Energy

Procedia
EnergyProcedia
Procedia 100(2009)
Energy
(2008)295302
000000

www.elsevier.com/locate/XXX
www.elsevier.com/locate/procedia

GHGT-9

High Temperature Gas Separation Membranes in Coal

Gasification
a*
a

J. C. Diniz da Costa, b G.P. Reed and cK. Thambimuthu

The University of Queensland, FIMLab Films and Inorganic Membrane Laboratory, Division of Chemical Engineering, Brisbane Qld 4072,
Australia.
b
Centre for Low Emission Technology, Pullenvale Qld 4069, Australia
Else vie r use only: Received date here; revised date here; accepted date here

Abstract
In this work we investigate the proof of concept of metal (Cobalt) doped silica membranes for H 2/CO 2 separation in single and
multi tube membrane modules, in addition to a membrane reactor (M R) configuration for the high temperature water gas shift
(WGS) reaction. The membranes were prepared by a sol-gel process using tetraethylorthosilicate (TEOS) in ethanol and H 2O 2
with cobalt nitrate hexahydrate (Co(NO 3)26H 2O). The membranes delivered high H 2/CO2 single gas selectivity up to 131. A multi
tube membrane module was tested up to 300 oC and 4 atmospheres for 55 days (1344 hours) for binary feed gas mixtures
containing H 2 and CO2 at 40:60 concentration ratio. The best membrane performance delivered H2 purity in excess of 98%. For
the high temperature water gas shift reaction and a ternary mixture of 40% (H 2), 40% (CO 2) and 20% (CO), which is equivalent
to 67.5% CO conversion, the membrane delivered a permeate stream containing 92.5% H 2. The membranes complied with a flux
temperature dependency mechanism, as H2 permeation had a positive energy of activation whilst CO 2 was negative. As a result,
H2 permeation and separation to other gases increased with temperature. In turn, this effect was combined with high CO
conversion for the water gas shift reaction, thus allowing for high throughput of H 2 production and separation in a single
processing step. The process integration provided by this work is potentially beneficial for the next generation of high
temperature processing unit operations in low emissions coal gasification.

2008Elsevier
ElsevierLtd.
Ltd.
rights
reserved
c 2009

AllAll
rights
reserved.
Keywords: IGCC; metal doped silica membranes; membrane reactors, hydrogen separation

1. Introduction
In this new paradig m of decarbonised energy economies, security of fuel supply still plays a major role in
sustainable development. These issues are socially, economically and polit ically sensitive around the globe, and in
particular for those economies driven by energy derived from fossil fuels. Govern ments are currently assessing a
basket of options to enhance the robustness of the energy mix for their resp ective countries. However, as demand for
energy is unlikely to reduce, many industrialised nations will continue to rely on coal as a primary energy supply,
which is likely to outlast gas and oil for centuries, based on current reserves and consumption rat es [1, 2]. To

* Corresponding author. Tel.: +61 7 3365 6960; fax: +61 7 3365 4199.
E-mail address: j.dacosta@uq.edu.au
doi:10.1016/j.egypro.2009.01.041

J.C.Author
Diniz da
Costa
et al. / Procedia
Energy Procedia
1 (2009)
295302
name
/ Energy
00 (2008)
000000

296

address the need to reduce CO2 and associated greenhouse gas emissions, research has intensified in the last decade
on post combustion, pre co mbustion and oxy-fuel as modes of operating coal power p lants with carbon capture.
The pre-combustion option via coal gasification is getting the attention of governments, industry and the research
community as an attractive alternative process to deliver electricity. Coal gasification is a flexib le process which can
be integrated with a combined cycle of gas and steam turbines to deliver power, called an Integrated Gasificat ion
Co mbined Cycle (IGCC). Th is process has a potential advantage over conventional power generation, as it also has
the capability for delivering synthetic fuels for industrial feed stock and hydrogen for the energy and transportation
sectors. An IGCC scheme for the production of hydrogen with CO 2 capture for bitu minous coals using the current
best available technology is shown in Figure 1. The process can be divided into four steps as follows: (1) air
separation & gasification, (2) gas cleaning, (3) gas conditioning, and (4) gas separation.
Flue Gas
Air separation
& gasification

Gas
Cleaning

Gas
Conditioning

(N2/H2O/O2)

Gas
Separation

CO2

SCR

2 stages

Air

Coal

CO2
compression

Steam

H2
ASU

O2

Gasifier

Cooling

Wet
Scrubber

Shift
reactor

Cooling

H2S & CO2

removal

Gas
turbine
Steam
turbine

Claus
plant

Steam

N2

N2

Figure 1- A conventional IGCC with carbon capture [3]

Using conventional technology, it is possible to build the IGCC p lant shown in Figure 1. There are several mature
technologies that can be employed as gas separation processes such as cryogenics, solvent extraction, adsorbents in
pressure or temperature swing adsorption, and low temperature polymeric memb ranes. Due to compliance with
sorption mechanisms, these technologies operate best at low temperatures (<50 o C). Examp les of gas separation
include physical solvents such as Selexol or Rectisol, zeo lite beds in pressure swing adsorption or polycarbonate,
polydimethylsilo xane PDMS membranes. The problem here is that gas separation follows downstream fro m gas
conditioning. The gasification of coal predominantly produces syngas (CO and H 2 ) with some remain ing
hydrocarbons, CO2 and water. The syngas requires further processing through the WGS reaction (Equation 1), in
order to maximise H2 production [4, 5].
CO + H2 O CO2 + H2

H =-41.2 kJ.mo l-1

(Equation 1)

The WGS reaction is exothermic and the conversion is limited by thermodynamic equilibriu m as the conversion to
H2 and CO2 decreases with increasing temperature. The reaction is therefore carried out in two stages, in high (350400o C) and low temperature (250-300o C) shift reactors with interstage cooling. In addition, further cooling to <50 o C
is required to reduce the temperature to the temperature acceptable by conventional gas separation technologies.
Cooling large stream of hot gases is capital intensive, and incurs a loss of power production; furthermore, these
processes are likely to deliver CO2 at reduced pressures, which will have to be re-co mpressed to > 100 at m for
transportation and storage. Hence, conventional processes will attract large energy penalties.
In order to reduce efficiency losses, an alternative is to separate gases at high er temperatures. In this case,
inorganic membranes derived fro m ceramics [6], silica [7-9], metal [10], and by further doping or alloying [11]
showed preferential H2 selectivity over CO2 at high temperatures (>200o C). These technologies can also operate in
memb rane reactor arrangements for the water gas shift reaction [12-14], which allow for shift ing the reactions to
higher conversions due to the extraction of hydrogen from the reaction chamber [15]. The advantage here is twofold;
1. CO2 is kept at high pressure thus reducing requirements for CO2 comp ression downstream.

J.C. Diniz
da Costa
al. / Energy
Procedia
1 (2009)
295302
Author
nameet/ Energy
Procedia
00 (2008)
000000

2.

297

As H2 is selectivity taken, the syngas stream is reduced by up 30-35% depending on the recovery rate.
Although CO2 will have to be cooled down prior to compression, the volumes are reduced thus requiring a
lower cooling duty.

Therefore, inorganic membranes and their incorporation in MRs are foreseen to be the technology of choice for
advanced IGCC plants as depicted in Figure 2.

Air separation &

gasification

Hot Gas
Cleaning

Gas Conditioning
& separation

Flue Gas

CO2

(N2/H2O/O2)

CO2
compression

Air

H2S
sorbent

Coal
Gasifier

ASU

O2
Steam

Cooling

Hot gas
filter
1 or 2 stages

Steam
Shift reaction &
CO2 removal

H2

SCR

Gas
turbine
Steam
turbine

Combined stage

N2

N2

Figure 2 An Advanced scheme for IGCC with carbon capture [3]

The scheme shown above in Figure 2 for gas separation and conditioning is being developed in Australia by The
University of Queensland as the research provider under the R&D program directed by the Centre for Lo w Emission
Technology (www.clet.net). The use of a higher temperature membrane allo ws further simp lification of the gas
cooling, conditioning and separation step, but this scheme requires a high temperature membrane tolerant to the
syngas. The R&D program is currently focusing on metal doped silica membranes. In this work, we report on proof
of concept and long term testing of memb ranes for H2 separation in a single memb rane and mult i tube membrane
module similar to the schematic shown in Figure 3, in addition to a membrane operating in a membrane reactor
(MR) arrangement for the high temperature W GS reaction. The membranes are tested for several temperature and
pressure regimes, for single gas and binary mixtures, and for gas separation during WGS react ions, to determine
their performance in terms of permeation and gas separation properties.

Figure 3 Schemat ic of a mu lti tube membrane module for H2 and CO2 separation

298

J.C.Author
Diniz da
Costa
et al. / Procedia
Energy Procedia
1 (2009)
295302
name
/ Energy
00 (2008)
000000

2. Experi mental
Co mmercial D-alu mina tubes (OD-11mm Length-120mm) coated with a top -alumina layer were purchased from
Noritaki (Japan). Cobalt silica sol was prepared through the hydrolysis and condensation of tetraethyl orthosilicate
(TEOS) in ethanol and H2 O2 with cobalt nitrate hexahydrate (Co(NO3 )2 6H2 O) as described elsewhere [16]. A
solution of 42.2g TEOS in 600g ethanol was added to a second solution of 51.8g cobalt nitrate in 30%w/w aqueous
H2 O2 and vigorously stirred for 3 hours in an ice-cooled bath. The tubular alumina substrates were dip-coated at the
outer shell of the tube at a removal speed of 2cm.min -1 and an immersion time of 1 minute prior to removal. The
final tube membrane had a total of six layers which were sequentially calcined in air to 600C at a ramp rate of
0.7C.min -1 , and held at 600C for 4 hours. Finally the memb ranes were sintered in H2 at 500C for 15 hours to
reduce the cobalt oxide within the selective silica layer. Both ends of the tube were glazed to provide a good surface
for sealing. The membrane had a length of 60mm and an effective permeat ion area of 20cm2 .
The cobalt silica membranes were are also tested in a membrane reactor (M R) arrangement. The catalyst was
activated prior to experimentation following manufacturers protocols at 250C with 20% H2 and 20% H2 O in N2 at
20 ml.min -1 total flow. After activation, H2 was switched off and the H2 O/CO reactant flow was fed to the reactor,
with all gas flows controlled using high precision Cole Palmer rotameters. Water liquid flow was controlled by a
Bronkhurst liquid mass flow controller, prior to injection into an evaporator. The water to CO feed ratio 1:1 was
used in the experimental work.
In these experiments, the membrane and MR temperature was maintained within a muffle furnace using a
Eurotherm PID temperature controller. A back pressure valve maintained the feed pressure at the desired pressures
which was measured with a pressure transducer. Reactants, feed, retentate and permeate gases were tested in a
Shimad zu GC-2014 gas chromatograph using a Porapak Q colu mn with a thermal conductor detector (TCD) and
flame ionising detector (FID) to determine gas concentrations.

3. Results and Discussion

Figure 4 shows the single gas permeation results for the cobalt silica membrane. The permeation of mo lecules with
the smaller kinetic diameter such as He (d k = 2.6) and H2 (d k = 2.9) increased as a function of the temperature
whilst the opposite trend was observed for the larger molecules like CO 2 (d k = 3.3) and N2 (dk = 3.64). These
permeat ion trends as observed elsewhere [18-20] are indicative of activated transport or molecular sieving
mechanis m following a temperature dependency flu x equation derived fro m the model of transport through
microporous crystalline membrane proposed by Barrer [17]:

Jx

 E act dp
 DO K O exp

RT dx

(Equation 2)

where J is the flu x (mol.m-2 .s -1 ) through the membrane, Eact (kJ.mo l-1 ) is an activation energy, R the gas constant and
T the absolute temperature (K), Do and Ko are temperature independent proportionalities. The activation energies in
kJ.mo l-1 calculated fro m an Arrhenius relationship for single gas permeation were 8.0 (He), 16.4 (H2 ), -6.3 (N2 ) and
-6.3 (CO2 ).
A membrane module containing 4 cobalt doped silica memb rane tubes was subsequently assembled similarly to
the concept shown in Figure 3. Each tube membrane was individually tested with He and N 2 (see Figure 5) prior to
any H2 and CO2 permeation testing to verify their performance and whether leaks were occurring. Tests were
carried out up to 4 bar feed pressure and 400o C. Membrane 2 (M2) showed the best performance in terms of flu x
and selectivity. The He flu x reached 14.2 cc.cm-2 .min -1 and single gas He/N2 selectivity of 76 at the tested
conditions. It was found that increasing the pressure caused a small reduction of selectivity, in particular for M 4,

J.C. Diniz
da Costa
al. / Energy
Procedia
1 (2009)
295302
Author
nameet/ Energy
Procedia
00 (2008)
000000

299

Permeance ( mol . m -2 . s -1 . Pa -1 )

which was attributed to a small leak rate. In other words, the other membranes (M1, M3 and M4) could possibly
have higher selectivities if they were leak free. The module was dis mantled and re-assembled three times, and leak
rates reduced or increased each time. At this stage, the importance of certain aspects of the mechanical design of a
mu lti tube membrane module became very apparent. The tubes require precise align ment to allow for Kalrez
orings to deliver good sealing capabilit ies. In our init ial mu lti tube design, alignment tolerance did not meet this
mechanical design requirement.

1E-7
7E-8
5E-8
3E-8
2E-8

He

1E-8
7E-9
5E-9

H2

3E-9
2E-9
1E-9
7E-10
5E-10

N2

3E-10
2E-10
CO2

1E-10
40

60

80

100

120

140

Temperature ( o C )

160

16

80

14

70

12

60

10
8

He

N2

He / N 2 Selectivity

cc . cm -2 . min -1

Figure 4 Single gas permeance of gases as a function of temperature.

50
40
30

20

10

0
M1

M2

M3

M4

M1

M2

M3

M4

Figure 5 Mult i tube membrane (left) init ial gas permeation at 300o C and feed pressure at 4 bara (right) selectivity.
After one week o f testing, M2 and M4 started having a high gas leak rate. The seals were changed, but the high
leak rate remained. Hence, it was decided to continue the tests with M1 and M3 only. The membranes were tested
for 56 days (1344 hours) of operation between 100-300o C and feed pressures 2.8-4.0 bar. The feed gas concentration
was generally 35-40 vol.% H2 and 60-65 vol.% CO2 . The performance of the membrane M1 (Figure 6) shows the
permeate H2 purity increased with temperature, and preferab ly in excess of 200o C. The 100o C isotherm was shown
as an example of the effect of a very small leak on the H 2 purity of gases in the permeate stream as the feed pressure
increased. Again, these results highlight the molecular sieving mechanis m wh ich was regulated by the temperature
flu x dependency as discussed above. The membrane M 1 delivered almost 98% H 2 purity at a single pass.

J.C.Author
Diniz da
Costa
et al. / Procedia
Energy Procedia
1 (2009)
295302
name
/ Energy
00 (2008)
000000

300

H2 Permeate Concentration (%)

100

300oC
200oC

95

90
100oC

85

80

75

70

65
280

300

320

340

360

380

400

Feed Pressure (kPa absolute)

Figure 6 Isotherms of H2 purity as a function of total gas mixture pressure for 40:60 (H2 :CO2 ) mixture.
A single membrane tube was also tested in a MR arrangement at operating temperatures from 300 to 375o C as
shown in Figure 7a. CO conversion increased with temperature and started levelling off at 67.5% at 375 o C due to
equilibriu m limitations. Membrane seals failed after this temperature. The H 2 /CO molar ratio in the permeate stream
increased from 1.9 to 40.5 as a function of temperature, whilst the same ratio in the MR reaction chamber increased
fro m 0.32 to 2.05. There are t wo important aspects associated with these changes. As temperature increased, so did
CO conversion. This means that for every mole of CO reacted it generate d one mole of H2 (and CO2 ) resulting in a
higher H2 partial pressure in the MR reaction chamber. The permeation of H2 is directly proportional to the driving
force, which in this case is the H2 partial pressure difference between the MR reaction chamber and the permeate
stream. As the permeate stream was kept at 1 at m, then it was essential that the H2 partial pressure in the reaction
chamber to be above 1 atm. The second aspect relates to the membranes activation energy as described in Equation
2. This can be clearly observed in Figure 7b. The purity of H2 in the permeate stream increased from 60 to 92.5% as
the temperature increased from 300 to 375o C, respectively. In a single separation step, the MR was able to enrich H2
in the permeate stream by 39% at 300o C and 48.5% at 375o C as compared to the MR reaction chamber.
Notwithstanding that the H2 and CO2 concentration in the MR reaction chamber were equivalent, H 2 selectively
permeate through the membrane. Hence, higher temperature allowed for the comb ined effect of higher CO
conversion and H2 permeation to act synergistically.
70
90

65

55
50
45
40

CO Conversion (%)

35
30
25
20
15
10

Permeate Stream H2/CO molar ratio

80
70

MR Reaction Chamber H2/CO molar ratio

0
300 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375
o

C)

H2 Permeate Stream

60
50
40
30

Temperature (

Concentration ( vol/vol % )

60

H2 MR Reaction Chamber

20
300 305 310 315 320 325 330 335 340 345 350 355 360 365 370 375

Temperature ( o C )

Figure 7a (left) M R CO conversion and H2/CO ratios and 7b (right) H2 concentration in the permeate stream and
MR reaction chamber, both figures as a function of temperature.

J.C. Diniz
da Costa
al. / Energy
Procedia
1 (2009)
295302
Author
nameet/ Energy
Procedia
00 (2008)
000000

301

The flu x temperature dependency trend of H2 and CO2 permeation makes cobalt silica membranes attractive for
application in coal gasification processes. The H2 /CO2 single gas selectivity increased fro m 17 to 131, as the
temperature increased from 50 to 150o C, respectively. The positive energy of activation for H 2 permeation coupled
with the negative energy of activation for CO2 permeation of metal doped silica memb ranes provide a favourable
fundamental property for engineering design of gas separation at higher temperatures (up to 500 o C) such as those
required in a coal gasificat ion process. This point can be further supported by the mult i tube membrane module,
which delivered H2 purity in excess of 98% at 300o C and feed pressures ranging from 2.8 to 4 at mospheres. The
temperature testing was limited to 300o C as Kalrez orings start melt ing at about 315o C. The membranes were also
showed to be encouragingly robust, as they performed consistently for 55 days.
For the WGS react ion, the MR delivered lo wer gas purity for a gas mixture containing ~40% H 2 . For instance, a
CO conversion of 67.5 mo les at 375o C lead to the production of 67.5 mo les of H2 and CO2 , whist 32.5 mo les of CO
were not consumed. This gives approximately a ternary mixture of 40% (H 2 ), 40% (CO2 ) and 20%(CO). In this case
the membrane was able to deliver a permeate stream containin g 92.5% H2 , about 5.5% lower than the result
obtained for a binary mixture. These results strongly suggest that the separation of a ternary gas mixture is more
complex than a binary mixture, thus slightly reducing H2 purity in the permeate stream.
4. Conclusions
Metal (cobalt) doped silica membranes delivered h igh H2 /CO2 single gas selectivity up to 131. The flu x
temperature dependency allowed for a positive energy of activation for H2 permeation wh ilst the energy of
activation for CO2 was negative. A multi tube membrane module was tested up to 300o C and 4 at mospheres for 55
days(1344 hours) for b inary feed gas mixtures containing H 2 and CO2 at 40:60 concentration ratio. The best
memb rane performance for H2 purity was in excess of 98%. Fo r the high temperature WGS react ion, H2 selectively
permeated through the membrane; this effect also increased with temperature. Higher temperature also aided in CO
conversion which translated into more H2 (and CO2 ) being produced in the MR reaction chamber. Hence, higher
temperature allowed for the comb ined effect of higher CO conversion and higher H2 permeat ion. Nevertheless, the
MR delivered lower gas purity for a ternary gas mixture containing ~40% H 2 as compared to binary gas mixture
separation. For a ternary mixture of 40% (H2 ), 40% (CO2 ) and 20% (CO), which is equivalent to 67.5% CO
conversion, the memb rane delivered a permeate stream containing 92.5% H 2 , about 5.5% lo wer than the results
obtained for a binary mixture. These results strongly suggest that the separation of a ternary gas mixture is more
complex than a binary mixture. The purity of H2 in the permeate stream increased fro m 60 to 92.5% as the
temperature increased fro m 300 to 375o C, respectively.
Acknowledgements
The authors acknowledge financial support for project GS001 by the Centre fo r Low Emission Technology
(www.clet.net). Special thanks to Dr Shaomin Liu, Dr Suraj Gopalakrishnan, Mr Tsutomu Tasaki and Mr Scott
Battersby for their experimental work.
References
1.
2.
3.
4.
5.
6.

BP, BP statistical review of world energy, 2002.

R. H. Williams, Toward zero emissions for transportation using fossil fuel, in VIII Biennal conference on
transportantion, energy and environmental policy, 2001, Monterey, CA, USA.
J. C. Din iz da Costa, D. Sholl, G. Reed and K. Thamb imuthu, Gas separation in Coal Gasification, AICh E
Spring Meeting, Orlando Fl, USA, 24-28 April 2006, vol 1(2006) 142f.
S. Giessler, L. Jordan, J.C. Diniz da Costa, and G.Q.M. Lu, Performance of hydrophobic and hydrophilic silica
memb rane reactors for the water gas shift reaction, Separation and Purification Technology, 32 (2003) 255.
C. Wheeler, A. Jhalan i, E.J. Klein, S. Tu mmala, and L.D. Sch midt, The water-gas-shift reaction at short
contact times, Journal of Catalysis, 223 (2004) 191.
Y. S. Lin, M icroporous and dense inorganic membranes: current status and prospective, Sep. Purification
Tech., 25 (2001) 39.

302

7.

D. Lee, L. Zhang, S. T. Oyama, S. Niu, R. F. Saraf, Synthesis, characterization, and gas permeation properties
of a hydrogen permeable silica memb rane supported on porous alumina, J. Membrane Sci., 231 (2004) 117.
S. Morooka, K. Kusakabe, Inorganic membranes for gas separation at elevated temperatures, Industrial
Ceramics, 20 (2000) 15.
J. C. Diniz da Costa, G. Q. Lu, V. Rudolph, Y. S. Lin, Novel molecu lar sieve silica (M SS) membranes:
characterisation and permeation of single-step and two-step solgel membranes, J. Membrane Sci., 198 (2002)
9.
D. S. Sholl, Y. S. Ma, Dense Metal Membranes for Production of High Purity Hydrogen, MRS Bu llet in, 31
(2006) 770.
P. Kamakoti, B. D. Morreale, M. V. Ciocco, B. H. Howard, R. P. Killmeyer, A. Cugini, D. S. Sholl, Prediction
of hydrogen flu x through sulfur tolerant binary alloy memb ranes, Science, 307 (2005) 569 .
G. Barbieri, A. Brunetti, T. Granato, P. Bernardo, and E. Drioli, Engineering evaluations of a catalytic
memb rane reactor fo r the water gas shift reaction, Industrial and Eng ineering Chemistry Research, 44 (2005)
7676.
J. N. Armor, Applications of catalytic inorganic membrane reactors to refinery products, Journal of Memb rane
Science, 147 (1998) 217.
S. Battersby, D. Miller, M. Zed, J. Patch, V. Rudolph, M.C. Du ke, and J.C.d. Costa, Silica membrane reactors
for hydrogen processing, Advances in Applied Ceramics, 106 (2007) 29.
A. Julbe, D. Farrusseng, and C. Guizard, Po rous ceramic membranes for catalytic reactors - overview and new
ideas, Journal of Membrane Science, 181 (2001) 3.
S. Battersby, M. C. Duke, S. Liu, V. Rudolph, J. C. Din iz da Costa, Metal Doped Silica Membrane Reactor:
Operational effects of reaction and permeation for the water gas shift reaction, Journal of Me mbrane Science,
316 (2008) 46.
R. M. Barrer, Porous crystal membranes, J. Chem. Soc. Faraday Transaction., 86 (1990) 1123.
M.C. Duke, J.C. Din iz da Costa, G.Q. (Max) Lu, M. Petch, and P. Gray, Carbonised template mo lecular sieve
silica membranes in fuel processing systems: permeation, hydrostability and regeneration, J. Memb. Sci, 241
(2004) 325.
R.M. de Vos and H. Verweij, High-selectiv ity, high-flo w rate silica membranes for gas separation, Science,
279 (1998) 1710.
M. Kanezashi and M. Asaeda, Hydrogen permeation characteristics and stability of Ni-doped silica memb ranes
in steam at high temperature, J. Memb rane Sci, 271 (2006) 86.

8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.

J.C.Author
Diniz da
Costa
et al. / Procedia
Energy Procedia
1 (2009)
295302
name
/ Energy
00 (2008)
000000