Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
A study of the application of membrane-based reactive separation to the carbon dioxide methanation
(USC Thesis Other)
A study of the application of membrane-based reactive separation to the carbon dioxide methanation
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
A STUDY OF THE APPLICATION OF MEMBRANE-BASED REACTIVE
SEPARATIONS TO THE CARBON DIOXIDE METHANATION
by
Hyun Tae Hwang
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMICAL ENGINEERING)
May 2009
Copyright 2009 Hyun Tae Hwang
ii
Dedication
I dedicate this thesis to
my late father, who has supported me all the way
my mother, who offered me unconditional love
my wife, Soo-zin, who did more than her share around the house
and, my lovely daughter, Chae-eun.
It is also dedicated to my mother-in law, my father-in law and my family.
iii
Acknowledgements
I would like to express my profound gratitude, admiration and respect to my
advisors, Professor Muhammad Sahimi and Professor Theodore T. Tsotsis for their
guidance, encouragement and direction throughout this research work. This dissertation is
a result of their help, support and patience. Also, I would also like to thank Professor
Fokion Egolfopoulos for serving on my dissertation committee and for his helpful advice
during the period that the work for this dissertation was being done.
Thanks are due to the administrative staff of the Mork Family Department of
Chemical Engineering and Materials Science for all their help and support throughout my
graduate studies, and to all the fellow graduate students in our research group for their
help and fruitful discussions. Especially, I want to thank Aadesh Harale and Mitra
Abdollahi.
Finally, I would like to thank the members of my family for their support and
unconditional love, especially my wife and my daughter for patience, encouragement and
love. This dissertation would not have been possible without the support and love of my
family.
iv
Table of Contents
Dedication..................................................................................................................... ii
Acknowledgements...................................................................................................... iii
List of Tables ............................................................................................................... vi
List of Figures............................................................................................................. vii
Abstract........................................................................................................................xi
Chapter 1 : Introduction and Overview .........................................................................1
1.1 Motivation and background..............................................................................1
1.2 Removal and recovery of CO
2
..........................................................................6
1.3 Scope of the work ...........................................................................................10
Chapter 2 : Membrane Studies.....................................................................................12
2.1 Introduction.....................................................................................................12
2.2 Overview on membrane separation ................................................................12
2.3 Inorganic membranes......................................................................................14
2.4 Membrane performance evaluation ................................................................21
Chapter 3 : Reaction Studies........................................................................................46
3.1 Introduction.....................................................................................................46
3.2 Thermodynamics.............................................................................................46
3.3 Experimental setup..........................................................................................53
3.4 Catalyst...........................................................................................................54
3.5 Experimental reaction kinetics........................................................................55
Chapter 4 : Mathematical Modeling ............................................................................67
4.1 Introduction.....................................................................................................67
4.2 Isothermal modeling.......................................................................................67
4.3 Nonisothermal modeling.................................................................................70
Chapter 5 : Results and Conclusions ...........................................................................76
5.1 Introduction.....................................................................................................76
5.2 MR-Based Air Revitalization System (MARS)..............................................76
5.3 MR experiments (MARS)...............................................................................83
5.4 Simulation results (MARS).............................................................................87
5.5 Sumary and conclusions (MARS) ................................................................101
5.6 MR-Based In-Situ Resource Utilization System (MISRUS)........................102
v
5.7 Packed-bed reactor experiments (MISRUS).................................................105
5.8 Simulations with supported Pd membranes (MISRUS) ...............................109
5.9 Simulations with CMS membranes (MISRUS)............................................120
5.10 Sumary and conclusions (MISRUS).............................................................130
Nomenclature.............................................................................................................131
Bibliography ..............................................................................................................135
vi
List of Tables
Table 2.1: Physical properties and experimental conditions investigated .................. 25
Table 2.2: Single and mixed-gas permeance data for CMS membrane # 1................ 28
Table 2.3: Single and mixed-gas permeance data for CMS membrane # 2................ 29
Table 2.4: H
2
-CO
2
binary gas permeance data for CMS membrane # 3 .................... 30
Table 2.5: Comparison of permeance data for the feed side vs. the tube side
(CMS membrane # 4)................................................................................ 33
Table 2.6: Range of measured H
2
permeance and separation factors......................... 41
Table 2.7: Summary of published data for H2 permeability of Pd membranes ......... 45
Table 3.1: Hougen-Watson type model for the Sabatier reaction............................... 49
Table 5.1: Single and mixed-gas permeance data for CMS membrane # 7................ 86
Table 5.2: Base case and the range of conditions used in the simulations (MARS) .. 89
Table 5.3: Base case of conditions used in the simulations (MISRUS) ................... 111
Table 5.4: Base case of conditions used in the simulations (CMSM) ...................... 121
vii
List of Figures
Figure 1.1: A conventional CO
2
methanation system for ISRU................................... 2
Figure 1.2: A conventional ARS without a CO
2
reduction subsystem......................... 4
Figure 2.1: Transport mechanisms through microporous membranes........................ 17
Figure 2.2: Schematic of membrane reactor module.................................................. 22
Figure 2.3: Hydrogen permeance as a function of H
2
/CO
2
mol ratio in feed
(CMSM # 3).............................................................................................. 31
Figure 2.4: Hydrogen permeance as a function of H
2
/CO
2
mol ratio in feed
(CMSM # 4, # 5 and # 6) .......................................................................... 32
Figure 2.5: H
2
/CO (single gas) vs. H
2
permeance ...................................................... 34
Figure 2.6: H
2
/CO
2
(single gas) vs. H
2
permeance ..................................................... 35
Figure 2.7: H
2
/CH
4
(single gas) vs. H
2
permeance ..................................................... 36
Figure 2.8: H
2
/N
2
(single gas) vs. H
2
permeance........................................................ 37
Figure 2.9: H
2
/CO (mixed gas) vs. H
2
permeance...................................................... 38
Figure 2.10: H
2
/CO
2
(mixed gas) vs. H
2
permeance................................................... 39
Figure 2.11: H
2
/CH
4
(mixed gas) vs. H
2
permeance................................................... 40
Figure 2.12: H
2
/H
2
O (mixed gas) vs. H
2
permeance .................................................. 41
Figure 2.13: Flux of hydrogen as a function of () ( )
Fn P n
PP − ................................. 44
Figure 3.1: Thermodynamic equilibrium conversion and the enthalpy for the
Sabatier reaction (1 atm, H
2
/CO
2
= 4.0) ................................................... 47
Figure 3.2: Schematic of possible functions of a membrane in a reactor................... 52
Figure 3.3: Schematic of the experimental set-up ...................................................... 54
viii
Figure 3.4: CO
2
conversion as a function of W
c
/n
CO2
for various temperatures:
Range I (H
2
/CO
2
= 5.0 and P =1atm)........................................................ 57
Figure 3.5: CO
2
conversion as a function of W
c
/n
CO2
for various ratios of H
2
to CO
2
(T = 250
o
C and P = 1 atm)....................................................................... 58
Figure 3.6: CO
2
conversion as a function of W
c
/n
CO2
for various reactor pressures
(T = 250
o
C and H
2
/CO
2
= 5.0)................................................................. 59
Figure 3.7: Effect of O
2
concentrations in the feed on the catalyst activity
(H
2
/CO
2
= 4.0, P = 1 atm, W
c
/n
CO2
at 250
o
C = 38.9 g·h/mol and W
c
/n
CO2
at 300
o
C = 15.6 g·h/mol).......................................................................... 60
Figure 3.8: CO
2
conversion with 2 % of O
2
in the feed as a function of time on
stream (H
2
/CO
2
= 4.0, P = 1 atm, W
c
/n
CO2
at 250
o
C = 38.9 g·h/mol and
W
c
/n
CO2
at 300
o
C = 15.6 g·h/mol) ............................................................ 61
Figure 3.9: CO
2
conversion as a function of W
c
/n
CO2
for various temperatures:
Range II (H
2
/CO
2
= 4.0 and P =1atm) ...................................................... 64
Figure 3.10: CO
2
conversion as a function of W
c
/n
CO2
for various temperatures:
Range II (H
2
/CO
2
= 5.0 and P =1atm) .................................................... 65
Figure 3.11: Arrhenius plots: Range I and II .............................................................. 66
Figure 4.1: Schematic for the AR-MR system............................................................ 68
Figure 5.1: An ARS with a CO
2
reduction subsystem................................................ 79
Figure 5.2: An ARS with membrane reactor and hydrogen recovery system ............ 80
Figure 5.3: A schematic of the proposed MARS........................................................ 81
Figure 5.4: Effect of CO
2
concentration on its conversion
(W
c
/n
CO2
= 405 g·h/mol, S (H
2
) = 0.13, P
F
= P
P
= 1 atm).......................... 84
Figure 5.5: Effect of the pressure in the air feed side on CO
2
conversion
(W
c
/n
CO2
= 405 g·h/mol, S (H
2
) = 0.13, CO
2
% = 10, P
P
= 1 atm) ............ 85
Figure 5.6: A schematic of the MR-AR system.......................................................... 88
Figure 5.7: Effect of CO
2
permeance
(Symbol: membrane surface area, line: amount of catalyst)..................... 91
ix
Figure 5.8: Effect of separation factor of CO
2
/H
2
(T = 250 (
o
C), 5,000 CO
2
ppm, Ft = 50 (L/sec))
(Solid: amount of catalyst, hollow: surface area of membranes).............. 93
Figure 5.9: Effect of separation factor of CO
2
/H
2
(T = 250 (
o
C), 2,000 CO
2
ppm, Ft = 50 (L/sec))
(Solid: amount of catalyst, hollow: surface area of membranes).............. 94
Figure 5.10: Effect of the temperature in the reactor
(Symbol: membrane surface area, line: amount of catalyst)................... 96
Figure 5.11: Effect of target CO
2
concentration in the cabin
(Symbol: membrane surface area, line: amount of catalyst)................... 98
Figure 5.12: Effect of the air feed rate to the reactor
(Solid: amount of catalyst, hollow: surface area of membranes).......... 100
Figure 5.13: An ISRU system with a membrane reactor.......................................... 103
Figure 5.14: A schematic of the MR-ISRU system .................................................. 104
Figure 5.15: A cross section diagram of the reactor and insulations........................ 106
Figure 5.16: Temperature profiles with dimensionless distance from the inlet for
various W
c
/n
CO2
(T
o
= 250
o
C, H
2
/CO
2
= 4.0, P = 1 atm)...................... 108
Figure 5.17: CO
2
conversion as a function of W
c
/n
CO2
(T
o
= 250
o
C, H
2
/CO
2
= 4.0, P = 1 atm)................................................ 109
Figure 5.18: Effect of sweep ratio (S) (a) CO
2
conversion as a function of Da
(b) H
2
exiting the feed side (c) dimensionless temperature ( Г
F
) along
the length of the reactor (Solid lines correspond to MR, and dotted
line to PFR)........................................................................................... 113
Figure 5.19: Effect of Da·Pe (a) CO
2
conversion as a function of Da (b) H
2
exiting the feed side (c) dimensionless temperature ( Г
F
) along the
length of the reactor (Solid lines correspond to MR, and dotted line to
PFR) ...................................................................................................... 115
Figure 5.20: Effect of Г
o
F
(inlet feed side temperature) (a) CO
2
conversion as a
function of Da (b) H
2
exiting the feed side (c) dimensionless
temperature ( Г
F
) along the length of the reactor (Solid lines
correspond to MR, and dotted line to PFR) .......................................... 117
x
Figure 5.21: Effect of Г
o
P
(inlet permeate side temperature) (a) CO
2
conversion
as a function of Da (b) H
2
exiting the feed side (c) dimensionless
temperature ( Г
F
) along the length of the reactor (Solid lines
correspond to MR, and dotted line to PFR) .......................................... 119
Figure 5.22: Effect of feed configuration (a) CO
2
conversion as a function of Da
(b) H
2
exiting the feed side (c) dimensionless temperature ( Г
F
) along
the length of the reactor (Solid lines correspond to MR, and dotted
line to PFR)........................................................................................... 123
Figure 5.23: Effect of Г
o
P
for Г
o
F
= 0.80 (a) CO
2
conversion as a function of Da
(b) H
2
exiting the feed side (c) dimensionless temperature ( Г
F
) along
the length of the reactor (Solid lines correspond to MR, and dotted
line to PFR)........................................................................................... 125
Figure 5.24: Effect of Г
o
P
for Г
o
F
= 1.00 (a) CO
2
conversion as a function of Da
(b) H
2
exiting the feed side (c) dimensionless temperature ( Г
F
) along
the length of the reactor (Solid lines correspond to MR, and dotted
line to PFR)........................................................................................... 126
Figure 5.25: Effect of Г
o
P
for Г
o
F
= 1.20 (a) CO
2
conversion as a function of Da
(b) H
2
exiting the feed side (c) dimensionless temperature ( Г
F
) along
the length of the reactor (Solid lines correspond to MR, and dotted
line to PFR)........................................................................................... 127
Figure 5.26: Effect of H
2
O permeance (a) CO
2
conversion as a function of Da
(b) H
2
exiting the feed side (c) dimensionless temperature ( Г
F
) along
the length of the reactor (Solid lines correspond to MR, and dotted
line to PFR)........................................................................................... 129
xi
Abstract
The capture and utilization of CO
2
have significant potential applications in the
chemical and power generation industries, as well as in space applications. For the proper
performance of space life-support systems, for example, the removal from the cabin
atmosphere of the CO
2
produced by the inhabitants is required. For short-term flights,
CO
2
can be controlled by sorption on metal hydroxide. For long-term space applications,
however, continuous regenerative approaches are required, including pressure-swing
adsorption and membranes which, in addition to removing the CO
2
, may, potentially, also
allow for the recovery of oxygen. One of the approaches proposed is the use of the
methanation (Sabatier) reaction, in which the CO
2
catalytically reacts with hydrogen to
simultaneously produce methane and water. In space applications, one of the challenges
the application of catalytic reactor technology faces is the dilute concentrations of CO
2
which make its pre-concentration a required step, thus complicating the process train. In
this study, we investigate the application of a reactive separation technology, in which the
catalytic and separation steps are coupled in-situ through the use of high-temperature
membranes. Coupling reaction and separation provides added synergy, which enhances
the performance of both steps. Another potential application of the Sabatier reaction
could emerge in In-Situ Resource Utilization (ISRU) on Mars. ISRU is a very important
new concept to be used to make human presence on Mars possible. This concept involves
utilizing raw resources from Mars atmosphere to create useful commodities such as
xii
oxygen and propellants like CH
4
. In this study, our current experimental and modeling
efforts in this area aiming to establish the feasibility of the proposed reactive separation
application for life-support and ISRU systems will be described.
1
Chapter 1 Introduction and Overview
1.1 Motivation and background
Mars is the fourth planet in the solar system, and is at its closest approach it is 80
million km away from Earth. In 1989, scientists were asked to estimate the price of a
manned Mars mission. They proposed a large spacecraft, that was supposed to be built in
orbit, that would send a group of astronauts in orbit around Mars, let them land on Mars,
from where they would return back into the craft, which would bring them back to earth.
This spacecraft was supposed to run on liquid hydrogen (H
2
) and liquid oxygen (O
2
). All
supplies were supposed to be brought along from Earth. NASA decided at that point, that
the cost was too high and soon after abandoned the manned Mars mission concept,
hoping to find a cheaper way to get there.
We now know that it is not necessary to take all supplies and equipment from
Earth to be able to travel to Mars, stay there and to come back. In-situ resource utilization
(ISRU) is an important new concept to make human presence on Mars and return
missions from there possible, both key objectives of NASA’s Mars exploration program.
The concept involves utilizing raw resources from the Mars atmosphere (composed,
approximately, of 95.3% CO
2
, 2.7% N
2
, 1.6% Ar, 0.13% O
2
, and 0.08% CO, and other
minor components) and the soil, and through a sequence of processing steps, to create
useful materials such as O
2
(either by direct capture from the atmosphere or through CO
2
decomposition) or iron from the soil. A key aspect of ISRU involves utilizing the CO
2
,
the main component of the Mars atmosphere, to generate CH
4
,
which
is
important as a
2
propellant, and as the starting feedstock for the production of a variety of other chemicals
(Brooks et al., 2007; Holladay et al., 2007; Hu et al., 2007).
Figure 1.1 shows a schematic of a conventional CO
2
methanation system to be
utilized as part of an ISRU system. This system will process CO
2
from the Martian
atmosphere to produce CH
4
, using the methanation reaction with H
2
shipped from Earth
or recovered by electrolysis from indigenous water. The principal product, CH
4
, will be
used as fuel for the return journey. In addition, H
2
O produced in the reaction (1.1) can be
converted via electrolysis to generate more H
2
(recycled to the process) and O
2
(for use as
an oxidant and for life support). Key to this process is the CO
2
catalytic hydrogenation
reaction, commonly also known as the methanation or Sabatier reaction, which proceeds
as follows:
22 4 2
42 CO H CH H O +←⎯→+ ( 165.4 / ) HkJmol Δ =− (1.1)
Figure 1.1 A conventional CO
2
methanation system for ISRU
3
Another potential application of the Sabatier reaction could emerge in the long-
duration manned space missions. For human presence on Mars or other planets to become
a reality, the development of highly reliable and efficient systems is required, which
provide basic life support provisions, such as food, water, and air for the crews during the
long space flights (Sridhar et al., 2004). A key function that space life-support systems
perform is the removal of metabolic CO
2
(Wieland, 1994) from the atmosphere in the
living quarters; otherwise, the CO
2
levels in the closed cabin environment would rise to
unacceptable levels.
Carbon dioxide is a colourless gas. As it is fully oxidized, it is not very reactive
and is, in particular, not flammable. At concentrations of 2,500 ppm to 5,000 ppm, CO
2
may cause headaches to some people. At extremely high levels of 100,000 ppm, people
lose consciousness in ten minutes, and at 200,000 ppm, CO
2
can lead to death (Greiner,
1995). The original Air Revitalization System (ARS) in the Freedom Space Station was
designed, for example, to maintain CO
2
levels at 2000 ppm or less; the life-support
system for the space suit worn during Extra-Vehicular Activities (EVA) is designed to
maintain CO
2
levels at or below 5000 ppm.
A schematic of a proposed conventional ARS system (Raatschen and Preiss,
2001) is shown in Figure 1.2. Major components of the ARS include a CO
2
removal
subsystem, an Oxygen Generation Subsystem (OGS), a trace contaminant control
subsystem (not shown), and a regenerative Water Recovery System (WRS). CO
2
, which
is removed from the cabin by the CO
2
removal subsystem, is vented overboard. H
2
,
which is generated with O
2
by the OGS, is also vented overboard. There will be other air
4
and water losses, such air leaks through spacecraft seals, EVA air loss in airlock
operation, and possible water loss from EVA cooling device. These O
2
losses can be
made up from electrolysis of water carried with the spacecrafts. Water lost in life support
functions can be supplied from the stored water.
Venting and discarding the CO
2
overboard is an easy solution, however, it carries
with it the downside of the loss of oxygen. The methanation reaction (1.1) provides again
the opportunity to recover part of this oxygen, which otherwise must be carried in the
form of H
2
O as supplies from earth.
Figure 1.2 A conventional ARS without a CO
2
reduction subsystem
CO
2
is, of course, a normal constituent of the Earth’s atmosphere, currently at a
350-400 ppm level. CO
2
is emitted by natural and human-induced activities. The most
5
common natural source is respiration. CO
2
emissions, due to human activities, are
attributed to three major causes: transportation, industry, and power plants. Carbon
dioxide is an essential ingredient in the cycle of life on earth. Plants directly use CO
2
in
the process of photosynthesis, where, combined with water, it is converted into sugars
and oxygen. Plants then use the sugars to fuel their growth, and animals breathe in the
oxygen and exhale CO
2
. In the 20th century, scientists realized that gases in the
atmosphere cause a "greenhouse effect," which affects the earth's temperature. In 1938,
Callendar argued that the level of CO
2
was climbing and raising the global temperature,
but most scientists found his arguments implausible at that point. In the early 1960’s,
Keeling measured the level of CO
2
in the atmosphere, and found that it was rising fast.
Researchers then began to take an interest, trying to understand how the level of CO
2
had
changed in the past, and how the level was influenced by chemical and biological forces
(Keeling, 1960; Keeling and Rakestraw, 1960). They found that the gas plays a crucial
role in climate change.
During the 1990-2004 period, total global emissions of greenhouse gases by
human activities increased by about 25%. Over the same period, total fossil fuel
combustion emissions of CO
2
increased about 28% world-wide. In 2004, alone, global
greenhouse gas emissions increased by almost 4%. This increase was mainly caused by a
5% increase of CO
2
emissions from fossil fuel combustion (corresponding to about 1300
megatons of CO
2
), especially in electricity generation (a 6% increase), and by an increase
in the industrial production of CO
2
(a 5% increase). More than half of the increase in the
production of CO
2
that is generated by fossil fuel combustion originates in China,
6
where CO
2
emissions from electricity production and industry use increased by 17% and
22% (corresponding to about 350 and 250 megatons) respectively (CO
2
emissions from
fuel combustion 1971-2004, in Part III: Greenhouse gas emissions, pp. III.1-III.41. 2006,
International Energy Agency (IEA), Paris). In addition, CO
2
emissions from other sources,
such as cement production, contributed to this increase (Olivier et al., 2005).
Since carbon emissions increase with economic growth, it is difficult to decrease
carbon emissions only by energy savings resulting by the more efficient uses of energy.
Consequently, CO
2
must be separated and recovered prior to being emitted to the
atmosphere, and methods to achieve this goal must be, therefore, developed. The
methanation reaction (1.1) again provides a direct route for CO
2
beneficiation and
utilization in combination with a renewable or non CO
2
producing method to produce
hydrogen (e.g., via nuclear energy).
1.2 Removal and recovery of CO
2
Different methods for CO
2
removal have already been studied for use in space
habitats and suits. The following is a description of the most pertinent techniques to
remove or recover CO
2
from the cabin or from the Mars atmosphere.
1.2.1 Lithium Hydroxide
Lithium Hydroxide (LiOH) is the most commonly used CO
2
sorbent for use in
expendable devices. The presence of water vapor is important to the functioning of the
LiOH beds. Chemisorption of CO
2
is thought to take place via a two-step reaction, in
which lithium hydroxide monohydrate is first formed by the exothermic reaction (Boryta
7
and Maas, 1971),
22
( ) ( ) ( )( 60.9 / ) LiOH s H O g LiOH H O s H KJ mol +→ ⋅ Δ=− (1.2)
followed by the endothermic formation of lithium carbonate.
22 23 2
2 ( ) ( ) ( ) 3 ( )( 16.0 / ) LiOH H O s CO g Li CO s H O g H KJ mol ⋅+ → + Δ= (1.3)
Significantly, water is required on both sides of these reaction equations. For the
net reaction,
223 2
2 ( ) ( ) ( )( 44.9 / ) LiOH CO g Li CO s H O s H KJ mol +→ + Δ=− (1.4)
two moles of water are liberated for each mole of CO
2
that is chemisorbed.
1.2.2 Silver Oxide
Silver oxide reacts with CO
2
in the presence of water to yield silver carbonate.
2
22 23
220
o
HO
C
Ag O CO Ag CO
⎯⎯ ⎯ →
+
←⎯ ⎯ ⎯
(1.5)
The chemisorption reaction can be reversed at elevated temperature. In addition to
silver oxide, magnesium oxide and zinc oxide are also known as viable candidates for use
as CO
2
reversible chemisorbents in life support systems (Hart et al., 1992; Nacheff et al.,
1989). To date, however, neither of these oxides has been shown to load to theoretical
capacity following thermal regeneration. A composite chemisorbent composed of a
mixture of silver oxide and zinc oxide has also shown been investigated.
1.2.3 Molecular sieves
Zeolite molecular sieves are commonly used for CO
2
control. The material is
porous and has a very large internal surface area. Gases are selectively removed from the
gas stream by the material sorting out gas molecules of a certain size range and holding
them by adsorption to its internal surface. Several molecular sieves of varying pore size
8
are manufactured which are able to separate CO
2
from various mixtures of gases (Barker
et al., 1991; Hyun et al., 1999). Although their ability to adsorb CO
2
is far below that of
chemical adsorbents, such as LiOH, molecular sieves have the advantage that can be
regenerated for repeated usage.
One apparent disadvantage in using molecular sieves is the need to dry the air
prior to its introduction into the canister. This is necessary since the sieves have an even
greater affinity for water than for CO
2
(Xu et al., 2005).
1.2.4 The Bosch reaction
The Bosch or otherwise known as the carbonization reaction (Holmes et al., 1972),
is described by the following equation:
22 2
2 2 ( 571.1 / ) CO H C H O H KJ mol +→+ Δ =− (1.6)
It is a reaction that occurs over an iron catalyst at high temperatures in the range
of 600 to 800
o
C. This reaction produces carbon, instead of methane, and also water that
must be continually removed to drive the reaction to the right. Further close examination
has shown that several side reactions also take place simultaneously, such as reaction 1.1,
as well as
22 2
CO H CO H O +→ + (1.7)
The efficiency of the Bosch reaction is dependent upon the design of the reactor
and the ability to conserve heat in the recirculating loop. The exit gases leaving the
reactor must be cooled in order to condense the water, and then recirculated through the
reactor making it necessary to reheat them. Theoretically, there is no hydrogen lost in the
Bosch process, and therefore there is no need for hydrogen makeup. However, the high
9
power requirement of the Bosch reactor, together with the weight of the carbon removal
components and the recirculating equipment are disadvantages of this reaction (Bunnel et
al., 1991).
The reliability of the Bosch reaction is inferior to that of the Sabatier reaction
simply because of the large number of dynamic components involved. In addition, the
iron catalyst used in the Bosch reactor is continually depleted because of the formation of
iron carbides at the high reaction temperature conditions. The catalyst must therefore be
continually replaced, and the free carbon and iron carbide removed from the system loop.
1.2.5 The Sabatier reaction
The catalytic reaction of CO
2
with H
2
to produce CH
4
and water (equation (1.1))
is often called the Sabatier reaction. The reaction takes place at the relatively low
temperature range of 150 to 350
o
C, depending on the catalyst selected to promote the
reaction, with lower temperatures (150 to 200
o
C) being possible for Ru-based catalysts
(Lunde and Kester, 1972; Lunde and Kester, 1974). The Sabatier reaction is an
exothermic process and, therefore, may require active cooling of the reactor to prevent it
from exceeding the low favorable reaction temperatures.
For each mole of CO
2
reduced, this system produces two moles of water and one
mole of CH
4
. For the conventional ARS, as noted above, the CH
4
is vented into space.
For this reason, only half of the H
2
consumed in this reaction can be recovered (e.g., by
electrolysis of the product water), while the rest is wasted in space. For ISRU and CO
2
capture/sequestration applications that is not a problem, however.
10
1.3 Scope of the work
Removal and utilization of CO
2
are essential for the success of manned missions
to Mars, and the Sabatier reaction appears promising for application in advanced ARS
and ISRU schemes. In this study, the application of membrane-based reactive separation
technology to the Sabatier reaction will be investigated. Coupling reaction and separation
using a membrane provides synergy for these applications, which enhances the
performance of both steps. Specifically, the Sabatier reaction provides the “chemical
pumping” needed to enhance the performance of the membranes, and the membranes
supplies the concentrated CO
2
needed to enhance the reactor performance.
The feasibility of applying the membrane reactor (MR)-based Sabatier technology for
ARS and ISRU applications is discussed in this dissertation. For the ISRU application,
for example, the key challenge is the exothermicity of the Sabatier reaction making the
internal temperature control of the unit a challenging task, therefore, requiring thermal
optimization of the process to enhance its performance. Membrane reactors provide the
additional option of optimizing the reactor temperature through the use of the sweep gas
stream on the permeate side, which in addition, can be used to increase the conversion by
favorably shifting the reaction equilibrium. The advantages and disadvantages that the
MR-based systems offer for this as well as the ARS application are studied and discussed
in this Thesis
The outline of this dissertation is as follows. This Chapter (Chapter 1) discusses
the motivation and overview for the research. In Charter 2, the selective membranes used
in this research are studied. We use both carbon molecular sieve membranes (CMSM)
11
and palladium membranes, which are characterized through both single as well as mixed
gas permeation measurements. The effects of temperature, pressure, and composition on
the permeation characteristics of these membranes are also investigated. In Chapter 3, the
kinetics of the CO
2
methanation reaction on a commercial Ni catalyst are investigated. In
Chapter 4, the mathematical model developed for the MR-based Sabatier reaction system
is presented. In Chapter 5, the experimental results of the CO
2
methanation in the MR
system are discussed and compared with the model predictions.
12
Chapter 2 Membrane Studies
2.1 Introduction
In this chapter, a brief overview of membrane separation terminology and of
various H
2
selective-membranes is provided. In this study we have investigated the use of
CMS and palladium membranes, prepared by Media & Process Technology, Inc. These
membranes are characterized through single and mixed-gas permeation tests under
various conditions.
2.2 Overview on membrane separation
The permeability of a membrane towards gases is a function of membrane
properties (its physical and chemical structure), the nature of the permeating species (size,
shape, and polarity), and the interaction between membrane and these species (Burggraaf
and Keizer, 1991; Hsieh, 1991; Hsieh, 1996; Sanchez and Tsotsis, 2002; Stern, 1994).
The gas flux (J) through the membrane is defined by the following expression:
V
J
A t
Δ
=
⋅
(2.1)
where ∆V is the volume of the permeated gas, A is the membrane area, and t is the time of
peremation. The permeance (U) through the membrane is defined by equation (2.2) below,
and is related to the permeability (P) by equation (2.3):
J
U
p
=
Δ
(2.2)
PU L = ⋅ (2.3)
13
where ∆p is the partial pressure difference between the upstream and downstream side of
the membrane, and L is the membrane thickness. Permeances can be defined as a
reciprocal resistance against mass transport through the porous membrane (Uhlhorn et al.,
1992a; Uhlhorn et al., 1992b). Since a multilayer membrane can be considered as a
number of resistances in series, the overall permeance, U, of a membrane is related to the
permeances of different layers of the membrane and of the support permeance in the
following way:
sup 1 2
11 1 1
port layer layer
UU U U
= ++ +⋅⋅⋅ (2.4)
The permeability P of a given species is a measure of the rate at which a penetrant
traverses the membrane, and it generally relates to the transport rate through the
membranes as well as the affinity of a given gas towards the membrane surface.
Membranes utilized in separations need to possess both a high selectivity, towards a
given species as well as a high permeation rate. The permselectivity or ideal separation
factor, α, between two given components is simply the ratio of two gas permeabilities
1
1,2
2
P
P
α
⎡ ⎤
=
⎢ ⎥
⎣ ⎦
(2.5)
The actual separation factor (for mixed-gas separation) is defined as:
12
1,2
21
y x
y x
α
⋅
=
⋅
(2.6)
where x
1
and x
2
are the mole fractions of species 1 and 2, respectively, in the feed, and y
1
and y
2
are the mole fraction of species 1 and 2, respectively, in the permeate stream.
The temperature-dependence of permeability for a given gas penetrant in a
membrane is generally described by the Arrhenius expression:
14
0
exp
p
E
PP
RT
− ⎛⎞
=⋅
⎜⎟
⎝⎠
(2.7)
where P
0
is a pre-exponential factor and E
p
is the apparent activation energy for
permeation..
2.3 Inorganic membranes
Inorganic membrane can be classified into two groups, based on their structure,
namely: porous and dense (Hsieh, 1991). In porous inorganic membranes, a porous thin
top layer is supported on a porous metal or ceramic support, which provides mechanical
strength, but typically offers a minimal mass-transfer resistance. Alumina, carbon, glass,
silicon carbide, titania, and zeolite membranes are mainly used as porous inorganic
membranes supported on different substrates, such as α-alumina, γ-alumina, zeolite, or
porous stainless steel. The porous inorganic membranes can be further classified, based
on their pore size such as microporous(< 2 nm) or mesoporous (2nm< to < 50nm) or
macroporous (> 50 nm)), or based on their structure (e.g., symmetric, characterized by a
homogeneous, uniform structure throughout the membrane, or asymmetric, consisting of
one or more layers with a different structure).
Palladium (and its alloys with ruthenium, nickel or other metals from groups VI to
VIII) and silver are common examples of dense metallic membranes. Palladium-based
membranes are permeable only to H
2
, while silver and other materials (e.g., yitria-
stabilized zirconia) are permeable only to oxygen (Lee and Choi, 2007; Shu et al.,
1991). These membranes have high selectivity, but generally a low permeability. Recently,
15
attempts are being made to improve the permeability of the dense membranes by creating
asymmetric structures where a thin metallic layer is supported on an underlying porous
membrane support using various techniques (Gu and Zhong, 2006; Sun et al., 2006).
As shown in Figure 2.1, there are four main transport mechanisms by which gas
separation using porous inorganic membranes takes place (Hassan et al., 1995). The basis
of these transport mechanisms, are differences in molecular weight (Knudsen diffusion),
surface interactions (surface diffusion and capillary condensation), and the size of the
molecules (molecular sieving) to be separated.
Knudsen diffusion occurs in the gas phase by transport through membrane pores
having diameters (d) smaller than the mean free path dimensions of the molecules ( λ) in
the gas mixture (i.e. the Knudsen number ( λ/d), is much greater than one). As a result, the
movement of molecules inside the narrow pore channels takes place through collision of
the diffusing molecules with the surface (wall) rather than with each other. Since the
driving force for transport is the partial pressure of the gas species, Knudsen transport can
occur either by concentration or by pressure gradients. The relative permeation rate of
each component is inversely proportional to the square root of its molecular weight.
In the surface diffusion mechanism, the diffusing species adsorb on the walls of
the pore, and then readily transport across the surface in the direction of decreasing
surface concentration. Rao et al. (Rao and Sircar, 1996) took advantage of the surface
diffusion mechanism through a type of membrane, called the Selective Surface Flow
membranes, in order to separate molecules with larger molecular weight and with larger
polarity and polarizability which are selectively adsorbed on the membrane surface from
16
their gas mixtures with lighter molecules(Rao and Sircar, 1996). The adsorbed species on
the membrane pores can also drastically reduce or eliminate the transport of non-
selectively adsorbed molecules across the pore by reducing the size of accessible void
space through the pore (Rao and Sircar, 1996). The concentration of adsorbed species
depends upon the temperature, pressure, and the nature of the surface. The interaction
between a gas and the pore can also be induced by modification of the adsorbent layers.
Multilayer diffusion occurs when species adsorb in several layers. The multilayer
diffusional flux is generally much larger than the gas phase diffusional flux (Uhlhorn et
al., 1992a; Uhlhorn et al., 1992b). Capillary condensation also blocks the pore, thus
preventing gas transport of other components of the gas/vapor mixture. Both aspects can
result in increased selectivities. The condensation pressure depends on the pore size and
shape and also the strength of the interaction between the fluid and the pore walls.
In the molecular sieving mechanism, the separation is caused by passage of
smaller molecules of a gas mixture through the pores of porous inorganic membranes,
while the larger molecules cannot enter into these pores and a selective separation, based
upon size exclusion is thus observed.
17
A
B
C
D
Figure 2.1 Transport mechanisms through microporous membranes (A: Knudsen
diffusion, B: Surface diffusion, C: Capillary condensation, D: Molecular sieving)
2.3.1 CMS-based membranes
Carbon-based membranes are usually prepared by the pyrolysis or the
decomposition of an organic/polymeric material, which produces a dense layer having a
carbon microporous sieving network. Microporous granular carbon materials have, of
course, been used commercially for years for molecular size separation by adsorption
(Shirley and Lemcoff, 2002), however their use as membranes is a most recent
18
development. (Harale et al., 2007; Kita et al., 1997).
2.3.2 Pd-based membranes
Palladium was first identified as a highly hydrogen permeable material in the 19th
century and it is still used for high-performance H
2
-separation applications today.
Palladium has advantages over other membrane materials because of its catalytic surface,
high H
2
permeability, infinite H
2
selectivity, temperature stability, and corrosion
resistance.
Palladium membranes have been used for many years for the production of pure
hydrogen for laboratory use, and more recently, the use of these membranes has been
advocated for membrane reactors. In the latter application, due to the infinite selectivity
for hydrogen possessed by these membranes, removal of this gas via the membrane can
lead to enhanced reaction yields in equilibrium limited reactions where H
2
is a product.
Consequently, many experimental investigations have been conducted on reactions of
this type, including, for example, dehydrogenation of hydrocarbons, decomposition
reactions and steam reforming, together with the associated water gas shift reaction
(WGSR).
Palladium alloys are often preferred over pure Pd membranes, because pure
palladium tends to become brittle after repeated cycles of hydrogen adsorption and
desorption. Schmitz et al. (1997, 1998), for example, investigated the permeation of H
2
in
membranes made of palladium, palladium-silver, stoichiometric titanium-nickel alloy
coated with nickel, membranes of vanadium with a coating of copper and palladium, and
a double-layer membrane of Pd-Ag allow coated with vanadium. The results showed that
19
the Pd-Ag membranes had the best permeation rate. Pd–Ag alloys are currently used to
prepare commercial permeator tubes for hydrogen purification and separation.
Although the conventional multi-tube dense Pd/Ag alloy modules, as used for the
laboratory purification of hydrogen, can also be used for catalytic membrane reactors, its
wall thickness generally limits the flux obtainable. Therefore, composite membranes, in
which a thin film of Pd or Pd/Ag is deposited onto a porous support, have been developed
for catalytic membrane reactors (Sun et al., 2006). These films may be deposited by a
variety of techniques, including magnetron sputtering, electroless plating, chemical vapor
deposition or other techniques, with the produced films being typically between 5 and 20
mm thick. At these thicknesses, adequate flux levels for hydrogen permeation appropriate
for the removal of product hydrogen from an equilibrium limited reaction can be attained.
Hydrogen permeation through palladium membranes is usually described to occur
by the following six steps (Ward and Dao, 1999): (i) adsorption of H
2
on the feed side
surface, (ii) dissociation of H
2
molecules into H atoms on the same surface, (iii)
dissolution of the H atoms into the metal bulk, (iv) hydrogen bulk diffusion to the
opposite membrane surface (permeate side), (v) formation of H
2
molecules from H atoms
on the permeate side, and (vi) desorption of H
2
from the permeate side surface.
Co-existing gas species often damage palladium membranes and decrease its
hydrogen permeation rate (Musket, 1976). Contaminants such as sulfur, chlorine and
arsenic chemically react with palladium, leading to a collapse of the membrane structure.
Antoniazzi et al. and Ali et al. (Ali et al., 1994) reported that sulfur poisoned Pd
membrane sites, which were still active for the dissociation of hydrogen.
20
Various studies (Amano et al., 1990; Hara et al., 1999; Jorgensen et al., 1997;
Mcbride and Mckinley, 1965) on CO addition to H
2
during permeation through Pd
membranes (Pd alloy membrane) have shown a deterioration in hydrogen permeation to
occur at temperatures below about 573 K.. Li et al. (2000) observed a CO retardation
effect at temperatures up to 653 K when using 5 % CO, with 10 μm-thick palladium film
deposited on the stainless steel support (Li et al., 2000). Gallucci et al. (2007) reported
that the negative influence of CO decreases with increasing temperature and at 623 K the
H
2
permeation flux trend for a (CO/H
2
) mixture was found to be the same with that for a
(N
2
/H
2
) mixture. This fact indicates that, for surface interactions, the influence of CO on
H
2
permeation depends both on the CO concentration in the mixture and on the
temperature. The negative effect of CO on the H
2
permeation is also found when CO is
used as sweep gas and pure hydrogen is used in the lumen side of the reactor.
Hou and Hughes (2002) also reported that steam and CO have significant
inhibitive effects on H
2
permeation through the membrane due to their competitive
adsorption with hydrogen on the metallic surface at temperatures less than 623 K,
whereas, CO
2
displayed only a slight inhibitive effect at these temperatures. The effect
increased with the amount of steam, CO or CO
2
in the mixed feed, and decreased with the
temperature (Hou and Hughes, 2002).
For hydrocarbons, Jung et al. (2000) reported that a hydrogen stream containing
methane and propane had a negligible effect on H
2
permeation, while a mixture with
propylene resulted in a gradual decrease of H
2
permeation with time due to deposition of
carbonaceous matter, which formed from the decomposition of propylene, accumulated
21
on the membrane surface. Similar phenomena have been observed for palladium
membranes exposed to a continuous mixed feed of H
2
with alcohols (Dannetun et al.,
1996).
In order to prevent the deactivation of membrane and catalyst by coking, steam is
usually mixed in the feed of a dehydrogenation reactor. However, more recent studies
have revealed that steam had a much more pronounced effect on membrane properties
than carbon monoxide at temperatures below 653 K (Li et al., 2000). Steam played two
competing roles, i.e. decreasing the permeation rate by competitive adsorption on sites
which are active for hydrogen dissociation, and increasing it by removing the
carbonaceous matter deposited on the membrane.
According to the study by Li et al. (2000), the H
2
permeation rate is lower during
addition of steam than during the addition of CO. They concluded that the effect of the
addition of steam on H
2
permeation is due to the adsorption of steam on the palladium
membrane surface and that the adsorption of steam is reversible. The restoration of the H
2
permeation rate, after stopping the steam addition, indicates that steam did not change
any of the permeation properties of the Pd membrane.
2.4 Membrane performance evaluation
Figure 2.2 shows the schematic of the membrane module used in our studies. A
thermowell, bored through the module, is used to install the thermocouples. The feed
flows through the outer side of the membrane tubes. Three thermocouples are installed to
control the temperature in the reactor. Two thermocouples are installed using the
22
thermowell. A third thermocouple is installed on the top, as shown in the figure. This hole
is also used to fill in the catalyst and adsorbent in the reactor.
membrane
membrane
Shell side inlet
Tube side
(sweep side)
inlet
Tube side
(sweep side)
outlet
Thermo-well
Shell side outlet
Thermo-well catalyst or quartz
Figure 2.2 Schematic of membrane reactor module
2.4.1 Membrane permeation
As previously note, the two most important characteristics of membranes are the
permeance and separation factor. Permeance is a measure of the gas flow rate per unit
area per unit pressure difference. A more fundamental unit is the permeability, which is
the permeance multiplied by the thickness of the membrane. In most cases, the thickness
of the membrane is not known very accurately however, and, therefore the permeance is a
more practical quantity. The separation factor is meaningful only with respect to a
mixture of two gases. The transport of gases through membranes depends on the pore
diameter, and it can also be affected by temperature. However, measuring pore diameters
23
that are smaller than 2 nm is difficult. Therefore, it is critically important to be able to
follow the changes in the transport mechanisms of different gases during pore diameter
reduction, in order to help determine the extent to which pores have been reduced.
The details of the experimental set-up used in the membrane characterization
studies are discussed in Chapter 3. Two types of hydrogen selective membranes have
been investigated, the CMS membranes and palladium membranes. In this study, tubular-
type membranes were used and characterized through single and mixed-gas permeation
measurements. The membranes were installed in the tubular stainless steel reactor
module, shown in Figure 2.2. Graphite seals along with compression fittings were used to
seal the area between the membrane and reactor module.
The reactor module was installed in the furnace, and isothermal conditions were
maintained in the system with the help of three temperature controllers and heating coils.
The exit and inlet flow lines were insulated to avoid water condensation in the lines.
Pressure was measured using two transducers located at the inlet and outlet for the outer
cell, which is called the feed side and another pressure transducer, is located at the exit of
the inner cell, called as membrane side. Feed gas flow rates were controlled using mass
flow controllers and gas cylinder pressure regulators. When measuring the permeation of
water vapor through the membrane, the water was introduced through a syringe pump
into the evaporator where it was vaporized, and introduced into the reactor. A back-
pressure regulator is used to control the pressure on feed/outer side. An atmospheric
pressure is maintained on the membrane/permeate side. The flow rates of the retentate
and permeated gases were measured using a soap bubble-flow meter and the water vapor
24
permeated through the membrane was determined by measuring its weight as trapped by
water adsorbent.
The compositions of the separated gases in the mixed gas measurements were
analyzed using a mass spectrometer. In order to prevent the system from being
contaminated from other gases, the feed and permeate sides prior to the experiments are
swept using argon or nitrogen for more than 20 min. The temperature in the system is
raised under an argon or nitrogen environment.
To determine the permeability and the separation factors of the membranes used, a
series of experiments were carried out for both single, and gas mixtures. Generally ideal
and mixed-gas separation factors are different from each other. This is due to the fact that,
during mixed-gas permeation the flux values are influenced by the multicomponent
interactions, which often impact the overall behavior. In this study the permeabilities and
separation factors of the membranes have been determined by feeding to the reactor a
mixture of gases containing all the species present during the reaction (H
2
, CO, CO
2
, CH
4
,
and H
2
O). The effect of temperature and pressure on the membrane performance was also
investigated.
The mixed-gas permeabilities (U
j
) were obtained by simultaneously solving the
governing equations for flow through the porous membrane for each component. The
permeabilities were fitted as parameters using either feed- or permeate-side flow rates
measured using the bubble-flow meter and the partial pressures, obtained using the exit
stream composition analyzed with the help of online mass spectrometers, as previously
noted.
25
Table 2.1 presents the physical dimensions of the CMS membranes and palladium
membranes tested, as well the operating conditions used in this study.
Table 2.1 Physical properties and experimental conditions investigated
Carbon molecular sieve
membranes
Palladium membranes
Inner diameter (mm) 3.5 3.5
Outer diameter (mm) 5.7 5.7
Length (mm) 254 60
Temperature (°C) 120-250 250-450
Feed side pressure (psi) 30-75 10-60
Gases investigated H
2
, CO, CO
2
, CH
4
, H
2
O, Ar, N
2
2.4.2 CMS based membrane data
Mass transfer through a porous membrane is described by the following equation:
() =−
FP
jj j j
FU P P (2.8)
where F
j
is the molar flux,
F
j
P the partial pressure of component j on the feed side,
P
j
P the partial pressure of component j on the permeate side, and U
j
the membrane
permeance for component j.
Table 2.2 presents the single (H
2
, CO, CO
2
, CH
4
, H
2
O, N2) and mixed-gas (two
different compositions were used, namely H
2
: CO: CO
2
: CH
4
= 5:1:1:1 and H
2
: CO: CO
2
:
26
CH
4
: H
2
O = 2.5:1:1:1:2.5) permeances and ideal separation factors (defined as the
permeance of H
2
divided by the permeance of the corresponding gas) obtained for one of
our CMS membranes (# 1). Table 2.3 shows the corresponding data for a different CMS
membrane (# 2). For the most of the CMS membranes investigated the separation factors
obtained from the mixtures were somewhat different than those obtained from the single
components. Specifically, H
2
permeance reduces a bit while CO and CO
2
permeances
somewhat increase. Table 2.4 shows the H
2
-CO
2
binary permeance data as a function of
composition for a third CMS membrane (# 3). As can be seen from Table 2.4, CO
2
composition in the gas mixture does have an effect on the H
2
permeance. The normalized
H
2
permeance (mixed-gas H
2
permeance/pure gas H
2
permeance) was fitted as a function
of H
2
molalr ratio on the feed-side. Figure 2.3 shows the normalized H
2
permeance data
for feed pressures of 45 and 75 psi; also shown on the same figure is the equation
obtained by fitting the permeance data for the CMS membrane # 3. Figure 2.4 shows the
corresponding data for other CMS membranes (# 4, # 5 and # 6).
The membrane performance may also be different depending on whether the
gases are fed, either on the membrane layer or the support side. The effect on the
membrane performance of feeding the gas on a different side was also investigated. Table
2.5 shows a comparison between the permeance data among experiments where the gases
are either fed on the support or the membrane layer sides, with data measured for CMS
membrane # 4. For the single gases there is no significant difference in the experiments
when the gases are fed on different sides. For the mixed gas experiments there are some
effect when changing the feed position. The permeances are generally higher when the
27
mixed gas is fed into the tube side where the carbon layer is deposited rather than the
permeate side. However, the separation factors remain fairly unaffected. Among the
various gases it is the water permeance for which the permeance is most affected by the
feeding direction.
28
29
Table 2.3 Single and mixed-gas permeance data for CMS membrane # 2
Mixed Gas
Pure Gas
H
2
:CO:CO
2
:H
2
O =
4.0 : 1.0 : 1.0 : 1.0
250
o
C/ ΔP = 64.7 psi 250
o
C/ ΔP = 64.7 psi
Permeance
S.F.
Based on H
2
Permeance
S.F.
Based on H
2
H
2
1.6847 1.0 1.6916 1.0
CO 0.0483 34.9 0.0578 29.2
CO
2
0.1069 15.8 0.1182 14.3
H
2
O - - 1.9005 0.9
N
2
0.0352 47.9 - -
Permeance unit: m
3
/(m
2
·hr·bar)
30
Table 2.4 H
2
-CO
2
binary gas permeance data for CMS membrane # 3
Temperature = 250°C
Binary (H
2
+ CO
2
), Total Flow = 30 cc/sec
Δ P Permeance
Gas
(psi)
Mol Fraction
m
3
/(m
2
·hr·bar)
H
2
0.00 0.0000
CO
2
45.5
1.00 0.0286
H
2
0.05 0.1700
CO
2
45.6
0.95 0.0280
H
2
0.10 0.1708
CO
2
44.8
0.90 0.0289
H
2
0.15 0.4308
CO
2
45.3
0.85 0.0258
H
2
0.20 0.5566
CO
2
45.8
0.80 0.0262
H
2
0.30 0.6439
CO
2
46.4
0.70 0.0278
H
2
0.40 0.6667
CO
2
45.6
0.60 0.0308
H
2
0.50 0.6838
CO
2
45.5
0.50 0.0274
H
2
0.70 0.6909
CO
2
45.8
0.30 0.0289
H
2
0.80 0.7133
CO
2
44.7
0.20 0.0287
H
2
0.85 0.7287
CO
2
47.2
0.15 0.0281
H
2
0.90 0.7510
CO
2
45.1
0.10 0.0251
H
2
1.00 0.7810
CO
2
44
0.00 0.0000
31
H
2
/CO
2
ratio
02 468 10
H
2
permeance (normalized)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
45 psi
75 psi 2
1
6.416, 6.493
0.95
ax
y
bx
ab
r
⋅
=
+⋅
==
=
Figure 2.3 Hydrogen permeance as a function of H
2
/CO
2
molar ratio in the feed (CMSM
# 3)
32
H
2
/CO
2
ratio
01 2345
H
2
permeance (normalized)
0.0
0.2
0.4
0.6
0.8
1.0
CMSM # 4 (65 psi)
CMSM # 5 (70 psi)
CMSM # 6 (30 psi)
2
1
7.05, 8.51
0.94
ax
y
bx
ab
r
⋅
=
+⋅
==
=
Figure 2.4 Hydrogen permeance as a function of the H
2
/CO
2
molar ratio in the feed
(CMS # 4, # 5 and # 6)
33
34
Figures 2.5-2.12 summarize all the CMS membranes, characterized through the
single and mixed-gas permeation experiments in this study. The single and mixed-gas
separation factors for various gases for all the membranes tested are plotted as a function
of H
2
permeance. Table 2.6 shows the range of H
2
permeance and separation factors for
the CMS membrane we investigated. The CMS membranes investigated showed high
selectivity for CO, CH
4
, N
2
and Ar and high steam permeance.
H
2
permeance (m
3
/m
2
/bar/h)
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Separation factor of H
2
/CO
0
20
40
60
80
100
Shell-side feeding
Tube-side feeding
Figure 2.5 H
2
/CO (single gas) vs. H
2
permeance
35
H
2
permeance (m
3
/m
2
/bar/h)
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Separation factor of H
2
/CO
2
0
10
20
30
40
Shell-side feeding
Tube-side feeding
Figure 2.6 H
2
/CO
2
(single gas) vs. H
2
permeance
36
H
2
permeance (m
3
/m
2
/bar/h)
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Separation factor of H
2
/CH
4
0
20
40
60
80
Shell-side feeding
Tube-side feeding
Figure 2.7 H
2
/CH
4
(single gas) vs. H
2
permeance
37
H
2
permeance (m
3
/m
2
/bar/h)
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Separation factor of H
2
/N
2
0
20
40
60
80
Shell-side feeding
Tube-side feeding
Figure 2.8 H
2
/N
2
(single gas) vs. H
2
permeance
38
H
2
permeance (m
3
/m
2
/bar/h)
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Separation factor of H
2
/CO
0
20
40
60
80
100
Shell-side feeding
Tube-side feeding
Figure 2.9 H
2
/CO (mixed gas) vs. H
2
permeance
39
H
2
permeance (m
3
/m
2
/bar/h)
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Separation factor of H
2
/CO
2
0
10
20
30
40
Shell-side feeding
Tube-side feeding
Figure 2.10 H
2
/CO
2
(mixed gas) vs. H
2
permeance
40
H
2
permeance (m
3
/m
2
/bar/h)
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Separation factor of H
2
/CH
4
0
20
40
60
80
Shell-side feeding
Tube-side feeding
Figure 2.11 H
2
/CH
4
(mixed gas) vs. H
2
permeance
41
H
2
permeance (m
3
/m
2
/bar/h)
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Separation factor of H
2
/H
2
O
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Shell-side feeding
Tube-side feeding
Figure 2.12 H
2
/H
2
O (mixed gas) vs. H
2
permeance
Table 2.6 Range of measured H
2
permeance and separation factors
Pure Gas Mixed Gas
H
2
permeance (m
3
/m
2
/h/bar) 0.34-2.61 0.32-2.55
CO 30.0-80.9 6.5-40.9
CO
2
10.8-29.3 5.9-29.0
CH
4
52.2-77.2 23-78.1
N
2
40.0-62.5
Ar 35.9-94.2
S.F.
based on H
2
H
2
O 1.1-6.3 0.6-2.1
42
2.4.3 Pd membrane data
The flux of the hydrogen through a palladium membrane is generally described by
the following equation (known also as the Sieverts equation):
22 2 2 2 2 2
/
,
(( ) ( ) ) (( ) ( ) )
p
ERT
Fn Pn Fn Pn
HHo H H H H H
JU e P P U P P
−
=−= − (2.9)
where
2
H
J [ mol/m
2
·s] is the H
2
permeation flux,
2
F
H
P [atm] its partial pressure in the feed-
side,
2
P
H
P [atm] its pressure on the permeate-side, n the pressure exponent,
2
H
U [ mol/m
2
·atm·s] the membrane permeance of H
2
, U
H2,o
[mol/m
2
·atm·s] the pre-
exponential factor, and E
p
[J/mol] the activation energy for transport.
For thick membranes (X
M
> 100 μm), the limiting resistance is assumed to be the
transport of hydrogen atoms through the palladium. Under these conditions, the surface
reaction is considered to be very fast and the dissolved hydrogen atoms at the surface of
the palladium are in equilibrium with the hydrogen gas on either side of the membrane.
The concentration of hydrogen atoms in the palladium can then be related to the
hydrogen partial pressure via the Sieverts equation, with n=0.5, which reflects the
dissociation of the gaseous hydrogen molecule into two hydrogen atoms that diffuse into
the metal, where an ideal solution of hydrogen atoms in palladium is formed.
Values of n greater than 0.5 are commonly reported in ultra-thin, supported
palladium membrane studies (Peachey et al., 1996; Yan et al., 1994), where it is possible
that the validity of the diffusion limited hydrogen transport mechanism assumption is
questionable. n values in the range of 0.5 and 1.0 may be attributable to a more complex
transport mechanism involving surface effects as well as the hydrogen diffusion process
43
through the membrane.
Figure 2.13 shows permeation data with one of the supported Pd membranes
prepared by our team. The membrane obeys well Equation (2-11) with the exponent n =
0.75 and E
p
= 7.96 KJ/mol. Reported activation energies for transport for Pd/ceramic
composite membranes range from 6 – 12 KJ/mol (Gobina and Hughes, 1994; Guo et al.,
2003; Uemiya et al., 1991). These data are utilized as the base case in the simulations in
Chapter 5. Table 2.7 summarizes the literature data for H
2
peremation through palladium
and composite palladium membranes.
44
P
F
n
- P
P
n
(atm
n
)
0.0 0.5 1.0 1.5 2.0 2.5 3.0
J
H2
(mol/s/m
2
)
0.0
0.1
0.2
0.3
0.4
0.5
250
o
C
300
o
C
350
o
C
400
o
C
450
o
C
n = 0.75
Figure 2.13 Flux of hydrogen as a function of ( ) ( )
Fn P n
PP −
45
Table 2.7 Summary of published data for H
2
permeation through palladium membranes
Membrane X
M
( μm)
T
(K)
n E
(kJ/mol)
Reference
Pd 10-150 623-773 0.68 11.92 (Hurlbert and
Konecny, 1961)
Pd-Ag 800-2025 500-900 0.5 12.81 (Holleck, 1970)
Pd 940 769-1219 0.5 20.50 (Katsuta et al., 1979)
Pd 1000 623-1173 0.5 13.81 (Morreale et al., 2003)
Pd 1000 623-1173 0.62 13.41 (Morreale et al., 2003)
Pd/Al
2
O
3
0.5-5 673 0.6-0.7 (Xomeritakis and Lin,
1998)
Pd-Ag 25 623 0.5 (Amandusson et al.,
2000)
Pd-Ag/Al
2
O
3
4.5-6.4 673 0.76 7.92 (Uemiya et al., 1991)
Pd-Ag/glass 6 673-773 0.5 7.17 (Gobina and Hughes,
1994)
Pd-Ag/Al
2
O
3
8.6 673-773 0.5 8.22 (Guo et al., 2003)
Pd/Al
2
O
3
523-723 0.75 7.96 this study
46
Chapter 3 Reaction Studies
3.1 Introduction
In this chapter, the reaction kinetics for the CO
2
methanation reaction are
investigated. The effect of temperature, pressure, and composition on the reaction rate is
studied and discussed here. A brief literature review on the reaction and the proposed
reaction mechanisms is also presented.
3.2 Thermodynamics
The thermodynamic data for the CO
2
methanation reaction, shown in Figure 3.1,
indicate that it is highly exothermic and that equilibrium conversions starts to decrease at
temperatures above 300
o
C.
The reaction has received significant attention in the literature since the 1950's
(Karn et al., 1965; Lunde and Kester, 1974) because of its significance for producing
methane from products of coal gasification. The reverse reaction is called steam
reforming, and is the commercial method for H
2
manufacture.
47
Figure 3.1 Thermodynamic equilibrium conversion and the heat of the reaction for the
Sabatier reaction (1 atm, H
2
/CO
2
= 4.0)
Ruthenium and nickel were found to be the most active catalysts for promoting
the reaction (Lunde and Kester, 1973; Lunde and Kester, 1974). Particularly suitable for
the methanation of CO
2
are Ru-based catalysts (the problem with Ru catalysts, of course,
is their cost). These catalysts display good activity at relatively low temperatures, which
are favorable with respect to the equilibrium conversion of the strongly exothermic
reaction, and form small amounts of methane even at room temperature. In addition, Ru-
based catalysts (Darensbourg and Ovalles, 1986; Marwood et al., 1997) are also more
resistant to deactivation, and produce no carbon (as the ratio of H
2
/CO
2
falls below 3.5,
carbon formation becomes thermodynamically favorable for the reaction at increasingly
higher temperatures). The reaction mechanism over Ru-based catalysts is complex, but
48
fairly well understood.
The reaction kinetics were studied by Vlasenko et al.(1955) and Dew et al. (1961).
They found a first-order dependence on CO
2
over the temperature range of (125 - 350
o
C),
and the concentration range of (0.05 - 0.4%) CO
2
in H
2
at 1 atm. Based on their studies
with a Ru/TiO
2
catalyst at 383 K, Marwood et al. (1997) reported that the reaction
mechanism involves the presence of a formate species adsorbed on the metal/support
interface, which is a precursor to adsorbed CO, and which appears to be a key
intermediate in the pathway to CH
4
. They reported this precursor to be a formate species
adsorbed at the metal/support interface.
Hougen-Watson type rate equations have been considered by Phungquach and
Rouleau (1976) for the CO
2
methanation reaction involving combinations of the possible
adsorbed species and various controlling steps. Generally these models ignore the reverse
reaction, and do not consider desorption to be rate controlling. Only four models, which
consider the surface reaction between the adsorbed species as the rate-determining step,
seem to fit the experimental data well (Phungquach and Rouleau, 1976; Rotaru and
Blejoiu, 2001), and are thought to be most adequate in describing the real mechanism of
the process. The rate determining steps and the corresponding rate expressions for the
four models are listed in Table 2.1 (Rotaru and Blejoiu, 2001). It this table, L stands for
an active center on the catalyst surface, k is the kinetic constant, K
i
is the adsorption
equilibrium thermodynamic constant for species i, and p
i
the partial pressures of species i.
Phungquach and Rouleau showed, for example, that the rate controlling step for the
reaction between CO
2
and H
2
on Ru catalyst (cylindrical alumina tablets, 3.2 mm in
49
diameter and thickness, impregnated with 0.5% Ru) is the surface reaction between
adsorbed CO
2
and H
2
to produce CH
4
and water vapor in the gas phase (Phungquach and
Rouleau, 1976).
Because of the high cost of Ru-based catalysts, Ni-based catalysts have also been
considered for the reaction. Rotaru and Blejou (2001) , for example, investigated the CO
2
methanation on a Ni/Al
2
O
3
catalyst, and also found that Model I, listed in Table 2.1,
showed the best fit with their experimental results. This is also in good agreement with
the results of Binder and White (1950) and of Dew et al. (1955). Herwijnen et al. (1973)
studied the kinetics of CO
2
and H
2
at atmospheric pressures on a supported Ni catalyst at
a partial pressure of CO
2
below 0.02 atm and for the temperature range of (200 - 230
o
C).
They found that the adsorption of CO
2
on the catalyst surface is the rate-determining step
under their operating conditions, but they also noted that at higher concentrations of CO
2
,
surface reaction or a desorption step probably controls the rate
Table 3.1 Hougen-Watson type model for the Sabatier reaction
Rate controlling step Rate equation Model
22 4 2
425 CO L H L C H H O L ⋅+ ⋅ ⇔ + +
22 2 2
22 2
44
5
2
(1 )
CO H C O H
CO H C O
kK K p p
r
Kp K pH
=
++
I
22 42
424 CO L H L C H L H O L ⋅+ ⋅ ⇔ ⋅+ +
22 2 2
22 2 2 4 4
44
5
(1 )
CO H CO H
CO CO H H CH CH
kK K p p
r
Kp K p K p
=
++ +
II
22 4 2
42 3 CO L H L C H H O L L ⋅+ ⋅ ⇔ + ⋅ +
22 2 2
22 2 2 2 2
44
5
(1 )
CO H C O H
CO CO H H HO HO
kK K p p
r
Kp K p K p
=
++ +
III
22 42
422 COLHL CHL HOLL ⋅+ ⋅ ⇔ ⋅ + ⋅+
22 2 2
22 2 2 4 4 2 2
44
5
(1 )
CO H C O H
CO CO H H CH CH H O H O
kK K p p
r
Kp K p K p K p
=
++ + +
IV
50
There have also been a few prior membrane reactor (MR) applications to the
Sabatier reaction, as part of the worldwide effort in membrane catalysis since the 1980’s.
The use of membranes in chemical reactors is motivated principally by the synergy that is
created by the preferential permeation of products (or reactants), leading to higher
conversion and/or selectivity, and a potentially reduced downstream separation load.
There has been an intense, worldwide effort on membrane reactors since the 1980’s and
these efforts have been summarized in a number of recent review articles (Armor, 1989;
Harold et al., 1994; Hsieh, 1991; Sanchez and Tsotsis, 2002; Saracco and Specchia, 1994;
Shu et al., 1991; Zaman and Chakma, 1994). Incorporating inorganic membranes into
reactors has also been surveyed for a number of reactions. A number of groups are
working to bring this technology to a demonstrated commercial application, particularly
for catalytic membrane reactors.
Both inorganic and polymeric membranes are used for reactive separation
operations; however, applications of polymeric membranes are limited due to operating
temperature constraints. The development of inorganic membranes having consistent
quality and narrow pore size distribution, in recent years, paved the way for the
application of membranes in high temperature reactors. As opposed to polymeric
membranes, the inorganic membranes are characterized by high resistance to temperature
and corrosive environments, and good mechanical stability. The use of membranes in
chemical reactors is motivated principally by the equilibrium shift caused by selective or
preferential permeation of reaction products, leading to a higher conversion in a single
pass. The equilibrium shift also allows attaining a given conversion at less severe
51
conditions of temperature and pressure. As reaction and permeation proceed
simultaneously, the separation of product can be accomplished in the reactor unit itself, or
at least the downstream separation load is reduced.
The membrane reactors are usually operated in parallel or in a cross-flow mode,
with the reactants fed on one side and the products typically exiting on the permeate side,
where often vacuum or a sweep gas is applied. The membrane is generally inert, and the
reactor may be packed with catalyst (or the catalyst may be fluidized). On occasion, the
membrane itself acts as a catalyst or catalyst is impregnated in the membrane. Figure 3.2
schematically identifies many of the major generic functions performed by a membrane
in a reactor. Theyincludes(Sirkar et al., 1999);
• Separation of products from the reaction mixture
• Separation of a reactant from the mixed-gas stream for introduction into the
reactor
• Controlled addition of one or more reactants
• Nondispersive phase contacting (with reaction at the interface or in the bulk
phases)
• Segregation of a catalyst (or a cofactor) in a reactor
• Immobilization of a catalyst in (or on) a membrane
• Membrane is the catalyst
• Membrane is the reactor
• Solid-electrolyte membrane supports the electrodes, conducts ions, and achieves
the reactions on its surfaces
52
• Transfer of heat
• Immobilizing the liquid reaction medium
Figure 3.2 Schematic of possible functions of a membrane in a reactor
In a study by Ohya et al. (1997), an MR using a SiO
2
, water-permeable membrane
together with a commercial 0.5% Ru/alumina catalyst was investigated at somewhat
elevated pressures (0.2 Mpa) (higher pressures result in higher equilibrium conversions).
In the experiments, H
2
and CO
2
mixtures (H
2
/CO
2
feed ratio in the range of 1-5) were fed
on one side of the membrane over the catalyst. The H
2
O produced during the reaction
permeated selectively through the membrane, hence resulting in increased conversion. An
53
interesting MR application to CO
2
capture was reported by Nishiguchi et al., in which
CO
2
was catalytically hydrogenated to produce CH
4
and water. The CH
4
was then fed into
an MR utilizing a Pd membrane and operating at 500
o
C,
to be converted over a Ni/SiO
2
catalyst into graphitic carbon and hydrogen. In this two-stage reactor system, over 70%
of the CO
2
was converted into graphitic carbon.
3.3 Experimental set-up
A schematic of the lab-scale experimental apparatus is shown in Figure 3.3. The
reactor is made of stainless steel and has an internal diameter of 3.175 cm, and a length of
25.4 cm (when it operates as an MR, a membrane is inserted in the middle of the reactor,
and the catalyst is packed in the volume exterior to the membrane; for kinetic rate
measurements no membrane is present and the full reactor volume is occupied by the
catalyst, diluted by inert quartz particles). The reactor is heated by a three-zone furnace
using three individual temperature controllers and three thermocouples installed at three
different locations in the bed. Back-pressure regulators are used to control the pressure.
The gases exiting the reactor flow through a condenser and a moisture-trap (in order to
separate out the moisture), and their flow rates are then measured by a bubble flow-meter.
A small slip-stream from the reactor was analyzed for its composition with a mass
spectrometer. Synthetic gas feed mixtures are prepared from gas cylinders using mass
flow controllers.
54
Figure 3.3 Schematic of the experimental set-up
3.4 Catalyst
The catalyst utilized is a commercial Ni-based catalyst (KATALCO 11-4),
purchased in the form of cylindrical pellets (5.4 mm in diameter and 3.6 mm length) from
the Johnson-Matthey Co. The bulk density of the pellet was measured as 1180 kg/m
3
. The
catalyst contains 35 wt. % NiO, 4 wt. % MgO and various proprietary promoters
combined with a calcium aluminate cement support. For each experimental run, a certain
amount of crushed catalyst (with particle sizes in the range of 850~1,000 µm) were
mixed with inert quartz particles of the same size, at an amount sufficient to fill the
reactor space. Prior to the catalytic reaction, the catalyst was pretreated for 4 h with H
2
and steam diluted with N
2
at 3 atm and 300
o
C.
55
3.5 Experimental reaction kinetics
Finding the parameters of kinetic equations (kinetic parameters) is a necessary
and important step in the construction of a kinetic model for a complex heterogeneous
catalytic reaction. The resulting kinetic model can be used in the design and optimization
of industrial-scale reactors and chemical technologies. The kinetic parameters for a
chosen kinetic model are most often derived from the results of experiments carried out
under steady-state conditions in either perfectly mixed (gradientless ) or in plug-flow
reactors. For this purpose, it is important to use experimental data obtained under kinetic
control, when the reaction rate is independent of external and internal heat and mass
transfer limitations.
3.5.1 Kinetics at lower temperature (Range I, T (
o
C) < 300)
The catalytic reaction between CO
2
and H
2
over a catalyst was investigated in a
reactor, under isothermal conditions. The following range of experimental conditions
were studied: temperature from 225 to 300
o
C, total pressure from 1 to 3 atm, feed
composition (H
2
to CO
2
ratio) from 4:1 to 5:1, and total flow rate from 2.4 to 12.0 ml/s.
Before investigating the kinetics, blank tests without catalyst were carried out to see
whether the membrane acts as catalyst for the reaction; no effect of the membrane (or the
reactor walls) on the reaction was observed. In between each run (for a given set of
conditions), the catalyst bed was treated for 20 min in H
2
in order to maintain its activity
(however, this was a preventive measure, and no catalyst deactivation was observed in
any of the runs). Typical experimental results, in terms of CO
2
conversion vs.
2
/
cCO
Wn (
c
W is the total weight of the catalyst),
are shown in Figures 3.4 ~ 3.6. The
56
isothermal reaction data were analyzed using the following, Hougen-Watson type rate
equation (Phungquach and Rouleau, 1976; Rotaru and Blejoiu, 2001),
22 2 2
22 2 2
44
5
(1 )
(1 )
CO H CO H
CO CO H H
kK K P P
r
KP K P
β =−
++
(3.1)
where the approach-to-equilibrium coefficient β is defined as
42
22
2
4
()
1
()
CH H O
eq CO H
PP
KP P
β
⋅
=
⋅
(3.2)
In the rate expression (3.1), k is the reaction rate constant, P
i
the partial pressure
of component i, K
i
the surface adsorption equilibrium constant for component i, and K
eq
the overall Sabatier reaction equilibrium constant. The constants in the rate equations
were estimated using nonlinear least-square fitting of the experimental data (the fitted
data lines are also shown in Figures 3.4-3.6). The following three equations provide the
temperature dependence of the rate constant k and the adsorption equilibrium constants
2
CO
K and
2
H
K , calculated by fitting the experimental data.
11
113497.4
1.064 exp( ) kE
RT
−
= (3.3.a)
2
7
69691.8
9.099 exp( )
CO
KE
RT
−
= (3.3.b)
2
4
39942.0
9.6104 exp( )
H
KE
RT
−
= (3.3.c)
57
W
c
/n
CO2
[ g/(mol CO
2
/h)]
0 20406080 100
CO
2
Conversion
0.0
0.2
0.4
0.6
0.8
1.0
225
o
C
250
o
C
300
o
C
Fit
Figure 3.4 CO
2
conversion as a function of
2
c
CO
W
n
for various temperatures (Range I)
(H
2
/CO
2
= 5.0 and P =1atm)
58
W
c
/n
CO
2
(g/(mol CO
2
/h))
0 2040 6080 100
CO
2
conversion
0.0
0.2
0.4
0.6
0.8
1.0
H
2
/CO
2
= 4.0
H
2
/CO
2
= 5.0
Fit
Figure 3.5 CO
2
conversion as a function of
2
c
CO
W
n
for various ratios of H
2
to CO
2
(T = 250
o
C and P = 1 atm)
59
W
c
/n
CO
2
(g/(mol CO
2
/h))
0 2040 6080 100
CO
2
conversion
0.0
0.2
0.4
0.6
0.8
1.0
1 atm
2 atm
3 atm
Fit
Figure 3.6 CO
2
conversion as a function of
2
c
CO
W
n
for various reactor pressures
(T = 250
o
C and H
2
/CO
2
= 5.0)
Practical membranes (including the CMS membranes utilized in this study) will
permit some of the O
2
to leak into the catalyst bed. The effect of oxygen on reactor
conversion, therefore, was also studied. Figure 3.7 presents the reactor conversion as a
function of the oxygen concentration in the feed (up to 2%) for two different
temperatures: 250 and 300
o
C. During the experiments the molar flow rate of CO
2
was
kept constant while the O
2
concentration (synthetic air is used in these experiments) in
the feed was varied. O
2
seems to not have much of an effect on reactor conversion. Figure
60
3.8, for example, shows a 40 h test in the presence of 2 % O
2
. No apparent catalyst
deactivation is evident.
O
2
concentration in feed (%)
0.00.5 1.01.5 2.0
CO
2
conversion
0.4
0.5
0.6
0.7
0.8
250
o
C
300
o
C
Figure 3.7 Effect of O
2
concentrations in the feed on the catalyst activity (H
2
/CO
2
= 4.0,
P = 1 atm,
20
/
cCO
Wn at 250
o
C = 38.9 g·h/mol and
20
/
cCO
Wn at 300
o
C = 15.6 g·h/mol)
61
Time (hr)
0 10203040 50
CO
2
Conversion
0.4
0.5
0.6
0.7
0.8
250
o
C
300
o
C
Figure 3.8 CO
2
conversion with 2 % of O
2
in the feed as a function of time on stream
(H
2
/CO
2
= 4.0, P = 1 atm,
20
/
cCO
Wn at 250
o
C = 38.9 g·h/mol and
20
/
cCO
Wn at 300
o
C =
15.6 g·h/mol)
3.5.2 Kinetics at higher temperature (Range II, 325 < T (
o
C) < 350)
Diffusion of species through the pores of solid catalysts is very often the rate-
limiting step in catalytic reactions. Diffusion limitations also have a very important effect
on the selectivity in multiple reaction systems. Since the kinetic rate expressions are
usually nonlinear, it is usually very difficult to determine the effect of transport processes
on the observed rates by solving the controlling differential equations of the model.
62
A number of criteria were derived to predict the importance of diffusion
limitations on the overall rate of catalytic reactions (Bischoff, 1967; Guha and
Narsimha.G, 1972; Hutching.J and Carberry, 1966; Petersen, 1965; Schneide.P and
Mitschka, 1966; Weisz and Prater, 1954). The criterion developed by Hudgins is
applicable for reactions having rate expressions which are different from power-law type
(Hudgins, 1968).
At low temperatures, reaction rates are generally kinetically controlled. At higher
temperatures, they can become sufficiently fast for diffusion effects to become important.
Under these conditions, diffusion controls the overall rate of reaction. In so-called
Arrhenius plots this is nicely seen as a changing slope of the rate versus 1/T dependency,
which can be used as an indication for the presence of limitations. Good examples of this
behavior can be found in the literature, for example the Fischer-Tropsch synthesis of
middle distillates (Post et al., 1989) and the catalyzed gasification of carbon (Bernardo
and Trimm, 1979). This also demonstrates the applicability of the theory. Diffusion
limitations can be overcome to some extent by the use of a smaller particle size to
provide a higher geometric surface area and enhance mass transfer.
Changing activation energies are, however, not always indicator for the presence
of limitations. The approach of thermodynamic equilibrium in the case of exothermic
reactions can also cause this phenomenon, as for hydrogenation reactions (Bernardo and
Trimm, 1979). Also, changes in the rate determining steps and catalyst deactivation might
be causes.
In this study, the catalytic reaction was investigated at higher temperatures under
63
isothermal conditions. The following range of experimental conditions were studied:
temperature from 300 to 400
o
C, total pressure 1 atm, feed composition (H
2
to CO
2
ratio)
from 4:1 to 5:1, and total flow rate from 2.5 to 18.0 ml/s. The experimental results are
shown in Figures 3.9 ~ 3.10. Significant reduction of activities was observed at the
temperature above 375
o
C.
At low temperatures (Range I), the observed reaction rate constant can be
calculated from the conversion in the absence of transport limitation and the Arrhenius
plot results in straight lines of which the slope corresponds with the apparent activation
energy of 113.5 KJ/mol. At higher temperatures (Range II) this value deviates from the
straight line and this deviation is caused by internal diffusion limitation. The apparent
activation energy derived from the slope within Range II of Figure 11 is 11.3 KJ/mol. and
is significantly lower than 113.5 KJ/mol.
11262.5
104.2exp( ) k
RT
−
= (3.4.a)
2
3
33023.9
2.190 exp( )
CO
KE
RT
−
= (3.4.b)
2
2
18112.5
5.916 exp( )
H
KE
RT
−
= (3.4.c)
64
W/F
CO2
(g-cat/(mol CO
2
/hr))
0 10203040506070
CO
2
conversion
0.0
0.2
0.4
0.6
0.8
1.0
300
o
C
335
o
C
350
o
C
Fit
Figure 3.9 CO
2
conversion as a function of
2
c
CO
W
n
for various temperatures (Range II)
(H
2
/CO
2
= 4.0 and P =1atm)
65
W/F
CO2
(g-cat/(mol CO
2
/hr))
0 10203040506070
CO
2
conversion
0.0
0.2
0.4
0.6
0.8
1.0
300
o
C
325
o
C
335
o
C
350
o
C
Fit
Figure 3.10 CO
2
conversion as a function of
2
c
CO
W
n
for various temperatures (Range II)
(H
2
/CO
2
= 5.0 and P =1atm)
66
1/RT
0.00019 0.00020 0.00021 0.00022 0.00023 0.00024 0.00025
ln (k)
-4
-2
0
2
4
6
8
T (K)
500 550 600
Range I Range II
~586 K (313
o
C)
Figure 3.11 Arrhenius Plots: Range I and II
67
Chapter 4 Mathematical Modeling
4.1 Introduction
In this Chapter, the mathematical model developed for describing the membrane
reactor for the CO
2
methanation reaction is presented. This chapter is divided into two
parts. In part (A) the isothermal modeling is discussed. In part (B) the nonisothermal
model is described.
4.2 Isothermal modeling
An isothermal mathematical model was developed in order to predict the behavior
of the membrane reactor at steady state. A schematic of the MR is shown in Figure 4.1. In
the MR, the catalyst is packed in the exterior of the membrane (signified by the
superscript F, or the feed side), without any catalyst being present in the membrane
interior (signified by the subscript P, or the permeate side). To simplify the analysis, it is
assumed that the reactor operates isothermally under plug-flow, ideal gas law conditions,
that external mass-transfer resistances are negligible for the membrane as well as for the
catalyst, and that catalyst internal diffusion limitations are accounted for by the overall
rate coefficients.
68
Figure 4.1 A schematic of MR system
Mass transfer through the membrane is described by the following equations:
(( ) ( ) )
Fn P n
jj j j
FU P P =− (4.1)
where F
j
is the molar flux,
F
j
P the partial pressure of component j on the feed side,
P
j
P the partial pressure of component j on the permeate side, U
j
the membrane permeance
for component j and n the pressure exponent. For porous membranes, like the CMS
membranes, the pressure exponent is assumed to be 1, while exponent values in the range
of 0.5 and 1.0 are measured for the Pd-based membranes. It is also assumed that Pd
membranes are highly permeable to H
2
, and are virtually impermeable to other gases.
The mass balance for the reactor packed with methanation catalyst is described by
the following equations for CO
2
, H
2
, CH
4
, H
2
O, O
2
, and N
2
:
69
Feed (reactor) side: (( ) ( ) ) (1 )
F
j Fn P n
mj j j j v c c
n
UP P r
V
α υεβρ
∂
=− − + −
∂
(4.2)
Permeate (sweep) side: (( ) ( ) )
P
j Fn P n
mj j j
n
UP P
V
α
∂
=−
∂
(4.3)
j=1, 2….6
at 0,
jjo
Vnn = =
In Equations (4.2) and (4.3),
F
j
n is the molar flow rate for species j on the feed
side,
P
j
n the molar flow rate for species j on the permeate side, V the reactor volume
variable,
m
α the membrane area per feed-side reactor volume,
j
υ the stoichiometric
coefficient of component j,
v
ε the bed porosity on the feed side,
c
β the fraction of the
solid volume occupied by catalyst (
c
β = 1 when no inert quartz particles are present),
c
ρ the catalyst density, and r the overall Sabatier reaction rate.
The pressure drop in the feed side is calculated using the Ergun equation (the
pressure drop in the permeate side is ignored),
2
6
()
10
F F
FF
cP
fG dP
dV g A d ρ
−
−= (4.4)
0
at 0,
FF
VPP ==
3
Re
1 150(1 )
1.75
()
vv
F
v
f
N
εε
ε
⎛⎞⎛ ⎞ −−
=+
⎜⎟⎜ ⎟
⎝⎠⎝ ⎠
(4.5)
Re
500(1 )
F
v
N ε <−
Re
/
FFF
P
NdG μ =
where
F
P
is
the feed side pressure,
0
F
P the inlet feed side pressure,
F
A the cross-
70
sectional area for the feed side,
F
μ the viscosity,
P
d the particle diameter in the feed side,
FFF
Gu ρ = the superficial mass flow velocity in the feed side, with
F
u being the average
velocity of the fluid, and
F
ρ the average density, and
c
g the gravity conversion factor
equal to 1 in SI units.
The CO
2
conversion is defined by the following equation:
20 2 2
2
20
()
ex ex
FF P
CO CO CO
CO F
CO
nn n
X
n
−+
= (4.6)
where
20
F
CO
n is the inlet molar flow rate of CO
2
and
2ex
F
CO
n and
2ex
P
CO
n are the CO
2
molar
flow rates at the exit of the reactor feed and permeate sides correspondingly.
4.3 Nonisothermal modeling
A nonisothermal mathematical model was also developed in order to describe the
behavior of the membrane reactor based on the following assumptions:
(1) Plug-flow conditions with negligible radial temperature/concentration gradients
(2) Constant bed porosity in the axial and radial directions
(3) Catalytically inactive membranes
(4) Ideal gas behavior
In this model, gas permeances through CMS membranes are assumed to be
constant with the reactor temperature. Equation (4.1) is replaced by the following
equation for the Pd membranes which incorporates the temperature dependence on
temperature:
71
22 2 2 2 2 2
/
,
(( ) ( ) ) (( ) ( ) )
p
ERT
Fn Pn Fn Pn
HHo H H H H H
FU e P P U P P
−
=−= − (2.11)
where
2
H
F [ mol/m
2
·s] is the H
2
permeation flux,
2
F
H
P [atm] its partial pressure in the feed-
side,
2
P
H
P [atm] its pressure on the permeate-side, n the pressure exponent,
2
H
U [ mol/m
2
·atm·s] the membrane H
2
permeance, U
H2,o
[mol/m
2
·atm·s] the pre-
exponential factor, and E
p
[J/mol] the activation energy for transport.
Equations (4.2)-(4.5) used for the isothermal model still apply. One, in addition,
requires the following energy balance equations, expressed as:
Feed-side:
() ( )(1) ( )( )( )
F
FFPFP
jPj R v c c m m j Pj
T
nC H r U T T T T FC
V
εβρ α γα
∂
=−Δ − − − + −
∂
∑∑
(4.7)
Permeate-side:
() ( ( )( )((1) ))
P
PFPFP
jPj m m j Pj
T
nC U T T T T FC
V
αγα
∂
=−+− −
∂
∑∑
(4.8)
, at 0
FF P P
oo
TT T T V == = ( γ = 1 for F
j
< 0 and γ = 0 for F
j
> 0)
where
Pj
C [J/mol·K] is the heat capacity of component j,
R
H Δ [J/mol] the heat of the
Sabatier reaction, U [J/s·m
2
·K] the overall heat transfer coefficient through the membrane,
F
T
[K] the feed-side temperature and
P
T [K] the permeate-side temperature.
The rate for the Sabatier reaction in the above equation is described by the
Hougen-Watson type rate expression discussed in Chapter 3:
22 2 2
22 2 2
44
5
(1 )
(1 )
CO H CO H
CO CO H H
kK K P P
r
KP K P
β =−
++
(3.1)
where the “approach-to-equilibrium” coefficient β is defined as
72
42
22
2
4
()
1
()
CH H O
eq CO H
PP
KP P
β
⋅
=
⋅
(3.2)
exp( )
a
o
E
kk
RT
=− ;
2
22
,
exp( )
H
HHo
E
KK
RT
=− ;
2
22
,
exp( )
CO
CO CO o
E
KK
RT
=− (4.9)
where k [mol/g·s] is the reaction rate constant, P
i
[atm] the partial pressure of component
i, K
i
[atm
-1
] the surface adsorption equilibrium constant for component i, and K
eq
[atm
-2
]
the overall Sabatier reaction equilibrium constant.
The overall heat-transfer coefficient U for the tubular membrane is given as
1
1/ ( / )ln( / ) ( / )/
P im o i i o F
U
hrk rr rr h
=
++
(4.10)
where k
m
[J/s·m·K] is the thermal conductivity of the porous ceramic membrane, h
F
and
h
P
[J/s·m
2
·K ] are the heat transfer coefficients of the feed-side and the permeate-side
respectively. r
i
[m] and r
o
[m] are inner diameter and outer diameter of the membrane.
The membrane utilized in this study consists of a macroporous ceramic substrate with a
thin film layer deposited on it. In Equation (4.10), the thermal conductivity is taken to be
equal to that of the ceramic substrate, since the thermal conductivity of the film is higher
than that of the ceramic substrate and the thickness of the ceramic support is typically
more than two orders of magnitude greater than that of the thin film layer. The thermal
conductivity of a porous material is strongly dependent of the porosity of the material.
The theoretical model proposed by Aivazov and Domashnev (Aivazov and Domashnev,
1968), well correlates the thermal conductivity of porous ceramics with the porosity
according to Equation (4.11):
2
,
1
1
m
mo
k
km
ε
ε
−
=
+
(4.11)
73
where k
m,o
[J/s·m·K] is the thermal conductivity of a porous free ceramic material, ε is
the porosity, and m is a constant, which is taken to be equal to 3.0 as regarded in
Sugawara and Yoshizawa (Sugawara and Yoshizawa, 1962). h
F
and h
P
can be estimated
from the following equations (Stephan, 1959; Wakao and Funazkri, 1978).
1/3 0.6
Pr Re
()[2 1.1( ) ( ) ]
FF F
F
p
k
hNN
d
=+ ,
Pr
FF
F P
F
C
N
k
μ
≡ (4.12)
1.33
Re Pr
0.3
Pr Re
0.0677( (2 / ))
( )[3.657 ]
210.1((2/))
PP
i F
P PP
ii
NN r L k
h
rNNrL
=+
+
,
Re
2
P
P i
P
rG
N
μ
≡ ,
Pr
P P
P P
P
C
N
k
μ
≡ (4.13)
where
Pr
N is the Prandtl number, and
Re
N is the Reynolds number.
Equations (4.1)-(4.9) can be written in dimensionless form by defining the
following variables and groups:
FF
ojo
j
nn =
∑
;
P P
ojo
j
nn =
∑
;
FF F
oj
j
nn y ≡
∑
;
P FP
oj
j
nn y ≡
∑
F
j F
j F
o
n
y
n
≡ ;
P
j P
j F
o
n
y
n
≡ ;
R
V
V
η ≡
F
F
F
o
P
P
Φ≡ ;
P
P
F
o
P
P
Φ≡ ;
F
F
ref
T
T
Γ≡ ;
P
P
ref
T
T
Γ≡
F
j FF
j F
j
j
y
P
y
=Φ
∑
;
P
j P P
j P
j
j
y
P
y
=Φ
∑
;
P
o
F
o
n
S
n
≡
* R
R o
R
H
H
H
Δ
Δ≡
Δ
;
2
2
2
* H
H o
H
U
U
U
≡ ;
2
* Pj
Pj o
PH
C
C
C
≡ ;
*
o
U
U
U
≡ ;
* eq
eq o
eq
K
K
K
≡ ;
*
o
r
r
k
≡
F
F
F
o
u
u
ξ ≡ ;
F
j F
j F
o
P
x
P
≡ ;
2
j
j
H
MW
MW
α ≡ ;
2
2
6
()
10
F
oHR
FF
cP ref
fuMWV
gA d RT
−
Ξ≡
Γ
74
a
a
ref
E
RT
γ ≡ ;
2
2
H
H
ref
E
RT
γ ≡ ;
2
2
CO
CO
ref
E
RT
γ ≡ ;
2
()
oF
eq o
KP κ ≡ ;
42
22
2
2
*2 4
()
()
()
() ( )
F
FF j
CH H O j
FF F
eq CO H
y
yy
Kyy
β
κ
=
Φ
∑
exp( )
a
o
kk
γ
=−
Γ
;
2
22
,
exp( )
H
HHo
KK
γ
=−
Γ
;
2
22
,
exp( )
CO
CO CO o
KK
γ
=−
Γ
()
o
RRref
HHT Δ≡Δ ;
22
()
o
PH PH ref
CC T ≡ ; ()
o
ref
kkT ≡ ;
22
()
o
H Href
UU T ≡
,,
(, , )
oFP
ref o base o base
UUT n n ≡ ; ()
o
eq eq ref
KK T ≡
2
0
()
F
oFn
mR H o
n
Pe
VU P α
≡ ;
(1 )
o
vc cR
F
o
kV
Da
n
εβρ −
≡ ;
2
o
mR
Fo
oPH
VU
St
nC
α
≡ ;
2
o
R
o
ref PH
H
B
TC
−Δ
≡ (4.14)
The dimensionless equations corresponding to Equations (4.2)-(4.8) are
22 2
*
*
(( ) ( ) )
FP F
HH H j Fn P n
FP
jj
jj
Uy y y
Da r
Pe y y η
∂
=− Φ − Φ + ⋅
∂
∑∑
(4.15)
22 2
*
(( ) ( ) )
FP P
HH H j Fn P n
FP
jj
jj
Uy y y
Pe y y η
∂
=Φ− Φ
∂
∑∑
(4.16)
1, 0
FP
jj
yy Sat η == =
2
()
F
FF F
jj
x ξ α
η
∂Φ
=−Ξ Φ
∂
∑
, 1at 0
F
η Φ== (4.17)
22 2 2
** *
**
*
*
()
()
(( ) ( ) )( )
F
FP R
FF
jPj j Pj
jj
oF P
HPH H H
Fn P n F P
FF P
jPj j j
jj j
HDaBr StU
yC yC
UC y y
Pe y C y y
η
γ
ΔΓ⋅ ⋅ ⋅ ∂Γ
=− Γ−Γ
∂
+Φ−ΦΓ−Γ
∑∑
∑∑ ∑
(4.18)
75
22 2 2
*
*
*
*
()
(1 )
(( ) ( ) )( )
P
FP
P
jPj
j
oF P
HPH H H
Fn P n F P
PF P
jPj j j
jj j
StU
yC
UC y y
Pe y C y y
η
γ
∂Γ
=Γ−Γ
∂
−
+Φ−ΦΓ−Γ
∑
∑∑ ∑
(4.19)
, at 0
FF P P
oo
η Γ=Γ Γ =Γ =
where
ref
T is
the reference temperature and
F
o
Γ and
P
o
Γ are the dimensionless inlet
temperatures of the feed- and the permeate-side respectively.
76
Chapter 5 Results and discussions
5.1 Introduction
In this chapter, the MR experimental and modeling results are described and
discussed. As previously noted, the CO
2
methanation reaction was experimentally
investigated using a commercial Ni-based catalyst, and hydrogen-permselective
nanoporous CMS membranes and supported palladium membranes. The isothermal and
nonisothermal models presented in Chapter 4 are used to analyze experimental data of
relevance for ARS and ISRU applications respectively. With the aid of this model the
effect of the various membrane and other parameters is discussed. The performance of the
MR is compared with that of the more conventional PFR system.
5.2 MR-Based Air Revitalization System (MARS)
As discussed previously, for human presence on Mars or other planets to become
a reality, the development of highly reliable and efficient systems is required, which
provide basic life support provisions, such as food, water, and air for the crews during the
long space flights. A key function space life-support systems perform is the removal of
metabolic CO
2
from the atmosphere in the living quarters; otherwise, the CO
2
levels in
the closed cabin environment would rise to unacceptable levels. Carbon dioxide in earth's
atmosphere is considered a trace gas currently occurring at an average concentration of
about 385 ppm. At high concentrations, however, CO
2
begins to cause respiratory
difficulties (ASHRAE standard 62-1999, Ventilation for acceptable indoor air quality). As
77
noted in Chapter 1, the original Air Revitalization System (ARS) in the Freedom Space
Station was designed, for example, to maintain CO
2
levels at 2000 ppm or less; the life-
support system for the space suit worn during Extra-Vehicular Activities is designed to
maintain CO
2
levels at or below 5000 ppm.
As discussed in Chapter 1, different methods for CO
2
removal have been studied
for use in space habitats and suits. CO
2
adsorption by LiOH, for example, is the current
standard approach, but for long-term flights the method presents challenges, since LiOH
is not easily regenerable. Other similar systems, based on CO
2
chemisorption, such as
those using silver oxide, and solid phase amines, also face the same challenge (Boryta
and Maas, 1971; Hart et al., 1992; Khatri et al., 2006). Molecular sieves, such as zeolites,
can trap CO
2
in their pores by physical adsorption instead, and are more amenable to
regeneration, e.g., via pressure-swing adsorption (PSA), at a cost, nevertheless, of greater
process complexity (Chou and Chen, 2004; Walton et al., 2006). Supported liquid-phase
membranes (SLM) have a promise for CO
2
removal from the air in the life support
system but, one disadvantage of SLM’s is their instability due mainly to loss of the
membrane liquid (Okabe et al., 2007; Steinwandel et al., 1999). A membrane-based
absorption system using hollow fibers has been discussed as a promising way of CO
2
removal for spacecraft application. However, though good CO
2
selectivity was obtained
for this process, generally the permeation rates achieved were rather low (Eckhard et al.,
1996; Feron et al., 1997).
The challenge for all these systems, however, particularly in environments where
no source of O
2
is available, is that CO
2
removal is not accompanied by O
2
loss. In this
78
context, an approach that has been proposed to remove CO
2
, without complete loss of the
O
2
, is through the use of the methanation (Sabatier) reaction, in which CO
2
reacts
with
H
2
to produce CH
4
and water,
22 4 2
42 CO H CH H O +←⎯→+ ( 165.4 / ) HkJmol Δ =− (1.1)
A conventional ARS system was discussed in Chapter 1. A schematic of a more
advanced ARS is shown in Figure 5.1, where a CO
2
Reduction Subsystem (CRS) is now
added to the conventional ARS shown in Figure 1.1. In the CRS, CO
2
, which is recovered
from the CO
2
removal subsystem, catalytically reacts with H
2
, which is co-generated with
O
2
from the OGS, and water and CH
4
are the products of the reaction. Water then flows
to the water recovery system and eventually is fed to the OGS for O
2
generation. CH
4
,
another product of the CRS, is vented overboard along with unreacted CO
2
, H
2
and other
gases. By adding the CRS subsystem, some of the O
2
in CO
2
can be recovered as water.
Figure 5.2 shows an advanced ARS whereby the CO
2
removal subsystem and the
CO
2
reduction subsystem in Figure 5.1 are replaced by the membrane reactor. This MR-
based ARS (MARS) is envisioned to convert CO
2
and H
2
into water and CH
4
, with its
design depending on what one does with the CH
4
produced. A number of investigators
have, in recent years, both in the context of In-situ Resource Utilization (ISRU) but also
in CO
2
sequestration applications, studied the decomposition of CH
4
into carbon and H
2
(Jung et al., 2007; Venugopal et al., 2007).
42
2 ( 75.6 / ) CH C H H KJ mol →+ Δ = (5.1)
In the MARS in Fig. 5.2 another membrane reactor is added to recover H
2
from the CH
4
that is decomposed through reaction 5.1. For CO
2
removal reactor, a CO
2
permselective
79
membrane is required, while a H
2
permselective membrane is needed for the reactor in
the CH
4
decomposition subsystem.
Figure 5.1 An ARS with a CO
2
reduction subsystem
80
Figure 5.2 An ARS with membrane reactor and hydrogen recovery system
Though the MARS presented in Fig. 5.2 is ideal, since it allows for the complete
recovery of hydrogen (barring incidental losses) the operation of the second Pd
membrane-based MR is rather challenging in a space-flight environment, as it requires
high temperatures and elaborate reactor designs (e.g., fluidized beds) to deal with the
produced carbon, and is more appropriate for ISRU rather than for ARS applications.
Here, we assume, therefore, that no secondary conversion of CH
4
takes place, and that the
CH
4
is vented instead overboard. This simpler MARS is shown in Figure 5.3, and is
proposed as part of a “closed-loop,” life-support system, designed to control the humidity,
CO
2
, and CH
4
levels in the cabin, and to minimize the initial required supply of oxygen.
81
The system comprises of the MR for removing CO
2
, the oxidation systems for both sides
of the MR (to convert H
2
or CH
4
into water and/or CO
2
), and the water condensation and
electrolysis systems. In the reactor, CO
2
, which is permeated from the air, catalytically
reacts with H
2
, which is generated (together with O
2
) from the water electrolysis system.
The permeate effluent contains the unreacted H
2
and CO
2
, the CH
4
and H
2
O products
from the reaction, and O
2
that has permeated through the membrane. The H
2
is further
oxidized in the H
2
oxidation system with the aid of additional supplemental O
2
, while the
H
2
O produced is condensed, and the unreacted CO
2
, CH
4
and O
2
, as well as some of the
H
2
O that remains in the vapor phase are vented overboard.
Figure 5.3 A schematic of the proposed MARS
82
One person needs, typically, ~850 g of O
2
/day, and exhales roughly 1 kg of
CO
2
/day [29]. For space travel, the oxygen is conventionally produced using an
electrolyzer, which splits water into O
2
and H
2
. When adsorption on metal hydroxides is
used to remove the exhaled CO
2
, 6 astronauts will need ~2.1 tons of water annually for
their oxygen supply. Without any oxygen recovery (through the Sabatier reaction), there
is no use for the H
2
and CO
2
produced, and 2.2 tons of CO
2
, and 233 kg of H
2
must,
therefore, be vented into space.
Interestingly, absent reaction (1.1) to recover the H
2
, an optimal conversion for
reaction (1.1) exists. For every mol of CO
2
one produces, one consumes on the average
1.168 mol of O
2
. Therefore, 2.236 mol of H
2
O (per mol of CO
2
exhaled) are needed to
provide, through electrolysis, the O
2
required for breathing. Only part of this H
2
O can be
recovered through the Sabatier reaction. The limitation here arises from the fact that the
H
2
required must also be produced on board by water electrolysis. This limits the
Sabatier reaction conversion to 58.4%, and implies that 1.168 mol of H
2
O stored aboard
must be used to produce the H
2
utilized in the Sabatier reaction. On the other hand, using
the Sabatier reaction means that 1.168 mol of H
2
O (per mol of exhaled CO
2
) are saved,
corresponding to 1,046 Kg of H
2
O annually for the 6 astronauts example (these
calculations ignore any other incidental losses of O
2
and H
2
O overboard). During the
simulations, we maintain the CO
2
:CH
4
composition ratio in the stream exiting the reactor
side equal to 41.6:58.4, corresponding to a conversion (of the CO
2
transported from the
tube side) equal to 58.4%.
83
5.3 MR experiments (MARS)
To validate the model used in the design simulations, and the applicability of the
rate expressions, a series of MR experiments with a CMS membrane were also carried
out (the low conversions are due to limitations with the size of our laboratory system,
which accommodates only one CMS membrane). MR experiments were carried out in
which the feed gas consisted of a mixture of CO
2
and N
2
. The flow rate of N
2
in the feed
was varied while that of CO
2
was set to 3.21 x 10
-2
(mol/h) to investigate the effect of
CO
2
concentration on the MR conversion, as shown in Figure 5.4 (other experimental
conditions are indicated in the figure captions). Figure 5.5 shows the effect of feed side
pressure. Shown in the same figures are the simulation results using Equations (4.2)-(4.5),
the experimental reaction rates (Equations (3.1)-(3.3)), and the experimentally measured
mixed-gas membrane permeances. In order to determine the permeation characteristics of
the membrane used in this simulation, a series of experiments were carried out using the
mixtures of gases consisting of CO
2
, H
2
, CH
4
, H
2
O and N
2
and those data is listed in
Table 5.1. The gas permeances generally depend on the mixed-gas compositions. The
error bars in the simulated values are due to the experimental uncertainty in the transport
measurements for the membrane properties. Agreement between the experiments and the
model predictions is satisfactory, particularly given the various simplifying assumptions
in the model.
84
CO
2
concentration in feed (%)
0 5 10 15 20
CO
2
Conversion
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
250
o
C Experimental
250
o
C Simulated
300
o
C Experimental
300
o
C Simulated
Figure 5.4 Effect of CO
2
concentration on its conversion (
2
/
cCO
Wn = 405 g·h/mol,
S (H
2
) = 0.13, P
F
= P
P
= 1 atm)
85
Feed side pressure, P
F
(atm)
1.0 1.5 2.0 2.5 3.0 3.5 4.0
CO
2
Conversion
0.00
0.05
0.10
0.15
0.20
0.25
0.30
250
o
C Experimental
250
o
C Simulated
300
o
C Experimental
300
o
C Simulated
Figure 5.5 Effect of the pressure in the air feed side on CO
2
conversion (
2
/
cCO
Wn = 405
g·h/mol, S (H
2
) = 0.13, CO
2
% = 10, P
P
= 1 atm)
86
87
5.4 Simulation results (MARS)
The MARS proposed here (see Figure 5.6) is ideally suited for membranes that
are only permeable to CO
2
, and allow no other fixed gases to permeate through, since as
previously noted that significantly simplifies the system design.
The membrane properties,
used in the simulations, instead, are more typical of CMS membranes developed
collaboratively between USC and Media and Process Technology. They show relatively
high permeance towards H
2
, CO
2
and H
2
O, but mostly exclude other fixed gases (e.g.,
CH
4
, CO, N
2
, etc.) from permeating through. The isothermal model described in Chapter
4 is used to simulate the behavior of the MR in ARS system due to the small fraction of
CO
2
in the reactor composition. The range of performance characteristics for the
membrane utilized in the simulations are listed in Table 5.2, which also contains values of
the other parameters (as well as their range) used in the simulations. Typically, values of
all the parameters (other than the membrane area, catalyst weight and the sweep ratio) are
fixed. The sweep ratio is then varied, and the required weight of the catalyst and
membrane area are calculated for the reactor to convert an amount of CO
2
equal to that
exhaled by the crew, while maintaining the conversion of the CO
2
transported to the shell
side equal to 58.4%.
88
Figure 5.6 A schematic of the MR-AR system
89
Table 5.2 Base case and the range of conditions used in the simulations (MARS)
Parameter Base Value Range of Values
reactor side pressure (P
F
), atm 1.07 -
air feed side pressure (P
P
), atm 1.00 -
reactor temperature (T),
o
C 250 225 - 300
feed flow of air (F
Po
), (m
3
/h) 180 -
CO
2
permeance (U
CO2
), m
3
/(m
2
·bar·h) 1.2 0.8 – 5.0
CO
2
/H
2
separation factor 5.0 2.5 - ∞
H
2
/CH
4
separation factor 60.0 -
H
2
/H
2
O separation factor 1.0 -
H
2
/N
2
separation factor 79.4 -
H
2
/O
2
separation factor 84.9 -
CO
2
concentration in air feed (y
CO2,P0
), ppm 5,000 3,000 – 5,000
CH
4
concentration in air feed (y
CH4,P0
), ppm 1,000 -
H
2
O concentration in air feed (y
H2O,P0
), ppm 8,000 -
H
2
O concentration in the sweep (y
H2O,F0
), ppm 8,000 -
catalyst density ( ρ
c
), g/m
3
1.18 x 10
6
-
length of membrane (L), m 1.0 -
inner diameter of membrane (D
i
), m 0.0035 -
outer diameter of membrane (D
o
), m 0.0057 -
cross-sectional flow area (A
F
), m
2
0.14 -
90
Figure 5.7 shows the effect that varying the CO
2
permeance has on the MARS
performance. Increasing the permeance significantly decreases the membrane area
requirements. This is because, for the same sweep ratio and other conditions, a more
permeable membrane makes it easier to remove the CO
2
from the air flow. On the other
hand, increasing the permeance (while leaving the separation factors the same at their
base-case values) also means that more H
2
will back-diffuse into the tube side, which has
a negative impact on the amount of catalyst that one requires, as noted previously. The
changes in catalyst weight are, however, significantly smaller than the decreases in the
membrane surface area requirements.
91
Sweep (H
2
) ratio
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Membrane surface area (m
2
)
0
20
40
60
80
Amount of catalyst (Kg)
0
10
20
30
40
50
60
U
CO2
= 0.8
U
CO2
= 1.2
U
CO2
= 1.6
U
CO2
= 2.0
U
CO2
= 10.0
Figure 5.7 Effect of CO
2
permeance (Symbol: membrane surface area, line: amount of
catalyst)
92
Figures 5.8 and 5.9 show the effect that varying the CO
2
/H
2
permeance ratio
(while maintaining the other parameters at their base case values) has on the MARS
performance. Increasing the CO
2
/H
2
permeance ratio decreases the membrane area
requirements, while increasing the required amount of catalyst (these differences are
more pronounced for lower target cabin CO
2
concentrations as shown in Figure 12). This
can be explained by the fact that a decreasing CO
2
/H
2
permeance ratio (while keeping the
CO
2
permeance constant) significantly reduces the permeation and loss of H
2
to the tube
side. Decreasing the amount of H
2
loss makes CO
2
transport more effective and, therefore,
requires less membrane area to transport to the shell-side a given amount of CO
2
(the
permeance of other gases such as CH
4
, N
2
, O
2
, H
2
O also decrease but this effect is not as
significant as the loss of H
2
). On the other hand, under reaction conditions with excess H
2
,
its loss has a beneficial effect, as it implies increased residence times in the reactor
93
Sweep (H
2
) ratio
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Membrane surface area (m
2
)
0
20
40
60
80
Amount of catalyst (Kg)
0
10
20
30
40
50
60
S.F. CO
2
/H
2
= 2.5
S.F. CO
2
/H
2
= 5.0
S.F. CO
2
/H
2
= 10.0
S.F. CO
2
/H
2
= infinite
Figure 5.8 Effect of separation factor of CO
2
/H
2
(T = 250 (
o
C), 5,000 CO
2
ppm, Ft = 50
(L/sec) (Solid: amount of catalyst, hollow: surface area of membranes)
94
Sweep (H
2
) ratio
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
Membrane surface area (m
2
)
60
80
100
120
140
160
180
200
Amount of catalyst (Kg)
0
10
20
30
40
50
60
70
S.F. CO
2
/H
2
= 2.5
S.F. CO
2
/H
2
= 5.0
S.F. CO
2
/H
2
= 10.0
S.F. CO
2
/H
2
= infinite
Figure 5.9 Effect of separation factor of CO
2
/H
2
(T = 250 (
o
C), 2,000 CO
2
ppm, Ft = 50
(L/sec)) (Solid: amount of catalyst, hollow: surface area of membranes)
95
Figure 5.10 shows the effect of varying the reactor temperature has on MARS
performance. Temperature has a very complex, and yet not well understood effect on
CMS membrane transport, and we have opted in these simulations to assume that
membrane properties remain invariant with temperature, and as a result the membrane
area requirements do not change with temperature. On the other hand, temperature
significantly impacts the catalyst amount that is required. Since the MARS is operated far
from equilibrium, increasing the temperature significantly decreases the amount of
catalyst that is required.
96
Sweep (H
2
) ratio
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Membrane surface area (m
2
)
0
10
20
30
40
50
60
Amount of catalyst (Kg)
0
20
40
60
80
100
120
140
225
o
C
250
o
C
300
o
C
Figure 5.10 Effect of the temperature in the reactor (Symbol: membrane surface area,
line: amount of catalyst)
97
Figure 5.11 presents the effect on MARS performance and design requirements of
target CO
2
concentration in cabin. Allowing for higher CO
2
cabin concentrations has a
beneficial effect on the membrane area requirements. This is to be expected, since with
the higher target CO
2
concentration in the cabin it is easier to remove CO
2
from the air
flowing through the membrane. The decreased catalyst weight is due to the complex
reaction kinetics as expressed by Equation (3.1).
98
Sweep (H
2
) ratio
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Membrane surface area (m
2
)
0
20
40
60
80
100
120
140
160
Amount of catalyst (Kg)
0
10
20
30
40
50
60
2,000 CO
2
ppm
3,000 CO
2
ppm
5,000 CO
2
ppm
Figure 5.11 Effect of target CO
2
concentration in the cabin (Symbol: membrane surface
area, line: amount of catalyst)
99
Figure 5.12 shows the effect of air feed rate on the amount of catalyst and
membrane surface area required to reach the target CO
2
concentration in the cabin. The
required membrane surface area decreases with increasing air feed rate. Increase of the
air feed rate impacts the CO
2
amount removed from the cabin air which results in
reduction in the membrane surface area required. The required catalyst amount goes up
with increasing the air feed rate to the reactor. This could be easily explained by the
decrease of space time for the reactants.
100
Sweep (H
2
) ratio
0.0 0.5 1.0 1.5 2.0
Membrane surface area (m
2
)
0
20
40
60
80
100
120
140
160
Amount of catalyst (Kg)
0
50
100
150
200
250
Ft = 50 L/sec
Ft = 75 L/sec
Ft = 100 L/sec
Figure 5.12 Effect of the air feed rate to the reactor (Solid: amount of catalyst, hollow:
surface area of membranes)
101
5.5 Summary and conclusions (MARS)
The capture and utilization of CO
2
has significant potential applications in the
chemical and power generation industries, as well as in space applications where for the
proper performance of space life-support systems, the removal of the CO
2
from the cabin
atmosphere is required. In this thesis we have studied a continuous regenerative approach
which, in addition to removing the CO
2
, may potentially also allow for the recovery of
oxygen. It involves the use of the methanation (Sabatier) reaction, in which the CO
2
reacts catalytically with H
2
to simultaneously produce CH
4
and water. We have
investigated, in particular, the application of a reactive separation technology, in which
the catalytic and separation steps are coupled in situ through the use of high-temperature
membranes. Coupling reaction and separation provides added synergy, which enhances
the performance of both steps. In this study, the experimental efforts, which involve
catalyst and membrane characterization, and are used to validate experimentally a reactor
model are first described. The model was subsequently utilized to investigate and
establish the feasibility of the proposed reactive separation application for life-support
systems, in terms of the required membrane area and catalyst weight, which are shown to
be modest, and to depend sensitively on the membrane characteristics, and the conditions
prevailing in the cabin.
5.6 MR-based in-situ resource utilization system (MISRUS)
As previously noted, in situ resource utilization is a concept used for increasing
the efficiency of space missions by utilizing indigenous resources found, for example, on
102
the moon or Mars for producing propellants and consumables for life, in order to reduce
the amount of material that must be brought from Earth. Specifically, ISRU applied to the
manufacture of the propellant for the return journey, may reduce the earth-to-orbit mass
by 20-45%, thereby increasing the cost-effectiveness of the mission. NASA plans to
establish chemical plants on Mars prior to the arrival of the first astronauts. These
chemical plants will process CO
2
from the Martian atmosphere to make CH
4
, using the
Sabatier reaction. To H
2
required, is either to be generated from the electrolysis of water,
which is thought to exist on Mars in the form of subsurface ice, or otherwise brought as
supplies in cargo missions.
Figure 5.13 shows an advanced ISRU system, proposed by us, with a membrane
reactor utilized for the Sabatier reaction. In this system CO
2
is fed to the reactor side
while H
2
is fed to the membrane side. H
2
permeated through the membrane reacts with
CO
2
which is fed to the reactor side. H
2
O, one of the products, is removed from the
reactor side to shift the equilibrium (a membrane that has both a high H
2
and H
2
O
selectivity is ideally suited for this system).
103
Figure 5.13 An ISRU system with a membrane reactor
In this Chapter, the feasibility of applying the MR-based Sabatier technology for
the MISRUS is studied. The driver for the development of this MR-based technology is
the promise for improved reactor yield and better thermal management and energy
utilization. As previously noted, the Sabatier reaction is highly exothermic and
equilibrium limited, its conversion starting to significantly decrease at temperatures
above 300
o
C. A practical challenge in conventional Sabatier reactor design is to remove
the heat produced by the exothermic reaction and maintain a relatively low process
temperature without generating hot spots or quenching the reaction. Another challenge is
the potential for having to use raw H
2
feedstock streams that may contain substantial
amounts of impurities (e.g., CH
4
, CO, and O
2
) that could interfere with the reaction.
Again, using membranes makes it possible to use such streams without requiring a
104
substantial pre-treatment. Coupling reaction and separation once more potentially
provides the synergy needed to substantially improve reactor performance.
In this study, we make use of a commercial Ni-based methanation catalyst for
which we have determined the reaction-rate expression experimentally (see Chapter 3).
We are also using supported Pd membranes and CMSM that are developed
collaboratively between USC and Media and Process Technology, Inc. In this Chapter, a
systematic analysis of the performance of the MR-based ISRU (MISRU) system by
means of numerical simulations is provided. A schematic of the MR-ISRU system is
shown in Figure 5.14 (note that this figure is virtually the same with Figure 5.6). The
influence of main process variables such as the feed temperature, the sweep gas flow rate,
and key dimensionless groups on the reactor performance with respect to the CO
2
conversion were investigated and are discussed below.
Figure 5.14 A schematic of the MR-ISRU system
105
5.7 Packed-bed reactor experiments (MISRUS)
To validate the model used in the design simulations we also carried a series of
packed-bed reactor experiments. For these experiments, we used the laboratory
experimental system shown in Figure 3.2, which has been described in Chapter 3. The
reactor is heated by a furnace using temperature controllers and thermocouples installed
in the bed. The reactor is surrounded by a layer of insulation so that there is a small
temperature difference between the inside of the reactor and the surrounding furnace
temperature. When the reaction starts heating the reactor is stopped while the pre-heater
zone is controlled to maintain the temperature of the feed stream to the reactor. Another
movable thermocouple is placed in the reactor side to observe the temperature profile
along the reactor. The feed gas consisted of a mixture of CO
2
and H
2
only. The total flow
rate of mixture was varied from 30 to 62.5 ml/s while the ratio of H
2
/CO
2
was set to 4.
Pressure was kept under atmospheric conditions with the inlet temperature being 250
o
C
and 8 g of catalyst was used in the experiments. Cross section of the reactor and
insulations used in the experiments is shown in Figure 5.15.
106
Figure 5.15 A cross section diagram of the reactor and insulations
For the simulation of these experiments, the Equations (4.12), (4.10) and (4.7) are
modified to the following equations. The heat transfer coefficient of air at the reactor
surface can be estimated by the following equation (Kato et al., 1968).
0.25 0.25 0.25 Pr
Pr
Pr
( )(0.683 )[ ]
(0.861 )
air
air Gr
k N
hNN
LN
=
+
2
e
Gr
gTL
N
β
ν
Δ
≡ ,
Pr
P air
air
C
N
k
μ
≡ (5.2)
where
Pr
N is the Prandtl number, and
Gr
N is the Grashof number.
F
μ is the viscosity,
air
k the thermal conductivity of the air, C
P
the heat capacity,
c
g the gravity conversion
factor,
e
β the volume expansivity, T Δ the difference between
S
T and
bulk
T , L the
characteristic length (for cylinder
2
D
L
π
= ), and ν the kinematic viscosity.
The overall heat-transfer coefficient U
R
for the reactor module given as
107
1
1/ ( / )ln( / ) ( / )ln( / ) ( / )ln( / ) ( / ) /
R
F si s so si si i i so si b b i si b air
U
hr k r r r k rr r k rr r r h
=
++ + +
(5.3)
where h
F
and h
air
[J/s·m
2
·K ] are the heat transfer coefficients of the gas stream in the
reactor outdoor air. k
s
, k
i
, k
b
are the thermal conductivities of stainless steel (reactor
module), insulation (glass wool) and insulation brick, respectively. r
s
, r
i
, r
b
are the
diameters of the reactor, glass wool and insulation brick. i and o represent inner and outer
diameter of each material.
Eq. (4.7) can be modified to the following equation
() ( )(1) ( )
S
jPj R v c c R R
T
nC H r U T T
V
εβρ α
∂
=−Δ − − −
∂
∑
(5.4)
where
R
α the surface area of the reactor per reactor volume, T
[K] the feed-side
temperature and
S
T [K] the surface temperature of the reactor.
Figure 5.15 shows the temperature profiles in the reactor with the dimensionless
distance from the inlet. Shown in the same figure are the simulation results using the
Equations previously derived and the experimental reaction rates. In this calculation, two
different sets of kinetic parameters (reaction rate and adsorption constants) were applied
according to the range of the reactor temperature (Eq. (3.3) and (3.4)). CO
2
conversion
calculated by the numerical simulation is compared with experimental data in Figure 5.16.
As shown in Figure 5.16 and 5.17 the experimental data and simulated conversion and
the temperature profiles. The temperature profiles match fairly well for the first half of
the reactor length but there is a systematic error of ~ 20
o
C for the latter part of the reactor.
These systematic differences also reflect themselves in the conversion, with the measured
108
values being systematically higher that the simulated values. The reasons for these
differences are stil being investigated. Nevertheless, since both experiments gave similar
qualitative trends and the differences in the temperature profiles are not that substantial
the kinetics and heat transfer Equations used for modeling the data were utilized for
simulating the behavior of the MISRU system, see below.
Dimensionless distance from the inlet
0.0 0.2 0.4 0.6 0.8 1.0
Temperature (
o
C)
240
260
280
300
320
340
360
380
400
W/F
CO2
= 4.0
W/F
CO2
= 5.0
W/F
CO2
= 6.2
W/F
CO2
= 8.3
Symbol: Experimental
Line: Simulated
Figure 5.16 Temperature profiles with dimensionless distance from the inlet for various
W
c
/n
CO2
(T
o
= 250
o
C, H
2
/CO
2
= 4.0, P = 1 atm)
109
W/n
CO2
[g*h/mol]
246 8 10
CO
2
conversion
0.0
0.1
0.2
0.3
0.4
0.5
Measured
Simulated
Figure 5.17 CO
2
conversion as a function of W
c
/n
CO2
(T
o
= 250
o
C, H
2
/CO
2
= 4.0, P = 1
atm)
5.8 Simulations with supported Pd membranes (MISRU)
The behavior of membrane reactor was calculated by utilizing the properties of
the Pd membrane studied in Chapter 2. The membrane properties and the kinetic
parameters used in the simulation are those indicated in Table 5.3. As noted above, one of
the challenges in Sabatier reactor design is the potential for having to use raw H
2
feedstock streams that may contain substantial amounts of impurities (e.g., CH
4
, CO and
O
2
) that could interfere with the reaction. Using Pd membranes makes it possible to use
such streams without requiring a substantial pre-treatment, due to the high H
2
110
permselectivities of these membranes.
Here we analyze the membrane reactor and compare its behavior with respect to
that of the conventional PFR in terms of a number of dimensionless parameters including
the sweep ratio (S), the dimensionless temperature in the reactor-side (
F
o
Γ ), the
dimensionless temperature in the permeate-side (
P
o
Γ ), the Damköhler number (Da), and
the product of the Da and the Pe (DaPe) numbers. The sweep ratio (S) is defined as the
ratio of the molar flow rate of H
2
utilized in the permeate-side to that of CO
2
in the feed-
side. When comparing the PFR and the MR, the overall molar flow ratios of H
2
and CO
2
are set equal to each other, i.e., S for the MR is equal to the feed molar ratio in the PFR.
For a given catalyst, Da varies proportionally to (W
c
/F
CO2
), where W
c
is the weight of
catalyst utilized and F
CO2
is the feed molar flow rate of CO
2
. For a given Da, the DaPe
product is inversely proportional α
m
, i.e., the (membrane surface/reactor volume) ratio.
111
Table 5.3 Base case of conditions used in the simulations (MISRUS)
parameter value dimension
B 17.27 - (base case)
Da 1,000 - (base case)
E
p
7962.9 J/mol
n 0.75 -
F
o
n
3.47 x 10
-5
mol/s
Pe 1.8 - (base case)
F
o
P
1.0 atm
P
o
P
1.0 atm
r
i
0.00175 m
r
o
0.00285 m
S 6.0 - (base case)
St 83.1 - (base case)
F
o
Γ
1.0 - (base case)
P
o
Γ
1.0 - (base case)
ref
T
573.15 K
2
, H o
U
0.713 mol/m
2
·s·atm (base case)
m
α
5.49 m
-1
(base case)
ε
0.420 -
v
ε
0.515 -
112
Figure 5.18 (a) and (b) show the CO
2
conversion and the H
2
exiting the feed side
as a function of Da for various sweep ratios. Figure 5.18 (c) shows the dimensionless
temperature ( Г
F
) along the length of the reactor. The CO
2
conversion and the H
2
exiting
the feed side increase with Da, and S. For all the sweep ratios studied, the conversion of
the MR is lower than that of the PFR. One advantage that the MR provides, however, is
that while for the PFR case, in order to recycle the unreacted hydrogen it must be
separated out of the effluent mixture also containing CH
4
, H
2
O and CO
2
, for the MR the
unreacted hydrogen contains no other impurities and can be directly recycled back into
the reactor. This advantage of the MR system over the PFR is particularly true for a range
of Da, where the MR shows higher CO
2
conversion with less unreacted H
2
in the feed
side as compared to the PFR. The temperature in the feed side for PFR sharply increases
near the reactor entrance, while slow increase of the reactor temperature is observed for
MR. Increasing the sweep ratio increases the permeation of H
2
to the reactor side and
results in increased CO
2
conversion.
113
Figure 5.18 Effect of sweep ratio (S)
114
The effect of varying the Da·Pe( α
m
) is shown in Figure 5.19. The CO
2
conversion
is a strong function of the membrane area one utilizes. Increasing α
m
significantly reduces
the Da beyond which conversion levels off, and has an impact on the MR temperature
profiles which increase as α
m
increases.
115
Figure 5.19 Effect of Da·Pe
116
The effect of varying the inlet feed temperatures (
o
Γ ) is shown in Figure 5.20.
Lowering
o
Γ impacts positively the equilibrium conversion, but it is not beneficial for
the H
2
permeation and reaction rates. Nevertheless, apart from an initial range of Da
numbers, using lower
o
Γ leads to enhanced CO
2
conversions and significantly improved
temperature profiles.
117
Figure 5.20 Effect of Г
o
F
(inlet feed side temperature)
118
The effect of varying the inlet temperature in the permeate-side (while
maintaining the inlet feed temperature unchanged at its base value) is simulated in Figure
5.21. Lowering
P
o
Γ results in both higher CO
2
conversions, and in lower temperature
profiles. However, since the membrane may be damaged by large differences in the
temperature between the two sides one must be careful to maintain such differences
rather small.
119
Figure 5.21 Effect of Г
o
P
(inlet permeate side temperature)
120
5.9 Simulation with CMS membranes (MISRUS)
In a study by Ohya et al. (1997), previously discussed in Chapter 3, an MR using
a SiO
2
, water-permeable membrane was investigated at somewhat elevated pressures (0.2
Mpa). In the experiments, H
2
and CO
2
mixtures were fed on one side of the membrane
over the catalyst. The H
2
O produced during the reaction permeated selectively through
the membrane, hence resulting in increased conversion. The CMS membranes
investigated by us also show high steam permeability, as discussed in Chapter 2. In this
section the behavior of the membrane reactor utilizing such membranes is simulated. The
membrane properties, used in the simulations are typical of the CMS membranes, which
are highly permeable to H
2
and steam, and values of the various parameters used in the
simulations are given in Table 5.4. In the remainder of this Chapter, the membrane reactor
is analyzed and its behavior is compared to that of the conventional PFR.
121
Table 5.4 The base case of conditions used in the simulations (CMSM)
parameter value dimension
B 21.05 - (base case)
Da 50.0 - (base case)
n 1.00 -
2,
F
CO o
n
1.39 x 10
-4
mol/s
2,
F
H o
n
5.56 x 10
-4
mol/s
Pe 1.38 - (base case)
F
o
P
1.0 atm
P
o
P
1.0 atm
S 0.25 - (base case)
St 82.85 - (base case)
F
o
Γ
1.0 - (base case)
P
o
Γ
1.0 - (base case)
ref
T
523.15 K
U
H2
1.920 m
3
/(m
2
·bar·h)
U
CO2
0.089 m
3
/(m
2
·bar·h)
U
CH4
0.009 m
3
/(m
2
·bar·h)
U
H2O
1.680 m
3
/(m
2
·bar·h)
m
α
5.49 m
-1
(base case)
122
It is interesting to study the influence of feed configuration, i.e., the effect of the
distributing the H
2
feed between the feed- and permeate-side of the MR. Figure 5.22 (a)
and (b) shows, for example, the CO
2
conversion and the H
2
exiting the feed side as a
function of Da while varying the ratio of H
2
fed to the feed-side to that fed on the
permeate side, while keeping the total H
2
feed constant. Figure 5.22 (c) shows and
dimensionless temperature ( Г
F
) along the length of the reactor. The behavior of the PFR
is also shown in this figure. The CO
2
conversion and the H
2
exiting the feed side increase
with Da, and R
H2
(ratio of
2, 2,
/
FP
HH ). At the low Da range, the H
2
consumed by the
reaction is higher than the H
2
transferred from the permeate side. For this reason, the
unreacted H
2
exiting the feed side decrease with Da at the low Da range. Then, it
increases gradually since the H
2
permeation along the membrane increases while the H
2
consumption via the reaction reduces. The PFR prevails over the MR with respect to CO
2
conversion attained, while the advantage of the MR system over the PFR in the ability to
utilize the hydrogen more effectively. The outlet temperature in the reactor increases with
R
H2
because it is proportional to the CO
2
conversion for the exothermic reaction.
123
Figure 5.22 Effect of feed configuration
124
The effect of varying the inlet temperature in the permeate side is simulated in
Figures 5.23-5.25. CO
2
: H
2
feed ratio of 1:4 is introduced in the feed side, and H
2
is used
as the sweep gas. Lowering
P
o
Γ reduces the reaction rate at the reactor entrance, but it
finally enhances the CO
2
conversion by reducing the feed-side temperature. Increased
CO
2
conversion results in reducing the unreacted H
2
. Figures 5.23-5.25 compare the
performances of the MR for three different inlet temperatures in the feed-side with that of
the PFR. The higher the reactor temperature is the faster the PFR reach stable state with
less CO
2
conversion. The effect of
P
o
T is more significant at higher
F
o
T as shown in
these figures.
125
Figure 5.23 Effect of Г
o
P
for Г
o
F
= 0.80
126
Figure 5.24 Effect of Г
o
P
for Г
o
F
= 1.00
127
Figure 5.25 Effect of Г
o
P
for Г
o
F
= 1.20
128
One of the advantages of the MR provides is that the reactant conversion can be
enhanced by removing the products thorough shifting of the equilibrium toward the
products. The permeance of H
2
O, which is one of the products of the CO
2
methanation, is
high for the CMS membrane utilized in this simulation. Figure 5.26 shows the effect of
membrane properties, expressed in terms of the H
2
O permeance. CO
2
conversion
increases with an increase in the H
2
O permeance, but the effect does not seem to be that
significant as seen in the figure. This is because the H
2
O removal from the reactor side is
not enough due to the low feed side pressure.
129
Figure 5.26 Effect of H
2
O permeance
130
5.10 Summary and conclusions (MISRUS)
In this study, the Sabatier reaction taking place in a MR was investigated by
numerical simulations. Parametric analysis was carried out using a non-isothermal model
in order to systematically investigate the influence of various process parameters on the
reactor performance in terms of the CO
2
conversion, unreacted H
2
exiting the feed side,
and reactor temperature. The results indicate clearly the advantage that the MR system
provides in terms of efficient H
2
utilization over the conventional PFR system. Low
temperature is desirable to get higher equilibrium conversion, while high temperature
takes advantage of the increase of H
2
permeation to the feed-side. Then, optimization of
inlet temperature is required to get a better performance for the MR. It was found that
CO
2
conversion can be enhanced by increasing membrane surface area and H
2
permeance.
MR show good promise here in terms of simultaneously attaining high conversion, while
allowing for proper thermal energy management and utilization.
131
Nomenclature
A
F
cross-sectional area for the reactor side (m
2
)
B adiabatic temperature rise (dimensionless)
C
pj
heat capacity of component j (J/mol·K)
Da Damkohler number (dimensionless)
E
a
activation energy for the reaction (J/mol)
E
j
adsorption activation energy for component j (J/mol)
E
p
activation energy of the membrane (J/mol)
F
j
molar flux for component j (mol/m
2
·s)
G
F
superficial mass flow velocity in the reactor side (g/m
2
·s)
H
j
enthalpy (J/mol)
J
j
molar flux for component j (mol/m
2
·s)
K
j
adsorption equilibrium constant for component j (atm
-1
)
K
eq
equilibrium constant (atm
-2
)
L length of membrane (m)
r G
N
Grashof number (dimensionless)
Pr
N
Prandtl number (dimensionless)
Re
N
Reynolds number (dimensionless)
P
j
partial pressure of component j (atm)
P
F
feed (reactor) side pressure (atm)
P
j
F
partial pressure of component j in the reactor side (atm)
P
P
permeate (air feed) side pressure (atm)
132
P
j
P
partial pressure of component j in the air feed side (atm)
Pe Peclet number (dimensionless)
R ideal gas constant (J/mol.K)
S sweep ratio
St Stanton number (dimensionless)
T absolute temperature (K)
F
o
T
inlet feed side temperature (
o
C)
T
F
feed side temperature (
o
C)
T
P
permeate side temperature (
o
C)
T
S
surface temperature of the reactor (
o
C)
ref
T
reference temperature (
o
C)
U overall heat transfer coefficient (J/s·m
2
·K)
U
H2
H
2
permeance (mol/m
2
·atm·s)
U
H2,o
pre-exponential factor of H
2
permeance (mol/m
2
·atm·s)
U
j
permeance for component j (mol/m
2
·atm·s)
V reactor side volume (m
3
)
V
R
total volume of the reactor (m
3
)
2
CO
X
carbon dioxide conversion (dimensionless)
d
p
particle diameter in the reactor side (m)
f friction factor (dimensionless)
g
c
gravity conversion factor
h
F
heat transfer coefficient in the feed side (J/s·m
2
·K)
h
P
heat transfer coefficient in the permeate side (J/s·m
2
·K)
133
k reaction rate constant (mol/g·s)
k
m
thermal conductivity of membrane (J/s·m·K)
n pressure exponent
20
CO
n
inlet flow rate of CO
2
(mol/s)
n
j
molar flow rate for component j (mol/s)
n
j
F
molar flow rate for component j in the feed side (mol/s)
n
j
P
molar flow rate for component j in the permeate side (mol/s)
n
o,base
total inlet flow rate for base case (mol/s)
n
o
F
total inlet flow rate in the feed side (mol/s)
n
o
P
total inlet flow rate in the permeate side (mol/s)
r overall reaction rate expression (mol/g.s)
r
i
inner diameter of membrane (m)
r
o
outer diameter of membrane (m)
u
F
superficial flow velocity on the reactor side (m/s)
Subscripts
0 entrance condition
eq equilibrium
ex exit condition
j chemical species
Superscripts
F reactor side
134
P air feed (permeate) side
Greek Letters
m
α membrane area per feed side reactor volume (m
-1
)
β equilibrium coefficient
c
β fraction of solid volume occupied by catalyst
e
β
volume expansivity
v
ε
bed porosity in the reactor side
F
μ viscosity of the fluid (g/m·s)
ρ
c
catalyst density (g/m
3
)
ρ
F
average density of the fluid (g/m
3
)
τ thickness of membrane (m)
ν
j
stoichiometric coefficient of component j
135
Bibliography
Aivazov, M.I. and Domashnev, I.A., 1968. Influence of porosity on the conductivityof hot
pressed titanium nitride specimens. Poroshkovaya Metallurgiya (Soviet P/M and Metal
Ceramics), 8(9): 51-54.
Ali, J.K., Newson, J.K. and Rippin, D.W.T., 1994. Deactivation and regeneration of Pd–
Ag membrane for dehydrogenation reactions. Journal of Membrane Science, 89(1-2):
171-184
Amandusson, H., Ekedahl, L.G. and Dannetun, H., 2000. The effect of CO and O-2 on
hydrogen permeation through a palladium membrane. Applied Surface Science,
153(4): 259-267.
Amano, M., Nishimura, C. and Komaki, M., 1990. Effects of high concentration CO and
CO2 on hydrogen permeation through the palladium membrane. Materials
Transactions, JIM 31(5): 404-408.
Armor, J.N., 1989. Catalysis with Permselective Inorganic Membranes. Applied Catalysis,
49(1): 1-25.
Barker, R.S., Russel, M.R. and Whitmer, L.R., 1991. Mathematical Modelling of a Four-
Bed Molecular Sieve with CO2 and H2O Collection, SAE Technical Paper Series No.
911470, 21st International Conference on Environmental Systems, San Francisco, CA.
Bernardo, C.A. and Trimm, D.L., 1979. Kinetics of Gasification of Carbon Deposited on
Nickel Catalysts. Carbon, 17(2): 115-120.
Bischoff, K.B., 1967. An Extension of General Criterion for Importance of Pore
Diffusion with Chemical Reactions. Chemical Engineering Science, 22(4): 525-&.
Boryta, D.A. and Maas, A.J., 1971. Factors Influencing Rate of Carbon Dioxide Reaction
with Lithium Hydroxide. Industrial & Engineering Chemistry Process Design and
Development, 10(4): 489-&.
Brooks, K.P., Hu, J.L., Zhu, H.Y. and Kee, R.J., 2007. Methanation of carbon dioxide by
hydrogen reduction using the Sabatier process in microchannel reactors. Chemical
Engineering Science, 62(4): 1161-1170.
Bunnel, C.T., Boyda, R.B. and Lee, M.G., 1991. Optimization of the Bosch CO2
Reduction Process, SAE Technical Paper Series No. 911451, 21st International
Conference on Environmental Systems, San Francisco, CA.
136
Burggraaf, A.J. and Keizer, K. (Editors), 1991. "Synthesis of Inorganic Membranes,” in
Inorganic Membranes: Synthesis, Characteristics, and Applications. Van Nostrand
Reinhold, New York, 10-63 pp.
Chou, C.T. and Chen, C.Y., 2004. Carbon dioxide recovery by vacuum swing adsorption.
Separation and Purification Technology, 39(1-2): 51-65.
Dannetun, H., Wilzen, L. and Petersson, L.G., 1996. The study of alcohol induced
hydrogen permeation in a catalytic membrane reactor. Surface Science, 358(1-3): 804-
808.
Darensbourg, D.J. and Ovalles, C., 1986. Catalytic Carbon-Dioxide Methanation by
Alumina-Supported Mononuclear and Polynuclear Ruthenium Carbonyls. Inorganic
Chemistry, 25(10): 1603-1609.
Eckhard, F., Feron, P.H.M. and Witt, J., 1996. Evaluation of a membrane based carbon
dioxide absorber for spacecraft ECLS applications, SAE Technical Paper Series No.
961369, 26th International Conference on Environmental Systems, Monterey, CA.
Feron, P.H.M., Jacobs, P., Paul, P.G., Savage, C.J. and Witt, J., 1997. Integrated CO2 and
humidity control by membrane gas absorption, Sixth European Symposium on Space
Environmental Control Systems.
Gobina, E. and Hughes, R., 1994. Ethane Dehydrogenation Using a High-Temperature
Catalytic Membrane Reactor. Journal of Membrane Science, 90(1-2): 11-19.
Greiner, T., 1995. Indoor air quality: Carbon monoxide and carbon dioxide. Extension
Publication No. AEN-125, Iowa State University.
Gu, Y.J. and Zhong, S.H., 2006. Preparation and application of supported palladium-
modified polyimide-silica hybrid membrane for selective separation of H-2. Chinese
Journal of Catalysis, 27(3): 250-254.
Guha, B.K. and Narsimha.G, 1972. Control Regimes in Experimentation of
Heterogeneous Kinetics. Chemical Engineering Science, 27(4): 703-&.
Guo, Y., Lu, G., Wang, Y. and Wang, R., 2003. Preparation and characterization of Pd–
Ag/ceramic composite membrane and application to enhancement of catalytic
dehydrogenation of isobutane. Separation and Purification Technology, 32(1-3): 271-
279.
Hara, S., Sakaki, K. and Itoh, N., 1999. Decline in hydrogen permeation due to
concentration polarization and CO hindrance in a palladium membrane reactor.
Industrial & Engineering Chemistry Research, 38(12): 4913-4918.
137
Harale, A., Hwang, H.T., Liu, P.K.T., Sahimi, M. and Tsotsis, T.T., 2007. Experimental
studies of a hybrid adsorbent-membrane reactor (HAMR) system for hydrogen
production. Chemical Engineering Science, 62(15): 4126-4137.
Harold, M.P. et al., 1994. Catalysis with Inorganic Membranes. Mrs Bulletin, 19(4): 34-
39.
Hart, J.M., Borghese, J.B., Chang, C.H. and Cusick, R.J., 1992. Portable Life Support
System Regenerative Carbon Dioxide and Water Vapor Removal by Metal Oxide
Absorbents Preprototype Hardware Development and Testing, SAE Technical Paper
Series No. 921299, 22nd International Conference on Environmental Systems, Seattle,
WA
Hassan, M.H., Way, J.D., Thoen, P.M. and Dillon, A.C., 1995. Single-Component and
Mixed-Gas Transport in a Silica Hollow-Fiber Membrane. Journal of Membrane
Science, 104(1-2): 27-42.
Holladay, J.D. et al., 2007. Microreactor development for Martian in situ propellant
production. Catalysis Today, 120(1): 35-44.
Holleck, G.L., 1970. Diffusion and Solubility of Hydrogen in Palladium and Palladium-
Silver Alloys. Journal of Physical Chemistry, 74(3): 503-&.
Holmes, R.F., Keller, E.E. and King, C.D., 1972. Bosch Co2 Reduction Unit Research
and Development. Mechanical Engineering, 94(10): 61-&.
Hou, K. and Hughes, R., 2002. The effect of external mass transfer, competitive
adsorption and coking on hydrogen permeation through thin Pd/Ag membranes.
Journal of Membrane Science, 206(1-2): 119-130.
Hsieh, H.P., 1991. Inorganic Membrane Reactors. Catalysis Reviews-Science and
Engineering, 33(1-2): 1-70.
Hsieh, H.P. (Editor), 1996. "Membrane and Membrane Processes,” in Inorganic
Membranes for Separation and Reaction. Elsevier, Amsterdam, 1-13 pp.
Hu, J.L., Brooks, K.P., Holladay, J.D., Howe, D.T. and Simon, T.M., 2007. Catalyst
development for microchannel reactors for martian in situ propellant production.
Catalysis Today, 125(1-2): 103-110.
Hudgins, R.R., 1968. A General Criterion for Absence of Diffusion Control in an
Isothermal Catalyst Pellet. Chemical Engineering Science, 23(1): 93-&.
Hurlbert, R.C. and Konecny, J.O., 1961. Diffusion of Hydrogen through Palladium.
Journal of Chemical Physics, 34(2): 655-&.
138
Hutching.J and Carberry, J.J., 1966. Influence of Surface Coverage on Catalytic
Effectiveness and Selectivity . Isothermal and Nonisothermal Cases. Aiche Journal,
12(1): 20-&.
Hyun, S.H., Song, J.K., Kwak, B.I., Kim, J.H. and Hong, S.A., 1999. Synthesis of ZSM-5
zeolite composite membranes for CO2 separation. Journal of Materials Science,
34(13): 3095-3103.
Jorgensen, S.L., Morgensen, G., Lehoman, T.P., Hyldoft, J. and Nielsen, P.E.H., 1997.
Application of Pd-membranes for the production of pure hydrogen in methanol based
fuel-cell powered vehicles, 4th ESF Workshop on catalytic membrane reactors, Oslo.
Jung, J.U., Nam, W., Yoon, K.J. and Han, G.Y., 2007. Hydrogen production by catalytic
decomposition of methane over carbon catalysts in a fluidized bed. Korean Journal of
Chemical Engineering, 24(4): 674-678.
Karn, F.S., Shultz, J.F. and Anderson, R.B., 1965. Hydrogenation of Carbon Monoxide
and Carbon Dioxide on Supported Ruthenium Catalysts at Moderate Pressures.
Industrial & Engineering Chemistry Product Research and Development, 4(4): 265-&.
Kato, H., Nishiwak.N and Hirata, M., 1968. On Turbulent Heat Transfer by Free
Convection from a Vertical Plate. International Journal of Heat and Mass Transfer,
11(7): 1117-&.
Katsuta, H., Farraro, R.J. and McLellan, R.B., 1979. The diffusion of hydrogen in
palladium. Acta Metall, 27 1111-1114.
Keeling, C.D., 1960. The Concentration and Isotopic Abundances of Carbon Dioxide in
the Atmosphere. Tellus, 12(2): 200-203.
Keeling, C.D. and Rakestraw, N.W., 1960. The Concentration of Carbon Dioxide in the
Atmosphere. Journal of Geophysical Research, 65(8): 2502-2502.
Khatri, R.A., Chuang, S.S.C., Soong, Y. and Gray, M., 2006. Thermal and chemical
stability of regenerable solid amine sorbent for CO2 capture. Energy & Fuels, 20(4):
1514-1520.
Kita, H., Maeda, H., Tanaka, K. and Okamoto, K., 1997. Carbon molecular sieve
membrane prepared from phenolic resin. Chemistry Letters(2): 179-180.
Lee, J.M. and Choi, G.M., 2007. Oxygen-permeating zirconia membrane: The effect of
thickness and surface coating. Journal of the European Ceramic Society, 27(13-15):
4219-4222.
139
Li, A., Liang, W. and Hughes, R., 2000. The effect of carbon monoxide and steam on the
hydrogen permeability of a Pd/stainless steel membrane. Journal of Membrane Science,
165(1): 135-141.
Lunde, P.J. and Kester, F.L., 1972. Kinetics of Carbon-Dioxide Methanation on a
Ruthenium Catalyst. Abstracts of Papers of the American Chemical Society, 164(Aug-
S): 6-&.
Lunde, P.J. and Kester, F.L., 1973. Rates of Methane Formation from Carbon-Dioxide
and Hydrogen over a Ruthenium Catalyst. Journal of Catalysis, 30(3): 423-429.
Lunde, P.J. and Kester, F.L., 1974. Carbon-Dioxide Methanation on a Ruthenium Catalyst.
Industrial & Engineering Chemistry Process Design and Development, 13(1): 27-33.
Marwood, M., Doepper, R. and Renken, A., 1997. In-situ surface and gas phase analysis
for kinetic studies under transient conditions - The catalytic hydrogenation of CO2.
Applied Catalysis a-General, 151(1): 223-246.
Mcbride, R.B. and Mckinley, D.L., 1965. A New Hydrogen Recovery Route. Chemical
Engineering Progress, 61(3): 81-&.
Morreale, B.D. et al., 2003. The permeability of hydrogen in bulk palladium at elevated
temperatures and pressures. Journal of Membrane Science, 212(1-2): 87-97.
Musket, R.G., 1976. Effects of Contamination on Interaction of Hydrogen Gas with
Palladium - Review. Journal of the Less-Common Metals, 45(2): 173-183.
Nacheff, M.S., Chang, C.H., Colombo, G.V. and Cusick, R.J., 1989. Metal Oxide
Regenerable Carbon Dioxide Removal System for an Advanced Portable Life Support
System, SAE Technical Paper Series No. 891595, 19th Intersociety Conference on
Environmental Systems, San Diego.
Okabe, K., Matsumiya, N. and Mano, H., 2007. Stability of gel-supported facilitated
transport membrane for carbon dioxide separation from model flue gas. Separation and
Purification Technology, 57(2): 242-249.
Olivier, J.G.J. et al., 2005. Recent trends in global greenhouse gas emissions: regional
trends 1970-2000 and spatial distribution of key sources. Env. Sc., 2: 81-99.
Peachey, N.M., Snow, R.C. and Dye, R.C., 1996. Composite Pd/Ta metal membranes for
hydrogen separation. Journal of Membrane Science, 111(1): 123-133.
Petersen, E.E., 1965. A General Criterion for Diffusion Influenced Chemical Reactions in
Porous Solids. Chemical Engineering Science, 20(6): 587-&.
140
Phungquach, T.Q. and Rouleau, D., 1976. Kinetics of Reaction between Carbon-Dioxide
and Hydrogen over a Ruthenium Catalyst in a Continuous Stirred Tank Reactor.
Journal of Applied Chemistry and Biotechnology, 26(10): 527-535.
Post, M.F.M., Vanthoog, A.C., Minderhoud, J.K. and Sie, S.T., 1989. Diffusion
Limitations in Fischer-Tropsch Catalysts. Aiche Journal, 35(7): 1107-1114.
Raatschen, W. and Preiss, H., 2001. Potential and benefits of closed loop ECLS systems
on the ISS. Acta Astronautica, 48(5-12): 411-419.
Rao, M.B. and Sircar, S., 1996. Performance and pore characterization of nanoporous
carbon membranes for gas separation. Journal of Membrane Science, 110(1): 109-118.
Rotaru, P. and Blejoiu, S.I., 2001. Kinetics and mechanism of CO2 methanation on a
nickel catalyst. Journal of the Indian Chemical Society, 78(7): 343-351.
Sanchez, J. and Tsotsis, T.T., 2002. Catalytic membranes and membrane reactors. Wiley-
VCH.
Saracco, G. and Specchia, V., 1994. Catalytic Inorganic-Membrane Reactors - Present
Experience and Future Opportunities. Catalysis Reviews-Science and Engineering,
36(2): 305-384.
Schneide.P and Mitschka, P., 1966. Effect of Internal Diffusion on Catalytic Reactions.
Chemical Engineering Science, 21(5): 455-&.
Shirley, A.I. and Lemcoff, N.O., 2002. Air separation by carbon molecular sieves.
Adsorption-Journal of the International Adsorption Society, 8(2): 147-155.
Shu, J., Vanneste, A. and Kaliaguine, S., 1991. Catalytic Palladium-Based Membrane
Reactors - a Review. Canadian Journal of Chemical Engineering, 69(5): 1036-1060.
Sirkar, K.K., Shanbhag, P.V. and Kovvali, A.S., 1999. Membrane in a reactor: A
functional perspective. Industrial & Engineering Chemistry Research, 38(10): 3715-
3737.
Sridhar, K.R., Iacomini, C.S. and Finn, J.E., 2004. Combined H2O/CO2 solid oxide
electrolysis for mars in situ resource utilization. Journal of Propulsion and Power,
20(5): 892-901.
Steinwandel, J., Jehle, W., Wagner, B. and Staneff, T., 1999. Method and apparatus for
removing carbon dioxide.
141
Stephan, K., 1959. Wärmeübergang und druckabfall bei nicht ausgebildeter
laminarströmung in rohren und ebenen spalten. Chem.-Ing.-Tech.(Chemie Ingenieur
Technik ), 31: 773-778.
Stern, S.A., 1994. Polymers for Gas Separations - the Next Decade. Journal of Membrane
Science, 94: 1-65.
Sugawara, A. and Yoshizawa, Y., 1962. An Experimental Investigation on Thermal
Conductivity of Consolidated Porous Materials. Journal of Applied Physics, 33(10):
3135-&.
Sun, G.B., Hidajat, K. and Kawi, S., 2006. Ultra thin Pd membrane on alpha-Al2O3
hollow fiber by electroless plating: High permeance and selectivity. Journal of
Membrane Science, 284(1-2): 110-119.
Uemiya, S., Matsuda, T. and Kikuchi, E., 1991. Hydrogen Permeable Palladium Silver
Alloy Membrane Supported on Porous Ceramics. Journal of Membrane Science, 56(3):
315-325.
Uhlhorn, R.J.R., Keizer, K. and Burggraaf, A.J., 1992a. Gas-Transport and Separation
with Ceramic Membranes .1. Multilayer Diffusion and Capillary Condensation.
Journal of Membrane Science, 66(2-3): 259-269.
Uhlhorn, R.J.R., Keizer, K. and Burggraaf, A.J., 1992b. Gas-Transport and Separation
with Ceramic Membranes .2. Synthesis and Separation Properties of Microporous
Membranes. Journal of Membrane Science, 66(2-3): 271-287.
Venugopal, A. et al., 2007. Hydrogen production by catalytic decomposition of methane
over Ni/SiO2. International Journal of Hydrogen Energy, 32(12): 1782-1788.
Wakao, N. and Funazkri, T., 1978. Effect of Fluid Dispersion Coefficients on Particle-to-
Fluid Mass-Transfer Coefficients in Packed-Beds - Correlation of Sherwood Numbers.
Chemical Engineering Science, 33(10): 1375-1384.
Walton, K.S., Abney, M.B. and LeVan, M.D., 2006. CO2 adsorption in Y and X zeolites
modified by alkali metal cation exchange. Microporous and Mesoporous Materials,
91(1-3): 78-84.
Ward, T.L. and Dao, T., 1999. Model of hydrogen permeation behavior in palladium
membranes. Journal of Membrane Science, 153(2): 211-231.
Weisz, P.B. and Prater, C.D., 1954. Interpretation of Measurements in Experimental
Catalysis. Advances in Catalysis, 6: 143-196.
142
Wieland, P.O., 1994. Designing for human presence in space: An introduction to
environmental control and life support systems. RP-1324, NASA Marshall Space
Flight Center: MSFC, Alabama.
Xomeritakis, G. and Lin, Y .S., 1998. CVD synthesis and gas permeation properties of thin
palladium/alumina membranes. Aiche Journal, 44(1): 174-183.
Xu, X.C., Song, C.S., Miller, B.G. and Scaroni, A.W., 2005. Influence of moisture on
CO2 separation from gas mixture by a nanoporous adsorbent based on
polyethylenimine-modified molecular sieve MCM-41. Industrial & Engineering
Chemistry Research, 44(21): 8113-8119.
Yan, S.C., Maeda, H., Kusakabe, K. and Morooka, S., 1994. Thin Palladium Membrane
Formed in Support Pores by Metal-Organic Chemical-Vapor-Deposition Method and
Application to Hydrogen Separation. Industrial & Engineering Chemistry Research,
33(3): 616-622.
Zaman, J. and Chakma, A., 1994. Inorganic Membrane Reactors. Journal of Membrane
Science, 92(1): 1-28.
Abstract (if available)
Abstract
The capture and utilization of CO2 have significant potential applications in the chemical and power generation industries, as well as in space applications. For the proper performance of space life-support systems, for example, the removal from the cabin atmosphere of the CO2 produced by the inhabitants is required. For short-term flights, CO2 can be controlled by sorption on metal hydroxide. For long-term space applications, however, continuous regenerative approaches are required, including pressure-swing adsorption and membranes which, in addition to removing the CO2, may, potentially, also allow for the recovery of oxygen. One of the approaches proposed is the use of the methanation (Sabatier) reaction, in which the CO2 catalytically reacts with hydrogen to simultaneously produce methane and water. In space applications, one of the challenges the application of catalytic reactor technology faces is the dilute concentrations of CO2 which make its pre-concentration a required step, thus complicating the process train. In this study, we investigate the application of a reactive separation technology, in which the catalytic and separation steps are coupled in-situ through the use of high-temperature membranes. Coupling reaction and separation provides added synergy, which enhances the performance of both steps. Another potential application of the Sabatier reaction could emerge in In-Situ Resource Utilization (ISRU) on Mars. ISRU is a very important new concept to be used to make human presence on Mars possible. This concept involves utilizing raw resources from Mars atmosphere to create useful commodities such as oxygen and propellants like CH4. In this study, our current experimental and modeling efforts in this area aiming to establish the feasibility of the proposed reactive separation application for life-support and ISRU systems will be described.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
The use of carbon molecule sieve and Pd membranes for conventional and reactive applications
PDF
Studies of transport phenomena in hydrotalcite membranes, and their use in direct methanol fuel cells
PDF
A flow-through membrane reactor for destruction of a chemical warfare simulant
PDF
Fabrication of nanoporous silicon carbide membranes for gas separation applications
PDF
Process intensification in hydrogen production via membrane-based reactive separations
PDF
A hybrid adsorbent-membrane reactor (HAMR) system for hydrogen production
PDF
On the use of membrane reactors in biomass utilization
PDF
Methanol synthesis in a membrane reactor
PDF
A process-based molecular model of nano-porous silicon carbide membranes
PDF
Hydrogen storage in carbon and silicon carbide nanotubes
PDF
Development of a novel heterogeneous flow reactor: soot formation and nanoparticle catalysis
PDF
Integrated capture and conversion of carbon dioxide from air into methanol and other C1 products
PDF
Electrochemical pathways for sustainable energy storage and energy conversion
PDF
Reactive and quantum molecular dynamics study of materials: from oxidation and amorphization to interfacial charge transfer
Asset Metadata
Creator
Hwang, Hyun Tae
(author)
Core Title
A study of the application of membrane-based reactive separation to the carbon dioxide methanation
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Chemical Engineering
Publication Date
05/07/2009
Defense Date
10/21/2008
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
air revitalization system,in situ resource utilization,membrane reactor,methanation,nonisothermal modeling,OAI-PMH Harvest,Sabatier reaction
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Sahimi, Muhammad (
committee chair
), Tsotsis, Theodore T. (
committee chair
), Egolfopoulos, Fokion N. (
committee member
)
Creator Email
htt7303@gmail.com,htt7303@hanmail.net
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m2206
Unique identifier
UC1477174
Identifier
etd-Hwang-2653 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-246643 (legacy record id),usctheses-m2206 (legacy record id)
Legacy Identifier
etd-Hwang-2653.pdf
Dmrecord
246643
Document Type
Dissertation
Rights
Hwang, Hyun Tae
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
air revitalization system
in situ resource utilization
membrane reactor
methanation
nonisothermal modeling
Sabatier reaction