of pure N2, with conditions selected on
the basis of the protocol followed for multicycle tests in the
literature.11,13,25-29 However, under realistic conditions,
high CO2 partial pressure and a temperature of about 950 o
during calcination are needed to produce a near pure CO2
stream suitable for storage or use in an alternative chemical
process.31
Sorption Enhanced Reforming Experiments. A reaction
system of this type is inherently in an unsteady state because
2593

Energy Fuels 2010, 24, 25892595

: DOI:10.1021/ef9014058

Martavaltzi et al.

Figure 7. Outlet H2 and CO2 concentration (dry basis) response

curve. Reaction conditions: 1 atm, 650 C, molar H2O/CH4 = 3.4.
Figure 6. Cyclic sorption capacity of CaO-Ca12Al14O33 (85:15 wt)
derived from calcium acetate (sorption: 690 C, 30 min, 15% CO2;
desorption: 850 C, 5 min, 100% N2).

experimental results of SR (open symbols), SER (closed

symbols), and the equilibrium concentrations (solid lines).
The steam reforming results are close to the equilibrium
values, indicating that the catalyst activity at the relevant
conditions is high enough. The effluent gas composition
indicates constant hydrogen concentration at near 77%
and a CO2 concentration at 10-11% throughout the course
of the experiment.
The new sorbent material CaO-Ca12Al14O33 (85:15) was
then mixed with the reforming catalyst and introduced to the
reactor. The reaction conditions are summarized in Table 1.
The composition of the reactor exit stream in H2 and CO2
from the first sorption-desorption cycle of an experiment at
650 C and a H2O/CH4 feed ratio of 3.4 is illustrated in
Figure 7, along with the corresponding results from the
experiment without CO2 sorbent. The equilibrium values
for the SER and SR system reactions are also included. A key
feature of this figure is the close proximity of the experimental and equilibrium values in both the prebreakthrough
and postbreakthrough periods. The effluent gas composition
of the SER experiments follows a different trend than that of
the SR. In the SER experiment, more than 93% of the
methane was converted following the thermodynamic predictions. The low CO2 and CO contents (2-3% and 3-4%,
respectively) observed in the prebreakthrough period prove
the efficiency of the sorbent. Moreover, the high hydrogen
concentration (>92%) demonstrates the enhancing effect of
the sorbent on the conversion of methane to the desired
product. As the sorbent started saturating, the hydrogen
concentration decreased, whereas CO and CO2 increased,
finally equilibrating to values almost equal to those of SR.
The hydrogen concentration during the prebreakthrough
period of the first, third, and 13th cycles is shown in Figure 8.
The minor differences in the profiles between the first and the
13th cycles point indirectly to the adequate stability of the
sorbent and the catalyst under the examined conditions. The
durations of the prebreakthrough and breakthrough periods
provide an indication of the stability of the sorbent. Figure 9
presents the normalized prebreakthrough and breakthrough
times for the first, third, and 13th cycles. The normalized
time is defined as the ratio of the time of cycle i to the time of
the first cycle. A clear reduction of the prebreakthrough
period and an evident increase of the breakthrough period
can be observed. The reduction in the prebreakthrough
period is most probably a result of the decrease of the

of the progressive saturation of the fixed amount of sorbent

present in the reactor. The sorbent is progressively saturated,
and its enhanced effect on the equilibrium gradually diminishes. A typical reactor response curve is characterized
by three regimes. The first one, where the product composition is almost steady, is usually referred to as the prebreakthrough period.4,32 During this period, the three reactions
of the systemreforming, shift, and carbonationare all
active, and the concentration of the valuable H2 is at its
maximum value while the undesired CO2 and CO attain their
minimum values. The duration of this period is a function of
the sorbent properties (sorption capacity and sorption
kinetics), its mass, and the operating conditions. For a
certain material and under the same conditions, the duration
of the prebreakthrough period would decrease as the result
of the deactivation of the sorption sites of the material.
Prebreakthrough is followed by the breakthrough period,
which begins when the leading edge of the surface carbonation reaction front reaches the exit of the fixed bed of the
solid. At this point, CO2 removal becomes less effective, and
as a result CO conversion in the shift reaction is limited.
Under the same conditions, the duration of this period is
affected by the diffusion of CO2 through the generated
CaCO3 layer and by the available space in the interior of
the material for the generation of the larger (compared to
CaO) specific volume CaCO3.24 Increase of the breakthrough time implies sorbent morphology degradation, thus
deactivation. When the carbonation rate approaches zero,
the postbreakthrough period is reached and the system
behaves as it would in the absence of the CO2 sorbent. Only
the reforming and shift reactions are active, and the product
composition is determined by the kinetics and thermodynamics of these reactions.
As mentioned in the Experimental Section, a commercial
steam reforming catalyst was used in the tests. The long-term
efficiency of the catalyst was tested and confirmed under
conditions far from those leading to equilibrium composition. Moreover, steam reforming experiments without a
sorbent were carried out to test the activity of the commercial
catalyst at 650 C and a steam to methane molar ratio equal
to 3.4 and were used as a reference for the experiments with a
mixed bed of catalyst and sorbent. Figure 7 presents the
2594

Energy Fuels 2010, 24, 25892595

: DOI:10.1021/ef9014058

Martavaltzi et al.

Figure 8. Prebreakthrough and postbreakthrough hydrogen concentration (dry basis) for the 1st, 3rd, and 13th reformingregeneration cycles in SER. Reaction conditions: 1 atm, 650 C,
molar H2O/CH4 = 3.4.

Figure 10. Comparison of SER and TGA sorption capacity of

CaO-Ca12Al14O33 (85:15wt) for multicycle testing.

results are presented together with the TGA results. It should

be stressed that this figure should not be used for direct
comparison of the sorption capacity in TGA and SER
experiments, as the operating conditions are not similar. It
however provides a qualitative indication of the trend in
stability. Two areas with different slopes can be seen for both
the TGA and the SER curves. The SER capacity decreased for
the first 5 cycles while attaining an almost flat profile for the
remaining cycles in agreement with the TGA capacity curve.
The amount of CO2 emitted during the prebreakthrough
period of the first cycle of the SER experiments is 67% less than
the amount of total CO2 that would have been emitted in
conventional steam reforming. Therefore, the application of
the SER process not only increases the hydrogen purity but also
reduces the number of necessary downstream processing steps.
Conclusions
Previously reported results showed that high purity hydrogen can be produced in a single processing step via the
sorption-enhanced steam-methane reforming using CaO as
a CO2 sorbent. However, the decrease in sorbent conversion
detected in most of the multicycle studies limits the economic
viability of sorption enhanced reforming.
This study demonstrates that CaO-Ca12Al14O33 (85:15 wt)
synthesized using calcium acetate as a CaO precursor may be
effectively used as a CO2 sorbent. More than 93% of the
methane is converted at the low temperature of 650 C and
1 bar of pressure with a steam to methane molar ratio equal to
3.4. The high hydrogen purity, >92%, of the reforming outlet
stream compared to the 77% hydrogen concentration in the
case of the conventional steam reforming reveals the high
importance of the sorbents presence. At the same time, with
the use of CaO-Ca12Al14O33, the necessity of downstream
purification steps decreases due to the low CO and CO2
contents (2-3% and 3-4%) at the reformer exit. Multicycle
tests show a moderate decrease in the CaO-Ca12Al14O33
sorption capacity, providing incentive for further improvement and evaluation under more realistic process conditions.

Figure 9. Normalized prebreakthrough and breakthrough times

for the 1st, 3rd, and 13th reforming-regeneration cycles in SER.
Reaction conditions: 1 atm, 650 C, H2O/CH4 = 3.4.

external surface sorption sites of the material. On the other

hand, the slight increase in the breakthrough duration, as
shown in Figure 9, indicates not a loss but a transformation
of some of the external surface sites to less accessible ones.
Overall, after a certain number of cycles, only part of the
sites was permanently lost, resulting in a moderate loss of
CaO-Ca12Al14O33 sorption capacity (15%). The stability of
CaO-Ca12Al14O33 (75:25) in methane sorption enhanced
reforming has been also demonstrated in the literature.11
However, direct comparison with the present results is not
possible due to the differences in the material composition
and operating conditions.
The CO2 capture effectiveness of the new material under
real reaction conditions was also determined on the basis of
rough calculations of the mass of the CO2 desorbed during
the regeneration in each cycle. These calculations, based on
the integration of the desorption curves of the CO2 analyzer,
showed that 3 g of the material captured approximately 1.3
and 1.15 g of CO2 at the first and fifth cycles, respectively.
These values account for a 44% and a 38% sorption capacity, as shown in Figure 10, where the SER experimental

Acknowledgment. The authors are grateful to Ms. Evangelia

Tousiaki for her help in the synthesis of the materials, Mr. Dimitris
Sergiadis for his help in thermodynamic analysis, and Dr. Lori
Nalbandian and Dr. Eleni Heracleous for the fruitful discussions.
2595