Simultaneous production of hydrogen and synthesis gas by combining water splitting with partial oxidation of methane in a hollow-fiber membrane reactor

Article abstract of DOI:10.1002/anie.200803899

(Figure Presented) Splitting water: Oxygen-permeable perovskite hollow-fiber membranes are used to produce hydrogen and synthesis gas from water and methane (see picture). The process yields valuable compounds, and provides insight into the interplay of c

Full text of DOI:10.1002/anie.200803899

Angewandte
Chemie
DOI: 10.1002/anie.200803899
Hydrogen Production
Simultaneous Production of Hydrogen and Synthesis Gas by
Combining Water Splitting with Partial Oxidation of Methane in a
Hollow-Fiber Membrane Reactor**
Heqing Jiang, Haihui Wang,* Steffen Werth, Thomas Schiestel, and Jꢀrgen Caro
Hydrogen is gaining more and more attention because it is
regarded as an important future fuel. Today, hydrogen is
mainly produced from nonrenewable natural gas and petro-
leum. With concerns over worldwide energy demands and
global climate change, alternative sources must be found.
Obviously, water is recommended as the ideal source for the
generation of large amounts of hydrogen.[1] In addition to
electrolysis, recently several new processes, such as photo-
voltaic–photoelectrochemical water splitting[2–4] and one-step
or multistep thermochemical water splitting[5,6] based on
focused solar[7–9] or nuclear[10] heat, have been developed.
Although water dissociation into oxygen and hydrogen is
conceptually simple [Eq. (1)], efficient hydrogen production
the poor oxygen permeability of these membranes, and only a
modest rate of 0.6 cm3 minꢁ1 cmꢁ2 could be obtained at
16838C.[19] To obtain a higher hydrogen production rate, one
possible way is to increase the rate of oxygen removal.
Balachandran et al. fed hydrogen on the permeation side of a
MIECM to consume the permeated oxygen.[21,22] In this case,
a high oxygen partial pressure gradient across the membrane
was established and a maximum hydrogen production rate of
10.0 cm3 minꢁ1 cmꢁ2 at 9008C was obtained.[21,22] However, the
amount of hydrogen produced was equivalent to the amount
consumed, which makes it unpractical. Therefore, it would be
attractive for both the feed and permeate stream to yield a
valuable compound.
Herein, for the first time we report the simultaneous
production of hydrogen and synthesis gas in a hollow-fiber
perovskite MIECM reactor using methane to consume the
permeated oxygen (see Figure 1). At high temperatures,
1
ð1Þ
H2O Ð H2 þ = O2
2
from water remains difficult as a result of the low equilibrium
constant of Kp ꢀ 2 ꢀ 10ꢁ8 at the relatively high temperature of
9508C.[11]
However, the hydrogen production rate can be signifi-
cantly improved by extracting oxygen with an oxygen-
permeable membrane. One possible type of membrane for
the efficient removal of oxygen at high temperatures is a
mixed oxygen-ion- and electron-conducting membrane
(MIECM) with high oxygen permeability.[12–16] Early studies
demonstrated the possibility of hydrogen production through
direct water decomposition by using MIECMs at the
extremely high temperature of 1400–18008C.[17–20] However,
the hydrogen production rate was relatively low because of
Figure 1. The concept of simultaneous production of hydrogen and
synthesis gas by combining water splitting with POM in a perovskite
oxygen-permeable hollow-fiber membrane.
[*] Prof. Dr. H. Wang
School of Chemistry & Chemical Engineering
South China University of Technology
Wushan Road, Guangzhou 510640 (P.R. China)
Fax: (+86)20-87110131
water dissociates into hydrogen and oxygen on the membrane
surface of the core side. Oxygen permeates from the core to
the shell side of the hollow fiber, where it is consumed by the
partial oxidation of methane (POM) to form synthesis gas
E-mail: hhwang@scut.edu.cn
H. Jiang, Prof. Dr. J. Caro
Institute of Physical Chemistry and Electrochemistry, Leibniz
University of Hanover (Germany)
according to CH4 + 1= O2!CO + 2H2. Thus, when operating
2
under equilibrium-controlled conditions, the water dissocia-
tion proceeds continuously as the oxygen is continuously
consumed by the POM to produce synthesis gas. The
advantage of this process is to give pure hydrogen as well as
synthesis gas, which can be used to synthesize a wide variety
of valuable hydrocarbons (for example, by Fischer–Tropsch
synthesis) and oxygenates (methanol).
A novel BaCoxFeyZr1ꢁxꢁyO3ꢁd (BCFZ) perovskite hollow-
fiber membrane was applied for the in situ removal of the
oxygen that is produced by high-temperature water splitting,
Dr. S. Werth
Uhde GmbH, Dortmund (Germany)
Dr. T. Schiestel
Fraunhofer Institute of Interfacial Engineering and Biotechnology
(IGB), Stuttgart (Germany)
[**] We gratefully acknowledge the financial support of the BMBF for
project 03C0343A under the auspices of ConNeCat and H.W. thanks
the NSFC (No.2070620) for financial support.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200803899.
Angew. Chem. Int. Ed. 2008, 47, 9341 –9344
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9341

Communications
because it exhibits a high oxygen permeation rate[23–25] and has
already been used, for example, in the production of oxygen-
enriched air[26,27] and the partial oxidation of hydrocar-
bons.[28,29] Figure 2 shows the influence of temperature on
magnitude higher than the rate of a nonreactive sweep gas
such as He.
From the above discussion it follows that the hydrogen
production rate is related to three reactions: water splitting,
oxygen transport through the membrane, and POM. When
the temperature is increased, the equilibrium constant of the
endothermic water splitting will increase according to the
Vanꢂt Hoff equation. So the equilibrium is shifted towards
water dissociation and more hydrogen can be produced.
Moreover, with increasing temperature, both the rate of the
POM and the permeability of the BCFZ hollow-fiber
membrane will increase. Thus, the hydrogen production rate
becomes higher with increasing temperature.
With increasing concentration of steam in the feed gas on
the core side, the equilibrium partial pressures of oxygen and
hydrogen will increase, thus providing a higher driving force
for oxygen permeation. As a result, the amount of hydrogen
obtained is expected to increase. As shown in Figure 3, the
Figure 2. H2 production rate on the core side as a function of temper-
ature. Core side: flow rate FH O =30 and FHe =10 cm3 minꢁ1; shell side:
2
50 cm3 minꢁ1 (FHe =45, FNe =3, and FCH =2 cm3 minꢁ1). Amount of
4
packed Ni/Al2O3 catalyst: 0.8 g. Effective membrane area: 0.86 cm2.
the hydrogen production rate. This rate increases as the
temperature rises from 800 to 9508C and a hydrogen flux
of 3.1 cm3 minꢁ1 cmꢁ2 was obtained at 9508C. It was also
found that the hydrogen production rate is very low
(< 0.026 cm3 minꢁ1 cmꢁ2 at 9008C, see Figure S1 in the Sup-
porting Information) if a sweep gas, such as He, is used on the
permeate side.
Clearly, the hydrogen production rate depends directly on
the rate of oxygen removal from the water dissociation
system. The oxygen transport can be described by the Wagner
equation for the oxygen flux j(O2) [Eq. (2)],[30,31] where sel and
Figure 3. H2 production rate on the core side at different steam
concentrations. Core side (steam + He): 40 cm3 minꢁ1; shell side:
50 cm3 minꢁ1 (FHe =45, FNe =3, and FCH =2 cm3 minꢁ1). Amount of
4
packed Ni/Al2O3 catalyst: 0.8 g. Effective membrane area: 0.86 cm2.
T=9508C.
@R T ln a
@x
1
sel sion
O2
jðO2Þ ¼ ꢁ
rmO with rmO
¼
ð2Þ
2
2
2
sel þ sion
ð4 FÞ
hydrogen production rate increased from about 0.2 to
3.1 cm3 minꢁ1 cmꢁ2 when the steam concentration was
changed from 5 to 75 vol.%. In addition, the hydrogen
production rate can be enhanced by increasing the methane
concentration (Figure 4), because a higher oxygen partial
pressure gradient across the perovskite membrane is estab-
lished, which leads to a higher oxygen permeation rate.
At the same time as hydrogen is produced on the core side
of the BCFZ hollow fiber by water dissociation, synthesis gas
can be produced on the shell side. Figure 5 shows the methane
conversion and the CO selectivity for feeding different
methane concentrations. When the amount of the feed gas
methane was small, most of the methane was totally oxidized
to give CO2 and H2O. On feeding more methane, the methane
conversion decreased gradually and the CO selectivity
increased. When 2 cm3 minꢁ1 methane diluted by
48 cm3 minꢁ1 inert gases was fed to the shell side, about
70% methane conversion and 60% CO selectivity were
obtained. However, for high methane concentrations the
calculated selectivities became lower again. This is probably
sion are the electronic and ionic conductivities, respectively, F
the Faraday constant, and 5mO the gradient of the chemical
2
potential of oxygen across the membrane. Assuming that sel @
sion and that @lnaO /@x can be approximated by ln(p0O /pO00 )/L
2
2
with p0O and p0O0 denoting the oxygen partial pressures on2the
shell and core2 sides and L the membrane thickness, the
2
Wagner equation reads [Eq. (3)]:
ln ðp0O =p0O0
Þ
R T
ð4 FÞ
2
2
ð3Þ
jðO2Þ ¼ ꢁ
2 sion
L
The oxygen partial pressure p0O on the permeate side can
be reduced much more efficiently2when feeding methane in
combination with a Ni-based catalyst (Sꢁd Chemie AG) to
the shell side. The permeated oxygen on the shell side is then
consumed very quickly by the oxidation of methane. As a
result of the increased oxygen partial pressure gradient across
the membrane, a higher hydrogen production rate of about
2.3 cm3 minꢁ1 cmꢁ2 at 9008C is obtained, which is two orders of
9342 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 9341 –9344

Angewandte
Chemie
reforming with hydrogen-selective Pd membranes has been
developed since 1992 by Mitsubishi Heavy Industries and
Tokyo Gas. The membrane reactor operates at 5508C and
produces 40 m3 (STP) hydrogenhꢁ1.[32] However, the Pd/Pd
alloy membrane can be poisoned by CO; therefore, a
subsequent water-gas shift stage is needed to convert CO to
CO2, which makes the operation more complex. On the other
hand, in the hydrogen-selective membrane reactor, only H2
rather than synthesis gas can be obtained. In our concept, we
can get both pure hydrogen, with an industrially interesting
flux (ca. 3 m3 hꢁ1 mꢁ2) without further water-gas shift reaction,
and synthesis gas. Compared to conventional steam reforming
and hydrogen-selective membrane-reactor-based steam
reforming, our multifunctional reactor contributes to the
concept of process intensification.
In conclusion, it is possible to produce significant amounts
of hydrogen from water splitting at around 9008C by using a
novel BCFZ oxygen-permeable hollow-fiber membrane. By
combining high-temperature water splitting with POM, not
only hydrogen but also synthesis gas can be obtained. This
process presents new insight into the interplay of catalysis and
separation in a membrane reactor. Besides, abundant raw
materials, such as water and methane (natural gas), were
used, which is of broad interest.
Figure 4. H2 production rate on the core side at different methane
concentrations. Core side: FH O =30 and FHe =10 cm3 minꢁ1; shell side:
2
50 cm3 minꢁ1 (FNe =3 and FCH + FHe =47 cm3 minꢁ1). Amount of
4
packed Ni/Al2O3 catalyst: 0.8 g. Effective membrane area: 0.86 cm2.
T=9508C.
Experimental Section
The dense BCFZ perovskite hollow-fiber membranes were manufac-
tured by phase-inversion spinning followed by sintering.[23–29] The
sintered fiber had a wall thickness of around 0.17 mm with an outer
diameter of 1.10 mm and an inner diameter of 0.76 mm. Figure S2 in
the Supporting Information shows a schematic diagram of the
membrane reactor used in this study. Two ends of the hollow fiber
were coated with Au paste. After sintering at 9508C, a dense Au film
that was not permeable to oxygen was obtained. Such an Au-coated
hollow fiber was sealed by a silicon rubber ring and the uncoated part
(3.0 cm), which was permeable to oxygen, was kept in the middle of
the oven thus ensuring isothermal conditions. A mixture of steam and
He was fed to the core side and a mixture of CH4, Ne, and He was fed
to the shell side. A Ni-based catalyst (Sꢁd Chemie AG) was packed
around and behind the hollow-fiber membrane. The CH4, He, and Ne
flow rates were controlled by gas mass-flow controllers (Bronkhorst).
The H2O flow was controlled by a liquid mass-flow controller
(Bronkhorst) and completely evaporated at 1808C before it was fed
to the reactor. All gas lines to the reactor and the gas chromatograph
were heated to 1808C. The concentrations of the gases at the exit of
the reactor were determined by an online gas chromatograph
(Agilent 6890). Assuming that the oxygen from water splitting on
the core side was totally removed and the flow rate at the outlet was
equal to that at the inlet, the H2 production rate after steam
condensation in the retentate on the core side was calculated from the
total flow rate Fcore (cm3 minꢁ1), the hydrogen concentration c(H2),
and the effective membrane area S (cm2) based on Equation (4).
Figure 5. Conversion X of methane and selectivity S of CO on the shell
side at different methane concentrations. Core side: FH O =30 and
2
F
F
He =10 cm3 minꢁ1; shell side: 50 cm3 minꢁ1 (FNe =3 and FCH
+
4
He =47 cm3 minꢁ1). Amount of packed Ni/Al2O3 catalyst: 0.8 g.
Effective membrane area: 0.86 cm2. T=9508C.
caused by the formation of coke, which is expected to occur
under these reaction conditions.
The net reaction in our concept is conventional methane
steam reforming according to H2O + CH4!3H2 + CO with a
H2/CO ratio of 3, which is unsuitable for the methanol or
Fischer–Tropsch syntheses. However, in our process we
obtain pure hydrogen on the core side as well as synthesis
gas with a H2/CO ratio of around 2 on the shell side. Synthesis
gas with such a ratio is usually produced by oxygen-blown
autothermal reforming, which requires a costly oxygen
separation plant. Therefore, our concept provides a new
way to obtain a Fischer–Tropsch synthesis gas. Furthermore,
the pure hydrogen produced by our method can be used to
operate the hydrocracking step in the product refinery section
of the Fischer–Tropsch plant.
Fcore cðH2Þ
ð4Þ
JðH2Þ ¼
S
The CH4 conversion X(CH4) and the CO selectivity S(CO) on the
shell side were calculated as Equations (5) and (6), where F(i) is the
Hydrogen can also be produced with a hydrogen-selective
membrane reactor in which methane steam reforming takes
place with a reforming catalyst. The technology of steam
ꢀ
ꢁ
FðCH4,outÞ
FðCH4,inÞ
XðCH4Þ ¼ 1ꢁ
ꢂ 100 %
ð5Þ
Angew. Chem. Int. Ed. 2008, 47, 9341 –9344
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
9343

Communications
FðCO,outÞ
SðCOÞ ¼
[16] C. Chen, S. J. Feng, S. Ran, D. C. Zhu, W. Liu, H. J. M.
Bouwmeester, Angew. Chem. 2003, 115, 5354 – 5356; Angew.
Chem. Int. Ed. 2003, 42, 5196 – 5198A.
[17] B. Cales, J. F. Baumard, High Temp. High Pressures 1982, 14, 681.
[18] J. Lede, F. Lapicque, J. Villermaux, B. Cales, A. Ounalli, J. F.
Baumard, A. M. Anthony, Int. J. Hydrogen Energy 1982, 7, 939 –
950.
[19] H. Naito, H. Arashi, Solid State Ionics 1995, 79, 366 – 370.
[20] Y. Nigara, K. Watanabe, K. Kawamura, J. Mizusaki, M.
Ishigame, J. Electrochem. Soc. 1997, 144, 1050 – 1055.
[21] U. Balachandran, T. H. Lee, S. Wang, S. E. Dorris, Int. J.
Hydrogen Energy 2004, 29, 291 – 296.
[22] U. Balachandran, T. H. Lee, S. E. Dorris, Int. J. Hydrogen Energy
2007, 32, 451 – 456.
[23] T. Schiestel, M. Kilgus, S. Peter, K. J. Caspary, H. H. Wang, J.
Caro, J. Membr. Sci. 2005, 258, 1 – 4.
[24] H. H. Wang, P. Kꢃlsch, T. Schiestel, C. Tablet, S. Werth, J. Caro,
J. Membr. Sci. 2006, 284, 5 – 8.
[25] H. H. Wang, T. Schiestel, C. Tablet, M. Schroeder, J. Caro, Solid
State Ionics 2006, 177, 2255 – 2259.
[26] H. H. Wang, S. Werth, T. Schiestel, J. Caro, Angew. Chem. 2005,
117, 7066 – 7069; Angew. Chem. Int. Ed. 2005, 44, 6906 – 6909.
[27] C. Hamel, A. Seidel-Morgenstern, T. Schiestel, S. Werth, H. H.
Wang, C. Tablet, J. Caro, AIChE J. 2006, 52, 3118 – 3125.
[28] H. H. Wang, C. Tablet, T. Schiestel, S. Werth, J. Caro, Catal.
Commun. 2006, 7, 907 – 912.
[29] J. Caro, K. J. Caspary, C. Hamel, B. Hoting, P. Kꢃlsch, B.
Langanke, K. Nassauer, T. Schiestel, A. Schmidt, R. Scho-
mꢅcker, A. Seidel-Morgenstern, E. Tsotsas, I. Voigt, H. H.
Wang, R. Warsitz, S. Werth, A. Wolf, Ind. Eng. Chem. Res. 2007,
46, 2286 – 2294.
[30] M. Schroeder, Phys. Chem. Chem. Phys. 2005, 7, 166 – 172.
[31] R. Merkle, J. Maier, H. J. M. Bouwmeester, Angew. Chem. 2004,
116, 5179 – 5183; Angew. Chem. Int. Ed. 2004, 43, 5069 – 5073.
[32] R. Dittmeyer, J. Caro in Handbook of Heterogeneous Catalysis,
Vol. 4 (Eds.: G. Ertl, H. Knꢃzinger, F. Schꢁth, J. Weitkamp),
Wiley-VCH, Weinheim, 2008, p. 2221.
ð6Þ
ꢂ 100 %
FðCH4,inÞꢁFðCH4,outÞ
flow rate of species i on the shell side, calculated based on the
measured concentration of the respective species and the total flow
rate measured by a bubble flowmeter.
Received: August 7, 2008
Published online: October 23, 2008
Keywords: heterogeneous catalysis · hydrogen · perovskite ·
.
synthesis gas · water chemistry
[1] J. A. Turner, Science 2004, 305, 972 – 974.
[2] O. Khaselev, J. A. Turner, Science 1998, 280, 425 – 427.
[3] C. Srinivasan, Curr. Sci. 2006, 90, 756 – 757.
[4] D. Lu, T. Takata, N. Saito, Y. Inoue, K. Domen, Nature 2006, 440,
295 – 297.
[5] O. Coskuner, E. A. A. Jarvis, T. C. Allison, Angew. Chem. 2007,
119, 7999 – 8001; Angew. Chem. Int. Ed. 2007, 46, 7853 – 7855.
[6] Y. Tamaura, R. Uehara, N. Hasegawa, H. Kaneko, H. Aoki,
Solid State Ionics 2004, 172, 121 – 124.
[7] H. Ishihara, N. Hasegawa, H. Aoki, H. Kaneko, A. Suzuki, Y.
Tamaura, Solid State Ionics 2004, 172, 117 – 119.
[8] P. Ritterskamp, A. Kuklya, M.-A. Wꢁstkamp, K. Kerpen, C.
Weidenthaler, M. Demuth, Angew. Chem. 2007, 119, 7917 – 7921;
Angew. Chem. Int. Ed. 2007, 46, 7770 – 7774.
[9] H. Aoki, H. Kaneko, N. Hasegawa, H. Ishihara, A. Suzuki, Y.
Tamaura, Solid State Ionics 2004, 172, 113 – 116.
[10] X. Wu, K. Onuki, Tsinghua Sci. Technol. 2005, 10, 270 – 276.
[11] S. Ihara, Bull. Electrotech. Lab. 1977, 41, 259 – 280.
[12] R. Hartung, H.-H. Mꢃbius, Chem. Ing. Tech. 1968, 40, 592 – 600.
[13] Y. T. Liu, X. Y. Tan, K. Li, Catal. Rev. Sci. Eng. 2006, 48, 145 –
198.
[14] J. Pꢄre-Ramirez, B. Vigeland, Angew. Chem. 2005, 117, 1136 –
1139; Angew. Chem. Int. Ed. 2005, 44, 1112 – 1116.
[15] Z. P. Shao, W. S. Yang, Y. Cong, H. Dong, J. H. Tong, G. X.
Xiong, J. Membr. Sci. 2000, 172, 177 – 188.
9344 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 9341 –9344