Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
1
SrCeO3-BASED PROTONIC CONDUCTORS FOR HYDROGEN PRODUCTION AND SEPARATION BY WATER GAS SHIFT, STEAM REFORMING, AND CARBON DIOXIDE
REFORMING REACTIONS
By
JIANLIN LI
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
2009
2
© 2009 Jianlin Li
3
To my grandmother, parents and sisters and friends who encouraged and supported me in good times and bad
4
ACKNOWLEDGMENTS
The completion of the research work for this dissertation would be impossible without the
assistance of so many people. First and foremost I would like to thank my advisor, Dr. Eric D.
Wachsman, who guided me through my entire course of study and encouraged me to reach a
higher level of success. I am truly grateful for his generosity in sharing his knowledge and
insights with me.
I would like to thank Florida Institute for Sustainable Energy and the National Aeronautics
and Space Administration for funding NASA3-2930. I would also like to thank my committee
members Dr.Mark Orazem, Dr. David Norton, Dr. Scott Perry and Dr. Ying Shirley Meng
(replacement for Dr. Wolfgang Sigmund) for their valuable time and contributions to my
dissertation.
In addition, I would like to thank Dr. Heesung Yoon and Dr. Takkeun Oh for their valuable
discussion and assistance. Thanks to Dr. Sean Bishop and Dr. Martin Van Assche for their
valuable comments, editing and friendship. Furthermore, I want to thank the following people for
their expert advice, assistance and valuable friendship: Dr. Keith Duncan, Dr. Xin Guo, Dr.
Guojing Zhang, Dr. Yanli Wang, Dr. Aijie Chen, Dr. Cynthia Kan, Dr. Jeremiah Smith, Mr. Eric
Macam, Mr. Dohwon Jung, Mr. Bryan Blackburn, Mr. Danijel Gostovic, Mr. Nicholas Vito, Dr.
Matthew Camaratta, Dr. Shobit Omar, Dr. Jinsoo Ahn, Mr. Eric Armstrong, Mr. Dongjo Oh, Mr.
Byungwook Lee, Mr. Kangtaek Lee, Dr. Briggs White, and our secretary Mrs. Jennifer Tucker.
Most of all, many thanks to my parents and other family members for their unselfish
support and love throughout my life. Their trust in me has made me a strong and confident
person. I also would like to thank my friends Dr. Qi Wei, Mrs. Rongrong Liu, Dr. Xiaomin Lv,
Mr. Tianyuan Deng, Mr. Ting Zhu, Mr. Zhiliang Kong, Mrs. Xing Zhang, Mr. Hanneng Li, and
others which are far too many to mention for their priceless friendship.
5
TABLE OF CONTENTS
page
ACKNOWLEDGMENTS.................................................................................................................... 4
LIST OF TABLES................................................................................................................................ 8
LIST OF FIGURES .............................................................................................................................. 9
ABSTRACT ........................................................................................................................................ 12
CHAPTER
1 INTRODUCTION....................................................................................................................... 14
2 LITERATURE REVIEW ........................................................................................................... 19
2.1 Hydrogen Production Technologies................................................................................. 19 2.1.1 Thermochemical Processes ................................................................................... 19 2.1.2 Electrolytic Processes ............................................................................................ 20 2.1.3 Photolytic Processes .............................................................................................. 20
2.2 Hydrogen Separation Membranes .................................................................................... 21 2.3 Proton Conducting Materials ............................................................................................ 22 2.4 Structure of SrCeO3........................................................................................................... 23 2.5 Proton Transport in SrCeO3 .............................................................................................. 24 2.6 Hydrogen Permeation ....................................................................................................... 26 2.7 Hydrogen Membrane System Design .............................................................................. 27
3 FABRICATION OF SUPPORTED TUBULAR SrCe 0.9Eu 0.1O3-δ and SrCe0.7Zr0.2Eu0.1O3-δ THIN FILM MEMBRANES ................................................................... 38
3.1 Introduction ....................................................................................................................... 38 3.2 Fabrication of Supported Thin Film Membranes ............................................................ 39
3.2.1 Materials Synthesis ................................................................................................ 39 3.2.2 NiO-SCZ82 Slurry for Support ............................................................................ 39 3.2.3 SrCe0.7Zr0.2Eu0.1O3-δ Thin Film Membranes on NiO-SCZ82 Support ............... 40
4 HIGH TEMPERATURE SrCe0.9Eu0.1O3 -δ PROTON CONDUCTING MEMBRANE REACTOR FOR H2 PRODUCTION USING THE WATER GAS SHIFT REACTION ..... 50
4.1 Introduction ....................................................................................................................... 50 4.2 Experimental...................................................................................................................... 52 4.3 Results and Discussion ..................................................................................................... 53
4.3.1 Thermodynamic Calculation ................................................................................. 53 4.3.2 Experimental Conversion ...................................................................................... 54 4.3.3 H2 Production ......................................................................................................... 57
4.4 Conclusions ....................................................................................................................... 58
6
5 STABILITY OF SrCe1-xZrxO3-δ UNDER WATER GAS SHIFT REACTION CONDITIONS............................................................................................................................. 67
5.1 Introduction ....................................................................................................................... 67 5.2 Experimental...................................................................................................................... 68 5.3 Results and Discussion ..................................................................................................... 69
5.3.1 Stability under Wet CO Conditions ...................................................................... 69 5.3.2 Decomposition Mechanism................................................................................... 71 5.3.3 Hydrogen Permeability ......................................................................................... 72
5.4 Conclusions ....................................................................................................................... 73
6 HYDROGEN PERMEATION OF THIN SUPPORTED SrCe0.7Zr0.2Eu0.1O3-δ MEMBRANES UNDER DIFFERENT OXYGEN PARTIAL PRESSURE .......................... 82
6.1 Introduction ....................................................................................................................... 82 6.2 Experimental...................................................................................................................... 82
6.2.1 Membrane Fabrication .......................................................................................... 82 6.2.2 Membrane Morphology......................................................................................... 83 6.2.3 Membrane Permeation .......................................................................................... 83
6.3 Result and Discussion ....................................................................................................... 83 6.3.1 Heat Treatment ...................................................................................................... 83 6.3.2 Flow Rate Effect on H2 Permeation ..................................................................... 84 6.3.3 H2 Permeation as a Function of Thickness .......................................................... 84 6.3.4 Effect of Temperature, H2 and H2O Partial Pressure in the Feed Side on H2
Permeation............................................................................................................. 85 6.3.5 Activation Energy .................................................................................................. 86 6.3.6 Long Term Stability .............................................................................................. 87
6.4 Conclusions ....................................................................................................................... 87
7 SrCe0.7Zr0.2Eu0.1O3-δ-BASED HYDROGEN TRANSPORT WATER GAS SHIFT REACTOR................................................................................................................................... 95
7.1 Introduction ....................................................................................................................... 95 7.2 Experimental...................................................................................................................... 95 7.3 Results and Discussion ..................................................................................................... 96
7.3.1 Heat Treatment of the Membranes ....................................................................... 96 7.3.2 H2O/CO Effect on CO Conversion....................................................................... 96 7.3.3 H2O/CO Effect on H2 Production ......................................................................... 97 7.3.4 H2O/CO Effect on H2 Production and H2/CO ..................................................... 98 7.3.5 Flow Rate Effect on WGS Reaction..................................................................... 99 7.3.6 CO Concentration Effect on WGS Reaction ....................................................... 99 7.3.7 Long Term Stability ............................................................................................ 100
7.4 Conclusions ..................................................................................................................... 100
7
8 HIGH TEMPERATURE SrCe0.7Zr0.2Eu0.1O3-δ MEMBRANE REACTOR FOR H2 PRODUCTION AND SEPARATION USING THE STEAM REFORMING OF METHANE ................................................................................................................................ 112
8.1 Introduction ..................................................................................................................... 112 8.2 Experimental.................................................................................................................... 113 8.3 Results and Discussion ................................................................................................... 114
8.3.1 Thermodynamic Calculation Results ................................................................. 114 8.3.2 Experimental Results ........................................................................................... 116
8.3.2.1 Influence of CH4/H2O on the SRM .................................................... 116 8.3.2.2 Influence of CH4 concentration on the SRM ..................................... 117 8.3.2.3 Influence of total flow rate on the SRM ............................................. 118 8.3.2.4 Influence of the H2 membrane reactor on the SRM .......................... 118 8.3.2.5 Long term stability ............................................................................... 119
8.4 Conclusions ..................................................................................................................... 119
9 HIGH TEMPERATURE SrCe0.7Zr0.2Eu0.1O3-δ PROTON CONDUCTING MEMBRANE REACTOR FOR CARBON DIOXIDE REFORMING OF METHANE .... 134
9.1 Introduction ..................................................................................................................... 134 9.1.1 Carbon Dioxide Reforming of Methane (CDRM) ............................................ 134 9.1.2 Membrane Reactors for the CDRM ................................................................... 135 9.1.3 Reaction Mechanism and Kinetics ..................................................................... 136
9.2 Experimental.................................................................................................................... 138 9.3 Results and Discussion ................................................................................................... 139
9.3.1 CH4/CO2 Effect on Conversion, H2/CO and H2 Production ............................. 139 9.3.2 Flow Rate Effect on Conversion, H2/CO and H2 Production ........................... 142 9.3.3 CH4/CO2/H2O Effect on XCH4, XCO2, H2/CO and H2 Production ..................... 143
9.4 Conclusions ..................................................................................................................... 144
10 CONCLUSIONS AND FUTURE WORKS ........................................................................... 155
10.1 Conclusions ..................................................................................................................... 155 10.2 Future Work ..................................................................................................................... 157
LIST OF REFERENCES ................................................................................................................. 160
BIOGRAPHICAL SKETCH ........................................................................................................... 169
8
LIST OF TABLES
Table page 2-1 Properties of relevant hydrogen selective membranes [14]................................................. 30
2-2 Conductivities of potential proton conducting membranes [19]. ........................................ 31
2-3 Structural parameters of SrCeO3. .......................................................................................... 32
5-1 Intensity ratios between the strongest peaks of CeO2 and SCZ82. ..................................... 81
5-2 Tolerance factors of SrCe1-xZrxO3-δ. ..................................................................................... 81
6-1 Activation energy as a function of H2 partial pressure under dry H2 and H2/3% H2O conditions balanced by Ar. .................................................................................................... 94
6-2 Activation energy as a function of H2O partial pressure with a constant H2 flow rate of 20 cm3/min. ........................................................................................................................ 94
9
LIST OF FIGURES
Figure page 2-1 Proton conductivities of various oxides [37]. ....................................................................... 33
2-2 Structure of SrCeO3 A) (001) projection and B) AO12 (blue) and BO6 (green) ................ 34
2-3 XRD pattern of SrCeO3 [46] ................................................................................................. 35
2-4 Predominant proton transfer between oxygen sites (shown by arrows) in the CeO6 octahedra of orthorhombically distorted BaCeO3 and SrCeO3 ........................................... 36
2-5 Comparison four categories setups of cross-flow operation [14]. ...................................... 37
3-1 XRD patterns of as-calcined SCZ82 and SCZE721 samples at 1300 oC ........................... 42
3-2 DV-E Viscometer ................................................................................................................... 43
3-3 Viscosity of NiO-SCZ82 slurry as a function of shear rate. ................................................ 44
3-4 Schematic process flow chart for fabrication of SCZ721 thin film membranes on NiO-SCZ82 supports.............................................................................................................. 45
3-5 Tape caster for making ceramic green tapes ........................................................................ 46
3-6 Process sequence for fabricating one end closed green body supports .............................. 47
3-7 Pictures of tubular SCZE721 thin film membrane coated on the inner side of NiO (or Ni)-SCZ82 support at each processing step ......................................................................... 48
3-8 SEM images of the NiO-SCZ82 and SCZE72 ..................................................................... 49
4-1 Morphology of thin film membranes and experimental setup ............................................ 59
4-2 Thermodynamic equilibrium of WGS under A) H2O/CO =1/1 and B) H2O/CO =2/1 ...... 60
4-3 Blank reference effluent gas composition as a function of temperature under .................. 61
4-4 Catalytic effluent gas composition as a function of temperature under ............................. 62
4-5 Catalytic effluent gas composition with in situ H2 removal as a function of temperature for H2O/CO =2/1 feed gas ................................................................................ 63
4-6 Temperature dependence of XCO under 3% CO + 3 % H2O and 3% CO + 6% H2O ........ 64
4-7 H2 production under 3% CO + 6% H2O as a function of temperature for three reactor configurations. ........................................................................................................................ 65
10
4-8 H2 yield and syngas H2/CO ratio as a function of temperature under 3% CO + 6% H2O and with in situ H2 removal ........................................................................................... 66
5-1 Temperature profile and photograph of the membranes ...................................................... 74
5-2 XRD pattern of SrCeO3-δ after exposure to 2.8% CO and 5.6% H2O for 19 h .................. 75
5-3 XRD Pattern and their lattice parameters of SrCe1-xZrxO3-δ as-calcined at 1300 oC ......... 76
5-4 XRD Pattern of SrCe1-xZrxO3-δ after exposure to 2.8% CO and 5.6% H2O for 19 h at 800 oC...................................................................................................................................... 77
5-5 XRD Pattern of SrCe0.8Zr0.2O3-δ after exposure to 2.8% CO and 5.6% H2O for 19 h ....... 78
5-6 XRD Pattern of SrCe0.8Zr0.2O3-δ after stability experiment at different atmospheres at 800 oC...................................................................................................................................... 79
5-7 H2 permeation as a function of time under 5% CO and 3% H2O at 900 oC ....................... 80
6-1 SrCe0.7Zr0.2Eu0.1O3-δ membrane and experimental setup ..................................................... 88
6-2 H2 permeation and H2 recovery as a function of feed flow rates ........................................ 89
6-3 H2 permeation vs thickness at 900oC .................................................................................... 90
6-4 H2 permeation as a function of H2 partial pressure and temperature .................................. 91
6-5 H2 permeation as a function of feed steam concentration and temperature. ...................... 92
6-6 H2 permeation as a function of time...................................................................................... 93
7-1 Membrane morphology and experiment setup ................................................................... 101
7-2 Gas compositions of the reactor side effluent as a function of temperature..................... 102
7-3 XCO as a function of temperature ........................................................................................ 104
7-4 H2 production as a function of temperature ........................................................................ 105
7-5 H2 yield and H2/CO in the reactor side effluent as a function of temperature ................. 107
7-6 The XCO, H2 production and H2/CO in the reactor side effluent as a function of flow rates under 900 oC ................................................................................................................ 109
7-7 The XCO, H2 production and H2/CO in the reactor side effluent as a function of CO concentrations with H2O/CO=2/1 ....................................................................................... 110
7-8 The performance of the membrane reactor as a function of time under 900 oC .............. 111
11
8-1 Membrane morphology and experimental setup ................................................................ 121
8-2 Influence of CH4/H2O on XCH4 under thermodynamic equilibrium ................................. 122
8-3 Influence of CH4/H2O concentrations on XCH4 under thermodynamic equilibrium with CH4/H2O=1/2 and Ar as the diluent ........................................................................... 123
8-4 Thermodynamic calculation of carbon formation as a function of temperature and CH4/H2O [130] ..................................................................................................................... 124
8-5 Influence of CH4/H2O on XCH4............................................................................................ 125
8-6 Influence of CH4/H2O on SCO, SCO2 and H2/CO in reactor side effluent .......................... 126
8-7 Influence of CH4/H2O on H2 production ............................................................................ 127
8-8 Influence of CH4 concentration on SRM ............................................................................ 128
8-9 Influence of total flow rate on SRM ................................................................................... 129
8-10 Influence of reactor configurations on SRM. ..................................................................... 132
8-11 The performance of the membrane reactor as a function of time under 850 oC .............. 133
9-1 Membrane morphology and experimental setup ................................................................ 145
9-2 XCH4 and XCO2 as a function of temperature and CH4/CO2 ............................................... 146
9-3 SH2 and SCO as a function of temperature and CH4/CO2 .................................................... 147
9-4 H2 production as a function of temperature and CH4/CO2. ............................................... 148
9-5 H2/CO in the reactor side effluent as a function of temperature and CH4/CO2................ 149
9-6 XCH4, XCO2, SH2 and SCO as a function of total flow rate. .................................................. 150
9-7 H2/CO in reactor side effluent as a function of total flow rate. ......................................... 151
9-8 H2 production as a function of total flow rate. ................................................................... 152
9-9 XCH4 and XCO2 as a function of temperature. ...................................................................... 153
9-10 H2 production and H2/CO as a function of temperature. ................................................... 154
10-1 A SrCe0.7Zr0.2Eu0.1O3-δ thin film membrane coated on graphite-SrCe0.8Zr0.2O3-δ substrate. ............................................................................................................................... 159
12
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy
SrCeO3-BASED PROTONIC CONDUCTORS FOR HYDROGEN PRODUCTION AND
SEPARATION BY WATER GAS SHIFT, STEAM REFORMING, AND CARBON DIOXIDE REFORMING REACTIONS
By
Jianlin Li
August 2009 Chair: Eric Wachsman Major: Materials Science and Engineering
Hydrogen has been considered as an ideal energy carrier for a clean and sustainable energy
future. New ceramic membranes have potential to reduce the syngas (a mixture of hydrogen and
carbon monoxide) cost by 30-50% and incorporate hydrogen production and separation into one
unit. SrCe1-x-yZryEuxO3-δ has been investigated to maximize hydrogen production and enhance
stability. 10 at% europium was used to fabricate tubular micro-cracking free membranes. 20 at%
zirconium was used to enhance the stability of SrCe0.9Eu0.1O3-δ.
Supported SrCe0.7Zr0.2Eu0.1O3-δ thin film membranes on NiO-SrCe0.8Zr0.2O3-δ substrates
were developed. Hydrogen permeation flux through these membranes was proportional to the
transmembrane Hydrogen partial pressure gradient with a 1/4 dependence and controlled by bulk
diffusion. A maximum Hydrogen permeation of 0.23 and 0.21 cm3/cm2 min was obtained for the
33 μm thick SrCe0.7Zr0.2Eu0.1O3-δ membrane at 900 oC with 100% H2 and 97% H2/3% H2O as the
feed gases, respectively. Hydrogen permeation was stable under wet H2, and conditions of WGS
reaction, steam reforming of methane (SRM), and carbon dioxide reforming of methane
(CDRM).
13
Thermodynamic equilibrium calculations were carried out for WGS reaction and SRM.
Hydrogen production and separation through WGS reaction, SRM and CDRM with
SrCe0.7Zr0.2Eu0.1O3-δ membranes were investigated. In situ removal of hydrogen through
hydrogen membranes moves the reaction toward the products side resulting in higher conversion
and hydrogen yield. 77% and 44% increase in the CO conversion for the WGS reaction was
achieved compared to the thermodynamic calculation data under 900 oC with H2O/CO = 1/1 and
2/1, respectively. 73% and 42% enhancement in the hydrogen production was achieved
simultaneously. For the SRM, the hydrogen membrane increased both the CH4 conversion and
total hydrogen production by 15% at 900 oC compared to the conventional reactor with only Ni
catalyst.
Whereas the H2/CO in the syngas product from the SRM is too high to produce liquid fuels
through the Fischer-Tropsch process, it is too low from the CDRM. However, an appropriate
value can be obtained by combining the SRM and CDRM. The H2/CO between 700 oC to 900
oC, for instance, is between 1.9-1.7 and 2.5-2.0 for CH4/CO2/H2O = 2/1/1 and 2/1/1.5,
respectively.
14
CHAPTER 1 INTRODUCTION
H2 is perceived as an ideal energy carrier for a clean and sustainable energy future. It is the
simplest common element consisting of only one proton and one electron. It is the most abundant
chemical-energy resource in the world. However, it is not a primary source of energy as it occurs
only in nature in combination with other elements, primarily with oxygen in water and with
carbon, nitrogen and oxygen in living materials and fossil fuels.
Although generation of H2 from renewable energy sources has the potential to provide a
sustainable energy cycle, fossil fuels would provide a short- to medium-term solution to generate
H2 without additional adverse environmental impacts [1, 2]. The major source of H2 is steam
reformation of natural gas. Therefore, improvements in the efficiency and cost of H2 production
from natural gas are necessary in the near term. Gas separation membranes and membrane
reactors based on ion conducting ceramics may provide the technological advance necessary to
increase the efficiency and reduce the cost of H2 production from natural gas. However, other
sources of H2 must be developed for the envisioned H2 economy, and coal provides the greatest
U.S. domestic resource-based option. The U.S. DOE is developing a FutureGen plant based on
coal gasification, solid oxide fuel cells (SOFCs), and ion conducting membranes that will
produce H2 and electricity with zero emissions and carbon sequestration; thereby, not
contributing to global warming. The use of coal will help ensure America's energy security by
developing technologies that utilize a plentiful domestic resource.
Membrane reactor technology holds the promise to circumvent thermodynamic
equilibrium limitations by in situ removal of product species, resulting in improved chemical
yields. Mixed-conducting oxide-membrane technology presents the possibility for a dramatic
reduction in the cost of converting petroleum and coal derived feed stocks to H2 and other value-
15
added hydrocarbons. Some perovskite oxides such as SrCe1-xEuxO3-δ exhibit both ionic and
electronic (mixed) conductivity. Because of their significant electronic conductivity, these mixed
ionic-electronic conductors (MIECs) have an internal electrical short and the ionic species
selectively permeates through a dense film of the material under a differential partial pressure,
such as H2 permeation. The potential permeation rates of these materials are extremely high.
SrCeO3-δ is a protonic conductor with high protonic conductivity and relatively low
electronic conductivity. To be suitable for H2 separation, a membrane material must have
comparable protonic and electronic transference numbers. In addition, the proton transference
number must be much higher than the transference number for oxygen ion. To that end, my
colleague, Dr. Takkeun Oh, has investigated the effect of dopant concentration in SrCe1-xEuxO3-δ
(0.05≤x≤0.2) on ambipolar conductivity [3] and demonstrated that the maximum ambipolar
conductivity increases with temperature and Eu dopant concentration. However, it is difficult to
fabricate a tubular thin film membrane with En dopant concentration higher than 10 at% without
micro-cracking [4]. Therefore, 10 at% Eu dopant concentration was used in my work to maintain
mechanical stability.
My overall goal is to demonstrate the feasibility of producing H2 from hydrocarbon based
fuels using advanced proton conducting membranes. The objective of my research is to improve
the stability of SrCeO3-δ using Zr as a dopant; to fabricate tubular supported SrCe0.7Zr0.2Eu0.1O3-δ
thin film membranes; to measure the H2 permeation of the SrCe0.7Zr0.2Eu0.1O3-δ membrane; and
to incorporate H2 permeation and total production using this membrane through water gas shift
(WGS) reaction, steam reforming of methane (SRM) and carbon dioxide reforming of methane
(CDRM).
16
Fabrication of supported tubular thin film membranes: According to the Wagner
equation, when transport is bulk diffusion limited permeation through a MIEC membrane is
inversely proportional to thickness [5]. Therefore, our research has focused on the development
of thin film mixed protonic-electronic conducting membranes using porous tubular supports for
increased hydrogen production [6, 7]. SrCe0.9Eu0.1O3-δ and SrCe0.7Zr0.2Eu0.1O3-δ thin film
membranes were fabricated by tape casting followed by a rolling process (chapter 3).
SrCe0.9Eu0.1O3-δ membrane reactors for H2 production through WGS reaction: WGS
reaction converts CO and H2O to CO2 and H2. It is used to shift the CO/H2 ratio in the syngas
prior to Fischer-Tropsch synthesis and/or increase H2 yield. The WGS reaction is exothermic and
limited by thermodynamic equilibrium. When a H2 membrane reactor couples the H2 production
and separation together, continuous removal of H2 decreases the H2 concentration in the reaction
system and moves the reaction forward. As a result, CO conversion and H2 yield can be
increased. The thermodynamic equilibrium of the WGS reaction was calculated. The WGS
reaction was investigated under three reactor configurations and as a function of temperature and
H2O/CO (chapter 4).
Stability improvement of SrCe0.9Eu0.1O3-δ: The H2 permeation of SrCe0.9Eu0.1O3-δ is
stable under wet H2 atmospheres but degrades under dry H2 conditions [7]. It is unstable under
the WGS reaction conditions as well. Therefore, the stability of this material needs to be
improved. Zr has been used to improve the chemical stability of BaCeO3-δ system [8-12] and
SrCe0.95Yb0.05O3-δ [13]. The stability of SrCe0.9Eu0.1O3-δ was improved with zirconium dopant.
The stability of SrCe0.8Zr0.2O3-δ was investigated under different atmospheres. CO2 was found to
cause the decomposition of SrCe0.8Zr0.2O3-δ (chapter 5).
17
H2 permeation properties of the SrCe0.7Zr0.2Eu0.1O3-δ thin film membranes: The H2
permeation properties of the SrCe0.7Zr0.2Eu0.1O3-δ membranes are not known yet. It was
investigated as a function of H2 partial pressures, feed flow rates, steam partial pressures and
temperature. The activation energy of the permeation process was discussed. The long term
stability of H2 permeation under wet H2 and the conditions of WGS reaction and SRM was
investigated as well (chapter 6).
SrCe0.7Zr0.2Eu0.1O3-δ based H2 transport WGS reactor: Chapter 4 compares the CO
conversion and H2 yield under different reactor configurations and they are significantly
improved with the H2 membrane reactor. In chapter 7, the SrCe0.7Zr0.2Eu0.1O3-δ effect on the
WGS reaction was investigated in details as a function of temperature, H2O/CO, CO
concentration, and CO feed flow rates. A long term stability experiment was carried out as well.
SrCe0.7Zr0.2Eu0.1O3-δ membrane reactors for H2 production through SRM: Currently,
the major H2 is produced from SRM. The SRM reaction is limited by thermodynamic
equilibrium. It needs to be carried out at high temperature to achieve high CH4 conversion.
Therefore, the SRM reaction is highly capital intensive. Catalytic ceramic membranes supply an
option to incorporate H2 separation and SRM into one unit which can increase the CH4
conversion or decrease the operating temperature. In chapter 8, the thermodynamic equilibrium
of the SRM was calculated. The SrCe0.7Zr0.2Eu0.1O3-δ membrane effect on the SRM was
investigated by comparing the performance under three different reactor configurations. The
performance of the SRM with the SrCe0.7Zr0.2Eu0.1O3-δ membrane was investigated as a function
of temperature, CH4/H2O, CH4 concentration, and CH4 feed flow rates. The membrane stability
under the SRM conditions was studied as well.
18
SrCe0.7Zr0.2Eu0.1O3-δ membrane reactor for H2 production through CDRM: While
SRM is the major process for H2 production, it produces a large amount of CO2 simultaneously.
0.3-0.4 million cubic meters of CO2 will be produced when one million cubic meters of H2 is
produced through a typical SRM H2 plant. Therefore, CO2 sequestration has drawn lots of
interest. The capture and disposal of CO2 costs a significant portion of the total cost of H2
production by the SRM process. The net cost of CO2 disposal, however, could be significantly
reduced if CO2 sequestration is accompanied by an enhanced product. CDRM provides one
solution to sequester CO2 and produce syngas simultaneously. In chapter 9, the performance of
the CDRM with the SrCe0.7Zr0.2Eu0.1O3-δ membrane was investigated as a function of
temperature, CH4/CO2, CH4 concentration, and CH4 feed flow rates. In addition, whereas the
H2/CO in the syngas from the SRM is too high to produce liquid fuels through the Fischer-
Tropsch process, it is too low in the syngas from the CDRM. Therefore, the SRM and CDRM
was combined to obtain appropriate H2/CO values.
19
CHAPTER 2 LITERATURE REVIEW
2.1 Hydrogen Production Technologies
H2 production processes are based on the separation of H2 from H2-containing feedstocks.
It can be produced using a variety of domestic energy resources - fossil fuels, such as coal and
natural gas; renewables, such as biomass, and renewable energy technologies, including solar,
wind, geothermal, and hydropower; nuclear power. H2 production technologies fall into three
general categories: thermochemical processes, electrolytic processes and photolytic processes.
2.1.1 Thermochemical Processes
Steam methane reforming (SMR): SMR is the most efficient and widely used process for
the production of H2. About 95% of the H2 in the United States is produced using this process (3-
25 bar, 700 oC-1000 oC).
In this process, high-temperature steam is used to extract H2 from a methane source such as
natural gas. This process consists of three steps: 1) reformation of the methane with high
temperature steam to obtain a syngas; 2) using a WGS reaction to form H2 and CO2, and 3)
purification. The reactions are listed below:
Step 1: CH4 + H2OCO + 3H2 1 6.205 −+=∆ kJmolH o (2-1)
Step 2: CO + H2OCO2 + H2 1 6.40 −−=∆ kJmolH o (2-2)
After the first two steps, a membrane is required to extract high-purity H2 from the H2 and
CO2 stream.
Partial oxidation: In this process, a fuel and oxygen are combined in proportions such that
the fuel is converted into a mixture of H2 and CO. There are several modifications of this
process, depending on the composition of the process feed and type of the fossil fuel used. Partial
oxidation of methane can be described by the following equation:
20
CH4+1/2O2CO+2H2 1 36 −−=∆ kJmolH o (2-3)
Coal gasification: Coal is converted into syngas by reacting coal with oxygen and steam
under high pressures and temperatures. Its gasification reaction may be represented by the
(unbalanced) reaction equation:
CH0.8+O2+H2OCO+CO2+H2+other species (2-4)
An advantage of this technology is that CO2 can be separated and captured more easily
from the syngas instead of being released into the atmosphere. If CO2 can be successfully
sequestered, H2 can be produced from coal gasification with near-zero greenhouse gas emission.
This technology is most appropriate for large-scale, centralized H2 production.
Other thermal processes: Other processes include (1) splitting water using heat from a
solar concentrator, and (2) gasifying or burning biomass (i.e., biological material, such as plants
or agricultural waste) to generate a bio-oil or gas, which is then reformed to produce H2.
2.1.2 Electrolytic Processes
Electrolysis: In electrolysis, electricity is used to split water (H2O) into H2 and oxygen.
The reaction takes place in a unit called an electrolyzer. There are three major electrolyzers:
polymer electrolyte membrane electrolyzer, alkaline electrolyzers, and solid oxide electrolyzers.
Current electrolysis systems are very energy intensive. The challenge is to develop low cost and
more energy efficient electrolysis technologies.
2.1.3 Photolytic Processes
Photolytic methods: In photolysis, sunlight is used to split water. Two photolytic
processes are being explored: (1) photobiological methods, in which microbes, when exposed to
sunlight, split water to produce H2, and (2) photoelectrolysis, in which semi-conductors, when
exposed to sunlight and submersed in water, generate enough electricity to produce H2 by
21
splitting the water. These processes offer long-term potential for sustainable H2 production with
low environmental impact.
2.2 Hydrogen Separation Membranes
H2 selective membranes can be classified into four categories: polymeric, metallic, carbon
and ceramic. Table 2-1 summarizes their properties [14].
Polymeric membranes are dense membranes, transporting species through the bulk of the
material. They cope well with high pressure-drops and are low cost. However, their operating
temperatures are limited to 90-100 oC. They are sensitive to certain chemicals and have weak
mechanical strength.
Palladium and palladium alloy membranes are the typical metallic membranes. They have
been studied and used as membrane reactors [15-17]. They have high H2 selectivity. However,
palladium-based membrane reactors have been operated at low temperatures of 300-500 oC [18].
They are highly sensitive to chemicals such as sulphur, chlorine and even CO. In addition,
palladium based membranes are expensive since palladium is a precious metal.
Carbon membranes separate H2 from other gases using small pores which only H2 can pass
through. They are usually used in non-oxidizing environments from 500 to 900℃. However,
they are difficult to fabricate and very brittle. Their selectivity is low, in the range of 4-20.
Ceramic membranes are a combination of a metal with a non-metal. They can be porous or
dense. Porous ceramic membranes generally are separation membranes on more porous ceramic
substrates. Their operating temperature is between 200 and 600 oC. One drawback for these
membranes is their poor hydrothermal stability. Dense ceramic membranes are also called proton
conducting membranes. The selectivity is very high since only H2 ions can transport through the
membranes. SrCeO3-δ and BaCeO3-δ are typical materials with an operating temperature from
22
600 to 900℃. Their chemical stability in the presence of certain species (e.g., CO2 and H2S) is a
major concern.
2.3 Proton Conducting Materials
There are numerous kinds of proton conducting membranes. Table 2-2 summarizes their
conductivities and operating temperatures [19]. Among those proton conducting materials,
perovskite (in form of ABO3, e.g., SrCeO3, BaCeO3) and related structures are of interest as
proton selective membrane materials [20-30]. In order to enhance the ionization of H2, cermet
membranes with a continuous metallic phase were also studied [31, 32]. In addition, complex
perovskites in the form of A2B1+x′B1-x″MxO6-δ (where A=Ba or Sr, B’=trivalent ion and
B”=pentavalent ion) or A3B1+x′B2-x″Mx2O9-δ (where A=Ba or Sr, B’=trivalent ion and
B”=pentavalent ion) [5, 33-36] have been developed to increase stability of the perovskite
oxides.
To be commercially useful, H2 separation membranes, the perovskite oxides should have
both high electronic and protonic conductivity, higher than 0.1 S cm-1 [19], and be stable in
operating conditions. While the electronic conductivities of SrCeO3 and BaCeO3 are relatively
low, they can be increased significantly by substituting Ce4+ with aliovalent ions (Y, Yb, Gd and
Eu) [3, 20]. The oxygen vacancies, typically created by acceptor doping to maintain
electroneutrality, play an important role for proton conduction.
Figure 2-1 shows the proton conductivities of several electronic-protonic conducting
ceramics [37]. Among those ceramics, BaZrO3-δ exhibits the highest proton conductivity.
However, its electronic conductivity is very low. To achieve high H2 permeability, the electronic
and protonic transference numbers should be comparable [27, 28]. Therefore, BaZrO3-δ is not a
good candidate for H2 separation membrane. BaCeO3-δ based oxides exhibit oxygen ion
23
conductivity comparable to their proton conductivity and thus are not proper for applications
where oxygen is present [38, 39]. Furthermore, BaCeO3 undergoes a complex sequence of phase
transitions [40], including a first order phase transition from orthorhombic to rhombohedral at
663-673 K where the two distinct oxygen sites become crystallographically equivalent. In
contrast, SrCeO3 undergoes no high temperature structural phase transitions up to 1273 K [41]. It
has high total conductivity and highest proton transference number which is due to a distorted
orthorhombic structure of SrCeO3-δ inhibiting oxygen ion conduction [42, 43]. Therefore,
SrCeO3-δ based oxides can be promising for selective H2 separation if their electronic
conductivity can be improved by a proper doping.
2.4 Structure of SrCeO3
SrCeO3 has an orthorhombic structure at room temperature. The lattice parameters of
SrCeO3 are a=6.126 Å, b=8.574 Å, c=6.000 Å. The theoretical density is 5.81 g/cm3 [44]. The
space group is Pnma (no. 62) [45], Z=4, with the A cation, Sr, situated on mirror planes (4c); the
B-site cation, Ce, situated on centers of inversion (4b); and two oxygen positions: O (1) on
mirror planes (4c), and O (2) in general positions (8d). The coordination numbers of Sr, Ce and
O are 12, 6 and 6, respectively. Table 2-3 shows the parameters of SrCeO3 [42] . The crystal
structure of SrCeO3 is shown in Figure 2-2. Its XRD pattern is shown in Figure 2-3 [46].
The conductivity of SrCe1-xAxO3 (A= aliovalent ions) has drawn great interest from
researchers. The typical conductivity is found to be between 10-2 to 10-3 S cm-1[23, 27, 28, 47-
50]. The thermal expansion coefficient and thermal conductivity of SrCeO3 are 1.11 × 10-5 K-1
from room temperature to 1273 K and 2.95 W m-1 K-1 at room temperature, respectively [51].
The heat capacity was determined in the temperature range of 373-1400 K as follows [44]:
CP (J mol-1 K-1)= 120.1+5.45× 10-3 T-1.26× 106 T-2 (2-5)
24
2.5 Proton Transport in SrCeO3
Vehicle mechanism and Grotthuss mechanism are the two mechanisms generally accepted
for proton diffusion in perovskite systems. Vehicle mechanism was first brought forward by
K.D.Kreuer et al. in 1982 [52]. According to this model, a proton does not migrate as H+ but as
+3OH , +
4NH , etc., bonded to a “Vehicle ” such as H2O, NH3 etc. The “unloaded” vehicles move
in the opposite direction. As for SrCeO3 based materials under wet H2 (H2 & H2O) atmosphere,
the vehicle is •OOH . The major proton incorporation reaction is shown below using Kröger-Vink
notation:
•×•• ↔++ OOO OHOVOH 22 (2-6)
With Kröger-Vink notation, ••OV represents oxygen vacancies, X
OO represents oxide ions
on an oxygen lattice site, and •OOH represents protons associated with oxide ions on an oxygen
lattice site. Then, •OOH is driven by H2 partial pressure gradient and transfers. The radii of O2-
and OH- are 1.32Å and 1.35Å, and they have to go over a saddle point in the diffusion process,
i.e., they need to overcome high energetic barriers. According to K.D. Kreuer et al., the
activation enthalpy for site exchange of an oxygen and an oxygen-ion vacancy is 0.8eV
(77.2kJ/mol) [53]. Hence, their diffusion coefficients are relatively slow and vehicle mechanism
usually takes place at high temperature.
Over the last decade, general agreement has been formed that protons transfer between
fixed oxygen sites via the Grötthuss mechanism at intermediate temperatures in ABO3
perovskites. Isotope effect (H+/D+) measurements of perovskite oxides have suggested that the
conduction mechanism is due to proton hopping between adjacent oxygen ions (Grötthuss
mechanism) rather than by hydroxyl ion migration (vehicle mechanism) [54-57].
25
The process consists of two steps, the proton hopping translation between O-O bonds and
reorientation of the hydroxide ion on the oxygen site. Outside the high H2 partial pressure
surface, a 1/2H2 would decompose into a proton, H+, and an electron, e-. In the material, H+ hops
from O2- to O2-and e- hops between cerium ions.
+−+ ↔+ 34 CeeCe (2-7)
Inside the low H2 partial pressure surface, the H+ combines with e- to form 1/2H2,
terminating the net H2 diffusion. The proton hopping translation from one oxygen site to another
oxygen site depends on the O-O bond length. A longer O-O bond favors formation of O-H bond,
but impedes the proton hopping; conversely, a shorter O-O bond favors the proton hopping, but
it is not favorable for hydroxide ion reorientation.
Molecular dynamic studies in both cubic and orthorhombic perovskites proton conducting
oxides support for the Grötthuss mechanism. Munch et al determined that the reorientation
process occurs relatively fast (10-12s) compared to the proton transfer process (10-9s) indicating
the proton transfer process is the rate limiting step [58, 59] . The activation energy for rotational
diffusion of protonic defects is small, 65 meV (0.63kJ/mol) and the activation enthalpy for
proton diffusion was 0.41 eV (39.6kJ/mol) [60].
In cubic perovskites, neighboring oxygen ions are treated as equivalent sites. In contrast,
oxygen ions must be treated differently in low symmetry orthorhombic perovskites (such as
SrCeO3). The most basic oxygen sites are O1 and O2 in SrCeO3 and BaCeO3, respectively [60].
Proton transfer between oxygen sites in BaCeO3 and SrCeO3 is shown in Figure 2-4 [61]. The
long-range transport between O2 sites in BaCeO3 should be easier than transport between O1 and
O2 sites in SrCeO3 because the O1 and O2 are chemically different in SrCeO3. This difference in
proton transport is a possible reason for lower conductivity in SrCeO3 than BaCeO3 [62].
26
2.6 Hydrogen Permeation
H2 permeates through H2 permeable membranes via ambipolar diffusion of protons and
electrons under a H2 chemical potential gradient [63]. The motion of electrons, the minority
carrier, gives rises to the H2 permeation by charge compensated transport of protons in the same
direction.
A few theoretical works on H2 permeation in a MIEC have been published [64-66]. In
MIECs, the flux of each charge carrier species, k, is driven by chemical and electrical forces.
Along the dimension of net transport (x= membrane thickness), assuming the bulk diffusion is
the rate limiting step, mass transfer rates per cross-sectional area for component k are give by
[67]:
)()()( 22 dx
dFzdx
dFzFzRT
CDJ kk
k
kk
k
kk
Kkk
φµσηση −−=∇−=∇−= (2-8)
The terms in the parenthesis represent the chemical and electrical potential gradient, and zke is
the charge of the species k. In open circuit conditions, the net current resulting from all fluxes is
zero:
∑ ∑∑ −−
=====
)(03
1 dxdez
dxd
ezeJzIi k
k
k
kkk
k
φµσ (2-9)
Equation (2-9) can be rearranged:
∑−=
ii
i
i
dxd
ezt
dxd µφ (2-10)
12 =++ −+ OHe ttt (2-11)
where ti is the transference number of species i.
We insert equation (2-10) into equation (2-8) and obtain:
27
)(22 ∑+−
=i
i
i
ik
k
k
kk dx
dztz
dxd
ezj µµσ (2-12)
Now, equilibria between neutral and charged species are introduced. For MIECs, it is natural to
consider transport by oxygen ions, protons and electrons.
−=+ 2'2 OeO eOO ddd µµµ 22 +=− (2-13)
'eHH += + eHH ddd µµµ −=+ (2-14)
We insert equations (2-13) and (2-14) into equation (2-12) for the flux of oxygen ions:
])(2)[(4 2
2
2 dxdt
dxdtt
dt
j HH
OHe
OtotO
µµσ++−
−
− ++−
= (2-15)
where σtot is the total conductivity. If local thermodynamic equilibrium is achieved,
)ln(21
21
222
0OOOO PRT+== µµµ
2
ln2 OO PdRTd =µ (2-16)
We insert equation (2-16) into equation (2-15) and obtain:
]ln4
ln)(8
[1J''2
'2
''2
'2
2222-2 22O ∫ ∫ −++− −+−= O
O
H
H
P
P
P
P HOHtotOeHOtot PdttF
RTPdtttF
RTL
σσ (2-17)
where L is sample thickness.
Following the same manner, we obtain the proton flux:
∫ ∫ ++−= −+−++
''2
'2
''2
'2
2'222 ln)(2
ln4
1J 22H
O
O
H
H
P
P
P
P HeOHtotOOHtot PdtttF
RTPdttF
RTL
σσ (2-18)
2.7 Hydrogen Membrane System Design
Currently there are two types of membrane configurations: flat and tubular [14]. The
building block of a membrane system is called the module. Module types based on flat
membranes are the plate-and-frame and spiral-wound modules. Tubular type membrane modules
28
are subdivided into tubular (diameters > 10 mm), capillary (0.5 mm < diameters < 10 mm) and
hollow fiber modules (diameters < 0.5 mm) [68]. Table 2-4 lists the packing densities of these
modules.
There are single stage and multistage membrane processes. A stage is formed by one or
more membrane modules assembled into an operating unit that provides a specific function
different from any other stages that may be utilized in the same process. Multistage membrane
systems are built to improve membrane system performance.
Single stage membrane process is the simplest membrane process. Membrane operations
can be subdivided into dead-end and cross-flow operations [14]. In dead-end operation, there is
no retentate stream. It is not preferred since non-permeating species in time become more
abundant on the feed side, resulting in concentration polarization and driving force. As a result,
the transport through the membrane decreases. Instead, flows run alongside the membrane in the
cross-flow operations. Deterioration of membrane flux in time is limited in this configuration.
The cross-flow operation can be distinguished into four categories: co-current, counter-current,
cross-flow with perfect permeate mixing, and perfect mixing [14].
In the co-current operation, feed and permeate flows run in the same direction, whereas,
they run in opposite direction in the counter-current operation. In the perfect permeate mixing,
the permeate is mixed to form one homogeneous permeate composition along the membrane
length coordinate. The perfect mixing setup results in homogeneous compositions in both the
feed side and permeate side. Figure 2-5 compares the configurations and driving forces of these
four categories of cross-flow operations. Generally, the membrane results are obtained in this
sequence due to the driving forces: counter-current > cross-flow with permeate mixing > co-
current > perfect mixing [69].
29
Membranes can selectively take away reaction products, thereby shifting the equilibrium to
the product side. If chemical reactions are carried out in a membrane module, the system is
called a membrane reactor [14]. Catalysts are usually necessary to carry out reactions and
accommodated inside the membrane reactors. Three types of arrangements are found to
accomplish this [69]: catalyst placed inside the feed stream, catalyst placed in a membrane top
layer, and catalyst placed inside the membrane itself. To place catalyst inside the feed stream is
easy to prepare and operate. In contrast, it is difficult to replace the catalyst placed in a
membrane top layer or inside the membrane itself since replacing the catalyst usually means
replacing the whole membrane.
30
Table 2-1. Properties of relevant hydrogen selective membranes [14].
Dense Polymer Micro porous
ceramic Dense metallic Porous carbon Dense ceramic
Temperature range
<100 oC 200-600 oC 300-600 oC 500-900 oC 600-900oC
H2 selectivity low 5-139 >1000 4-20 >1000 H2 flux ( 10-3 mol/m2s) at dP=1 bar
low 60-300 60-300 10-200 6-80
Stability issues Swelling, mechanical strength
Stability in H2O
Phase transition
Brittle, oxidizing
Stability in CO2
Poisoning issues
HCl, SOx, (CO2)
H2S, HCl, CO Strong adsorbing vapors, organics
H2S
Materials Polymers Silica, alumina, zirconia, titania, zeolites
Palladium alloy
Carbon Proton conducting ceramics
Transport Mechanism
Solution/ diffusion
Molecular sieving
Solution/ diffusion
Surface diffusion, molecular sieving
Solution/ diffusion (proton conduction)
Development status
Commercial by Air Products, Linde, BOC, Air Liquide
Prototype tubular silica membranes available up to 90 cm. Other materials only small samples
Commercial by Johnson Matthey; prototype membrane tubes available up to 60 cm
Small membrane modules commercial, mostly small samples (cm2) available for testing
Small samples available for testing
31
Table 2-2. Conductivities of potential proton conducting membranes [19].
32
Table 2-3. Structural parameters of SrCeO3. Atom X Y z Sr 0.2500 0.0116 -0.0447 Ce 0.0000 0.5000 0.0000 O1 0.2500 0.6059 0.0432 O2 -0.0558 0.7006 0.2988
Table 2-4. Packing densities of different hydrogen membrane modules [68]. Module Plate-and-frame Spiral-wound Tubular Capillary Hollow fiber
Packing Density (m2/m3)
100-400 300-1000 300 600-1200 30000
33
-6
-5
-4
-3
-2
-1
0 0.5 1 1.5 2 2.5
Logσ
(S/c
m)
1000/T (K-1)
BaZrO3
BaCeO3
SrTiO3
SrCeO3
CaZrO3
SrZrO3
PH2O
=30 hPa
Figure 2-1. Proton conductivities of various oxides [37].
34
A
B
Figure 2-2. Structure of SrCeO3 A) (001) projection and B) AO12 (blue) and BO6 (green).
35
20 30 40 50 60 70 80
Inte
nsity
(a.u
.)
(0 1
1)
(2 1
1)
(1 2
1)
(3 1
1)
(0 2
2)
(1 2
2)
(0 3
7)
(4 0
2)
(2 3
1)
(3 1
3)
(4 2
2)
(1 1
6)
(0 4
4)
(4 0
4)
2θ (o)
Figure 2-3. XRD pattern of SrCeO3 [46].
36
Figure 2-4. Predominant proton transfer between oxygen sites (shown by arrows) in the CeO6 octahedra of orthorhombically distorted BaCeO3 and SrCeO3. The degree of basicity is indicated by the color of the oxygen sites (purple = more basic) [61].
37
Figure 2-5. Comparison four categories setups of cross-flow operation [14].
38
CHAPTER 3 FABRICATION OF SUPPORTED TUBULAR SrCe 0.9Eu 0.1O3-δ AND SrCe0.7Zr0.2Eu0.1O3-δ
THIN FILM MEMBRANES
3.1 Introduction
Perovskite-type oxides such as SrCeO3-δ exhibit significant proton conductivities in H2-
containing atmospheres when oxygen vacancies and other charged defects are introduced by the
partial substitution of trivalent cations for Ce [70]. We previously investigated the effect of
dopant concentration in SrCe1-xEuxO3-δ (0.05≤x≤0.2) on ambipolar conductivity [3] and found
that the maximum ambipolar conductivity increases with temperature and Eu dopant
concentration. However, we also found that Eu dopant concentrations higher than 10 at% result
in mechanical instability. Therefore, 10 at% Eu dopant is used in my dissertation.
Compared to planar membranes, tubular membranes have much larger area and do not
require any high-temperature seals to isolate permeated gas from input gas. In addition, the
Wagner equation shows that when transport is bulk diffusion limited permeation through a
mixed ionic-electronic conducting (MIEC) membrane is inversely proportional to thickness [5].
Thus, many studies have focused on the fabrication of thin film membranes [5, 28, 64, 66, 71-
73]. Therefore, our research has focused on the development of thin film mixed protonic-
electronic conducting membranes using porous tubular supports for increased H2 production [6,
7]. SrCe0.9Eu0.1O3-δ and SrCe0.7Zr0.2Eu0.1O3-δ thin film membranes were investigated in my work.
The fabrication of SrCe0.9Eu0.1O3-δ membrane has been addressed in reference [74]. In this
chapter, I focus on the fabrication of SrCe0.7Zr0.2Eu0.1O3-δ membranes. Based on our previous
experience, NiO-SrCe0.8Zr0.2O3-δ was used to fabricate the support structure to maintain
mechanical integrity. Eu was eliminated from the support composition since electronic
39
conduction is not functionally necessary for the support. NiO was used to create porosity and to
serve as the catalyst, by reduction to Ni when the membrane was subsequently exposed to H2.
3.2 Fabrication of Supported Thin Film Membranes
3.2.1 Materials Synthesis
Polycrystalline SrCe0.8Zr0.2O3-δ and SrCe0.7Zr0.2Eu0.1O3-δ powders were prepared by
conventional solid-state reaction by mixing stoichiometric amounts of SrCO3 (99.9%, Alfa-
Aesar), CeO2 (99.9%, Alfa-Aesar) ZrO2 (99.9%, Alfa-Aesar) and Eu2O3 (99.9%, Alfa-Aesar)
powders, followed by ball milling and calcining at 1300 oC. Figure 3-1 shows their XRD
patterns. Both are orthorhombic structure.
3.2.2 NiO-SCZ82 Slurry for Support
Homogeneous and stable slurry is very necessary for tape casting. The NiO-SCZ82 slurry
was achieved by two stage ball milling. Firstly, 46.7 wt% NiO was mixed with SCZ82 powder
and dissolved in ethanol and toluene which served as solvents. To stabilize the slurry against
flocculation of the particles, a certain amount of solsperse (2400SC, Avecia) was used as a
dispersant and added into the solution. This solution was ball milled for 24hrs. Secondly, Binders
and plasticizers were added to the solution and the solution was ball milled for another 24 hrs.
PVB was chosen as the binder to provide plasticity of the solution. Plasticizer can soften the
binder and increase the flexibility of the green body. For tape casting process, the plasticizer
must be soluble in the same solution used to dissolve the binder. Specific combinations of binder
and plasticizer are used in tape casting process. Here, PEG and DBP were used as the
plasticizers.
Rheology plays an important part in the processing of ceramics from colloidal suspensions.
When the suspension is consolidated by casting methods, including slip and tape casting, the
suspensions are required to contain the highest possible fraction of particles to reduce the
40
shrinkage during drying the cast and to produce a consolidated powder form with high packing
density. In addition, the suspension should have a low enough viscosity to be poured. [75]
The rheological properties of the slurry can be characterized by viscosity, η, defined by
γτη /= (3-1)
where τ and γ are the shear stress and shear rate, respectively. There are a number of types of
rheological behavior of colloidal suspension:
(1) Newtonian: Viscosity is constant with change in shear rate. (2) Thixotropic: Viscosity decreases as shear rate increases and is also a time dependent. (3) Dilatant/shear thickening: Viscosity increases as shear rate increases. (4) Plastic: Viscosity decreases as shear rate increases after an initial threshold stress. (5) pseudo-plastic/shear thinning: Viscosity decreases as shear rate increases
The viscosity of the slurry is measured by DV-E Viscometer (Brookfield) (Figure 3-2).
The accuracy is guaranteed to be ± 1%. Figure 3-3 shows the viscosity as a function of shear rate
at 25 oC. LV3 spindle is used and % torque is between 10 and 100 for the whole measurement.
The viscosity of the NiO-SCZ82 slurry is plastic and decreases with increasing shear rate.
3.2.3 SrCe0.7Zr0.2Eu0.1O3-δ Thin Film Membranes on NiO-SCZ82 Support
Figure 3-4 shows the process flow design for the preparation of the SrCe0.7Zr0.2Eu0.1O3-δ
(SCZE721) thin film membranes on tubular NiO-SCZ82 supports. The tubular NiO-SCZ82
support was fabricated using tape-casting (Pro-Cast) followed by a rolling process. Figure 3-5
shows the tape caster. The process sequence for making a one end closed tubular-type green
body support is shown in Figure 3-6. After the green tubes were pre-sintered at 1100 oC,
SCZE721 was coated on the inner side of the supported by colloidal coating. Then the SCZE721
membranes were sintered at 1520 oC together with the NiO-SCZ82 supports to achieve dense
membranes. Figure 3-7 and Figure 3-8 show the photographs and morphology of the tubular
membranes at different processing steps. As shown in Figure 3-8, the pre-sintered structures (A
41
and B) are very porous. The sintered structure is pretty dense. There are still some pores on the
supported structure, but those pores are isolated and close pores. The thin film surface is crack-
free.
42
20 30 40 50 60 70 80
2θ (o)
Inte
nsity
(a.u
.)
(4 2
2)
(0 3
7)
(0 2
2)
(2 1
1)
(4 0
2)
(0 1
1)
(1 1
6)
(4 0
4)
(1 2
2)
(2 3
1)
(0 4
4)
SCZ82
SCZE721
(1 2
1)
(3 1
1)
(3 1
3)
Figure 3-1. XRD patterns of as-calcined SCZ82 and SCZE721 samples at 1300 oC.
43
Figure 3-2. DV-E Viscometer.
44
0
5 103
1 104
1.5 104
2 104
2.5 104
0 0.1 0.2 0.3 0.4 0.5 0.6
Visc
osity
(cP)
Shear Rate (s-1)
Figure 3-3. Viscosity of NiO-SCZ82 slurry as a function of shear rate.
45
Figure 3-4. Schematic process flow chart for fabrication of SCZ721 thin film membranes on NiO-SCZ82 supports.
Mix NiO, SrCe0.8Zr0.2O3-δ, and dispersant with ethanol
and toluene, ball mill
Add binder, plasticizers and ball mill
De-air
Tape casting
Rolling on rod
Pre-sinter
Coat SrCe0.7Zr0.2Eu0.1O3-δ on inner side of the pre-sintered
support
Final sinter
Mix SrCe0.7Zr0.2Eu0.1O3-δ, and dispersant in ethanol,
ball mill
46
Figure 3-5. Tape caster for making ceramic green tapes.
47
Figure 3-6. Process sequence for fabricating one end closed green body supports.
48
Figure 3-7. Pictures of tubular SCZE721 thin film membrane coated on the inner side of NiO (or Ni)-SCZ82 support at each processing step.
NiO-SCZ82 tubular green substrate
Pre-sintered NiO-SCZ82 substrate (SCZE721 was coated on the inner side of the substrate)
Sintered SCEZ721membrane on NiO-substrate
Reduced SCEZ721membrane on Ni-SCZ82 substrate
49
A B
C D
Figure 3-8. SEM images of the NiO-SCZ82 and SCZE721 A) surface of the pre-sintered NiO-SCZ82; B) surface of the as-coated SCZE721 thin film; C) cross section of the sintered membrane and D) thin film surface of the sintered membrane.
50
CHAPTER 4 HIGH TEMPERATURE SrCe0.9Eu0.1O3 -δ PROTON CONDUCTING MEMBRANE REACTOR
FOR H2 PRODUCTION USING THE WATER GAS SHIFT REACTION
4.1 Introduction
Syngas mixtures containing mostly H2 and CO are typically generated at elevated
temperatures via the conversion of natural gas, coal, biomass, petroleum and organic wastes [76].
The water-gas shift (WGS) reaction, equation (4-1), converts CO into CO2 and provides
additional H2.
H2O + CO H2 + CO2 6.40−=∆ oH kJmol-1 (4-1)
The WGS reaction is often used in conjunction with steam reforming of methane or other
hydrocarbons and is of central importance in the industrial production of H2, ammonia, and other
bulk chemical utilizing syngas [77]. It is an important method for further enhancing H2 yield
and/or to shift the H2/CO. This is especially important for synthesis gas derived from coal, which
tends to have a H2/CO of ~0.7 compared to the ideal of ~2 for the Fischer-Tropsch process.
The WGS reaction is an exothermic reaction. Thermodynamic equilibrium favors high
conversion of CO and steam to H2 and CO2 at low temperatures. Therefore, it is typically a two-
stage shift process, a high-temperature WGS and a low-temperature WGS, with each process
employing separate catalysts [78, 79]. In addition, a cooling step is necessary before the second
stage. U.S. Department of Energy for the production of H2 suggested an alternative concept by
carrying out the WGS reaction at high temperature in a H2-selective membrane reactor [80]. New
ceramic membranes have potential for cost reduction of syngas production by 30-50% [81] and
provide one solution to incorporate the WGS reaction and H2 separation into one unit.
Selectively continuous removal of H2 will drive the WGS reaction equilibrium forward. As a
result, the requirement to use a two-stage shift reaction and a cooling step can be eliminated and
51
the WGS reaction may be carried out at higher temperatures [19]. This would allow the WGS
reaction to be operated at low H2O/CO without the thermodynamic constraint [16, 82, 83]. In
addition, the reaction rate of the homogeneous WGS reaction at high temperature could be high
enough that permeation through the membrane could be the rate limiting step [84]. Therefore, the
need for the introduction of heterogeneous catalyst could be eliminated.
There is growing interest in the WGS reaction assisted by a catalytic membrane reactor.
The WGS reaction has been carried out under various operating conditions using porous Vycor
glass coated with ruthenium (III) chloride trihydrate. The highest CO conversion was 85% at 157
oC and at a permeate rate of 0.64 cm3/min. The CO conversion is lower than the equilibrium
valune (99.9%) at the same conditions [85]. Extensive research has been focused on Palladium
and palladium alloy membranes [15-17]. However, palladium-based membrane reactors have
been evaluated most extensively at low temperatures of 300-500 oC [18]. The Palladium
membranes are highly fragile due to thermal excursion in the presence of H2 which causes poor
durability. In addition, Palladium based membranes are usually expensive. An alternative for
membrane reactor is ceramic membranes. It has been pointed out that a minimum between the
efficiency penalty and system complexity in a conventional integrated gasification combined
cycle power plant is obtained when the H2 and CO2 is separated at high temperature using a
catalytic ceramic membrane reactor [85]. Most of the catalytic ceramic membrane reactors are
SrCeO3-δ and BaCeO3-δ based perovskite mixed protonic electronic conductors [26-28, 32].
SrCeO3-δ has high total conductivity and highest proton transference number compared to
BaCeO3-δ and SrZrO3-δ [49]. However, its electronic conductivity needs to be improved. We
previously successfully improved the electronic conductivity of SrCeO3-δ using Eu dopant
(0.05≤x≤0.2) [3] and selected SrCe0.9Eu0.1O3-δ to fabricate supported tubular thin membranes to
52
maintain mechanical stability [86, 87]. The tubular SrCe0.9Eu0.1O3-δ membrane coated on NiO-
SrCeO3-δ support was applied to incorporate the WGS reaction and H2 separation [87].
In this chapter, a WGS membrane reactor was developed using a SrCe0.9Eu0.1O3-δ tubular
membrane to incorporate the WGS reaction and H2 separation into one unit. Results from both
thermodynamic equilibrium calculation and experiment were compared to show the effect of the
ceramic membrane on the WGS reaction. The thermodynamic equilibrium of the WGS reaction
was calculated for H2O/CO = 1/1 and 2/1. The improved CO conversion and H2 yield using the
membrane reactor was further confirmed by carrying out the WGS reaction under three
situations: (1) blank reference, (2) with Ni catalyst, and (3) with Ni catalyst and in situ H2
removal. In addition, appropriate operating temperature region without carbon formation for
each H2O/CO was addressed since carbon formation is detrimental to the WGS reaction as it
causes catalyst deactivation.
4.2 Experimental
Polycrystalline SrCeO3-δ and SrCe0.9Eu0.1O3-δ powders were prepared by conventional
solid-state reaction by mixing stoichiometric amounts of SrCO3 (99.9%, Alfa-Aesar), CeO2
(99.9%, Alfa-Aesar) and Eu2O3 (99.9%, Alfa-Aesar) powders, followed by ball milling and
calcining at 1300 oC. A NiO-SrCeO3-δ tubular support was fabricated using tape-casting (Pro-
Cast) followed by a rolling process. The tubular support was sealed at one end and pre-sintered.
SrCe0.9Eu0.1O3-δ was coated on the inner side of the pre-sintered support. The tubular membranes
were finally sintered at 1450 oC. A detailed preparation process has been discussed in our
previous work [74].
The membrane tube is about 17 cm long and 0.48 cm in diameter (Figure 4-1 A). An SEM
image after experiment shows that the membrane is dense and ~23 µm thick on a porous support
53
(Figure 4-1 B). A thermal insulator was applied to the bottom of the membrane tube, forming an
insulating region to drop the temperature and allow O-ring sealing of the tube [6]. The area of the
membrane above the insulator zone is considered the active area and is about 12 cm2. The WGS
reaction was carried out from 600 oC to 900 oC under 3% CO + 3% H2O and 3% CO + 6% H2O
(total flow rate of 20 sccm balanced by Ar). Gas flow rates were controlled by mass flow
controllers.
Figure 4-1 C shows a photo of the experimental setup. A thermo couple was placed axially
at the middle of the membrane to control temperature. Argon was used as tracer to detect
leakage. The reactants, CO and H2O, were flowed into the quartz chamber and exposed to the Ni
catalyst on the outside of the membrane tube. The reactor side effluents were analyzed by gas
chromatography (GC) (Varian CP 4900). Helium was used as a sweep gas on the inner side of
the membrane. The permeated H2 together with Ar (leakage) were analyzed by a mass
spectrometer (Q100MS Dycor Quadlink).
4.3 Results and Discussion
4.3.1 Thermodynamic Calculation
The thermodynamic equilibrium conditions of the WGS reaction were calculated using
Thermocalc software [88] with a total pressure of 1 atm. Figure 4-2 (A) and Figure 4-2 (B) show
the temperature dependence on species mole fraction with feed H2O/CO = 1/1 and 2/1,
respectively. The mole fractions of the reactants, CO and H2O, increase with increasing
temperature, which is attributed to the exothermic nature of the WGS reaction. Thus
thermodynamic equilibrium moves to the reactant side at elevated temperature.
Achieving a carbon deposition free operating temperature region is very important for the
WGS reaction since carbon formation may block the pores of the porous support and lead to
catalyst deactivation as well as cracking of the membrane. It is shown from the thermodynamic
54
equilibrium results that carbon formation is in general favored at low temperature and low
H2O/CO. Carbon will not form at temperatures higher than 590 oC with H2O/CO = 1. This shifts
to 550 oC for H2O/CO = 2. This is in agreement with the results by Xue et al. [89]. They reported
the risk of carbon formation due to side reactions increased as the H2O/CO decreased. The
formation of carbon with H2O/CO = 1 was thermodynamically favorable over the entire
temperature range examined (up to 500 oC). However, a carbon-free operation condition was
achieved at temperatures higher than 230 oC with H2O/CO = 3.
Low temperatures also favor CH4 and H2O formation, which is clearly shown in Figure 4-2
A. If no side reactions are considered, the mole fractions of CO and H2O are equal to each other
with H2O/CO = 1/1. Similarly, the mole fractions of H2 and CO2 are the same. However, the
mole fraction of CO is lower than that of H2O at 710 oC and the mole fraction of H2 is less than
that of CO2. This is attributed to the consumption of H2 to form CH4 and H2O. Higher H2O/CO
can extend the operating temperature of the WGS reaction to lower temperature. When the
H2O/CO is increased to 2/1, the formation of CH4 and H2O occurs below 640 oC. Therefore, the
WGS reaction should be carried out at temperatures higher than 710 oC and 640 oC with H2O/CO
= 1/1 and 2/1, respectively.
4.3.2 Experimental Conversion
The CO conversion was measured under three situations: (1) blank reference, (2) with Ni
catalyst, and (3) with Ni catalyst and in situ H2 removal. In the first two situations, two different
gas compositions were applied: 3% CO + 3% H2O or 3% CO + 6% H2O, while maintaining a
constant flow rate of 20 sccm balanced by Ar. Only 3% CO + 6% H2O gas composition was
applied in situation (3). For the blank reference, CO and H2O were fed into an empty quartz
reactor. For the WGS reaction with Ni catalyst, the tubular membrane was installed in the quartz
reactor with the Ni in the porous support being exposed to the reactants and the permeated side
55
was blocked so that no produced H2 was removed. Therefore, the membrane only functioned as a
catalyst in this situation with the reactant gas mix passing by the Ni catalyst in the support tube
surface. For the WGS reaction with Ni catalyst and in situ H2 removal, the permeated side was
connected to a mass spectrometer, so that the permeated H2 concentration could be analyzed.
The mole fractions of H2, CO, H2O and CO2 under these three reactor configurations are
shown in Figure 4-3 to Figure 4-5. These mole fractions do not include the Ar diluent. The CH4
concentration was below detection limits and ignored here. For the blank reference, the mole
fraction of H2 is the same as that of CO2 under both H2O/CO = 1/1 and 2/1 (Figure 4-3). The
mole fractions of CO and H2O equal to each other for H2O/CO = 1/1 as well.
For WGS with the Ni catalyst (Figure 4-4), the mole fractions of species are similar with
the thermodynamic data shown in Figure 4-2. The mole fractions of H2 and CO2 agree with each
other at elevated temperature and there is similar deviation at low temperatures. This indicates
the WGS reaction approaches thermodynamic equilibrium in the presence of the Ni catalyst in
the tubular support.
The mole fractions in Figure 4-5 are just for the species in the reactor side effluent. The
permeated H2 is not included. It is important to point out that no Ar was observed in the
permeated gas so the membrane was leak free. The H2 and CO2 mole fractions were significantly
different comparing Figure 4-4 (B) and Figure 4-5. Both were under the WGS with Ni and
H2O/CO = 2/1, but the data in Figure 4-5 were achieved with in situ H2 removal. The CO2 mole
fraction was higher in Figure 4-5 and increased with increasing temperature due to the in situ
removal of H2 overcoming the inherent thermodynamic limitation at high temperature. In
addition, while the H2 mole fraction decreased with increasing temperature in Figure 4-4 (B), it
was almost independent of temperature in Figure 4-5.
56
CO conversion is a very important parameter in the WGS reaction and needs to be
maximized in order to increase H2 production. The CO conversion is defined as follows:
XCO (%)= %1002 ×inCO
outCO
FF
(4-2)
where outCOF
2 and in
COF are CO2 output flux and CO input flux, respectively.
Figure 4-6 shows the XCO temperature dependence for different reactor configurations with
a H2O/CO = 1/1 and 2/1, respectively and compares with their thermodynamic equilibrium data.
The thermodynamic conversion decreases as the reaction temperature increases. This is
consistent with the mole fraction decrease of H2 and CO2 as temperature increases in Figure 4-2.
At any temperature, the XCO under thermodynamic equilibrium increases with increasing
H2O/CO from 1/1 to 2/1. A higher feed steam concentration moves the reaction (4-1) forward,
resulting in higher XCO.
For the blank reference under both H2O/CO = 1/1 and 2/1, XCO was significantly lower
than the thermodynamic values. The reaction rate increased with increasing temperature and the
XCO approached theoretical at higher temperatures.
With the Ni catalyst, the XCO was comparable to and consistent with the thermodynamic
values. The slight deviation, especially at lower temperatures, was attributed to side reactions
which could take place during the WGS process [89].
According to the thermodynamic data, higher H2O/CO results in higher XCO. Therefore,
the effect of in situ H2 removal on the XCO was investigated under H2O/CO = 2/1 (Figure 4-6).
Compared to the XCO with only the Ni catalyst, much higher XCO was achieved especially at high
temperatures (46% increase at 900 oC). Since the permeated H2 lowered the H2 concentration in
the product stream it moved the reaction further toward the product side resulting in higher XCO.
57
Therefore, the H2 membrane can help overcome the thermodynamic limitations and improve
XCO. It simultaneously increases H2 yield as well. This can also be applied to reduce the CO level
in the H2 gas produced from hydrocarbon fuels for proton-exchange membrane (PEM) fuel cells.
4.3.3 H2 Production
H2 production for the blank reference increased with increasing temperature, in agreement
with the XCO under the same conditions as shown in Figure 4-6. Figure 4-7 shows the H2
production with H2O/CO = 2/1 as a function of temperature. The H2 production with only the Ni
catalyst decreased with increasing temperature consistent with the thermodynamic data due to
the exothermic nature of the WGS reaction. The H2 permeation flux increased with increasing
temperature due to the higher ambipolar conductivity of SrCe0.9Eu0.1O3-δ at elevated
temperatures [3]. In addition, the H2 production in the reactor side effluent was essentially
temperature independent with in situ H2 removal. The total H2 production with in situ H2
removal is the sum of the H2 in the reactor side effluent and the permeated H2. It increased with
increasing temperature. The improvement was more significant at elevated temperatures
compared to the thermodynamic value. A 46% increase in total H2 production was achieved at
900 oC.
The H2 yield is defined:
100% (%) yield H 22 ×= in
CO
outH
FF
(4-3)
where outHF
2 and in
COF are the H2 production and CO input flux, respectively.
The H2 yield was plotted in Figure 4-8 as well as the H2/CO in the reactor side effluent
with H2 in situ removal and H2O/CO = 2/1. The total H2 yield is the sum of the permeated H2
yield and the H2 yield in the reactor side effluent. The H2 yield was in similar trend with the H2
58
production (Figure 4-7). The permeated H2 yield and total H2 yield were 32% and 92% at 900
oC, respectively. The reactor side effluent consisted of H2 and CO2 rich gases together with the
residual CO and H2O. The H2/CO increased from 3.9 to 7.6 when the temperature increased from
700 to 900 oC.
4.4 Conclusions
WGS reaction is constrained by thermodynamic equilibrium limitations. A tubular
SrCe0.9Eu0.1O3-δ H2 transport WGS membrane reactor was fabricated. The XCO, H2 production,
H2 yield, and the H2/CO in the reactor side effluent increase significantly with the WGS
membrane. A 46% increase in XCO and total H2 yield was achieved at 900 oC under 3% CO and
6% H2O compared to the thermodynamic equilibrium calculation, resulting in a 92% single pass
H2 production yield and 32% single pass yield of pure permeated H2. These results demonstrate
the efficiency of H2 membranes for the WGS reaction.
59
A
B
C
Figure 4-1. Morphology of thin film membranes and experimental setup. A) Membrane tube, B) SEM image of sintered membrane cross section and C) Photo of the WGS reactor showing gas flow.
60
0
0.1
0.2
0.3
0.4
0.5
550 600 650 700 750 800 850 900 950
Mol
e Fr
actio
n
Temperature (oC)
CO2
H2
H2O
CO
CH4
C
A
0
0.1
0.2
0.3
0.4
0.5
550 600 650 700 750 800 850 900 950
Mol
e Fr
actio
n
Temperature (oC)
H2O
CO2
CO
CH4C
H2
B
Figure 4-2. Thermodynamic equilibrium of WGS under A) H2O/CO =1/1 and B) H2O/CO =2/1.
61
0
0.1
0.2
0.3
0.4
0.5
550 600 650 700 750 800 850 900 950
CO2
H2
CO
H2O
Mol
e Fr
actio
n
Temperature (oC) A
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
550 600 650 700 750 800 850 900 950
CO2
H2
CO
H2O
Mol
e Fr
actio
n
Temperature (oC) B
Figure 4-3. Blank reference effluent gas composition as a function of temperature under A) H2O/CO =1/1 and B) H2O/CO =2/1.
62
0
0.1
0.2
0.3
0.4
0.5
550 600 650 700 750 800 850 900 950
CO2
H2
CO
H2O
Mol
e Fr
actio
n
Temperature (oC) A
0
0.1
0.2
0.3
0.4
0.5
550 600 650 700 750 800 850 900 950
CO2
H2
CO
H2O
Mol
e Fr
actio
n
Temperature (oC) B
Figure 4-4. Catalytic effluent gas composition as a function of temperature under A) H2O/CO =1/1 and B) H2O/CO =2/1.
63
0
0.1
0.2
0.3
0.4
0.5
550 600 650 700 750 800 850 900 950
CO2
H2
CO
H2O
Mol
e Fr
actio
n
Temperature (oC)
Figure 4-5. Catalytic effluent gas composition with in situ H2 removal as a function of temperature for H2O/CO =2/1 feed gas.
64
0
20
40
60
80
100
550 600 650 700 750 800 850 900 950
X CO (%
)
Temperature (oC)
Figure 4-6. Temperature dependence of XCO under 3% CO + 3 % H2O and 3% CO + 6% H2O (○ blank reference (H2O/CO=1/1), ● blank reference (H2O/CO=2/1), – — thermodynamic data (H2O/CO=1/1), -- thermodynamic data (H2O/CO=2/1), □ WGS with catalyst (H2O/CO=1/1), ■ WGS with catalyst (H2O/CO=2/1), ♦ WGS with catalyst and in situ H2 removal (H2O/CO=2/1)).
65
0
0.1
0.2
0.3
0.4
0.5
0.6
550 600 650 700 750 800 850 900 9500
0.01
0.02
0.03
0.04
0.05
H2 P
rodu
ctio
n (c
m3 /m
in)
Temperature (oC)
H2 Flux (cc/cm
2min)
Figure 4-7. H2 production under 3% CO + 6% H2O as a function of temperature for three reactor configurations (-- Thermodynamic H2 production ● Catalytic H2 production without H2 removal, ▼ Pure permeated H2 through membrane ■ H2 production in the reactor side effluent with in situ H2 removal ♦ Total catalytic H 2 production with in situ H2 removal).
66
0
20
40
60
80
100
550 600 650 700 750 800 850 900 9500
2
4
6
8
10
Temperature (oC)
H2 y
ield
(%)
H2 /C
O
permeated H2
H2 in reactor side effluent
total H2 yield
reactor side effluent H2/CO ratio
Figure 4-8. H2 yield and syngas H2/CO ratio as a function of temperature under 3% CO + 6% H2O and with in situ H2 removal.
67
CHAPTER 5 STABILITY OF SrCe1-XZrXO3-δ UNDER WATER GAS SHIFT REACTION CONDITIONS
5.1 Introduction
In chapter 4, a supported SrCe0.9Eu0.1O3-δ thin film membrane was used to incorporate the
WGS reaction and H2 separation into one unit and significantly increased CO conversion and H2
yield was achieved [87]. We demonstrated that H2 flux for Ni-SrCeO3-δ supported SrCe1-xEuxO3-δ
(0.05≤x≤0.2) thin films is proportional to [PH2]1/4 under wet and dry H2 conditions [6]. The H2
flux of SrCe0.9Eu0.1O3-δ is stable under wet H2 atmospheres but degrades under dry H2 conditions
[7]. The degradation under dry H2 is caused by phase decomposition forming CeO2 under low
PO2 conditions. The tubular SrCe0.9Eu0.1O3-δ membrane is also unstable under hydrocarbon
conditions possibly due to phase change and/or coking below 700 oC as shown in Figure 5-1.The
temperature gradient in the reactor furnace as a function of axial distance and the set point
temperature is caused by our experimental setup for these tubular membranes [6]. In contrast, the
SrCe0.7Zr0.2Eu0.1O3-δ membrane maintains good integrity after exposure in methane with 18 %
steam. This indicates the partial substitution of Zr for Ce can improve the chemical stability of
SrCeO3-δ which is attributed to the higher stability of SrZrO3-δ against carbon dioxide than
SrCeO3-δ below 800 oC [90, 91]. In addition, Zr has been used to improve the chemical stability
of BaCeO3-δ system [8-12] and SrCe0.95Yb0.05O3-δ [13]. The thermo-chemistry of SrCeO3-δ [92-
94] and its stability under CO2 [93, 95] and water [96, 97] have been reported. However, the
stability of Zr-doped SrCeO3-δ hasn’t been systemically studied yet especially under various
hydrocarbon gas conditions.
In this chapter, four different zirconium dopant concentrations were examined to improve
the stability of SrCeO3-δ under WGS reaction atmospheres (CO/CO2, H2/H2O). The possible
SrCeO3-δ decomposition mechanism was investigated using SrCe0.8Zr0.2O3-δ (denoted SCZ82
68
hereafter) powder. We fabricated a supported SrCe0.7Zr0.2Eu0.1O3-δ (denoted SCZE721 hereafter)
thin film membrane using NiO-SCZ82 as the supported structure to increase H2 permeation.
Finally, its chemical stability was investigated by measuring the H2 permeation under WGS
conditions as a function of time.
5.2 Experimental
To investigate the zirconium dopant effect on the chemical stability under CO/CO2, five
different compositions (Zr=0, 2.5 at%, 5 at%. 10 at% and 20 at%) of SrCe1-xZrxO3-δ powders
were synthesized using conventional solid-state reaction by mixing stoichiometric amounts of
SrCO3-δ (99.9%, Alfa-Aesar), CeO2 (99.9%, Alfa-Aesar), and ZrO2 (99.7%, Alfa-Aesar)
powders, followed by ball milling and calcining at 1300 oC. These powders were exposed to
2.8% CO and 5.6% H2O (balanced by Ar) atmosphere at different temperatures for 19 hours. The
atmosphere finally became a mixture of CO, H2, H2O and CO2 due to the WGS reaction. In
addition, SCZ82 was exposed to 3.0% dry CO, 2.8% CO2/5.6% H2O, 3.0% dry CO2, 6.0% H2O,
and 2.8% H2/2.8% H2O atmospheres for 19 hours to investigate the decomposition mechanism.
Phase compositions were analyzed by X-ray diffraction (XRD) (XRD Philips APD 3720) with
CuKα radiation. The 2θ value chosen was from 20 to 80 degree with a step size of 0.03 degree/s.
A detailed preparation process of the supported thin film was discussed in chapter 3 and
our previous work [74]. For H2 permeation measurements, the tubular H2 membranes were
installed in a high temperature reactor apparatus which has been described previously [6]. H2
permeability dependence on time was investigated under 5.0% CO and 3.0% H2O atmosphere at
900 oC to further evaluate stability under operating conditions. The permeated gases were
analyzed by a mass spectrometer (Q100MS Dycor Quadlink).
69
5.3 Results and Discussion
5.3.1 Stability under Wet CO Conditions
Figure 5-2 displays the XRD patterns of SrCeO3-δ before and after the stability experiment.
The as-calcined powder had an orthorhombic perovskite structure. SrCeO3-δ phase remained
predominant at 900 oC, but decomposed into SrCO3 and CeO2 at temperatures between 600 oC
and 800 oC. XRD results show that SrCeO3-δ is unstable at low temperatures with respect to
SrCO3 under CO/H2O atmospheres, which is in agreement with Shirsat’s results [92].
Figure 5-3 (A) shows the XRD patterns of the as-calcined powder for the different
zirconium dopant concentrations at 1300 oC. No significant secondary phase was detected. All
the compositions were orthorhombic perovskite structure. The lattice parameters decrease with
increasing Zr dopant concentration (Figure5- 3 (B)) due to the smaller ionic radius of Zr4+. The
lattice parameters for SrCeO3-δ are a=8.575 Å, b=6.122 Å, c=6.000 Å, essentially identical to
Wei et al. results [98]. The lattice parameters are approximately a linear function of Zr dopant
concentration, following Vegard’s law. Figure 5-4 shows the XRD patterns of SrCe1-xZrxO3-δ at
800 oC after the stability experiment, exposed to 2.8% CO and 5.6% H2O for 19 hours. For Zr
dopant concentrations of 2.5 at% and 5 at%, the intensity of the CeO2 (111) reflection (2θ≈28.7o)
was comparable to the main peak of the perovskite (2θ≈29.7o), which indicates a significant
amount of CeO2 was formed. For Zr dopant concentration of 10 at% and 20 at%, the intensity of
CeO2 peak was much lower, showing higher stability under CO and CO2 atmospheres. This
demonstrates higher stability can be achieved with higher Zr dopant concentration. Therefore, 20
at% Zr dopant was used for further study.
To specify the temperature effect on SrCeO3-δ decomposition, the stability experiment was
carried out on SCZ82 at 800, 825, 850, 865, 875, 900 and 940 oC and their XRD patterns are
70
shown in Figure 5-5. The relative intensity of the strongest peaks of CeO2 and perovskite phase
is calculated and shown in table 5-1. The intensity of the CeO2 peak became much lower at
higher temperatures, with the relative intensity decreasing from 34% to 5% when temperature
was increased from 800 oC to over 865 oC. For the as-calcined SCZ82 powder, the relative
intensity of the secondary phase, CeO2, is 5%, which is in agreement with Mather’s results [97].
Generally, X-ray phase-pure material could not be obtained for nominally stoichiometric
SrCeO3-δ or Sr deficient samples synthesized by solid state reaction [42, 97]. At temperatures
higher than 865 oC, the relative intensity of CeO2 (111) is approximately equal to that in the as-
calcined powder, indicating the SCZ82 is stable at that temperature. In view of thermodynamics,
the formation reaction of SrCO3 and CeO2 from SrCeO3 and CO2 associates with a decrease in
entropy; thus, it is not favored at elevated temperatures. Therefore, to avoid SCZ82
decomposition, the WGS reaction should be carried out at temperatures higher than 865 oC.
The improvement in chemical stability of strontium zirconate can be explained by basicity
of the oxide and tolerance factor. A high basicity of the oxide is advantageous for the stability of
protonic defects but basic oxides are expected to react easily with acidic gases such as CO2 [99].
The stability against the formation of carbonates and hydroxides increases in the order cerate
zirconate with decreasing the stability of protonic defects.
The tolerance factor, t introduced by Goldschmidt, is a measure of the “cubic-ness” of the
perovskite [100]:
)(2 OB
OA
rrrrt+
+= (5-1)
where Ar , Br , Or are the ionic radii of the A-site cation, the B-site cation and the oxygen anion,
respectively. The tolerance factor has been used to explain the stability of many perovskites [95,
71
101-103]. When t equals unity, the structure is predicted to be perfect cubic. Lower values of t
correspond to lower symmetry. Using ionic radii found in Shannon’s paper [104], the tolerance
factor was calculated for SrCe1-xZrxO3-δ (x=0, 2.5%, 5%, 10% and 20%), Table 5-2. The
tolerance factor is less than 1 for all compositions studied but increases with increasing
zirconium dopant concentration.
The improvement of chemical stability with increasing tolerance factor is in accordance
with Yokokawa’s result [105]. They reported the stabilization energy of AIIMIVO3 perovskite,
which was related to the tolerance factor. Generally, the closer the value of the tolerance factor to
unity, the higher the chemical stability of the perovskite structure.
5.3.2 Decomposition Mechanism
The WGS atmosphere consists of CO, CO2, H2 and H2O gases. All of these gases might
react with the powder and cause decomposition. To identify which species, CO, CO2, H2 and
H2O, contribute to the decomposition, SCZ82 was exposed to 3% CO, 2.8% CO2/5.6% H2O, 3%
CO2, 6% H2O and, 2.8% H2/2.8% H2O, for 19 hours at 800 oC, respectively. Figure 5-6 shows
the XRD patterns. The pattern (a) is similar with that of the as-calcined powder, pattern (d) in
figure 5-3. No significant secondary phase was detected, indicating SCZ82 is stable under dry
CO atmosphere. Therefore, CO is unlikely to cause the decomposition. For patterns (b) and (c),
significant SrCO3 and CeO2 phases were detected and the intensity of CeO2 phase increased
dramatically compared to the pattern of the as-calcined powder. The intensity ratios between the
strongest peaks of CeO2 and perovskite phase in pattern (b) and (c) are, 41% and 47%,
respectively. They are higher than the ratio (34%) under 2.8% CO/5.6% H2O at 800oC. The
CO2% among these three conditions follows the relation: 2.8% CO/5.6% H2O < 2.8% CO2/5.6%
72
H2O < 3% CO2. The intensity ratio between the strongest peaks of CeO2 and the perovskite
phase follows this same order. Therefore, CO2 is responsible for the decomposition.
H2O and H2 may also contribute to decomposition under water gas shift (WGS) conditions.
Therefore, the effect of steam and wet H2 on the stability of SCZ82 was investigated. Patterns (d)
and (e) show the results after the stability experiment under 6% H2O and, 2.8% H2/5.6% H2O at
800oC, respectively. Both XRD patterns are similar with that of the SCZ82 as-calcined powder,
indicating SCZ82 is stable under these conditions. This is in agreement with Mather’s result
[106]. They reported SrCeO3 was stable under PH2O= 1.6 atm at much lower temperatures, 120-
174 oC. The formation reaction of CeO2 and Sr(OH)2 from SrCeO3 and H2O associates with a
decrease in entropy and is not favored at elevated temperatures. Therefore, steam and wet H2 do
not cause SCZ82 decomposition and CO2 plays the dominant role. Moreover, since no SrO and
Sr(OH)2 phases were detected under steam and wet H2 atmospheres, decomposition takes place
most likely by SrCeO3 reacting directly with CO2 through reaction (5-2).
2323 CeOSrCOCOSrCeO +↔+ (5-2)
5.3.3 Hydrogen Permeability
Figure 5-7 shows the H2 permeation versus time at 900 oC for SrCe09Eu0.1O3-δ and
SCZE721 membranes under 5% CO and 3% H2O atmospheres with a total flow rate of 20
cm3/min, balanced by Ar. In both cases, the H2 permeation increases in the first couple of hours,
which we attributed to the reduction of the NiO to Ni in the support structure. After that, the H2
permeation through the SrCe09Eu0.1O3-δ membrane degraded significantly with a degradation rate
of 1.8 %/hr. In contrast, the H2 permeation flux through the SCZE721 membrane was essentially
stable with a degradation rate of only 2.4×10-3 %/hr. The stable H2 permeation flux verifies the
73
stability enhancement by zirconium dopant. However, this comes at the expense of lower
permeation flux since the protonic conductivity of SrCeO3 is greater than SrZrO3 [37].
5.4 Conclusions
SrCeO3 is unstable under WGS atmospheres. CO2 is found to be the main cause of
decomposition, forming SrCO3 and CeO2. The H2 permeation flux of SrCe0.7Zr0.2Eu0.1O3-δ
membrane was essentially stable under WGS conditions at 900 oC. Its degradation rate is 1000
times lower than that of a SrCe0.9Eu0.1O3-δ membrane under the same conditions. This confirms
the stability of SrCeO3-δ can be improved by partially substituting Ce with Zr, which increases its
tolerance factor and decreases its basicity.
74
Figure 5-1. Temperature profile and photograph of the membranes A) temperature profile along the membrane [6] and B) Photograph of the SrCe0.9Eu0.1O3/Ni-SrCeO3 and SrCe0.9Eu0.1O3 /Ni-SCZ82 H2 membranes after exposure to methane with 18% steam.
A
B
75
Figure 5-2. XRD pattern of SrCeO3-δ after exposure to 2.8% CO and 5.6% H2O for 19 h
(*) perovskite phase, (o) CeO2 phase, (+) SrCO3 phase.
76
A
0 5 10 15 20
Latti
ce P
aram
eter
s (A
)
ac
b
Zr (at%)
Lattice Parameter (A
)
6.15
6.10
6.05
5.95
6.00
5.90
8.65
8.60
8.55
8.50
8.40
8.45
B
Figure 5-3. XRD Pattern and their lattice parameters of SrCe1-xZrxO3-δ as-calcined at 1300 oC A) XRD pattern and B) lattice parameters as a function of dopant concentrations.
77
Figure 5-4. XRD Pattern of SrCe1-xZrxO3-δ after exposure to 2.8% CO and 5.6% H2O for 19 h at
800 oC (o CeO2 phase, + SrCO3 phase).
78
Figure 5-5. XRD Pattern of SrCe0.8Zr0.2O3-δ after exposure to 2.8% CO and 5.6% H2O for 19 h (o CeO2 phase).
79
Figure 5-6. XRD Pattern of SrCe0.8Zr0.2O3-δ after stability experiment at different atmospheres at
800 oC (a) 3% CO, (b) 2.8% CO2 & 5.6% H2O, (c) 3% CO2, (d) 6% H2O, (e) 2.8% H2 & 5.6% H2O (o CeO2 phase, + SrCO3 phase).
80
0 2 4 6 8 10 12
H2 P
rodu
ctio
n (c
m3 /m
in)
Time (hrs)
SrCe0.9
Eu0.1
O3
(1.8 %/hr)
SrCe0.7
Zr0.2
Eu0.1
O3
(2.4*10-3 %/hr)
0.25
0.20
0.15
0.10
0.05
Figure 5-7. H2 permeation as a function of time under 5% CO and 3% H2O at 900 oC.
81
Table 5-1. Intensity ratios between the strongest peaks of CeO2 and SCZ82. Temperature (oC)
800 825 850 865 875 900 940 1300(as- calcined)
ICeO2/ISCZ82 34% 30% 24% 6% 5% 5% 5% 5% Table 5-2. Tolerance factors of SrCe1-xZrxO3-δ. x value 0 2.5% 5% 10% 20% t value 0.885 0.886 0.888 0.891 0.897
82
CHAPTER 6 HYDROGEN PERMEATION OF THIN SUPPORTED SrCe0.7Zr0.2Eu0.1O3-δ MEMBRANES
UNDER DIFFERENT OXYGEN PARTIAL PRESSURE
6.1 Introduction
In recent years, high-temperature mixed protonic and electronic conducting ceramics have
attracted considerable attention as potential membranes for H2 gas separation from hydrocarbon
fuels [21, 25, 71, 95]. Numerous SrCeO3-δ and BaCeO3-δ based perovskite oxides with
multivalent cation dopants have been reported [3, 26-30, 48, 107]. In order to enhance the
ionization of H2, cermet membranes with a continuous metallic phase were also studied [31, 32].
In addition, complex perovskites in the form of A2B1+x′B1-x″MxO6-δ or A3B1+x′B2-x″Mx2O9-δ [5,
33-36] have been developed to increased stability of the perovskite oxides. We previously
reported the H2 permeation of the SrCe0.9Eu0.1O3-δ [6] and demonstrated that its thermodynamic
stability was improved by partial substitution of Zr onto the Ce-site [108]. In chapter 5, the
stability of SrCeO3-δ was improved using 20 mol% Zr dopant even under WGS reaction
conditions [108, 109].
In this chapter, the H2 permeation of the SrCe0.7Zr0.2Eu0.1O3-δ thin film on NiO-
SrCe0.8Zr0.2O3-δ support was investigated in details as a function of temperature, H2 flow rate,
membrane thickness, and H2 and/or steam partial pressures. The activation energies under
different oxygen partial pressure were discussed. The long term stability of H2 permeation under
wet H2, WGS reaction, and SRM was investigated as well.
6.2 Experimental
6.2.1 Membrane Fabrication
A detailed preparation process of the SrCe0.7Zr0.2Eu0.1O3-δ thin film on NiO-SrCe0.8Zr0.2O3-
δ support has been discussed in chapter 3 and our previous work [74].
83
6.2.2 Membrane Morphology
The membrane tube was ~17 cm long and 0.48 cm in diameter (Figure 6-1 (A)). SEM
images after experiment show that the membrane is dense and pinhole free (Figure 6-1 (B)) on
the surface. It is ~33 µm thick (Figure 6-1 (C)). The effective area for H2 permeation in the
tubular membrane is 12 cm2.
6.2.3 Membrane Permeation
For H2 permeation measurements, the tubular H2 membranes were installed in a high
temperature reactor apparatus (Figure 6-1 (D)) which was previously described [74]. The outer
side of the membrane (feed) was exposed to H2 (99.999%) diluted to the desired concentration
using Ar (99.999%) with a 20 cm3/min total flow rate. The total flow rate was variable when the
flow rate effect on the H2 permeation was investigated. For wet gas flow, 3 vol % water vapor
was picked up by flowing the feed gas through a water bubbler at 25 oC. Greater concentration of
water vapor was achieved by gasifying a desired amount of water provided by a syringe pump.
The inner side (sweep) of the membrane was flushed with He at 20 cm3/min. The flow rates of
H2, Ar, and He were controlled by mass flow controllers. Ar was used as a tracer to determine
whether the membrane were pinhole/leak free [6, 74]. The components of the permeated gases
on the sweep side were measured using a mass spectrometer. The gases from the exhaust on the
feed side were measured using GC to verify the mass balance.
6.3 Result and Discussion
6.3.1 Heat Treatment
Before the as-sintered membrane was used for permeation, the NiO in the support had to
be reduced to create porosity. In this work, the outer side of the membrane was treated by 5
cm3/min H2 and 15 cm3/min Ar mixed with 3% H2O at 900 oC overnight. Meanwhile, 20
84
cm3/min He was fed to the inner side of the membrane. Finally, a flat H2 permeation was
observed indicating the NiO was completely reduced.
6.3.2 Flow Rate Effect on H2 Permeation
Figure 6-2 shows H2 permeation and recovery under different flow rates at 900 oC. The
feed gas composition was 48.5% H2, 48.5% Ar and 3 % H2O to maintain a constant feed H2
partial pressure. The H2 permeation increased initially with total feed flow rate but was
essentially constant when the feed flow rate was over 20 cm3/min. Tong et al [110] and Li et al.
[111] have observed similar results except that their H2 permeation reached a plateau level at
higher total feed flow rates. This is due to the higher permeability of their membranes. The
amount of the permeated H2 at lower feed flow rates can cause a larger decrease of H2 partial
pressure in the feed gas, resulting in a bigger drop in the driving force for H2 permeation. In
contrast, at higher feed flow rates, the same amount of permeated H2 only causes a smaller
decrease in the driving force. Therefore, the H2 permeation flux barely changed. H2 recovery was
defined as follows:
%100H feed
H permeated[%]recovery H2
22 ×= (6-1)
As shown in Figure 6-2, the H2 recovery decreased as the flow rate increased, probably due
to the decrease in the resident time. Thus, the opportunity of H2 atoms to be absorbed on the
membrane surface and diffused through the membrane is smaller. Consequently, the H2 recovery
decreased.
6.3.3 H2 Permeation as a Function of Thickness
Figure 6-3 shows the H2 permeation flux of SrZr0.2Ce0.7Eu0.1O3-δ as a function of thickness
and feed H2 concentration at 900 oC [112]. The H2 permeation flux linearly increases with
decreasing membrane thickness. This is consistent with the Wagner equation (6-2), which shows
85
that H2 permeation through a proton conducting membrane is inversely proportional to the
membrane thickness [5].
++−= ∫∫ •••••••
//2
/2
2/
//2
/2
2ln)(
2ln
41
22
H
H OO
O
O OOO
P
P HeVOHt
P
P OVOHtOH PdtttF
RTPdttF
RTL
J σσ (6-2)
This indicates that the H2 permeation of SrZr0.2Ce0.7Eu0.1O3-δ membranes is controlled by bulk
diffusion down to 17 micrometer thick, in agreement with the H2 permeation through the dense
SrCe0.95Yb0.05O3-δ membranes, which has been reported to be controlled by bulk diffusion at 950
K even for 2 µm films [73].
6.3.4 Effect of Temperature, H2 and H2O Partial Pressure in the Feed Side on H2 Permeation
Figure 6-4 shows the H2 permeation flux of the SrCe0.7Zr0.2Eu0.1O3-δ membrane as a
function of feed gas H2 partial pressure and temperature under dry H2 and 97% H2/3% H2O
conditions, respectively. fHP
2and P
HP2are the H2 partial pressure at the feed side and permeated
side. H2 permeation flux increased with temperature under both conditions which is attributed to
the increase in ambipolar conductivity of the SrCe0.7Zr0.2Eu0.1O3-δ membrane[3, 48]. In addition,
the H2 permeation was proportional to the transmembrane H2 partial pressure gradient with a 1/4
dependence.
When the membrane is exposed to a H2 atmosphere, protons and electrons are the
dominating defects [64]. We obtain equation (6-3) after integrating the Wagner equation
))()((1 4/14/122
PH
fHOH PP
LJ
O−∝• (6-3)
A maximum H2 permeation flux of 0.23 and 0.21 cm3/cm2 min was obtained at 900 oC for
100% H2 and 97% H2/3% H2O conditions, respectively. This is very close to the ~0.26 cm3/cm2
86
min maximum H2 permeation flux through the 30 µm thick SrCe0.7Zr0.2Eu0.1O3-δ membrane
under 97% H2/3% H2O conditions at 900 oC [112].
Compared to the dry H2 condition, the presence of H2O increases the O2 partial pressure
resulting in decrease in the electronic conduction of the SrCe0.7Zr0.2Eu0.1O3-δ. Therefore, the H2
permeation decreases. To further address the steam concentration effect, the H2 permeation was
investigated under three steam concentrations and dry H2 condition with a constant of H2 flow
rate of 20 cm3/min (Figure 6-5). It decreased consistently with increasing steam concentration.
The maximum H2 permeation flux decreased from 0.23 to 0.18 cm3/cm2 min when the steam
partial pressure increased from 0 to 30% at 900 oC.
6.3.5 Activation Energy
The activation energy can be obtained from the Arrhenius plot using the data in Figure 6-2
and Figure 6-4 and are listed in Table 6-1 and 6-2. For both dry and wet H2 conditions, the
activation energy decreases with increasing H2 partial pressure. In addition, when the ratio of
steam partial pressure to H2 partial pressure decreases, the activation energy decreases as well.
Since the H2 permeation is proportional to H2 partial pressure to the 1/4 power, the wet H2
condition corresponds to region V in the defect equilibrium diagram according to Song et al.
[27]. In this region, the electron concentration equals to the proton concentration and they are the
dominating defect species instead of oxygen vacancy. According to Guan et al. [72], the
activation energies of proton and electron mobility in perovskite are 0.4-0.6 eV and 1 eV,
respectively. In addition, the H2 permeation through SrCe0.9Eu0.1O3 was limited by electronic
conduction [6]. Therefore, the activation energies at high H2O concentration and low H2 partial
pressure are close to 1 eV. However, O2 partial pressure decreases as increasing H2 partial
pressure, resulting in an increase in electron concentration and electronic conduction. Therefore,
87
the H2 permeation is less limited by electronic conduction resulting in decrease in the activation
energy. Similarly, when the ratio of steam partial pressure to H2 partial pressure decreases, the
activation energy decreases.
6.3.6 Long Term Stability
The membrane’s long term stability was investigated by measuring the H2 permeation as a
function of time under wet H2, and conditions of WGS reaction and SRM (Figure 6-6). The H2
permeation was stable throughout the experiment for all three conditions. The H2 permeation is
~0.98 cm3/min under 5 cm3/min H2, 15 cm3/min Ar and 3% H2O at 850 oC, almost identical to
the result (~0.94 cm3/min) in Figure 6-2 (B). The stable H2 permeation under the conditions of
WGS reaction and SRM confirms the stability improvement of the SrCe0.7Zr0.2Eu0.1O3-δ using Zr
as a dopant.
6.4 Conclusions
H2 permeation through supported thin-film SrCe0.7Zr0.2Eu0.1O3-δ membranes was
investigated. Permeation flux was proportional to the transmembrane H2 partial pressure gradient
with a 1/4 dependence and controlled by bulk diffusion. A maximum H2 permeation flux of 0.23
and 0.21 cm3/cm2 min was obtained for the 33 μm thick SrCe0.7Zr0.2Eu0.1O3-δ membrane at 900
oC and 100% H2 and 97% H2/3% H2O in the feed gas, respectively. Permeation flux decreased
with increasing steam partial pressure. The activation energy decreased with increasing H2
partial pressure and/or decreasing steam partial pressure. Permeation flux through the
SrCe0.7Zr0.2Eu0.1O3-δ membrane was stable under wet H2, and conditions of WGS reaction and
SRM.
88
A
B C
D
Figure 6-1. SrCe0.7Zr0.2Eu0.1O3-δ membrane and experimental setup A) membrane after H2 permeation experiment, B) Surface morphology, C) Cross section and D) Experimental setup.
89
10
15
20
25
30
35
5 10 15 20 25 30 35 40 45Flow Rate (cm3/min)
J H2 (c
m3 /c
m2 m
in) Percentage (%
)
0.05
0.15
0.10
0.20
0.30
0.25
Figure 6-2. H2 permeation and H2 recovery as a function of feed flow rates.
90
0.01 0.02 0.03 0.04 0.05 0.06
J H2 (c
m3 /c
m2 m
in)
1/Thickness (µm-1)
24.3% H2
48.5% H2
72.8% H2
97.0% H2
0.40
0.35
0.30
0.25
0.20
0.15
0.10
Figure 6-3. H2 permeation vs thickness at 900oC.
91
0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.550
0.04
0.08
0.12
0.16
0.2
0.24
J H
2 (cm
3 /min
)
PH2f
1/4-PH2p
1/4 (atm1/4)
J H2 (cm
3/cm2 m
in)
900 oC
850 oC
800 oC
750 oC
700 oC
0.20
3.0
2.0
1.0
1.5
0.5
0
2.5
A
0.2 0.3 0.4 0.5 0.60
0.04
0.08
0.12
0.16
0.2
0.24
J H
2 (cm
3 /min
)
PH2f
1/4-PH2p
1/4 (atm1/4)
J H2 (cm
3/cm2 m
in)
900 oC850 oC
800 oC
750 oC
700 oC
0.20
3.0
2.5
2.0
1.5
1.0
0.5
0
B
Figure 6-4. H2 permeation as a function of H2 partial pressure and temperature A) under dry H2 condition and B) under 3% H2O.
92
650 700 750 800 850 900 950
dry H2
3% H2O+H
210% H
2O+H
230% H
2O+H
2
0
0.04
0.08
0.12
0.16
0.2
0.24J
H2 (c
m3 /m
in)
Temperature (oC)
J H2 (cm
3/cm2 m
in)
0.20
3.0
2.5
2.0
1.5
1.0
0.5
0
Figure 6-5. H2 permeation as a function of feed steam concentration and temperature.
93
Figure 6-6. H2 permeation as a function of time (a) under 8 cm3/min CH4, 12 cm3/min Ar and CH4/H2O=1:2, at 850 oC, (b) under 5 cm3/min H2, 15 cm3/min Ar and 3% H2O at 900 oC and (c) under 10 cm3/min CO, 10 cm3/min Ar and CO/H2O=1:2, at 900 oC.
94
Table 6-1. Activation energy as a function of H2 partial pressure under dry H2 and H2/3% H2O conditions balanced by Ar.
Dry H2 Ea (eV) H2/3% H2O Ea (eV) 25% 0.86 24% 0.95 50% 0.74 48% 0.84 75% 0.63 72% 0.73
Table 6-2. Activation energy as a function of H2O partial pressure with a constant H2 flow rate of 20 cm3/min.
30% H2O+70% H2 10% H2O+90% H2 3% H2O + 97% H2 100% H2 1.06 eV 0.91 eV 0.65 eV 0.58 eV
95
CHAPTER 7 SrCe0.7Zr0.2Eu0.1O3-δ-BASED HYDROGEN TRANSPORT WATER GAS SHIFT REACTOR
7.1 Introduction
In chapter 4, tubular SrCe0.9Eu0.1O3-δ membrane on NiO-SrCeO3-δ support was applied to
incorporate the WGS reaction and H2 separation. CO conversion and H2 yield were significantly
improved compared to the thermodynamic equilibrium. However, that work just focused on the
comparison of the CO conversion under three different reactor configurations. The CO
concentrations and flow rates were very low. In addition, this membrane is unstable under the
WGS reaction conditions [108]. In chapter 5, the chemical stability of SrCeO3-δ was significantly
improved using Zr dopant [108, 109]. H2 permeation property of tubular SrCe0.7Zr0.2Eu0.1O3-δ
membranes on NiO-SrCe0.8Zr0.2O3-δ supports [86] was discussed in chapter 6, In this chapter, the
effect of the SrCe0.7Zr0.2Eu0.1O3-δ membrane reactor on the WGS reaction was investigated in
various conditions. Its performance was investigated under high CO concentrations, high flow
rates, and different CO concentration and H2O/CO. Its long term stability under WGS reaction
condition was investigated as well.
7.2 Experimental
Polycrystalline SrCe0.8Zr0.2O3-δ and SrCe0.7Zr0.2Eu0.1O3-δ powders were prepared by
conventional solid-state reaction. NiO-SrCe0.8Zr0.2O3-δ tubular supports were fabricated by tape-
casting and rolling techniques. Detailed powder synthesis and membrane fabrication have been
discussed in chapter 3 and our previous works [74, 86].
The dense SrCe0.7Zr0.2Eu0.1O3-δ membrane used in this experiment was ~33 µm thick and
coated on NiO-SrCe0.8Zr0.2O3-δ support. Figure 7-1(A) shows the SEM image of the membrane
cross section after experiment. Based on our previous experience, its active area is about 12 cm2
96
[6, 86]. The influence of temperature, feed CO concentration, feed flow rate and H2O/CO were
evaluated in terms of CO conversion and H2 recovery.
The experimental setup is the same as that shown in Figure 4-1 (C). Figure 7-1 (B) shows
the schematic view of the membrane reactor. Outer side of the membrane (feed) was exposed to
CO and steam diluted to the desired concentration using Ar. Ar was also used as a tracer to
detect leakage. It is noted that no Ar leakage was detected during the experiment. Steam was
achieved by gasifying desired amount of water provided by a syringe pump. The inner side
(sweep) of the membrane was flushed with He at 20 cm3/min, in co-current flow with the feed
gas. The flow rates of CO, Ar and He were controlled by mass flow controllers. The reactants,
CO and H2O, were flowed into the quartz chamber and exposed to the Ni catalyst on the outside
of the membrane. The unreacted steam in the reactor side effluent was condensed by a cold trap
filled with ice prior to being analyzed by GC. The concentrations of the permeated H2 in the
sweep gas, He, were analyzed by a mass spectrometer.
7.3 Results and Discussion
7.3.1 Heat Treatment of the Membranes
Before the WGS reaction experiment, the membrane was heat treated to reduce the NiO in
the support by exposing it to 5 cm3/min H2 and 15 cm3/min Ar mixed with 3% H2O under 900 oC
overnight. This heat treatment has been described in reference [86].
7.3.2 H2O/CO Effect on CO Conversion
The temperature and H2O/CO effect on the WGS reaction were investigated. Figure 7-2
shows the mole fractions of the species in the reactor side effluent as a function of temperature
with various H2O/CO. The solid lines are experimental results and the dashed lines are
thermodynamic calculation data from reference [87]. Feed gas composition was 10 cm3/min CO,
10 cm3/min Ar and desired amount of steam based on the H2O/CO. To compare with the
97
thermodynamic data, Ar diluent is not included when the mole fraction was plotted. H2 and CO2
have the same profile for the thermodynamic calculation data regardless of the H2O/CO since
they are produced with the same rate. However, the mole fraction of CO2 is higher than that of
the H2 in experimental data due to the in situ removal of H2 through the H2 membrane. In
addition, the ratio of products, H2 and CO2, to reactants, H2O and CO, is greater for the
experimental results than for the thermodynamically calculated data. This indicates the WGS
reaction moves to the product side with the H2 membrane. A larger amount of CH4 than the
thermodynamic data was detected which is probably due to the Ni catalyst since Ni is an
effective catalyst for the methanation reaction [113, 114]. We didn’t detect CH4 in chapter 4 and
our previous work [87] because low concentration of CO was used and the CH4 concentration
was below detection limit.
CO conversion can be calculated from the mole fraction data and is plotted in Figure 7-3.
The CO conversion is defined in equation (4-2).
The dashed lines are thermodynamic calculation data from reference [87]. Based on the
thermodynamic calculation, the XCO decreases with increasing temperature due to the exothermic
nature of the WGS reaction. In contrast, it increases with increasing temperature in the
experimental results for any H2O/CO because of the in situ H2 removal through the H2
membrane. It also increases with increasing H2O/CO. A XCO of 83.6% and 90.2% was achieved
under 900 oC with H2O/CO = 1/1 and 2/1, respectively, 77% and 44% increase compared to the
thermodynamic calculation data.
7.3.3 H2O/CO Effect on H2 Production
Figure 7-4 shows the H2 production as a function of temperature under various H2O/CO
with 10 cm3/min CO, cm3/min Ar and desired amount of steam as feed gas. The H2 production
from thermodynamic calculation decreases with temperature. In contrast, the H2 permeation
98
increases with temperature due to the higher ambipolar conductivity of the membrane at elevated
temperature. The H2 flux in the reactor side effluent is not a significant function of temperature.
The H2 total production is the sum of permeated H2 and H2 flux in the reactor side effluent. It
increases with temperature. The H2 production curve is in a similar trend for any H2O/CO. The
permeated H2 decreases with increasing H2O/CO. A maximum H2 production of 8.1, 8.7 and 8.9
cm3/min was achieved under 900 oC with H2O/CO = 1/1, 1.5/1 and 2/1, respectively. Compared
to the thermodynamic calculation data, they are 73% and 42% improvement for H2O/CO = 1/1
and 2/1, respectively. If there is no side reaction, the improvement in XCO and H2 production
should be equal. However, the improvement of the XCO is higher compared to that of the H2
production. This is attributed to the methanation reaction consuming part of the produced H2,
which is clearly shown in Figure 7-2. In addition, the XCO is derived from GC measurement. The
H2 production is derived from both GC and MS measurement simultaneously. The XCO and H2
production data are pretty consistent within experimental error.
7.3.4 H2O/CO Effect on H2 Production and H2/CO
The H2O/CO effect on H2 yield and H2/CO in the reactor side effluent is plotted in Figure
7-5. The H2 yield is defined in equation (4-3). The H2 total yield is the sum of permeated H2
yield and H2 yield in the reactor side effluent. The H2 yield curves are in a similar trend with the
H2 production curves in Figure 7-4. Both the H2 yield and H2/CO increase with H2O/CO. The
H2/CO from thermodynamic calculation decreases with temperature, in agreement with the XCO.
When temperature is increased from 600 to 900 oC, it decreases from 1.3 to 0.9 and 2.7 to 1.7 for
H2O/CO = 1/1 and 2/1, respectively. In contrast, the H2/CO in the reactor side effluent increases
from 2.4 to 4.4 and 3.8 to 8.3 at the same temperature region for H2O/CO = 1/1 and 2/1,
respectively. The main composition of the products from the WGS reaction is syngas. The H2
99
membrane supplies an option to adjust the H2/CO prior to the Fischer-Tropsch process based on
desired products.
7.3.5 Flow Rate Effect on WGS Reaction
The flow rate effect on the WGS reaction was carried under 900 oC with a total flow rate
from 15 to 60 cm3/min and H2O/CO = 2/1. As shown in Figure 7-6, the XCO slightly decreases
with increasing flow rate, which is possibly due to the shorter residence time at higher flow rate.
Whereas the H2/CO in the reactor side effluent is not a significant function of total flow rate, the
permeated H2, H2 in the reactor side effluent, and H2 total production increases with total flow
rate.
7.3.6 CO Concentration Effect on WGS Reaction
Furthermore, the CO concentration effect on WGS reaction was investigated with a
H2O/CO = 2/1 under 850 and 900 oC. The total flow rate was 60 cm3/min and Ar was used as the
balanced gas. Figure 7-7 shows the H2 production, XCO, and H2/CO in the reactor side effluent as
a function of CO concentration. Compared to the results under 850 oC, a better performance was
obtained under 900 oC. For both temperatures, the XCO slightly decreases with increasing CO
concentration which can be explained by the percentage of the permeated H2 and H2 total
production. For example, the permeated H2/total H2 decreases from 14% to 8% under 900 oC
when CO concentration increases from 8.33% to 33.33%. Larger portion of produced H2
permeates through the membrane at lower CO concentration, moving the WGS reaction further
forward to the product side. As a result, the XCO increases. In contrast, the permeated H2, H2 in
the reactor side effluent, and H2 total production increases with CO concentration due to the
increase of the reactants amount.
100
7.3.7 Long Term Stability
Finally, the long term stability of the membrane was investigated. This experiment was
carried out under 900 oC with a feed gas composition of CO (10 cm3/min), H2O (20 cm3/min),
and Ar (10 cm3/min), respectively. As shown in Figure 7-8, the XCO is quite stable, only slightly
decreasing from 90% to 86% over 200 hours. Whereas the H2 total production slightly decreases
as the XCO decreases, the permeated H2 flux is essentially stable. The total carbon in gas phase is
the sum of detected CO, CO2 and CH4. It is essentially equal to the input amount (10 cm3/min),
indicating the carbon deposition is negligible. In addition, good integrity of the membrane
reactor remained after the experiment, showing high stability under the WGS reaction
conditions. This further confirms the stability of SrCe0.9Eu0.1O3-δ is improved by substitution 20
mol% Zr on Ce sites.
7.4 Conclusions
WGS reaction is constrained by thermodynamic equilibrium limitations. A tubular
SrCe0.7Zr0.2Eu0.1O3-δ H2 membrane reactor was fabricated. The XCO, H2 production, H2 yield and
the H2/CO in the reactor side effluent increased with increasing temperature and H2O/CO. A XCO
of 83.6% and 90.2% was achieved under 900 oC with H2O/CO = 1/1 and 2/1, respectively, 77%
and 44% increase compared to the thermodynamic calculation data. The respective improvement
in H2 production was 73% and 42%. In contrast to the XCO, the permeated H2, H2 in the reactor
side effluent and H2 total production increased with increasing flow rate and CO concentration.
The H2/CO in the reactor side effluent is variable through the SrCe0.7Zr0.2Eu0.1O3-δ H2 membrane.
This membrane is stable under the WGS reaction conditions.
101
A
B
Figure 7-1. Membrane morphology and experiment setup A) SEM image of the membrane and B) Schematic View of the membrane reactor.
102
0
0.1
0.2
0.3
0.4
0.5
650 700 750 800 850 900 950
Mol
e Fr
actio
nCO
2
H2
COH
2O
CH4
Temperature (oC)
CO2H
2
COH
2O
CH4
H2O/CO=1/1
A
650 700 750 800 850 900 9500
0.1
0.2
0.3
0.4
0.5
CO2
H2
H2O
CO
CH4
Temperature (oC)
H2O/CO=1.5/1
Mol
e Fr
actio
n
B
Figure 7-2. Gas compositions of the reactor side effluent as a function of temperature (dashed lines are from thermodynamic calculation) A) H2O/CO=1/1; B) H2O/CO=1.5/1 and C) H2O/CO=2/1.
103
0
0.1
0.2
0.3
0.4
0.5
650 700 750 800 850 900 950
Mol
e Fr
actio
n
CO2 H
2
CO
H2O
CH4
Temperature (oC)
CO2
H2
CO
H2O
CH4
H2O/CO=2/1
C
Figure 7-2.. Continued
104
Figure 7-3. XCO as a function of temperature.
40
50
60
70
80
90
100
650 700 750 800 850 900 950
X CO (%
)
Temperature (oC)
2/11.5/1
1/1
H2O/CO
Thermodynamic data (2/1)
Thermodynamic data (1/1)
105
0
2
4
6
8
10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
650 700 750 800 850 900 950
H2 P
rodu
ctio
n (c
m3 /m
in)
H2 permeation
thermodynamic H2 production
H2 total production
H2 flow in reactor side effluent
Temperature (oC)
H2 Flux (cm
3/cm2 m
in)
A
0
2
4
6
8
10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
650 700 750 800 850 900 950
H2 P
rodu
ctio
n (c
m3 /m
in)
Temperature (oC)
H2 Flux (cm
3/cm2 m
in)
H2 ptotal roduction
H2 flow in reactor side effluent
H2 permeation
B
Figure 7-4. H2 production as a function of temperature A) H2O/CO=1/1; B) H2O/CO=1.5/1 and C) H2O/CO=2/1.
106
0
2
4
6
8
10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
650 700 750 800 850 900 950
H2 P
rodu
ctio
n (c
m3 /m
in)
H2 permeation
thermodynamic H2 production
H2 total production
H2 flow in reactor
side effluent
Temperature (oC)
H2 Flux (cm
3/cm2 m
in)
C
Figure 7-4.. Continued
107
0
20
40
60
80
100
0
0.8
1.6
2.4
3.2
4
4.8
5.6
6.4
650 700 750 800 850 900 950
H2 y
ield
(%)
Temperature (oC)
total H2 yield
H2 in feed side effluent
H2/CO ratio in feed side effluent
permeated H2
thermodynamic H2/CO ratio
thermodynamic H2 yield H
2 /CO
4.0
A
0
20
40
60
80
100
650 700 750 800 850 900 9500
2
4
6
8
10
Temperature (oC)
H2 y
ield
(%) H
2 /CO
permeated H2
H2 in reactor side effluent
H2 total yield
H2/CO ratio in reactor side effluent
B
Figure 7-5. H2 yield and H2/CO in the reactor side effluent as a function of temperature A) H2O/CO=1/1; B) H2O/CO=1.5/1 and C) H2O/CO=2/1.
108
0
20
40
60
80
100
0
5
10
15
650 700 750 800 850 900 950
H2 y
ield
(%)
Temperature (oC)
H2 total yield
H2 in reactor side effluent
H2/CO ratio in reactor side effluent
permeated H2
thermodynamic H2/CO ratio
thermodynamic H2 yield H
2 /CO
C
Figure 7-5.. Continued
109
0
5
10
15
20
0
2
4
6
8
1080
85
90
95
100
10 20 30 40 50 60 70
H2 P
rodu
ctio
n (c
m3 /m
in)
Total Flow Rate (cm3/min)
permeated H2
H2/CO in reactor side effluent
H2 in reactor side effluent
total H2
CO conversion
H2 /C
OX
CO (%
)
Figure 7-6. The XCO, H2 production and H2/CO in the reactor side effluent as a function of flow rates under 900 oC.
110
0
5
10
15
20
0
2
4
6
8
1080
85
90
95
100
5 10 15 20 25 30 35
H2 P
rodu
ctio
n (c
m3 /m
in)
XC
O (%)
CO Concentration (%)
H2 /C
O
permeated H2
H2/CO in reactor side effluent
total H2
H2 in reactor side effluent
CO conversion
Solid symbol--900 oCHollow symbol--850 oC
Figure 7-7. The XCO, H2 production and H2/CO in the reactor side effluent as a function of CO concentrations with H2O/CO=2/1.
111
Figure 7-8. The performance of the membrane reactor as a function of time under 900 oC.
112
CHAPTER 8 HIGH TEMPERATURE SrCe0.7Zr0.2Eu0.1O3-δ MEMBRANE REACTOR FOR H2
PRODUCTION AND SEPARATION USING THE STEAM REFORMING OF METHANE
8.1 Introduction
Steam reforming of methane (SRM) has been the most important chemical process in
producing H2 and syngas [115]. About 95% of the H2 produced in the U.S. is through the SRM
process [1]. This process includes two reversible reactions [116]:
242 3 HCOCHOH +↔+ 206=∆ oH kJ/mol (8-1)
222 HCOCOOH +↔+ 41−=∆ oH kJ/mol (8-2)
The reforming reaction (8-1) is an endothermic reaction, thermodynamically favored by high
temperature and low pressure. The WGS reaction (8-2) is an exothermic reaction, independent of
pressure and favored by low temperature. Heat generated by the WGS reaction is not sufficient
for the reforming reaction (8-1), as can be seen from the overall reaction:
2242 4 2 HCOCHOH +↔+ 165=∆ oH kJ/mol (8-3)
The conventional industrial process is carried out in furnaces at about 850 oC, a few atmospheres
and with Ni/Al2O3 as the catalyst. 80% of CH4 conversion is achieved under these operating
conditions [116].
The SRM reaction is limited by thermodynamic equilibrium. It should be carried out at
high temperature to achieve high CH4 conversion. Therefore, the SRM reaction is highly capital
intensive accounting for about 70% of the total investment and operating cost in methanol
production based on natural gas [117]. Therefore, the development of a membrane based
separation process gives the possibility of increasing the CH4 conversion. Continuous removal of
H2 through H2 permeable membranes moves the equilibrium toward the products, resulting in
higher CH4 conversion. In addition, SRM in a membrane reactor becomes a transfer-limited
113
reaction related with membrane porosity and diffusivity, rather than an equilibrium-limited
reaction [118].
Enhanced performance of SRM with a palladium membrane was first reported by Oertel
et al. [119]. Previous research has been focused on Palladium-based membranes [119-124]. The
Palladium-based membranes are deposited on porous glass or metal substrates. However, the
porous glass substrate is fragile and difficult in connecting to metallic application [125]. The
rough surface of the porous metal substrate usually results in pinholes on the membrane [126].
Catalytic ceramic membranes supply another option to incorporate H2 separation and SRM into
one unit. The critical features for successfully membrane reactors are high separation selectivity,
high permeability and stability. In this chapter, SCZE721 membranes coated on the inner side of
tubular NiO-SCZ82 supports are used to incorporate H2 separation and SRM into one unit. The
separation selectivity of H2 is very high since they are dense membranes. Thin film membranes
are fabricated to improve permeability.
8.2 Experimental
The dense SCZE721 membrane used in this experiment was ~33 µm thick with an active
area of 12 cm2 [6, 86] and coated on a NiO-SCZ82 support. Figure 8-1 (A) shows the cross
section image of the membrane after experiment. The Influence of temperature, CH4/H2O, CH4
flow rate, and CH4 concentration on the SRM are investigated.
The experimental setup (Figure 8-1 (B)) is the same as that in reference [86] . The outer
side of the membrane (feed side) was exposed to CH4 and steam. Steam was achieved by
gasifying the desired amount of water provided by a syringe pump. The inner side (sweep side)
of the membrane was flushed with He at 20 cm3/min, in co-current flow with the feed gases. The
flow rates of CH4, Ar and He were controlled by mass flow controllers. The reactants were
114
flowed into the quartz chamber and exposed to the Ni catalyst on the outside of the membrane.
The unreacted steam in the reactor side effluent was condensed by a cold trap filled with ice
prior to being analyzed by GC. The concentrations of the permeated H2 in the sweep gas (He)
were analyzed by a mass spectrometer.
The membranes were heat treated in H2 to reduce NiO to Ni in the support before
experiments [86].
The conversion of CH4, the selectivity of H2 and CO, and the ratio of H2/CO in the reactor
side effluent were defined:
%1004
44
4 ×−
= inCH
outCH
inCH
CH FFF
X (8-4)
%1002
×+
= outCO
outCO
outCO
CO FFFS (8-5)
%1002
22 ×
+= out
COout
CO
outCO
CO FFF
S (8-6)
%1004
22 ×= in
CH
outH
FF
yieldH (8-7)
outCO
outH FFCOH //
22 = (8-8)
where iX , iS , iniF and out
iF (i=CH4, CO2, H2, and CO) are the conversion, selectivity, input and
output flux of i, respectively.
8.3 Results and Discussion
8.3.1 Thermodynamic Calculation Results
There are seven possible species in the SRM system: CH4, H2O, H2, CO, CO2, C and O2.
Three equations can be obtained through mass balance of C, H and O atoms. Four more
equations are needed to solve the seven unknowns. These equations can be obtained through the
115
relation between thermodynamic equilibrium constant (K) and Gibbs free energy of reactions
between 298-2000 K [127]:
KRTG ln−=∆ (8-9)
The following four reactions were used in the calculation:
224 3HCOOHCH +=+
)( 15.78ln25.22204920 JTTTGo −−=∆ (8-1)
222 HCOOHCO +=+
)( 3236000 JTGo +−=∆ (8-2)
24 2HCCH +=
)( 35.65ln25.2269120 JTTTGo +−=∆ (8-10)
OHOH 222 2/1 =+
)( 85.55247500 JTGo +−=∆ (8-11)
The influence of CH4/H2O and CH4 and H2O concentrations on the SRM was calculated with a
total pressure of 1 atm.
Figure 8-2 shows the influence CH4/H2O on the calculated XCH4 as a function of
CH4/H2O and temperature. As expected, the XCH4 is strongly dependent on both temperature and
CH4/H2O. For both with and without inert diluent, the XCH4 increases with increasing
temperature since the SRM reaction (equation (8-1) and (8-3)) is endothermic reaction. The XCH4
increases with decreasing CH4/H2O as well. Similar findings were reported by Liu [128] and
Rakas et al [129].
Figure 8-3 shows the influence of CH4 and H2O concentrations on the calculated XCH4 as a
function of temperature with CH4/H2O=1/2. The XCH4 increases with decreasing the CH4 and
116
H2O concentrations. The volume expands in the SRM process. Therefore, lower CH4 and H2O
concentrations move the reaction forward to the product side and increase the XCH4.
The condition for carbon formation was calculated as a function of CH4/H2O and
temperature and is shown in Figure 8-4. The carbon formation is suppressed by lower CH4/H2O.
The CH4/H2O should be below 0.6 to avoid possible coking. CH4/H2O=1/2 is used in the
following sections except for the discussion of the CH4/H2O effect on the SRM.
8.3.2 Experimental Results
8.3.2.1 Influence of CH4/H2O on the SRM
The effect of CH4/H2O on the SRM was investigated using min/ 4 34
cmF inCH = ,
min/ 16 3cmF inAr = with a desired amount of steam. Figure 8-5 shows the XCH4 as a function of
temperature and CH4/H2O. The XCH4 increases with increasing temperature and decreasing
CH4/H2O, in agreement with the thermodynamic calculation results. A XCH4 of 95.0%, 92.0%
and 89.3% was achieved at 900 oC with CH4/H2O/Ar=1/3/4, 1/2/4 and 1/1/4, respectively.
However, the experimental XCH4 is lower than thermodynamic calculation results, especially at
low temperatures. This indicates the SRM process is rate limited by kinetic reaction rate which is
possibly caused by the short residence time and/or the inadequacy of Ni catalyst. In this
experimental setup, the Ni catalyst is embedded in the porous substrate. The reactants, CH4 and
H2O, flow between the quartz reactor and the H2 membrane. It is possible that part of the
reactants flows out the system without being exposed to the Ni, especially with the volume
expansion in the SRM process. This situation will be solved by building a Ni catalyst bed
between the quartz reactor and the H2 membrane, which will be discussed in more detailed in
chapter 10. The lower experimental XCH4 than thermodynamic equilibrium results is common in
literatures. 80% of XCH4, for instance, was reported at 850 oC [116].
117
Figure 8-6 shows the SCO, SCO2 and H2/CO in the reactor side effluent as a function of
temperature and CH4/H2O. The SCO increases with increasing temperature whereas the SCO2 and
H2/CO decrease with increasing temperature which is ascribed to the higher XCH4 and exothermic
nature of the WGS reaction (equation (8-2)). The SCO decreases with decreasing CH4/H2O since
lower CH4/H2O moves the WGS reaction forward and produces more CO2 and H2.
Consequently, the SCO2 and H2/CO increase with decreasing CH4/H2O. If the effluent is used to
synthesize liquid fuels through the Fischer-Tropsch process, the H2/CO is too high since the ideal
value is ~2. However, it can be adjusted by combing the SRM and carbon dioxide reforming of
methane, which will be discussed in chapter 9.
The H2 production is shown in Figure 8-7 as a function of temperature and CH4/H2O. The
H2 permeation decreases with decreasing CH4/H2O due to the higher PO2 in lower CH4/H2O. It
increases with increasing temperature due to the higher ambipolar conductivity of the H2
membrane at higher temperatures. A maximum H2 permeation of 0.18 cm3/cm2 min was
achieved at 900 oC and CH4/H2O/Ar=1/1/4. The H2 flow in the reactor side effluent is not a
significant function of temperature and increases with decreasing CH4/H2O. The total H2 is the
sum of the H2 permeation and the H2 flow at the reactor side effluent. It increases with increasing
temperature and decreasing CH4/H2O, which is consistent with the XCH4 in Figure 8-5. A
maximum H2 production of 11.8 cm3/ min was achieved at 900 oC with CH4/H2O/Ar=1/3/4.
8.3.2.2 Influence of CH4 concentration on the SRM
The effect of CH4 concentration on the SRM was investigated with a total flow rate of 60
cm3/ min and CH4/H2O=1/2. Ar was used as the balanced gas. Figure 8-8 shows the XCH4, SCO,
H2/CO and H2 production as a function of CH4%. The XCH4 slightly decreases with increasing
CH4% in agreement with the thermodynamic calculation results (Figure 8-3).This can be
explained by the volume expansion of the SRM. Higher CH4% results in lower XCH4 and higher
118
PH2O. As a result, the SCO decreases with increasing CH4%. In contrast with the XCH4 and SCO, the
permeated H2, H2 flow in the reactor side effluent and total H2 production increase with CH4%
due to the increase of the reactants amount. The H2/CO increases with increasing CH4% as well
due to the higher H2 production and lower SCO. In addition, compared to the results under 850
oC, the XCH4, SCO and H2 production are higher at 900 oC whereas the H2/CO is lower. This is
consistent with the discussion in section 8.3.2.1.
8.3.2.3 Influence of total flow rate on the SRM
The total flow rate effect on the SRM was investigated from 700 to 900 oC with
CH4/H2O=1/2. As shown in Figure 8-9, the XCH4 decreases with increasing total flow rate due to
the shorter residence time at higher total flow rate. Lower XCH4 results in higher PH2O which
favors the WGS reaction and produces more CO2. Consequently, the SCO decreases with
increasing total flow rate as well. In contrast, the permeated H2, H2 flow in the reactor side
effluent and total H2 production increase with increasing total flow rate due to the increase of the
reactants amount. The H2/CO increases with increasing total flow rate as well due to the higher
H2 production and lower SCO. In addition, the XCH4, SCO and H2 production increase and the
H2/CO decreases with increasing temperature in agreement with the discussion in section 8.3.2.1.
8.3.2.4 Influence of the H2 membrane reactor on the SRM
To investigate the influence of the H2 membrane reactor on the SRM, experiment was
carried out with min/ 20 34
cmF inCH = and min/ 40 3
2cmF in
OH = under three reactor configurations:
(1) blank quartz reactor, (2) with Ni catalyst and (3) with Ni catalyst and in situ H2 removal. For
the blank quartz reactor, CH4 and H2O were fed into the empty quartz reactor. For the
configuration (2), a H2 membrane was installed in the quartz reactor but the permeated side
outlet was blocked so that no produced H2 was removed. Therefore, the membrane only
119
functioned as a catalyst in this situation. For the configuration (3), the permeated side of the H2
membrane was open and connected to a mass spectrometer, so that the permeated H2
concentration could be analyzed. These three configurations are similar with those configurations
discussed in the section 4.3.2.
As shown in Figure 8-10, the XCH4 in the blank quartz reactor is very low and increases
with increasing temperature which is due to the higher kinetic reaction rate at higher
temperature. The XCH4 of the configuration (3) is much higher than that of the configuration (2)
even though both XCH4 are lower than the thermodynamically calculated XCH4. The maximum
XCH4 of the configuration (3) is 89% at 900 oC, 15% increase compared to the 77.5% of the
configuration (2) under the same conditions. Similarly, the total H2 production is much higher in
the configuration (3). The maximum H2 production of configuration (3) and (2) are 57.6 and 50.3
cm3/min at 900 oC, 15% increase for the configuration (3) with in situ H2 removal. In contrast,
the SCO and H2/CO are higher in the configuration (2). This demonstrates the H2 membrane can
improve the SRM performance by increasing the XCH4 and H2 production.
8.3.2.5 Long term stability
The long term stability of the H2 membrane was investigated under 850 oC with
min/ 8 34
cmF inCH = and min/ 16 3
2cmF in
OH = . As shown in Figure 8-11, the XCH4, the H2
permeation and total H2 production are all quite stable over 70 hours. This demonstrates the high
stability of the H2 membrane under the SRM conditions.
8.4 Conclusions
The XCH4 under thermodynamic equilibrium increases with increasing temperature and
decreasing CH4/H2O and CH4%. The experimental XCH4 is lower than the thermodynamic data
and limited by kinetic reaction rate. However, the H2 membrane can still enhance the SRM
120
performance by increasing 15% of both the XCH4 and total H2 production at 900 oC compared to
that with only Ni catalyst. The experimental XCH4 increases with increasing temperature and
decreasing CH4/H2O, CH4% and total flow rate. In contrast with the H2/CO, the SCO increases
with increasing temperature and decreases with decreasing CH4/H2O and increasing CH4% and
total flow rate. In contrast with the total H2 production, the H2 permeation decreases with
decreasing CH4/H2O. The H2 permeation and production increase with increasing temperature,
CH4% and total flow rate. The SrCe0.7Zr0.2Eu0.1O3-δ hydrogen membrane is stable under the SRM
conditions.
121
A
B Figure 8-1. Membrane morphology and experimental setup A) Cross section of the membrane
after experiment and B) Experimental setup.
122
0
20
40
60
80
100
300 400 500 600 700 800 900 1000 1100
X CH
4 (%)
Temperature (oC)
CH4/H
2O
1/3
1/1
1/2
A
0
20
40
60
80
100
300 400 500 600 700 800 900 1000 1100
X CH
4 (%)
Temperature (oC)
CH4/H
2O/Ar
1/3/4
1/1/4
1/2/41/2/4
B
Figure 8-2. Influence of CH4/H2O on XCH4 under thermodynamic equilibrium A) without a diluent and B) with a diluent (Ar).
123
0
20
40
60
80
100
300 400 500 600 700 800 900 1000 1100
X CH
4 (%)
Temperature (oC)
CH4/H
2O/Ar
1/2/91/2/41/2/31/2/11/2/0
CH4 & H
2O
Concentration increasing
Figure 8-3. Influence of CH4/H2O concentrations on XCH4 under thermodynamic equilibrium
with CH4/H2O=1/2 and Ar as the diluent.
124
Figure 8-4. Thermodynamic calculation of carbon formation as a function of temperature and CH4/H2O [130].
125
70
75
80
85
90
95
100
650 700 750 800 850 900 950
X CH
4 (%)
Temperature (oC)
1/1/4
1/2/41/2/4
1/3/41/3/4
1/1/4
CH4/H
2O/Ar
Experimental
Thermodynamiccalculation
Figure 8-5. Influence of CH4/H2O on XCH4.
126
10
20
30
40
50
60
70
80
90
2.5
3
3.5
4
4.5
5
5.5
6
650 700 750 800 850 900 950
S CO &
SC
O2 (%
)H
2 /CO
Temperature (oC)
1/1/4
1/1/4
1/2/4
1/2/4
1/2/4
1/3/4
1/3/41/3/4
1/1/4
CH4/H
2O/Ar
6.0
5.0
4.0
3.0
Figure 8-6. Influence of CH4/H2O on SCO, SCO2 and H2/CO in reactor side effluent (solid symbol—SCO2, hollow symbol—SCO and dashed line—H2/CO).
127
650 700 750 800 850 900 950
H2 P
rodu
ctio
n (c
m3 /m
in)
1/1/4
1/1/4
1/1/4
1/2/4
1/2/4
1/2/41/3/4
1/3/4
1/3/4
CH4/H
2O/Ar
Temperature (oC)
12
11
10
9
8
2
1
0.20
0.160.18
0.140.120.100.080.06
H2 Perm
eation (cm3/cm
2 min)
Figure 8-7. Influence of CH4/H2O on H2 production (solid symbol—total H2 production, hollow symbol—H2 production in reactor side effluent and dashed line—H2 permeation).
128
65
70
75
80
85
90
95
3.2
3.4
3.6
3.8
4
4.2
4.4
5 10 15 20 25 30 35
X CH
4 & S
CO (%
)
H2 /C
O
CH4 Concentration (%)
XCH4
SCO
4.0
H2/CO
Solid symbol-900 oCHollow symbol-850 oC
A
0
10
20
30
40
50
60
5 10 15 20 25 30 350
0.8
1.6
2.4
3.2
4
4.8
H2 P
rodu
ctio
n (c
m3 /m
in)
CH4 Concentration (%)
Solid symbol-900 oCHollow symbol-850 oC
H2 Permeation
Total H2
H2 in reactor side effluent
H2 Perm
eation (cm3/cm
2 min)
4.0
B
Figure 8-8. Influence of CH4 concentration on SRM A) XCH4, SCO & H2/CO vs CH4% and B) H2 production vs CH4%.
129
70
75
80
85
90
95
0
10
20
30
40
50
60
20 30 40 50 60 70
X CH
4 & S
CO (%
)
H2 Production (cm
3/cm2 m
in)
Total Flow Rate (cm3/min)
XCH4
SCO H
2 permeation
H2 in reactor side effluent
Total H2
CH4/H
2O=1/2 & 900 oC
A
65
70
75
80
85
90
0
10
20
30
40
50
60
20 30 40 50 60 70
X CH
4 & S
CO (%
)
H2 Production (cm
3/cm2 m
in)
Total Flow Rate (cm3/min)
XCH4
SCO
H2 permeation
H2 in reactor side effluent
Total H2
CH4/H
2O=1/2 & 850 oC
B
Figure 8-9. Influence of total flow rate on SRM A) 900 oC, B) 850 oC, C) 800 oC, D) 750 oC, E) 700 oC and F) H2/CO in reactor side effluent.
130
60
65
70
75
80
85
90
0
10
20
30
40
50
60
20 30 40 50 60 70
X CH
4 & S
CO (%
)
H2 Production (cm
3/cm2 m
in)
Total Flow Rate (cm3/min)
XCH4
SCO
H2 permeation
H2 in reactor side effluent
Total H2
CH4/H
2O=1/2 & 800 oC
C
60
65
70
75
80
85
0
10
20
30
40
50
60
20 30 40 50 60 70
X CH
4 & S
CO (%
)
H2 Production (cm
3/cm2 m
in)
Total Flow Rate (cm3/min)
XCH4
SCO
H2 permeation
H2 in reactor side effluent
Total H2
CH4/H
2O=1/2 & 750 oC
D
Figure 8-9. Continued
131
55
60
65
70
75
80
0
10
20
30
40
50
60
20 30 40 50 60 70
X CH
4 & S
CO (%
)
H2 Production (cm
3/cm2 m
in)
Total Flow Rate (cm3/min)
XCH4
SCO
H2 permeation
H2 in reactor side effluent
Total H2
CH4/H
2O=1/2 & 700 oC
E
20 30 40 50 60 70
H2/C
O
Total Flow Rate (cm3/min)
700 oC
750 oC
800 oC
850 oC
900 oC
CH4/H
2O=1/2
6.0
5.5
5.0
4.0
3.0
4.5
3.5
F
Figure 8-9. Continued
132
0
20
40
60
80
100
4
4.5
5
5.5
650 700 750 800 850 900 950
X CH
4 & S
CO (%
), H
2 Pro
duct
ion
(cm
3 /min
)H
2 /CO
XCH4
XCH4
Total H2 production
Total H2 production
SCO
SCO
XCH4
H2/CO
H2/CO
4.0
Temperature (oC)
Thermodynamic XCH4
Figure 8-10. Influence of reactor configurations on SRM (solid symbol—with catalyst and H2 removal, hollow symbol—with catalyst and dashed line—blank quartz reactor).
133
Figure 8-11. The performance of the membrane reactor as a function of time under 850 oC.
134
CHAPTER 9 HIGH TEMPERATURE SrCe0.7Zr0.2Eu0.1O3-δ PROTON CONDUCTING MEMBRANE
REACTOR FOR CARBON DIOXIDE REFORMING OF METHANE
9.1 Introduction
Fossil fuels are likely to play a major role in H2 energy in the near to medium-term future
with their inherent advantages, such as availability, relatively low cost and the existing
infrastructure for delivery and distribution [1]. Currently, the majority of H2 is produced through
SRM process with significant CO2 emission. The global warming potential (GWP) of H2
production via the SRM process is estimated to be 13.7 kg CO2 (equiv.) per kg of net H2
produced (CO2 accounts for 77.6% of the system’s GWP) [131]. 0.3-0.4 million cubic meters of
CO2 will be produced when one million cubic meters of H2 is produced through a typical SRM
H2 plant. The amount of CO2 emission would be double if H2 is to be produced by coal
gasification [1]. Therefore, CO2 sequestration has drawn lots of interest. The capture and
disposal of CO2 costs about 25-30% of the total cost of H2 production by the SRM process [132].
The net cost of CO2 disposal, however, could be significantly reduced if CO2 sequestration is
accompanied by enhanced oil recover [133].
9.1.1 Carbon Dioxide Reforming of Methane (CDRM)
There is a growing interest in catalytic reforming of methane with carbon dioxide
(equation (9-1)) because of the great benefit to both the economy and the environment. The
reforming reaction is:
224 2 2 HCOCOCH +↔+ 247=∆ oH kJ/mol (9-1)
This conversion consumes two undesirable greenhouse gases, CO2 and CH4, to generate
syngas. The produced syngas has a low H2/CO ratio, i.e., 1:1 or less. A low ratio is preferred for
synthesis of oxygenated compounds and long-chain hydrocarbons [134-136]. It also introduces
135
the possibility of combining the steam reforming, partial oxidation, and dry reforming reactions
to get the desired H2/CO [137] for different applications. Several technologies have been applied
to CO2 reforming of CH4 including catalysis conversion [138-143], plasma conversion [144-146]
and combination of catalyst and plasma [147, 148]. Nickel is a typical component in catalytic
reforming catalysts due to its wide availability, low cost and high catalytic activity [149-152].
9.1.2 Membrane Reactors for the CDRM
While plenty of research has been focused on the catalysts, there are only a few works on
the membrane reactor effect on the CO2 reforming of CH4 [153-155]. With in situ removal of H2,
membrane reactors can increase conversion of CH4 and CO2 compared to traditional reactors, or
the reforming process can be operated at lower temperature. Most of those membrane reactors in
the literatures above are Pd based membranes and they are expensive. Here, we investigate the
CDRM using ceramic membranes.
SrCeO3-δ and BaCeO3-δ based perovskite oxides with multivalent cation dopants are
promising for H2 membranes and have been reported by several groups [5, 6, 26, 27, 47, 73, 86].
We previously reported the H2 permeation of SrCe0.9Eu0.1O3-δ and SCZE721 [6, 86, 112]. Zr was
used to improve the stability of SrCe0.9Eu0.1O3-δ [7, 108]. We also carried out the water gas shift
(WGS) reaction using a SCZE721 H2 membrane reactor and found CO conversion and H2
production were significantly increased compared to the thermodynamic calculation [156]. The
SCZE721 membrane was coated on the inner side of a tubular NiO-SCZ82 support. In this
chapter, we investigate the SDRM through H2 permeable SCZE721 membrane reactors in terms
of CH4 and CO2 conversion, CO and H2 selectivity, and H2/CO in syngas product. Whereas the
H2/CO in syngas product through the CDRM is 1/1 or below, it is much higher through the steam
136
reforming of methane (SRM). The ideal H2/CO is ~2 for Fischer-Tropsch process to produce
liquid fuels. Therefore, we adjust the H2/CO by combining the CDRM and SRM processes.
9.1.3 Reaction Mechanism and Kinetics
The reaction of CO2 and CH4 is expected to proceed by three steps: (1) dehydrogenation of
methane to form surface carbon and H2, (2) dissociative of adsorption of CO2 and H2, and (3)
reduction of CO2 to CO. Possible reaction mechanism has been proposed [135]:
)(4)(4 aHaCCH += (9-2)
)()()(2 aOaCOgCO += (9-3)
)()()( aCOaOaC =+ (9-4)
)()( gCOaCO = (9-5)
)()(2 2 gHaH = (9-6)
A simplified reaction sequence for the CO2 reforming may involve two irreversible steps,
namely, the activation of methane followed by the surface reaction with adsorbed oxygen atoms:
**34 * HCHCH +=+ (9-7)
**2
*3 * HCHCH +=+ (9-8)
***2 * HCHCH +=+ (9-9)
*** * HCCH +=+ (9-10)
*2)2/( 2** ++=+ HxCOOCH x (9-11)
COOCO +=+ *2 * (9-12)
It has been proved from experiments that CH4 promotes the dissociation of CO2 on
catalysts. The promotion is attributed to the effect of H2 in the decomposition of CH4. It has been
demonstrated that a small amount of H2 can significantly facilitate this process [157]. Assuming
137
the above effect, the reaction mechanism of the CO2 reforming of CH4 on Rh and Pd was
proposed as follows:
HCHCH += 34 (9-13)
OHCOHCO +=+2 (9-14)
OHCHOCH +=+ 34 (9-15)
HCHCH += 23 (9-16)
HCHCH +=2 (9-17)
xHCCH x += (9-18)
xHCOOCH x +=+ (9-19)
xHCOCOCH x +=+ 22 (9-20)
22 HH = (9-21)
OHOH 22 = (9-22)
The kinetics for the CO2 reforming of CH4 depends on the catalyst and the mechanism is
probably changing with temperature. Previous investigations of the CO2 reforming of CH4
mainly dealt with the screening test of catalysts. Little research has focused on the reaction
mechanism and the kinetics. Thus, no general expression has been derived so far.
Sakai et al. [158] considered that the half-order and zero-order dependence of the reaction
rates on the partial pressure of CO2 or CH4 suggested that CO2 participates in the reaction via the
dissociative adsorption mechanism, being described as equation (9-3), whereas CH4 participates
in the reaction as strongly adsorbed species dehydrogenated to CHx and (4-x)H. A rate equation
was obtained by Richardson and Paripatyadar [159] using linear regression analysis:
2)1/(44224242 CHCHCOCOCHCOCHCOr PKPKPPKKKR ++= (9-23)
138
A regression coefficient of 0.988 was from calculated and measured rates. Experimental results
and the model fit well with each other.
9.2 Experimental
The dense SCZE721 membrane used in this experiment was ~33 µm thick (Figure 9-1 (A))
with an active area of 12 cm2 [6, 86] and coated on a NiO-SCZ82 support. The influence of
temperature, feed flow rate, and CH4/CO2 and CH4/CO2/H2O were evaluated in terms of CH4 and
CO2 conversion, H2 production and H2/CO.
The experimental setup (Figure 9-1 (B)) is the same as that in reference [86] . The outer
side of the membrane (feed side) was exposed to CH4 and CO2 or/and steam. Steam was
achieved by gasifying the desired amount of water provided by a syringe pump. The inner side
(sweep side) of the membrane was flushed with He at 20 cm3/min, in co-current flow with the
feed gases. The flow rates of CH4, CO2 and He were controlled by mass flow controllers. The
reactants were flowed into the quartz chamber and exposed to the Ni catalyst on the outside of
the membrane. The produced and/or unreacted steam in the reactor side effluent was condensed
by a cold trap filled with ice prior to being analyzed by GC. The concentrations of the permeated
H2 in the sweep gas (He) were analyzed by a mass spectrometer.
The membranes were heat treated in H2 to reduce NiO to Ni in the support before
experiments [86].
The conversion of CH4 and CO2, the selectivity of H2 and CO, and the ratio of H2/CO in
the reactor side effluent were defined:
%1004
44
4 ×−
= inCH
outCH
inCH
CH FFF
X (9-24)
%1002
22
2×
−= in
CO
outCO
inCO
CO FFF
X (9-25)
139
%100)(2
44
2
2×
−×= out
CHin
CH
outH
H FFF
S (9-26)
%100)()(
2244
×−+−
= outCO
inCO
outCH
inCH
outCO
CO FFFFFS (9-27)
outCO
outH FFCOH //
22 = (9-28)
where iX , iS , iniF and out
iF (i=CH4, CO2, H2, and CO) are the conversion, selectivity, input and
output flux of i, respectively.
9.3 Results and Discussion
9.3.1 CH4/CO2 Effect on Conversion, H2/CO and H2 Production
The effect of CH4/CO2 on the CO2 reforming of CH4 was investigated using
min/ 10 34
cmF inCH = with a desired amount of CO2. Figure 9-2 shows the XCH4 and XCO2 as a
function of temperature and CH4/CO2. Both XCH4 and XCO2 increase with increasing temperature
since the CO2 reforming of CH4 is endothermic. The XCH4 increases with decreasing CH4/CO2 as
well. The XCH4 and XCO2 significantly depends on catalysts and their supports [135]. Various
results have been reported [138, 139, 151, 160-162] and our results are in the range in the
literatures. At 900 oC, the XCH4 is 87%, 89% and 93% for CH4/CO2 = 1/1, 1/1.5 and 1/2,
respectively.
The XCH4 and XCO2 should be equal to each other with CH4/CO2 = 1/1 without any side
reactions. However, the measured XCH4 is higher. This is ascribed to carbon deposition through
CH4 decomposition and/or the Boudouard reaction:
24 2 HCCH +↔ 75=∆ oH kJ/mol (9-29)
CCOCO +↔ 2 2 172−=∆ oH kJ/mol (9-30)
140
The carbon deposition is confirmed by the less detected total gas phase carbon amount,
sum of CH4, CO and CO2 in the output. It is 2-5% lower than the input amount for CH4/CO2 =
1/1. In contrast, the total gas phase carbon amount essentially remains constant for CH4/CO2 =
1/1.5 and 1/2 indicating that carbon deposition is negligible under these conditions. This is in
agreement with Wang’s results [135]. Carbon deposition is thermodynamically possible for
CH4/CO2 = 1/1 at temperature up to 870 oC at 1 atm. Lower CH4/CO2 can suppress carbon
deposition. The lowest temperature for CH4/CO2 = 1/1.5 and 1/2 are 760 oC and 710 oC,
respectively.
Controversial results of the XCH4 and XCO2 have been reported [151, 160, 163]. Higher
XCO2 is ascribed to the reverse water gas shift (RWGS) reaction:
OHCOCOH 222 +↔+ 6.40=∆ oH kJ/mol (9-31)
The difference may be due to different catalysts and higher operating temperature in this
work. Reaction (9-7) and (9-8) are highly dependent on catalysts. According to Sacco et al [164],
the primary source of surface carbon on Ni catalyst is CH4 indicating that most of the carbon
deposition is through Reaction (9-7). In addition, according to Gibbs free energy, RWGS and the
Boudouard reactions could not occur spontaneously over 820 oC [135] while CH4 decomposition
is favorable at high temperature. In addition, a part of the produced H2 permeates through the
membrane reactor, lowering the H2 partial pressure and further limiting the RWGS reaction. As a
result, the XCH4 is higher than the XCO2 due to carbon deposition with CH4/CO2 = 1/1.
A low CH4/CO2 not only suppresses the carbon deposition but also enhances the RWGS
reaction (Reaction 9-31), resulting in higher XCO2. For example, the XCO2 is 52.4% with
CH4/CO2 = 1/2 at 900 oC, higher than half of the XCH4 (93.2%).
141
Figure 9-3 shows the SH2 and SCO as a function of temperature and CH4/CO2. It is noted
that H2 total production is used in calculating the H2 selectivity. At any given temperature and
CH4/CO2, the SCO is higher than the SH2. This is due to the RWGS reaction consuming some
produced H2. The SH2 decreases with increasing temperature in contrast to the SCO since the
RWGS reaction is endothermic. Lower SH2 and higher SCO are obtained at lower CH4/CO2.
Lower CH4/CO2 means more amount of oxygen atoms are in excess, which eventually are in
forms of H2O, CO and CO2 through reaction (9-9). Therefore, SH2 decreases as CH4/CO2
decreases in contrast to the SCO.
Figure 9-4 shows the H2 production. H2 total production is the sum of the H2 permeation
and the H2 in the reactor side effluent. All of them increase with increasing temperature due to
the higher ambipolar conductivity of the membrane and higher XCH4 at higher temperatures. A
maximum H2 permeation of 2.2 cm3/min (~0.2 cm3/cm2 min) was achieved at 900 oC with
CH4/CO2 = 1/1. The H2 permeation decreases with decreasing the CH4/CO2. This is due to the
higher PO2 in the system, which causes lower electronic conductivity of the membrane material
and results in lower H2 permeation flux. A maximum H2 total production of 17.1 cm3/min was
achieved at 900 oC with CH4/CO2 = 1/2.
Figure 9-5 shows the H2/CO in the reactor side effluent. The H2/CO is less than 1 due to
the RWGS reaction. The H2/CO decreases with increasing temperature due to the endothermic
RWGS reaction and higher H2 permeation at elevated temperatures. The decrease in the H2/CO
with lower CH4/CO2 is ascribed to the RWGS reaction as well. When temperature is increased
from 700 oC to 900 oC, the H2/CO decreases from 1.00 to 0.91, from 0.95 to 0.83 and from 0.92
to 0.77 for CH4/CO2 = 1/1, 1/1.5 and 1/2, respectively.
142
9.3.2 Flow Rate Effect on Conversion, H2/CO and H2 Production
To prevent carbon deposition, operating temperatures must be high enough. However,
nickel carbide may form on the surface of Ni-based catalysts at high temperatures [135].
Accordingly, an upper-temperature limit is needed to prevent such formation. For example, the
optimum temperature is between 870 oC and 1040 oC with CH4/CO2 = 1/1 at 1 atm. As
previously described, lower CH4/CO2 can suppress carbon deposition and enhance the RWGS
reaction as well. Therefore, an intermediate CH4/CO2 = 1/1.5 was chosen to investigate the flow
rate effect on the XCH4 and XCO2, H2 production and H2/CO at 850 oC and 900 oC. The total flow
rates are 25, 37.5 and 50 cm3/min, respectively.
As shown in Figure 9-6, the XCH4, XCO2 and SCO are higher at 900 oC than 850 oC. The SH2
is lower at 900 oC. This agrees with the results in section 9.3.1. In addition, the XCH4 and XCO2
slightly decrease with total flow rate due to shorter residence time. Whereas the SCO increases
with increasing total flow rate, the SH2 decreases in a similar trend with Raybold’s results [154].
The permeated H2 percentage of the H2 total production decreases with increasing total flow rate.
It decreases from 13.2% to 9.3% and from 10.4% to 8.2% at 900 oC and 850 oC, respectively,
when the total flow rate is increased from 20 to 50 cm3/min (Figure 9-8). Higher H2 partial
pressure in the reactor side effluent favors the RWGS reaction. As a result, the H2 selectivity
decreases with increasing total flow rate.
The H2/CO in the reactor side effluent (Figure 9-7) increases with increasing total flow rate
due to the lower percentage of the H2 permeation in the total H2 production. It is also higher at
850 oC than 900 oC which is ascribed to the WGS reaction.
Figure 9-8 shows the H2 production as a function of total flow rate. The H2 production
increases with total flow rates since more mass of reactants react even though the XCH4 and XCO2
slightly decrease. The H2 permeation fraction in H2 total production decreases with increasing
143
total flow rate. A maximum H2 permeation of 3.1 cm3/min (~0.26 cm3/cm2 min) and a maximum
H2 total production 32.8 cm3/min were achieved at 900 oC with a total flow rate of 50 cm3/min.
9.3.3 CH4/CO2/H2O Effect on XCH4, XCO2, H2/CO and H2 Production
The theoretical H2/CO in syngas through reaction (9-1) is 1/1 and it is less than 1/1 with
this H2 membrane reactor due to the in situ removal of H2 and the RWGS reaction. The ideal
H2/CO is ~2 to produce liquid fuels through the Fischer-Tropsch process. In addition, the H2/CO
is high in syngas produced through SRM:
242 3 HCOCHOH +↔+ 206=∆ oH kJ/mol (9-32)
Therefore, H2O was added to CO2 and CH4, combining the CDRM and SRM, to increase
the H2/CO. CH4/CO2/H2O = 2/1/1 and 2/1/1.5 were investigated in terms of H2/CO and XCH4 and
XCO2 with min/ 20 34
cmF inCH = , min/ 10 3
2cmF in
CO = , and desired amount of steam.
Figure 9-9 shows the XCH4 and XCO2 as a function of temperature. Both the XCH4 and
XCO2 increase with temperature. The XCH4 increases with increasing H2O concentration as well.
A higher H2O concentration favors the SRM and enhances the XCH4. A higher H2O concentration
also means lower concentration of CH4 and CO2 since the CH4/CO2 is fixed at 2/1 and favors the
WGS reaction as well. As a result, the XCO2 decreases. The XCH4 and XCO2 are 85% and 70% at
900 oC with CH4/CO2/H2O = 2/1/1, which means more than half of the converted CH4 is through
the SRM (reaction (9-32)). The SRM is even more dominant at low temperature since it is less
endothermic and the WGS reaction is exothermic.
Figure 9-10 shows the H2 production and H2/CO as a function of temperature. The H2
production increases with temperature. The H2 flux in reactor side effluent and H2 total increase
with increasing H2O concentration as well due to the higher XCH4. In contrast, the H2 permeation
decreases with increasing H2O since higher H2O concentration means higher PO2. As a result, the
144
H2 permeation decreases. A maximum H2 permeation of 4.7 cm3/min (~0.39 cm3/cm2 min) and a
maximum H2 total production of 48.0 cm3/min were achieved at 900 oC with CH4/CO2/H2O =
2/1/1 and 2/1/1.5, respectively. The H2/CO in the reactor side effluent decreases with
temperature in similar trend as that in Figure 9-5 and is attributed to the RWGS reaction. The
H2/CO between 700 oC to 900 oC is between 1.9-1.7 and 2.5-2.0 for CH4/CO2H2O = 2/1/1 and
2/1/1.5, respectively. This demonstrates the H2/CO is variable in CH4/CO2/H2O system.
9.4 Conclusions
Carbon dioxide reforming of methane was investigated using a tubular
SrCe0.7Zr0.2Eu0.1O3-δ H2 membrane reactor. The XCH4 and XCO2, CO selectivity and H2
production increase with increasing temperature. 93.2% of XCH4 was achieved with CH4/CO2 =
1/2 at 900 oC. A maximum H2 permeation of 2.2 cm3/min (~0.2 cm3/cm2 min) and a maximum
H2 total production of 17.1 cm3/min were achieved at 900 oC with the CH4/CO2 = 1/1 and 1/2,
respectively. In contrast, the H2 selectivity and H2/CO decrease with increasing temperature.
The XCH4, XCO2, SH2, and H2/CO decrease with total flow rate. In contrast, the SCO and H2
production increases with increasing total flow rate. A maximum H2 permeation of 3.1 cm3/min
(~0.26 cm3/cm2 min) and a maximum H2 production 32.8 cm3/min were achieved at 900 oC with
a total flow rate of 50 cm3/min.
The H2/CO in the reactor side effluent for carbon dioxide reforming of methane is less than
1. However, it can be increased by adding steam to the system. The H2/CO between 700 oC to
900 oC is between 1.9-1.7 and 2.5-2.0 for CH4/CO2/H2O = 2/1/1 and 2/1/1.5, respectively.
145
A
B
Figure 9-1. Membrane morphology and experimental setup A) Cross section of the membrane after experiment and B) Experimental setup.
146
40
50
60
70
80
90
100
650 700 750 800 850 900 950
X CH
4 & X
CO
2 (%)
CH4/CO
2
Solid line-XCH4
Dashed line-XCO2
1/2
1/2
1/1.5
1/1.5
1/1
1/1
Temperature (oC)
Figure 9-2. XCH4 and XCO2 as a function of temperature and CH4/CO2.
147
92
93
94
95
96
97
98
99
100
650 700 750 800 850 900 950
S H2 &
SC
O(%
)
Temperature (oC)
CH4/CO
2
Solid line-SH2
Dashed line-SCO
1/2
1/1.5
1/1
1/2
1/1
1/1.5
Figure 9-3. SH2 and SCO as a function of temperature and CH4/CO2.
148
650 700 750 800 850 900 950
H2 p
rodu
ctio
n (c
m3 /m
in)
Temperature (oC)
CH4/CO
2
H2 permeation
solid symbol- total H2 production
hollow symbol- H2 in reactor side effluent
1/2
1/2
1/2
1/1.5
1/1.5
1/1.5
1/1
1/1
1/1
18
14151617
111213
103
2
1
Figure 9-4. H2 production as a function of temperature and CH4/CO2.
149
650 700 750 800 850 900 950
H2/C
O
Temperature (oC)
1/1
1/2
1/1.5
CH4/CO
2
1.00
0.95
0.90
0.85
0.80
0.75
Figure 9-5. H2/CO in the reactor side effluent as a function of temperature and CH4/CO2.
150
50
55
60
65
70
75
80
85
90
93
94
95
96
97
98
99
100
20 25 30 35 40 45 50 55
X CH
4 & X
CO
2(%) S
H2 &
SC
O (%
)
Total flow rate (cm3/min)
XCH4
900oC
XCH4
850oC
XCO2
900oC
XCO2
850oC
SH2
850 oC
SH2
900 oC
SCO
900 oC SCO
850 oC
CH4/CO
2=1/1.5
Figure 9-6. XCH4, XCO2, SH2 and SCO as a function of total flow rate.
151
0.82
0.83
0.84
0.85
0.86
0.87
0.88
20 25 30 35 40 45 50 55
H2/C
O
Total flow rate (cm3/min)
850 oC
900 oC
CH4/CO
2=1/1.5
Figure 9-7. H2/CO in reactor side effluent as a function of total flow rate.
152
8
9
10
11
12
13
14
20 25 30 35 40 45 50 55
H2 p
rodu
ctio
n (c
m3 /m
in)
H2p percentage of total H
2 production (%)
H2 permeation
solid symbol--900oChollow symbol--850oC
H2 in feed side effluent
CH4/CO
2=1/1.5
H2 total production
Total flow rate (cm3/min)
H2p
fraction
35.0
30.0
25.0
20.0
15.0
10.0
3.5
3.0
2.5
2.0
1.5
Figure 9-8. H2 production as a function of total flow rate.
153
10
20
30
40
50
60
70
80
90
650 700 750 800 850 900 950
X CH
4 & X
CO
2 (%)
Temperature (oC)
XCH4
XCO2
2/1/1.52/1/1
2/1/1
2/1/1.5
CH4/CO
2/H
2O
Figure 9-9. XCH4 and XCO2 as a function of temperature.
154
2
3
4
5
6
7
830
35
40
45
50
650 700 750 800 850 900 950
H2 p
rodu
ctio
n (c
m3 /m
in)
Temperature (oC)
H2 /C
O
H2 permeation
H2 total production
H2 in feed side effluent
H2/CO
solid symbol--2/1/1hollow symbol--2/1/1.5
CH4/CO
2/H
2O
2.6
1.5
2.0
Figure 9-10. H2 production and H2/CO as a function of temperature.
155
CHAPTER 10 CONCLUSIONS AND FUTURE WORKS
10.1 Conclusions
In this dissertation, H2 production and separation is demonstrated using supported
SCZE721 thin film membranes through WGS, SRM and CDRM processes. My conclusions are
as followings:
Supported SrCe0.9Eu0.1O3-δ and SCZE721 thin film membranes are successfully fabricated
by tape casting followed by a rolling step. NiO-SrCeO3-δ and NiO-SrCe0.8Zr0.2O3-δ are used as
the substrates to maintain mechanical integrity.
CO2 is the main reason for the decomposition of SrCe0.9Eu0.1O3-δ under hydrocarbon
conditions. However, the chemical stability of SrCe0.9Eu0.1O3-δ can be improved by partially
substituting Ce with Zr, which increases its tolerance factor and decreases its basicity. The H2
permeation of the SCZE721 membranes was essentially stable under WGS and SRM conditions.
H2 permeation through supported SCZE721 membranes is proportional to the
transmembrane H2 partial pressure gradient with a 1/4 dependence and controlled by bulk
diffusion with thickness down to 17 µm. A maximum H2 permeation flux of 0.23 and 0.21
cm3/cm2 min was obtained for the 33 μm thick SCZE721 membrane at 900 oC with the total flow
rate of 20 cm3/min and the feed gas composition of 100% H2 and 97% H2/3% H2O, respectively.
H2 permeation decreases with increasing steam partial pressure. The activation energy decreases
with increasing H2 partial pressure and/or decreasing steam partial pressure. Permeation flux
through the SCZE721 membrane is stable under wet H2, and conditions of WGS reaction and
SRM.
WGS reaction is exothermic and constrained by thermodynamic equilibrium limitations.
SrCe0.9Eu0.1O3-δ and SCZE721 thin film membranes incorporate H2 separation and WGS reaction
156
into one unit. Continuously removal of produced H2 moves the equilibrium toward the product
side, overcoming the thermodynamic equilibrium limitation. The XCO, H2 production, H2 yield
and the H2/CO in the reactor side effluent increase with increasing temperature and H2O/CO. A
XCO of 83.6% and 90.2% was achieved under 900 oC with H2O/CO = 1/1 and 2/1, respectively,
77% and 44% increase compared to the thermodynamic calculation data. The respective
improvement in H2 production was 73% and 42%. In contrast to the XCO, the permeated H2, H2
in the reactor side effluent and H2 total production increase with increasing flow rate and CO
concentration. The H2/CO in the reactor side effluent is variable through the SCZE721 H2
membrane. The SCZE721 membrane is stable under the WGS reaction conditions.
The XCH4 under thermodynamic equilibrium increases with increasing temperature and
decreasing CH4/H2O and CH4%. The experimental XCH4 is lower than the thermodynamic data
and limited by kinetic reaction rate. However, the H2 membrane can still enhance the SRM
performance by increasing 15% of both the XCH4 and total H2 production at 900 oC compared to
that with only Ni catalyst. The experimental XCH4 increases with increasing temperature and
decreasing CH4/H2O, CH4% and total flow rate. In contrast with the H2/CO, the SCO increases
with increasing temperature and decreases with decreasing CH4/H2O and increasing CH4% and
total flow rate. In contrast with the total H2 production, the H2 permeation decreases with
decreasing CH4/H2O. The H2 permeation and production increase with increasing temperature,
CH4% and total flow rate. The SCZE721 H2 membrane is stable under the SRM conditions.
The CH4 and CO2 conversion, CO selectivity and H2 production in the CSRM increase
with increasing temperature. 93.2% of CH4 conversion is achieved with CH4/CO2 = 1/2 at 900
oC. A maximum H2 permeation of 2.2 cm3/min (~0.2 cm3/cm2 min) and a maximum H2 total
production of 17.1 cm3/min are achieved at 900 oC with the CH4/CO2 = 1/1 and 1/2, respectively.
157
In contrast, the H2 selectivity and H2/CO decrease with increasing temperature. The CH4 and
CO2 conversion, H2 selectivity and H2/CO decrease with total flow rate. In contrast, the CO
selectivity and H2 production increases with increasing total flow rate. A maximum H2
permeation of 3.1 cm3/min (~0.26 cm3/cm2 min) and a maximum H2 production 32.8 cm3/min
were achieved at 900 oC with a total flow rate of 50 cm3/min. The H2/CO in the reactor side
effluent can be adjusted to desired value by combing the SRM and CDRM. The H2/CO between
700 oC to 900 oC is between 1.9-1.7 and 2.5-2.0 for CH4/CO2H2O = 2/1/1 and 2/1/1.5,
respectively.
10.2 Future Work
The objective of this research is to fabricate supported tubular thin film membranes by tape
casting followed by a rolling step, to improve the chemical stability of SrCe0.9Eu0.1O3-δ and to
incorporate H2 production and separation into one unit using SCZE721 membranes through
WGS reaction, steam reforming of CH4 and CO2 reforming of CH4. All these have been achieved
and demonstrated in chapter 3 to chapter 9. However, overall performance of the membrane
reactors can be improved in future work including material compositions and reactor design.
One future work is to modify the substrate composition. NiO-SCZ82 is the substrate in this
work. NiO is used to create porosity and to serve as the catalyst, by reduction to Ni when the
membrane is subsequently exposed to H2. However, Ni catalyst is embedded in the substrate. It
is difficult to replace the Ni since replacing the Ni usually means replacing the whole membrane.
The Ni particles are inevitable to grow during the sintering process of the membranes which
decreases its surface area. In addition, as shown in chapter 8, the XCH4 is lower than the
thermodynamic calculation result which is partially due to the inadequacy of the catalytic
activity. In addition, Ni is a catalyst for CH4 decomposition as well. The carbon deposited inside
158
the substrate will block the porous structure and even cause mechanically instability. Therefore, I
propose to substitute NiO with graphite as pore former and place Ni catalyst outside the
membrane. Ni catalyst can be synthesis separately to achieve optimized catalytic activity with
desired particle size and on different supports. In addition, much less amount of Ni is needed
compared to the amount used in the substrate above while maintaining enough activity for the
WGS, SRM and CDRM processes. Actually, I have successfully fabricated SCZE721 thin film
membranes coated on 30 vol% graphite/cellulose + 70 vol% SCZ82 substrates (Figure 10-1).
Another future work is to change the experimental setup and put the whole membrane into
the hot zone of the furnace. In the current setup, the membrane is installed at the bottom of the
quartz reactor. The bottom part of the tubular membrane is not in the hot zone and there is
temperature gradient along side with the membrane (Figure 5-1). This might cause mechanical
and chemical stability problem. Carbon deposition, for instance, is favorable at low temperature
in the WGS reaction, SRM and CDRM. Therefore, carbon deposition is more likely to happen in
that part of membranes outside the hot zone where the temperature is lower. In addition,
carbonate can be formed at low temperature which is detrimental to the chemical stability of the
membranes. One concern for the experimental setup change proposed above is the sealing. One
possible solution is sealing with metal rings with mechanic strings applying force on top.
159
Figure 10-1. A SrCe0.7Zr0.2Eu0.1O3-δ thin film membrane coated on graphite-SrCe0.8Zr0.2O3-δ substrate.
160
LIST OF REFERENCES
[1] N.Z. Muradov and T.N. Veziroglu, International Journal of Hydrogen Energy 30 (2005) (3), 225.
[2] S. Tosti, International Journal of Hydrogen Energy 28 (2003) (12), 1445.
[3] T. Oh, H. Yoon and E.D. Wachsman, Solid State Ionics submitted.
[4] T. Oh, Optimization of europium dopant concentration in strontium cerate mixed-conducting ceramic membrane for maximum hydrogen production, Ph. D dissertation, University of Florida, Gainesville, FL (2008), 84.
[5] S.J. Song, E.D. Wachsman, J. Rhodes, S.E. Dorris and U. Balachandran, Solid State Ionics 167 (2004) (1-2), 99.
[6] H. Yoon, T. Oh, J. Li, K. Duncan and E.D. Wachsman, Journal of the Electrochemical Society 156 (2009) (7), B791.
[7] T. Oh, J. Li, H. Yoon and E.D. Wachsman, Ionics Accepted.
[8] K. Katahira, Y. Kohchi, T. Shimura and H. Iwahara, Solid State Ionics 138 (2000) (1-2), 91.
[9] Z. Zhong, Solid State Ionics 178 (2007) (3-4), 213.
[10] K.H. Ryu and S.M. Haile, Solid State Ionics 125 (1999) (1-4), 355.
[11] C. Zuo, S. Zha, M. Liu, M. Hatano and M. Uchiyama, Advanced Materials 18 (2006) (24), 3318.
[12] L. Jingde, L. Wang, L. Fan, Y. Li, L. Dai and H. Guo, Journal of Rare Earths 26 (2008) (4), 505.
[13] J. Zhang, Z. Wen, J. Han, J. Wu, S. Huang and X. Zhu, Journal of Alloys and Compounds 440 (2007) (1-2), 270.
[14] S.C.A. KluitersECN-C--04-102, Petten. (2004).
[15] J. Shu, B.P.A. Grandjean and S. Kaliaguine, Catalysis Today 25 (1995) (3-4), 327.
[16] S. Tosti, A. Basile, G. Chiappetta, C. Rizzello and V. Violante, Chemical Engineering Journal 93 (2003) (1), 23.
[17] B.H. Howard, R.P. Killmeyer, K.S. Rothenberger, A.V. Cugini, B.D. Morreale, R.M. Enick and F. Bustamante, Journal of Membrane Science 241 (2004) (2), 207.
[18] A. Basile, E. Drioli, F. Santella, V. Violante, G. Capanneli and G. Vitulli, Gas separation & purification 10 (1996) (1), 53.
161
[19] J.W. Phair and S.P.S. Badwal, Ionics 12 (2006), 103.
[20] H. Iwahara, Solid State Ionics 86-88 (1996) (Part 1), 9.
[21] H. Iwahara, Solid State Ionics 125 (1999) (1-4), 271.
[22] H. Iwahara, T. Esaka, H. Uchida and N. Maeda, Solid State Ionics 3-4 (1981), 359.
[23] T. Hibino, K. Mizutani, T. Yajima and H. Iwahara, Solid State Ionics 57 (1992) (3-4), 303.
[24] D.A. Stevenson, N. Jiang, R.M. Buchanan and F.E.G. Henn, Solid State Ionics 62 (1993) (3-4), 279.
[25] S. Song, E.D. Wachsman, S.E. Dorris and U. Balachandran, In: E.D. Wachsman, K.S. Lyons, M. Carolyn, F. Garzon and M. Liu, Editors, Solid State Ionic Devices III, Electrochem. Soc. (2003), 456.
[26] H. Iwahara, H. Uchida, K. Ogaki and H. Nagato, Journal of The Electrochemical Society 138 (1991) (1), 295.
[27] S.J. Song, E.D. Wachsman, S.E. Dorris and U. Balachandran, Journal of The Electrochemical Society 150 (2003) (11), A1484.
[28] S.J. Song, E.D. Wachsman, S.E. Dorris and U. Balachandran, Journal of The Electrochemical Society 150 (2003) (6), A790.
[29] S.-J. Song, E.D. Wachsman, S.E. Dorris and U. Balachandran, Solid State Ionics 149 (2002) (1-2), 1.
[30] S.J. Song, E.D. Wachsman, J. Rhodes, S.E. Dorris and U. Balachandran, Solid State Ionics 164 (2003) (1-2), 107.
[31] S.J. Song, T.H. Lee, E.D. Wachsman, L. Chen, S.E. Dorris and U. Balachandran, Journal of The Electrochemical Society 152 (2005) (11), J125.
[32] G. Zhang, S.E. Dorris, U. Balachandran and M. Liu, Solid State Ionics 159 (2003) (1-2), 121.
[33] R.J. Phillips, N. Bonanos, F.W. Poulsen and E.O. Ahlgren, Solid State Ionics 125 (1999) (1-4), 389.
[34] S.J. Song, E.D. Wachsman, J. Rhodes, H.S. Yoon, K.H. Lee, G. Zhang, S.E. Dorris and U. Balachandran, Journal of Materials Science 40 (2005) (15), 4061.
[35] A.S. Nowick, Y. Du and K.C. Liang, Solid State Ionics 125 (1999) (1-4), 303.
[36] S. Tao and J.T.S. Irvine, Solid State Ionics 154-155 (2002), 659.
[37] K.D. Kreuer, Annual Review of Materials Research 33 (2003) (1), 333.
162
[38] N. Bonano, B. Ellis and M.N. Mahmood, Solid State Ionics 44 (1991) (3-4), 305.
[39] H. Iwahara, T. Mori and T. Hibino, Solid State Ionics 79 (1995), 177.
[40] A.V. Kuzmin, V.P. Gorelov, B.T. Melekh, M. Glerup and F.W. Poulsen, Solid State Ionics 162-163 (2003), 13.
[41] J. Ranlov and K. Nielsen, Journal of Materials Chemistry 4 (1994) (6), 867.
[42] K.S. Knight and N. Bonanos, Materials Research Bulletin 30 (1995) (3), 347.
[43] N. Bonanos, K.S. Knight and B. Ellis, Solid State Ionics 79 (1995), 161.
[44] S. Yamanaka, K. Kurosaki, T. Matsuda and S.-i. Kobayashi, Journal of Alloys and Compounds 352 (2003) (1-2), 52.
[45] J. Ranlov, B. Lebech and K. Nielsen, Journal of Materials Chemistry 5 (1995) (5), 743.
[46] R. Pankajavalli, K. Ananthasivan, S. Anthonysamy and P.R. Vasudeva Rao, Journal of Nuclear Materials 336 (2005) (2-3), 177.
[47] X. Qi and Y.S. Lin, Solid State Ionics 120 (1999), 85.
[48] T. Oh, H. Yoon and E.D. Wachsman, journal of the Electrochemical Society submitted.
[49] H. Iwahara, T. Esaka, H. Uchida, T. Yamauchi and K. Ogaki, Solid State Ionics 18-19 (1986) (Part 2), 1003.
[50] T. Higuchi, T. Tsukamoto, N. Sata, S. Yamaguchi, S. Shin and T. Hattori, Solid State Ionics 176 (2005) (39-40), 2963.
[51] S. Yamanaka, K. Kurosaki, T. Oyama, H. muta and M. Uno, Journal of the American Ceramic Society 88 (2005) (6), 1496.
[52] K.-D. Kreuer, A. Rabenau and W. Weppner, Angewandte Chemie International Edition in English 21 (1982) (3), 208.
[53] K.D. Kreuer, E. Schonherr and J. Maier, Solid State Ionics 70-71 (1994) (Part 1), 278.
[54] K.C. Liang, Y. Du and A.S. Nowick, Solid State Ionics 69 (1994) (2), 117.
[55] K.D. Kreuer, A. Fuchs and J. Maier, Solid State Ionics 77 (1995), 157.
[56] T. Scherban, R. Villeneuve, L. Abello and G. Lucazeau, Solid State Ionics 61 (1993) (1-3), 93.
[57] E. Matsushita and A. Tanase, Solid State Ionics 97 (1997) (1-4), 45.
[58] W. Munch, G. Seifert, K.D. Kreuer and J. Maier, Solid State Ionics 97 (1997) (1-4), 39.
163
[59] W. Munch, G. Seifert, K.D. Kreuer and J. Maier, Solid State Ionics 86-88 (1996) (Part 1), 647.
[60] K.D. Kreuer, Solid State Ionics 125 (1999) (1-4), 285.
[61] K.D. Kreuer, E. Schonherr and J. Maier, Phase separation and grain boundary proton conductivity in BaCeO3 based ceramics, 14th Riso International Society Materials Science (1993), 297.
[62] W. Munch, K.D. Kreuer, S. Adams, G. Seifert and J. Maier, Phase Transitions 68 (1999), 567.
[63] H. Rickert, Electrochemistry of solids. An introduction, Springer Verlag, Berlin-Heidelberg-New York (1982).
[64] T. Norby and Y. Larring, Solid State Ionics 136-137 (2000), 139.
[65] X. Qi, S. Cheng and Y.S. Lin, Modeling and experimental study of hydrogen permeaiton through proton conducting ceramic membranes, In: W.L.W. A. Ramanarayanan, and M. Mogensen, Editor, Ionic and Mixed Conducting Ceramics, Electrochemicl Society Proceeding Series, Salt Lake City, Utah (2002), 1.
[66] X. Qi and Y.S. Lin, Solid State Ionics 130 (2000) (1-2), 149.
[67] H.J. Bouwmeester and A.J. Burggraaf, Fundamentals of inorganic membrane science and technology, Elsevier Science, Amsterdam (1996).
[68] D.R. Noble and S.A. Stern, Membrane swparations technology-principles and applications, Elsevier Sience B. V., Amsterdam (1995).
[69] M. Mulder, Basic principles of membrane technology, Kluwer Academic Publishers, Dordrecht (2000).
[70] H. Iwahara, T. Yajima and H. Ushida, Solid State Ionics 70-71 (1994) (Part 1), 267.
[71] J. Guan, S.E. Dorris, U. Balachandran and M. Liu, Solid State Ionics 100 (1997) (1-2), 45.
[72] J. Guan, S.E. Dorris, U. Balachandran and L. Meilin, Solid State Ionics 110 (1998) (3-4), 303.
[73] S. Hamakawa, L. Li, A. Li and E. Iglesia, Solid State Ionics 148 (2002) (1-2), 71.
[74] H. Yoon, T. Oh, J. Li, K. Duncan and E.D. Wachsman, Journal of the American Ceramic Society Accepted.
[75] M.N. Rahaman, Ceramic processing and sintering, Marcel Dekker, Inc., New York (1995).
[76] I. Wender, Fuel Processing Technology 48 (1996), 189.
164
[77] Y. Li, Q. Fu and M. Flytzani-Stephanopoulos, Applied Catalysis B: Environmental 27 (2000) (3), 179.
[78] C. Rhodes, G.J. Hutchings and A.M. Ward, Catalysis Today 23 (1995) (1), 43.
[79] J. Pasel, R.C. Samsun, D. Schmitt, R. Peters and D. Stolten, Journal of Power Sources 152 (2005), 189.
[80] R.M. Enick, J. Hill, A.V. Cugini, K.S. Rothenberger and H.G. McIlvried, ACS Prepr. Fuel Chem. Div. 44 (1999), 919.
[81] C.A. Udovich, Studies in Surface Science and Catalysis 119 (1998), 417.
[82] A.K. Garg and L.C.D. Jonghe, Journal of Material Science 28 (1993) (13), 3427.
[83] D. Lee and S.T. Oyama, Journal of Membrane Science 210 (2002) (2), 291.
[84] F. Bustamante, R.M. Enick, R.P. Killmeyer, B.H. Howard, K.S. Rothenberger, A.V. Cugini, B.D. Morreale and M.V. Ciocco, AIChE Journal 51 (2005) (5), 1440.
[85] A. Basile, A. Criscuoli, F. Santella and E. Drioli, Gas Separation & Purification 10 (1996) (4), 243.
[86] J. Li, H. Yoon and E.D. Wachsman, Journal of Membrane Science to be submitted.
[87] J. Li, H. Yoon, T. Oh and E.D. Wachsman, Applied Catalysis B: Environmental submitted.
[88] B. Sundman, B. Jansson and J.-O. Andersson, Calphad 9 (1985), 153.
[89] E. Xue, M. O'Keeffe and J.R.H. Ross, Catalysis Today 30 (1996) (1-3), 107.
[90] T. Yajima, H. Kazeoka, T. Yogo and H. Iwahara, Solid State Ionics 47 (1991) (3-4), 271.
[91] T. Yajima, H. Suzuki, T. Yogo and H. Iwahara, Solid State Ionics 51 (1992) (1-2), 101.
[92] A.N. Shirsat, K.N.G. Kaimal, S.R. Bharadwaj and D. Das, Journal of Solid State Chemistry 177 (2004) (6), 2007.
[93] M.J. Scholten, J. Schoonman, J.C. van Miltenburg and H.A.J. Oonk, Solid State Ionics 61 (1993) (1-3), 83.
[94] G. Srikanth and V.V. Anil, Journal of The Electrochemical Society 140 (1993) (4), 1060.
[95] K.D. Kreuer, Solid State Ionics 97 (1997) (1-4), 1.
[96] W.T. Cameron and V.V. Anil, Journal of The Electrochemical Society 143 (1996) (4), 1386.
[97] D.C. Mather and J.R. Jurado, Bol. Soc. Esp, Ceram, Vidrio 42 (2003) (5), 331.
165
[98] X. Wei and Y.S. Lin, Solid State Ionics 178 (2008) (35-36), 1804.
[99] K.D. Kreuer, S. Adams, W. Munch, A. Fuchs, U. Klock and J. Maier, Solid State Ionics 145 (2001) (1-4), 295.
[100]V.M. Goldschmidt, T. Barth, G. Lunde and W. Zachariasen, Skrifter Norske Videnskaps-Akad. Oslo, Mat.-Nat. Kl. (1926), 117.
[101]W. Meng and A.V. Virkar, Journal of Solid State Chemistry 148 (1999) (2), 492.
[102]S.A. Gramsch and L.R. Morss, The Journal of Chemical Thermodynamics 27 (1995) (5), 551.
[103]V.B. Sanjeevani and V.V. Anil, Journal of The Electrochemical Society 146 (1999) (12), 4386.
[104]R.D. Shannon, Acta Crystallographica A32 (1976), 751.
[105]H. Yokokawa, N. Sakai, T. Kawada and M. Dokiya, Journal of Solid State Chemistry 94 (1991) (1), 106.
[106]G.C. Mather, F.M. Figueiredo, D.P. Fagg, T. Norby, J.R. Jurado and J.R. Frade, Solid State Ionics 158 (2003) (3-4), 333.
[107]J.M. Rhodes and E.D. Wachsman, In: E.D. Wachsman, W. Weppner, E. Traversa, M. Liu, P. Vanysek and N. Yamazoe, Editors, Solid State Ionics Devices II-Ceramic Sensors, The Electrochemical Society Proceedings Series, Pennington, NJ (2001), 137.
[108]J. Li, H. Yoon, T. Oh and E.D. Wachsman, Journal of the Electrochemical Society submitted.
[109]J. Li, H. Yoon, T. Oh and E.D. Wachsman, Stability of Zr-doped SrCeO3-[delta] under wet CO/CO2 atmospheres, In: E.D. Wachsman, F.H. Garzon, E. Traversa, R. Mukundan and A. Manivannan, Editors, Solid State Ionic Devices V, ECS Transactions, Washington DC (2008), 81.
[110]J. Tong, Y. Matsumura, H. Suda and K. Haraya, Separation and Purification Technology 46 (2005) (1-2), 1.
[111]A. Li, W. Liang and R. Hughes, Journal of Membrane Science 149 (1998) (2), 259.
[112]T. Oh, H. Yoon, J. Li and E.D. Wachsman, Journal of Membrane Science submitted.
[113]S. Takenaka, T. Shimizu and K. Otsuka, International Journal of Hydrogen Energy 29 (2004) (10), 1065.
[114]J. Sehested, S. Dahl, J. Jacobsen and J.R. Rostrup-Nielsen, The Journal of Physical Chemistry B 109 (2005) (6), 2432.
166
[115]F.A.N. Fernandes and A.B.S. Jr., Latin American Applied Research 36 (2006), 155.
[116]G. Barbieri, V. Violante, F.P.D. Maio, A. Criscuoli and E. Drioli, Industrial & Engineering Chemistry Research 36 (1997), 3369.
[117]J.R. Rostrup-Nielsen, Catalysis Today 18 (1993) (4), 305.
[118]E.M. Assaf, C.D.F. Jesus and J.M. Assaf, Brazilian Journal of Chemical Engineering 15 (1998), 160.
[119]M. Oertel, J. Schmitz, W. Weirich, D. Jendryssek-Neumann and R. Schulten, Chemical Engineering & Technology 10 (1987) (1), 248.
[120] J. Shu, B.P.A. Grandjean and S. Kaliaguine, Applied Catalysis A: General 119 (1994) (2), 305.
[121]Y.-M. Lin, S.-L. Liu, C.-H. Chuang and Y.-T. Chu, Catalysis Today 82 (2003) (1-4), 127.
[122]G. Barbieri and F.P.D. Maio, Industrial & Engineering Chemistry Research 36 (1997) (6), 2121.
[123]A. Bottino, A. Comite, G. Capannelli, R. Di Felice and P. Pinacci, Catalysis Today 118-222 (2006) (1-2), 214.
[124]S. Hara, K. Haraya, G. Barbieri and E. Drioli, Desalination 233 (2008) (1-3), 359.
[125]S. Uemiya, N. Sato, H. Ando, T. Matsuda and E. Kikuchi, Applied Catalysis 67 (1991), 223.
[126] D.-W. Lee, Y.-G. Lee, S.-E. Nam, S.-K. Ihm and K.-H. Lee, Journal of Membrane Science 220 (2003) (1-2), 137.
[127]D.R. Gaskell, Introduction to the thermodynamics of materials, 4th edition, Taylor & Francis, New York (2003).
[128]Z.-W. Liu, K.-W. Jun, H.-S. Roh and S.-E. Park, Journal of Power Sources 111 (2002) (2), 283.
[129]S. Rakass, H. Oudghiri-Hassani, P. Rowntree and N. Abatzoglou, Journal of Power Sources 158 (2006) (1), 485.
[130]T. Oh, Optimization of Europioum dopant concentration in strontium cerate mixed-conducting ceramic membrane for maximum hydrogen production, University of Florida, Gainesville (2008).
[131]S. P and m. M, Technical report NREL/TP-570-27637, NREL (2000).
167
[132]H. Adus, O. Kaarstad and M. Kowal, Decarbonization of fossil fuels: hydrogen as an energy carrier, Proceedings of the 11th World Hydrogen Energy Conference, Stuttgart, Germany (1996), 525.
[133]K. Blok, R. Williams, R. Katofsky and C. Hendriks, Energy 22 (1997), 161.
[134]J.R.H. Ross, A.N.J. van Keulen, M.E.S. Hegarty and K. Seshan, Catalysis Today 30 (1996) (1-3), 193.
[135]S. Wang and G.Q. Lu., Energy & Fuels 10 (1996), 896.
[136]I. Wender, Fuel Processing Technology 48 (1996) (3), 189.
[137]X. Xu and J.A. Moulijn, Energy & Fuels 10 (1996), 305.
[138]A. Valentini, N.L.V. Carreno, L.F.D. Probst, P.N. Lisboa-Filho, W.H. Schreiner, E.R. Leite and E. Longo, Applied Catalysis A: General 255 (2003) (2), 211.
[139]R. Martinez, E. Romero, C. Guimon and R. Bilbao, Applied Catalysis A: General 274 (2004) (1-2), 139.
[140] M.d.M.V.M. Souza, L. Clav, V. Dubois, C.A.C. Perez and M. Schmal, Applied Catalysis A: General 272 (2004) (1-2), 133.
[141] W. Nimwattanakul, A. Luengnaruemitchai and S. Jitkarnka, International Journal of Hydrogen Energy 31 (2006) (1), 93.
[142]S. Therdthianwong, A. Therdthianwong, C. Siangchin and S. Yongprapat, International Journal of Hydrogen Energy 33 (2008) (3), 991.
[143]X.E. Verykios, International Journal of Hydrogen Energy 28 (2003) (10), 1045.
[144]H. Le, L.L. Lobban and R.G. Mallinson, Catalysis Today 89 (2004) (1-2), 15.
[145] Y. Li, G.-h. Xu, C.-j. Liu, B. Eliasson and B.-z. Xue, Energy & Fuels 15 (2001) (2), 299.
[146] Y. Li, C.-J. Liu, B. Eliasson and Y. Wang, Energy & Fuels 16 (2002) (4), 864.
[147]B. Fidalgo, A. Dominguez, J.J. Pis and J.A. Menendez, International Journal of Hydrogen Energy 33 (2008) (16), 4337.
[148]Y. Li, Y. Wang, X. Zhang and Z. Mi, International Journal of Hydrogen Energy 33 (2008) (10), 2507.
[149].R. Rostrupnielsen and J.H.B. Hansen, Journal of Catalysis 144 (1993) (1), 38.
[150] X. Chen, K. Honda and Z.-G. Zhang, Applied Catalysis A: General 279 (2005) (1-2), 263.
[151] D. Liu, R. Lau, A. Borgna and Y. Yang, Applied Catalysis A: General 358 (2009) (2), 110.
168
[152]M. Rezaei, S.M. Alavi, S. Sahebdelfar, P. Bai, X. Liu and Z.-F. Yan, Applied Catalysis B: Environmental 77 (2008) (3-4), 346.
[153] D.A. Slade, A.M. Duncan, K.J. Nordheden and S.M. Stagg-Williams, Green Chemistry 9 (2007) (6), 577.
[154] T.M. Rayhold and M.C. Huff, AIChE Journal 48 (2002) (5), 1051.
[155] B.S. Liu and C.T. Au, Catalysis Letters 55 (2001) (1-3), 67.
[156] J. Li, H. Yoon and E.D. Wachsman, Applied Catalysis A: General to be submitted.
[157] A. Erdohelyi, J. Cserenyi, E. Papp and F. Solymosi, Applied Catalysis A: General 108 (1994) (2), 205.
[158] Y. Sakai, H. Saito, T. Sodesawa and F. Nozaki, Reaction Kinetics and Catalysis Letters 24 (1984) (3-4), 253.
[159] A.M. Gadalla and M.E. Sommer, Chemical Engineering Science 44 (1989), 2815.
[160] G.S. Gallego, F. Mondragon, J. Barrault, J.-M. Tatibouet and C. Batiot-Dupeyrat, Applied Catalysis A: General 311 (2006), 164.
[161] J. Wei and E. Iglesia, Journal of Catalysis 224 (2004) (2), 370.
[162] D.K. Kim, K. Stowe, F. Muller and W.F. Maier, Journal of Catalysis 247 (2007) (1), 101.
[163] D. Li, X. Li, M. Bai, X. Tao, S. Shang, X. Dai and Y. Yin, International Journal of Hydrogen Energy 34 (2009) (1), 308.
[164]A. Sacco, F.W.A.H. Geurts, G.A. Jablonski, S. Lee and R.A. Gately, Journal of Catalysis 119 (1989) (2), 322.
169
BIOGRAPHICAL SKETCH
Jianlin Li was born in a small village in Zhaoqing, China. As early as he could remember,
Jianlin always wanted to attend elementary school with his sister. In 1984, he couldn’t wait any
longer. He registered at Fuluo Elementary School and paid the tuition fee with the money, which
he borrowed from his grandmother. In 1991, Jianlin attended the Xinqiao Middle School where
he spent six years and living by himself until he graduated high school. Jianlin excelled in
mathematics, physics and chemistry. Jianlin was ranked 2nd in mathematics among the three
hundred thousand students taking the college entrance exam.
In 1997, Jianlin attended the University of Science and Technology of China in Hefei
which is about 1000 miles away from his hometown. It was his first time taking a train. He
couldn’t forget the 26-hour experience he had on the crowded train for the rest of his life. He was
extremely excited when he finally arrived after such a long trip. Jianlin lived by himself and only
went home once a year during the winter holidays. Another challenge was having to speak
Mandarin when your native language is Cantonese. Jianlin exposed him to different disciplines.
He took classes in mathematics, physics, chemistry, computer science, electronic engineering
and international trade. He received his dual bachelor’s degrees in materials chemistry and
electronic information engineering in June 2001. Then, he found himself more interested in
materials science and engineering. He pursued his master’s degree under the supervision of Dr.
Chusheng Chen and senior engineer Pinghua Yang. He received his Master of Engineering
degree in materials science in June 2004. During his master studies, he worked on hydrogen
separation using mesoporous silicon membranes and syngas production through partial oxidation
of methane using oxygen permeable membranes such as Ba0.5Sr0.5Co0.8Fe0.2O3-δ.
170
In August 2004, Jianlin joined Dr. Eric Wachsman’s group at the Department of Materials
Science and Engineering at the University of Florida.
It didn’t take long for Jianlin get used to the life in Gainesville, however, the language
barrier was a challenge at times. The last thing he wanted to do was to place an order by phone.
He really struggled with it for the first year.
Jianlin fell in love with college sports quickly, especially football. He had never seen a
football game before coming to America, but it only took him one season to become a football
fan. Jianlin felt lucky to witness the Gators win four national championships in basketball and
football in the last four years. He enjoyed the awesome experience cheering for the Gators in the
O’Dome and SWAMP. He also loved the experience of marching down University Avenue
every time the Gators were crowned with a national championship.
Jianlin was very active in student organizations. He was a two-term senator in the Student
Government at the University of Florida. He was the president of the Friendship Association of
Chinese Students and Scholars (FACSS) in 2005 and served as a consultant the following three
years. As president, Jianlin managed to get $15,000 from 14 sponsors to organize the 2006
FACSS Chinese New Year Show at the Phillips Center of Performing Arts. This is the biggest
event he ever organized and he felt proud of himself. In 2006, he was awarded the President of
the Year by their parent organization, the Volunteers for International Students Affairs (VISA).
Jianlin represented the University of Florida at the 6th Annual Mayor’s Summit in Tallahassee in
2006 and the 1st Florida International Leadership Conference in Ocala in 2007.
Jianlin enjoyed working with the group members in Dr. Wachsman’s group and
appreciated all the valuable discussion and comments from them. He received his Ph.D. from the
University of Florida in the fall of 2009.