University of Southern California Dissertations and Theses

A high efficiency, ultra-compact process for pre-combustion CO₂ capture

(USC Thesis Other)

A high efficiency, ultra-compact process for pre-combustion CO₂ capture

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 1

A High Efficiency, Ultra-Compact Process for
Pre-Combustion CO2 Capture

By
Mingyuan Cao

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMICAL ENGINEERING)

August 2019
2

Acknowledgements
First of all, I would express my most sincere gratitude and deepest appreciation towards my
Ph.D adviser Professor Theodore Tsotsis for your fatherly guidance, encouragement and support
throughout years of my Ph.D study. Your immense knowledge and expertise have been my
invaluable treasure for the academic research, and most importantly, your unconditional love,
tremendous patience and inspirational motivation have allowed me to overcome the weaknesses
in my life and build up my characters. I will always remember the most valuable lesson I have
learned from you in my future career and my life, i.e., to be a chemical engineer, I really need to
calm down and reflect on things in a systematic and methodical way, always do things carefully
and apply “critical thinking”. I believe your mentorship and advice will encourage and accompany
me for the rest of my life.
Beyond my Ph.D adviser, my sincere thanks also goes to Richard Prosser and Professor
Vasilios I. Manousiouthakis, though they are not my academic adviser, their guidance and
encouragement benefits me significantly throughout my life. I would thank Mr. Prosser for
teaching and training me with industrial knowledge which I could never learn from any books, and
I will never forget the time we spend together in GC Environmental. Inc and Santiago Canyon
Landfill Flare Station, those beautiful memories will stay with me for the rest of my life. I would
thank Professor. Manouiouthakis for teaching and guiding me with all the modeling and
mathematic knowledge, I feel I really benefit and learn a lot from working with you on those
projects and I greatly appreciate for that.
I am also greatly thankful to Nicole Kern and Tina Silva for being the nicest and kindest
people during my Ph.D study, which makes me believe that there are always good and nice people
like them staying right there supporting you, though sometimes the clouds and darkness in this
world might shroud your mind. I also appreciate the guidance and advice of Andy Chen throughout
my Master and Ph.D studies.
In addition, my sincere gratitude also goes to my “best buddies” and friends that I have
made during my Ph.D study at USC. I would specially like to thank my fellow Ph.D colleagues:
3

Dr. Sasan Dabir, Dr Zhongtang Li, Dr. Huanhao Chen, Dr. Ashkan Garshabi, Dr. Linghao Zhao,
and I would like to thank my TA colleague Dr. Lin Sun and my lab mate Dr. Dongwan Xu, etc. It
is a great pleasure to work with them and I will never forget the good time we spent together and
beautiful memories we have at USC. Our precious friendship will always shine like brightest stars
in the memory of Milky Way Galaxy.
In the end, I want to convey my deepest love and appreciation to my parents: Haihuan Liu
and Tianwai Cao for their endless love and support and sacrifices, without them I would have
never achieved anything I had today, and it would be impossible for me to realize my dream.

4

Table of Contents
Acknowledgements ......................................................................................................................... 2
Abstract ........................................................................................................................................... 6
Chapter 1. Introduction ................................................................................................................... 8
1-1. Energy Overview ................................................................................................................. 8
1-2. Hydrogen Energy ................................................................................................................. 9
1-3. Carbon Capture and CO2 Sequestration ............................................................................ 12
1-4. The Water Gas Shift (WGS) Reaction .............................................................................. 14
1-5. Our Research ..................................................................................................................... 15
Chapter 2: Preliminary MR-AR Studies -The Case without Steam Sweep* ................................ 20
2-1. Introduction ....................................................................................................................... 20
2-2. Experimental Section ......................................................................................................... 23
2-2-1. Materials ..................................................................................................................... 23
2-2-2. Experimental Set-up ................................................................................................... 24
2-2-3. Experimental Procedure ............................................................................................. 26
2-3. Results and Discussion ...................................................................................................... 29
2-3-1. Membrane Reactor Experiments ................................................................................ 29
2-4. Conclusions ....................................................................................................................... 39
5

Chapter 3: MR-AR Experiments with Steam Sweep .................................................................... 41
3-1. Introduction ....................................................................................................................... 41
3-2-2. Experimental Set-up and Procedure ........................................................................... 43
3-3. Results and Discussion ...................................................................................................... 47
3-3-1. Membrane Studies ...................................................................................................... 47
3-3-2. MR-AR Multiple-Cycle Run ...................................................................................... 58
3-4. Conclusions ....................................................................................................................... 64
Chapter 4: MR-AR Studies Using Oxygen-blown Coal Gasifier Off-gas .................................... 67
4-1. Introduction ....................................................................................................................... 67
4-3. Results and Discussion ...................................................................................................... 73
4-3-1. Membrane Studies ...................................................................................................... 73
4-3-2. MR-AR Multiple-Cycle Run ...................................................................................... 86
4-4. Conclusions ....................................................................................................................... 96
Chapter 5: Future Work ................................................................................................................ 99
References: .................................................................................................................................. 100

6

Abstract
The overarching objective of this lab-scale study is to prove the technical feasibility of the
membrane- and adsorption-enhanced water gas shift (WGS) process that employs a carbon
molecular sieve (CMS) membrane reactor (MR) followed by two adsorption reactors (ARs) in
parallel, operating alternately, utilized for the production of high-purity H2 with simultaneous CO2
capture during WGS reaction treating a coal gasifier off-gas. In the study whose results are
presented here, a commercial sour-shift WGS catalyst (Co/Mo/Al2O3) was employed in both the
MR and the AR. A CMS membrane was used in the MR, and a hydrotalcite adsorbent was used in
the AR. In a preliminary experimental study, we first investigated the integrated MR-AR system
without steam sweep in the MR’s permeate side. The experimental results show that membrane,
catalyst, and adsorbent all operated stably under the integrated gasification combined cycle
(IGCC)-relevant conditions. The stand-alone MR system displayed superior performance in CO
conversion and hydrogen purity compared with the conventional packed -bed reactor (PBR) and
the MR-AR reactor sequence demonstrated high performance superior to that of a PBR with near
100 % conversions attained while the ARs are functional (with an ultrapure hydrogen stream
exiting the AR and permeate-side hydrogen purities from the MR of ~75-80 %) under the IGCC-
relevant operating condition with pressure up to 25 bar and temperature of 250°C. These results
have experimentally validated the ability of the hybrid MR-AR process configuration to operate
stably and properly under the desired conditions and to intensify the efficiency of the WGS
reaction.
Further, we have also investigated the feasibility of performing the lab-scale experiments
with the integrated MR-AR system employing steam sweep in the MR’s permeate side. The system
was modified and upgraded in order to have the capability of utilizing steam sweep in the MR
permeate side and controlling the steam sweep pressure in a variable range (1-3 bar). This
integrated MR-AR system with steam sweep was then experimentally evaluated for the WGS
reaction under IGCC power generation conditions. The CMS membrane for the MR in the study
7

has displayed very robust and stable performance during the long-term run (over 500 hr run of H2S
exposure and under the 250°C temperature and up to 25 bar pressure environments) and
maintained high He/N2 selectivity (~126) over a total of 742 hours of operation during the MR-
AR experiments. We have experimentally tested the combined MR-AR with steam sweep system
in multiple-cycle runs (10-16 cycles), and the system has demonstrated superior performance to
that of a conventional PBR with high purities for the hydrogen product which can be directly
usable in a hydrogen turbine for power generation. In addition, we have carried-out parametric
studies for optimization of the operation of the integrated MR-AR system by investigating various
operating conditions for both the MR and the AR. Also, during the MR-AR with steam sweep
multiple-cycle run, the state of the membrane, catalyst in both the MR and the AR manifested
stable performance. We found that the AR performance (with respect to the catalyst and adsorbent
performance in the AR) takes a few cycles to settle to the eventual “steady-state” value. In general,
the membrane, catalyst and adsorbent are very robust and stable under the IGCC power generation
conditions (large concentration H 2S, high-temperature and high-pressure) environment during the
long-period MR-AR multiple-cycle run.
We have also investigated the feasibility of performing the lab-scale experiments with the
integrated MR-AR system employing oxygen-blown gasifier off-gas as the MR’s feed gas. The
CMS membrane for the MR in this study has demonstrated very robust and stable performance
during the long-term run (over 344 hr of H2S exposure and under the 250°C temperature and 25
bar pressure environments) and maintained high He/N2 selectivity (~170) during the MR-AR
experiments. We have also performed multiple-cycle runs for the integrated MR-AR system and
carried-out parametric studies for optimization of the operation of the integrated MR-AR system.
The membrane, catalyst and adsorbent are also very robust and stable during the MR-AR
experiments in multiple-cycle runs under the oxygen-blown gasifier off-gas environment.

8

Chapter 1. Introduction
1-1. Energy Overview
Energy usage worldwide continues to increase, with levels of demand projected to rise
further over the next three decades, led primarily by large increases from countries outside the
Organization for Economic Cooperation and Development (OECD). The U.S Energy Information
Administration (EIA) has reported, for example, that non-OECD countries, including India and
China, will account for more than 50% of the world’s total increase in energy consumption over
the 2012 to 2040 projection period [1]. The non-OECD countries’ contribution to a rapidly
increasing energy usage and consumption is due to their accelerated economic growth, as these
countries develop and their living standards improve. [2] The International Energy Outlook 2016
(IEO2016) projections for the worldwide energy demand are shown in Figure 1. In year 2012, the
overall world consumption of energy was 549 quadrillion British thermal units (BTU), and in 2020
is projected to rise to 629 quadrillion Btu. The number expands to 815 quadrillion Btu in 2040
equivalent to a 48 % increase over the 28 years period from 2012 to 2040.

9

Figure 1. World energy consumption, 1990-2040 (in quadrillion BTU)
Most of the world’s energy is provided today by fossil fuels. Because of concerns about
global warming caused by emissions from the burning of fossil fuel, and concerns about the
depletion of the world’s fossil fuel resources, renewable energy has become the world’s fastest-
growing source of energy, its consumption increasing at an average rate of 2.6 %/year, aided by
government incentive policies and supports [1]. Fossil fuels will continue to be the world’s largest
source of energy for the foreseeable future, however. It is projected [1], for example, that in 2040,
78 % of total world energy consumption will be still contributed by fossil fuels, including natural
gas and coal, and petroleum and other liquid fuels. The transportation and industrial sectors are
the main consumers of liquid fuels. According to the U.S EIA [1] during the 2012 to 2014 period,
the consumption of liquid fuels increased by 1.1%/year in the transportation sector and by 1%/year
in the industrial sector. Coal is the world second-largest source of energy behind petroleum and
other liquid fuels, with 59% of world coal consumption being attributed to electricity generation,
and 36 % of coal being used in the industrial sectors (both residential and commercial). Compared
with liquid fuels, coal is the slowest-growing energy source with a 0.6%/year average increases in
world total coal consumption from 2012 to 2040. This is due to the non-OECD countries’ policy
change to reduce coal consumption to remedy the air pollution in major urban areas. Despite these
trends, coal will continue to be a vital source of energy for many countries of the world, including
its two biggest economies China and the USA, for the foreseeable future.
1-2. Hydrogen Energy
Hydrogen is the most abundant element in the planet, stored in vast quantities in water,
crude oil, natural gas, etc. [3]. Hydrogen is colorless, odorless and fourteen times lighter than air;
Hydrogen molecules are smaller in size than most other molecules, which makes them diffuse
faster than most other gases. Hydrogen is an important chemical finding many important uses
today in the chemical and petroleum industries. It is also, potentially, a “green fuel” underpinning
the so called “hydrogen economy” [4], finding uses presently in fuel cells or in internal combustion
engines to power vehicles or in turbines for the production of electricity.
10

Hydrogen is a high-quality energy carrier with many potential benefits in terms of
diversifying today’s energy supply and reducing the emissions of global warming gases [3], since
when combusted it produces zero CO2 emissions. Compared with other fuels, hydrogen has highest
energy content with respect to its heat of combustion per mass. For example, the heat of
combustion per gr of hydrogen is 34 kcal, which is 3-4 times larger than that of petroleum, which
is 8.4-10.3 kcal/g [5]. As a gaseous fuel, however, its energy content on a volumetric basis is low,
which presents storage challenges, particularly for automotive applications, where the fuel tank
must be able to store sufficient hydrogen for an adequately long driving range. Another challenge
with hydrogen is that it is highly flammable in a wide range of concentrations and temperatures
when compared with other fuels, which presents safety problems with respect to its production,
transportation and storage [3]. The detailed energy-related properties of hydrogen are shown in
Table 1.
Most of the hydrogen on earth is in the form of water which cannot be used directly as a
fuel. Hydrogen can be produced directly from water via electrolysis with high efficiency around
70-80% [6]. Water electrolysis still requires a theoretical minimum of 237 kJ of electrical energy
input to dissociate each mole of water, which is the standard Gibbs free energy of formation of
water, but also often significantly more than that in order to compensate for entropic losses [7].
Thus, if one was to use hydrogen from water electrolysis as a fuel, it only makes sense to do so by
using renewable sources (e.g., solar or wind) or nuclear energy to produce the electricity needed
rather than from the burning of conventional fossil fuels.
Given the favorable economics and resource availability, conventional fossil fuels are
currently the most popular source of H2, which can be produced either via steam or autothermal
reforming of natural gas or from coal gasification, as seen in Figure 2. Unfortunately, both of the
aforementioned two processes result in the production of large quantities of CO2, which is a major
greenhouse gas [8]. Take the production of hydrogen from natural gas as an example: it involves
two steps, namely the methane steam reforming (SMR) and the water gas shift (WGS) reactions,
as displayed below.
𝐶𝐶𝐶𝐶 4 + 𝐶𝐶 2
𝑂𝑂 ↔ 𝐶𝐶 O + 3 𝐶𝐶 2
∆ 𝐶𝐶 298
𝑘𝑘 = 206.16 kJ/mol (R1)
11

𝐶𝐶 𝑂𝑂 + 𝐶𝐶 2
𝑂𝑂 ↔ 𝐶𝐶 𝑂𝑂 2
+ 𝐶𝐶 2
∆ 𝐶𝐶 298
𝑘𝑘 = −41.2 kJ/mol (R2)
Therefore, advanced technological approaches, as the one pursued in this Thesis, are required to
produce hydrogen from fossil fuels, while in the same time, being able to reduce the CO2 emission
via carbon capture and sequestration (CCS).

Figure 2. World hydrogen production structure for 2004.

Table 1. The energy-related properties of hydrogen compared with other fuels.

12

1-3. Carbon Capture and CO2 Sequestration
The continued increase in CO2 emissions from anthropogenic sources presents a key
challenge for the world today. The concentration of this powerful greenhouse gas in the
atmosphere continues to rise, which leads to raising concerns about global warming and climate
change. Currently, there are mainly three different technologies to reduce CO2 emissions from
power generation into the atmosphere: pre-combustion CO2 capture, post-combustion CO2 capture
and oxyfuel combustion (employing a O2/CO2 mixture) [9].
Although pre-combustion CO2 capture (which is the technology investigated in this Thesis)
and oxyfuel combustion are attracting significant recent attention, post-combustion capture is the
technology that has attracted the most attention over the past couple of decades. Absorption and
adsorption are the two key technologies for CO2 capture employed for post-combustion CO2
capture from the flue-gas in power plants (see Figure 3). Absorption is the most commonly utilized
approach employing alkanolamine-based solvents. However, this technology has several
drawbacks, such as a high equipment corrosion rate and most importantly high energy
consumption needed for regeneration [10]. Adsorption processes are also studied to, potentially,
overcome some of the challenges that absorption faces. Carbon dioxide capture by solid
adsorbents follows two different mechanisms, depending on the sorbent used, i.e., physical
adsorption and chemical adsorption. Physical adsorption occurs when using carbon-based, zeolite,
silica, and MOF sorbents. For chemical adsorption, one utilizes Li and Ca-based materials and
various amine-functionalized solid adsorbents. A CO2 capture adsorbent appropriate for a large-
scale power generation application, should satisfy several key requirements, such as being low-
cost, having low heat capacity, high CO2 adsorption capacity and rate, high CO2 selectivity, and
most importantly high thermal/chemical/mechanical stability under extensive cyclic operation
[10].
13

Figure 3. Flow diagram for CO 2 capture by absorption and adsorption[10].
CO2 capture employing solid adsorbents also involves the need for regeneration. There are
several approaches that are utilized (some of which are employed in this work) including
Temperature Swing Adsorption (TSA), Electrical Swing Adsorption (ESA), Pressure Swing
Adsorption (PSA), Vacuum Swing Adsorption (VSA), and a Pressure and Temperature Hybrid
Processes (PTSA). During TSA, CO2 is first adsorbed at a low temperature, and after the adsorbent
is saturated, the CO2 is desorbed by increasing the bed temperature. For PSA, adsorption first
takes place at a higher pressure to boost the CO2 partial pressure and, thus, to increase the CO2
adsorption rate, while desorption is performed at atmospheric pressure. In VSA, adsorption first
operates at atmospheric pressure and room temperature and the desorption takes place at vacuum
pressures. ESA is very similar to TSA except that the temperature is increased by conducting
electricity through the conductive adsorbents during the regeneration steps.
14

One of the challenges that post-combustion CCS faces is the very low concentration of
CO2 in the flue-gas, typically, <15 vol.% for coal combustion, and even less (<5 vol.%) for natural
gas combustion. This, then, results in a significant energy penalty resulting from using CCS
(primarily as a result of the need for solvent/sorbent regeneration). As a result, other CCS
approaches are also attracting significant attention. One such process, namely pre-combustion CCS
is the primary focus in this Thesis. This approach makes use of the WGS reaction, which is further
discussed below.
1-4. The Water Gas Shift (WGS) Reaction
The WGS reaction (eqn. R2 above) has been known for more than a century, and is widely
applied in the chemical industry for H2 production [11]. This reaction is exothermic and reversible,
and its conversion is equilibrium-controlled. Therefore, lower reaction temperatures lead to higher
CO conversion. On the other hand, employing higher temperatures accelerates the reaction kinetics.
As a result, in the industry, the WGS reaction is typically carried out in a sequence of two reactors
in series, i.e., a high-temperature shift (HTS) reactor followed by a low-temperature shift (LTS)
reactor. Cu/ZnO/Al2O3 is the catalyst of choice for the LTS reaction (which takes place in the
temperature range of 200–300°C), whereas a Fe2O3–Cr2O3 catalyst is used in the HTS reaction
(taking place in the temperature range of 320–450°C). Numerous experimental and modelling
efforts have been conducted to date to study different types of catalysts and to understand the
mechanism of the reaction [12, 13]. Further details can be found in recent publications by our
group [14-17].
The WGS reaction is employed together with the steam methane reforming reaction (R1
above), to enhance the content of the syngas produced, when H2 production (e.g., for ammonia
synthesis) is the key aim, and in that context it is among the most important reactions practiced
industrially today [18]. In the context of power generation, which is the key focus in this Thesis,
the importance of the WGS reaction is that it reduces the CO concentration (CO is a side product
in the generation of hydrogen in coal/biomass gasifiers) in the gasifier off-gas, while
simultaneously increasing its H2 concentration.
15

The equilibrium-limited WGS reaction is ideally suited for the application of novel reactive
separation technologies like membrane reactors (MR) and adsorptive reactors (AR). These reactors
(which are the subject matter of our own study) are favored with respect to the conventional
packed-bed reactor (PBR), because hydrogen is separated in-situ in the MR and carbon dioxide is
captured in-situ in AR, thus based on the Le Chatelier's principle shifting the equilibrium
conversion towards the products. Further, the ability to in situ separate the CO2 produced, makes
the use of such reactors ideal in the context of power generation with CCS.
1-5. Our Research
The research presented in this document involves the lab-scale demonstration of a highly-
efficient, low-temperature reactor process for the water gas shift reaction of coal-gasifier syngas
for pre-combustion CO2 capture in the context of Integrated Gasification Combined Cycle (IGCC)
power generation. In these plants, coal rather than being combusted directly it is first converted
into syngas (a gas mixture containing H2/CO/CO2 and smaller amounts of methane and trace
impurities) in a high-pressure gasifier in the presence of steam and oxygen [16]. In typical IGCC
power-plants (see Figure 4), this syngas is first cooled down to be treated in a cold-gas clean-up
unit (CGCU) to remove trace contaminants (e.g., H2S, NH3, volatile metals, etc.), and it is then
reheated-up to the appropriate temperature to be processed in a conventional packed-bed WGS
reactor (WGSR) in order to convert the CO in the syngas into CO2 and additional H2 (the CGCU
step is necessary to protect the conventional WGS catalysts).
16

Figure 4. Conventional IGCC Power Plant [19]
The exit stream from the WGSR is then subjected to an absorption step (typically, via the aid of
amine-based solvents) to remove the CO2, followed by its additional pressurization, needed for
transport and underground storage.
When compared to post-combustion CO2 capture (i.e., directly separating the CO2 from
flue-gas), pre-combustion CO2 capture in IGCC power plants is considered advantageous because
of the high partial pressures of CO2 involved (in the flue-gas, the CO2 is diluted by the N2 in the
combustion air). The overall IGCC process train is quite complex, however, resulting in a
substantial energy penalty [16]. And though employing a warm-gas clean-up unit (WGSU) – a
technology intensively pursued by US DOE - in lieu of the CGCU is a definite improvement
(potentially resulting in a lower parasitic power load.), even then the projected energy and capital
costs associated with this carbon capture and storage (CCS) system are considered prohibitive for
near-term market deployment without one relying on tax subsidies or the sale of CO2 to offset the
high cost of CCS.
A major part of the technical challenge pre-combustion CO2 capture via IGCC faces stems
from the need to employ the conventional WGS process itself, since it is technically quite complex,
17

requiring (in order to increase the CO conversion and hydrogen yield) using, typically, two reactors,
as noted above: one operating at high temperatures, known as the HTS reactor, and the other at
lower temperatures (typically<300
°
C), known as the LTS reactor [15, 20]. Coupled with the
requirement to use a substantial post-treatment process to separate the hydrogen from the unreacted
CO and the other components of the coal-gasifier syngas, and to separate and re-compress the CO2
for CCS, the resulting overall technical complexity handicaps the adaptation of the conventional
WGS technology (from existing large-scale reforming processes) to IGCC power generation from
coal with simultaneous pre-combustion CO2 capture.
A unique hybrid adsorbent-membrane reactor (HAMR) system, which combines a catalytic
adsorptive reactor with a hydrogen-selective membrane in a single reactor, that further improves
the technical and economic feasibility of reactive separation processes for hydrogen production
was originally proposed and studied by our group [21-25]. The HAMR system simultaneously
removes both products (H2 and CO2), thereby, achieving higher conversion and product yield, and
offers added synergy (between reaction and separation) and advantages over either the MR or the
AR systems. The HAMR concept has been widely applied since to enhance the WGS reaction
process to obtain high-purity hydrogen with simultaneous CO2 capture [22, 26-28].
The process configuration in this research pursued here (see Figure 5 for a schematic) for
the IGCC power generation application relies on the basic HAMR concept, i.e., enhancing reactor
yield and selectivity via the simultaneous removal from the reaction environment of both H2 and
CO2. It is the first known application of the hybrid MR-AR concept to the highly demanding high-
pressure and high-temperature IGCC process. It differs, however, in an important way from the
fully-integrated HAMR process previously investigated [21, 22, 24], because in the present system
the MR and AR units are physically separated, with the AR following the MR. Such an
experimental configuration provides important added flexibility for the IGCC application. for
which, in addition to efficient hydrogen production (a key focus for the past application [17, 22]),
CO2 recovery and purity are also important drivers. For example, the CMS membranes utilized
have an operational temperature threshold (~325
o
C), which places limitations on the regeneration
temperature for the adsorbent. Separating the two units allows one to separately optimize the
operating conditions for each individual unit (MR or AR) without undue concern for interference
18

with the operation of the other. In addition, it allows one to operate with the MR component at
steady state while the AR components are being dynamically cycled. This is a significant
advantage, as it lessens the burden on membrane seals and materials during operation at these
harsh conditions. (Process design calculations by our collaborators from UCLA, not presented here,
indicate that, in addition, the MR-AR system with the components separated provides for a more
efficient utilization of catalyst and membranes than the HAMR system).

Figure 5. The MR-AR Process
Such a hybrid system combining a membrane reactor and an adsorptive reactor in tandem
(with the AR following the MR, and the MR’s reject stream serving as the AR’s feed) can produce
an ultra-pure H2 product (without the need for using a post-processing step) continuously until the
adsorbent (in the AR unit) is saturated. Further, the potential use of a TSA regeneration scheme
rather than a PSA CO2 recovery step (as commonly practiced in stand-alone AR systems [29])
allows the recovery of CO2 at high pressures, thus requiring no additional re-compression step for
CO2 storage. This remarkable process intensification and simplicity exemplifies the potential of
this technology to significantly reduce the parasitic energy loses during pre-combustion CO2
capture.
19

This unique reactor configuration under study here can be viewed as a hybrid adsorption-
enhanced MR system. The individual, stand-alone MR and sorption-enhanced AR technologies
[29] proposed in the literature for hydrogen production (the latter primarily from methane steam
reforming) allow only one of the two ultimate reaction products (H2 or CO2), to be removed;
however, the reaction rate enhancement that results from removing both products (afforded by the
proposed hybrid MR-AR process) allows one to operate at much lower W/FCO (Kgcat/Kmol.hr),
thus significantly reducing catalyst weight usage requirements. In addition, the in-situ removal of
CO2 significantly improves hydrogen recovery.
In summary, the overarching objective of this lab-scale study, results of which are
presented here, is to prove the technical feasibility of the membrane and adsorption-enhanced
WGS reaction process that employs a carbon molecular sieve (CMS) membrane reactor followed
by an adsorption reactor for pre-combustion carbon dioxide capture, while demonstrating process
towards achievement of the overall fossil energy performance goals of 90 % CO2 capture rate with
95 % CO2 purity at a cost of electricity of 30% less than the baseline capture approaches. Such a
WGS system combining a MR and an AR in tandem can produce an ultra-pure hydrogen product
(without the need for using a post-processing step) continuously, until the adsorbent (hydrotalcite-
based) in the AR unit is saturated for regeneration via a pressure swing adsorption (PSA) and/or
temperature swing adsorption (TSA) operation.

20

Chapter 2: Preliminary MR-AR Studies -The Case
without Steam Sweep*
* Note that this chapter has been published as a paper from our group [30]
2-1. Introduction
As noted in the Introduction, hydrogen is an important “green fuel” underpinning the so-
called “hydrogen economy” [4]. Catalytic steam methane reforming and the water-gas-shift
reactions, the primary processes employed in H2 production, have as a result been attracting
renewed interest in recent years. A key challenge with H2 production via the SMR and WGS
processes, is that the H2 product contains a large concentration of CO2, a greenhouse gas
contributing to global warming [30-32].

As a result, conventional H2 production processes,
typically, utilize multiple separation and purification units to produce a pure H2 product [33].

Reactive separation processes (the best-known examples being catalytic membrane reactors and
adsorptive reactors) are a promising emerging technology, and have in recent years attracted
research interest, particularly for applications involving the production of H2 and of fine chemicals
[34, 35]. In such reactors, reaction and separation steps are integrated into a single unit, the
outcome, typically, being to attain conversions exceeding equilibrium for equilibrium-limited
reactions [21, 22, 29, 36]. Potential advantages of such reactors over their conventional
counterparts include: (i) increasing the conversion and the desired product yield; (ii) being able to
in-situ separate the undesired product (e.g., CO2 in H2 production) to produce a high-purity desired
product and, thus, reduce downstream purification requirements; (iii) providing cost and energy
savings [21, 22, 29].
MR’s utilizing nano-porous inorganic or metallic (e.g., Pd-alloy) membranes with high H2
selectivity show promise in H2 production. For example, a novel concept, termed the “one-box”
WGS-MR process, employing a carbon molecular sieve (CMS) membrane and a sulfided Co/Mo
catalyst, was proposed by our group to produce contaminant-free H2 [17] for application in power
generation. This WGS-MR achieved high conversion through the in-situ H2 removal from the
21

reaction mixture. Reviews of additional studies by this Group and others on using MR’s to carry
out the WGS reaction can be found in previous papers [14-16, 20].
ARs, employing the sorption-enhanced reaction process (SERP) concept, are also a
promising technology. They have been utilized in H2 production via the sorption-enhanced steam
reforming of methane [37, 38] or ethanol [36, 39], and in the sorption-enhanced WGS (SE-WGS)
reaction [30, 31, 40, 41] of reformate mixtures. The SE-WGS process is also a promising
technology to lower the cost of CO2 capture in power generation, in the context of Integrated
Gasification Combined Cycle (IGCC) power plants [30, 42, 43]. The SE-WGS reactor contains a
mixture of a WGS catalyst and a sorbent, for in-situ CO2 separation and capture to produce high-
purity H2 from syngas, to be used for power generation in PEM fuel cells and/or in turbines. The
sorbent plays several key roles: By in-situ removing the CO2 from the reacting mixture, it produces
a pure CO2 stream ready for sequestration, and a H2 product for use in power generation; and in
so doing, it increases the CO conversion by overcoming equilibrium limitations and, thereby,
improving energy efficiency [30, 41].
In the process under study here (see Figure 5 in Chapter 1), the MR’s reject stream serves
as the feed to the AR, and when properly designed the combined system can produce an ultra-pure
H2 product (without the need for using a post-processing step) continuously until the adsorbent (in
the AR unit) is saturated. At that point, the AR unit is taken off-line for regeneration via a pressure
swing adsorption (PSA) or a temperature swing adsorption (TSA) operation, and another freshly-
regenerated AR unit is brought on-line connected to the MR subsystem (the use of a TSA
regeneration scheme allows the recovery of CO2 at high pressures, thus requiring no added re-
compression step for CO2 storage). The system in Figure 5 (and the lab-scale system in this study)
incorporates two AR units, but optimal operation of the combined system is likely to dictate the
use of a higher number of units [22]. The H2 permeated through the CMS membranes (MR
permeate side) can be used directly by the Hydrogen Turbine in the IGCC plant for power
generation. For the ultra-pure H2 exiting the AR system there are two different options in the
context of the IGCC application: (i) use for power generation in PEM fuel cells, and (ii) re-mix
with the H2 stream from the permeate side of the MR for use in the Hydrogen Turbine. The latter
is the option our team is currently pursuing.
22

In the present combined MR-AR system there is still significant interaction and synergy
between the two individual units (MR and AR) that impacts overall system performance. For
example, the choice of operating variables for the MR (e.g., catalyst weight, membrane area,
reactor temperature and pressure, contact time, and purge rate) affect CO conversion and hydrogen
recovery/purity in the permeate stream, and the composition and flow rate of the reject stream,
which since it is the feed to the AR unit (Figure 5), in turn, impacts critically its performance.
Conditions that optimize the performance of each individual sub-system are not necessarily
optimal for the performance of the combined MR-AR system. For example, when operating the
MR alone, trying to meet various performance targets such as maximizing CO conversion, H2
recovery, and purity creates process challenges. For example, to maximize CO conversion and H2
recovery, one must use a large membrane area, which, however, can have a negative impact on
purity. One can improve the H2 purity by selecting a more highly perm-selective membrane, which
as the Robeson plot dictates [44] will have a smaller permeance, which, in turn, will negatively
impact CO conversion and H2 recovery. Using the MR and the AR in combination helps to
overcome most, if not all, these challenges. One, for example, can opt to focus on hydrogen purity,
and can design the MR component to attain such purity without worrying about simultaneously
meeting CO conversion and/or H2 recovery targets, since the AR unit that follows will be tasked
to help meet these goals. The same arguments also apply when using the AR system alone, in
comparison with the fully-integrated MR-AR system proposed here. Trying to meet
simultaneously CO2 recovery/purity targets together with carbon utilization (CO conversion) and
hydrogen recovery/purity goals is quite a daunting Task for such a reactor system [29]. The use of
the MR unit preceding the AR system, as proposed here, makes attaining such goals and
performance targets a reality.
The CMS membrane has been field-tested by this team for hydrogen separation from coal
and biomass-derived syngas [45], and has been used in the laboratory in a WGS-MR operating on
simulated coal- and biomass-gasifier off-gas [14-17, 20]. These membranes (in contrast to
competitive Pd-alloy type membranes) have, therefore, been proven robust in the real coal-gasifier
off-gas environment and under IGCC-relevant reactive conditions. The expectation, therefore,
23

prior to the initiation of this combined MR-AR study was that they will also function appropriately
in the simulated coal syngas environment, which is used in this lab-scale investigation.
The hydrotalcite adsorbent selected for the study has been studied by this team for CO2
separation from flue-gas and SMR reformate mixtures [21, 22, 24, 46] and, as noted above, its
CO2-affinity at the proposed operating pressure and temperature conditions has also been well-
documented in the literature [46-48]. It was also recently investigated in an AR study with
simulated coal gasifier off-gas at modest pressures [49, 50]. It has not been studied before, however,
under realistic pressure conditions (25-30 bar) or with mixtures relevant to the MR-AR process
(the reject side of the MR), and a key focus of our lab-scale study, therefore, was validating its
robustness in that environment. The primary focus, however, of this study has been investigating
at the lab-scale the performance of the novel hybrid MR-AR system and validating its applicability
for the proposed application.
In what follows, we first describe the lab-scale hybrid MR-AR system and its operation and
evaluation under realistic pressure and temperature conditions akin to the IGCC environment. We
then describe its preliminary experimental testing for simultaneous H2 production and CO2 capture
for a range of pressures up to 25 bar while employing a simulated coal-derived air-blown gasifier
off-gas. Specifically, the performance characteristics of the MR-AR system for a range of W/FCO
(weight of catalyst/molar flow rate of CO) conditions are presented.
2-2. Experimental Section
2-2-1. Materials
We have utilized in the MR a tubular CMS membrane with an inner diameter of 3.6 mm,
outer diameter of 5.6 mm, and a length of 254 mm, prepared by our industrial collaborator in this
project Media and Process Technology, Inc. (M&PT). The membrane consists of a thin nano-
porous CMS separation layer formed on the outside surface of a M&PT commercial asymmetric
mesoporous alumina ceramic tube. The specific CMS membrane utilized (termed as the candle-
filter membrane configuration) has only one end open to flow, for convenient installation and
24

sealing into the reactor (photographs of candle-filter CMS membranes and of membrane bundles
can be found in our recent paper [17]). A commercial Co/Mo/Al2O3 sour-shift catalyst provided
by Clariant (US) (whose physical and chemical properties are also reported in our recent paper
[17]) was used for both the MR and the AR units. A Mg-Al-CO3 layered double hydroxide (with
a Mg/Al molar ratio of 3:1) CO2 adsorbent was used in the AR; it was prepared by our team
(M&PT) using conventional approaches [51, 52].
2-2-2. Experimental Set-up
A schematic of the lab-scale hybrid MR-AR system is shown in Figure 6. It consists of three
sections:
(i) The MR section, which consists of the tubular stainless steel (SS) reactor (seated inside a
furnace with six separate heating zones) with a length of 25.4 cm and inside diameter of 3.2 cm.
To ensure isothermal reactor operation, the temperatures of the reactor is controlled with a six-
point thermocouple connected to PID controllers. Temperature isothermality is confirmed via a
thermocouple sliding inside a thermo-well imbedded in the reactor module. The tubular CMS
membrane, described in section 2-2-1, is sealed inside the reactor using graphite O’s rings and
compression fittings. 10 g of commercial Co/Mo/Al2O3 WGS catalyst mixed with ~90 g of crushed
quartz particles (with the same size in the range of 600-850 μm) are loaded into the annular space
in between the CMS membrane and the reactor body. The reason that we dilute the catalyst (and
adsorbent, see below) with inert quartz particles is so that we completely fill the MR (and AR)
volumes to avoid gas by-passing; this has the added benefit of making it more feasible to operate
under isothermal conditions by diluting the catalyst and/or adsorbent and adding more heat
capacity (further details about the MR section can be found in [17]).
(ii) The AR section, which consists of two different tubular stainless steel reactors with a length
of 14 cm and inside diameter of 3.8 cm, each AR being located inside its separate GC oven for
temperature control. Three two-point thermocouples (fabricated by the Thermometric Corporation)
are installed inside each AR to monitor the temperature at two radial and at three equi-distant axial
positions of the reactor. Both AR’s are loaded with an admixture containing the hydrotalcite
25

adsorbent (69 g), the Co/Mo/Al2O3 sour-shift WGS catalyst (10 g) and quartz (22 g), all with the
same particle size in the range of 600-850 μm. Additionally, the AR section is equipped with an
Argon gas cylinder, and its associated mass-flow controller (MFC), a high-pressure water syringe
pump, and a steam-generating unit (evaporator), specifically designed for generating an AR/steam
mixture at the desired pressure (e.g., 25 bar) for regenerating the adsorbent in the AR. A separate
steam-generating unit (high-pressure water syringe pump + evaporator) is installed in between the
MR and AR units for supplying additional water to the AR feed stream (i.e., the MR reject-side
stream) to adjust its H2O/CO ratio, if so desired.
(iii) The analysis section, which consists of a Gas Chromatograph (GC) for measuring the
concentrations of the MR exit gas streams, two separate Mass Spectrometers (residual gas
analyzers or RGA) capable of instantaneously analyzing the gas composition of the AR exit gas
streams, bubble flow-meters (BFM) for measuring the total MR exit dry-gas flow rates, condensers
to remove the water from the exit streams of the MR and AR, and traps to remove the H2S from
the same streams to protect the GC and RGA instruments.

Figure 6. Experimental set-up used in the MR-AR experiments (red lines are heat-traced to prevent
water from condensing; MR: Membrane Reactor; AR: Adsorptive Reactor; MFC: Mass Flow
Controller; BFM: Bubble Flow-meter; RGA: Residue Gas Analyzer; GC: Gas Chromatography;
BPR: Back Pressure Regulator;).
26

2-2-3. Experimental Procedure
To monitor the state of the CMS membrane before and after the MR-AR experiments, we
measured the single-gas permeances of the major syngas components (i.e., H2, CO, CO2, CH4, N2)
and of He, an inert test gas, at pre-determined temperatures and pressures. Specifically, to carry-
out these single-gas permeation tests the feed-side pressure of the MR was set to its appropriate
value with the aid of a BPR while the permeate-side pressure was typically maintained at
atmospheric conditions, and the permeate-side and reject-side flow rates were then measured with
a bubble flow-meter. In order to convert the flows to STP conditions (m
3
/h), the laboratory
temperature and pressure were also measured with a digital thermometer and an analog barometer,
respectively. The ideal gas selectivity of the CMS membrane, 𝛼𝛼 𝑖𝑖 𝑖𝑖 , is defined as the ratio of the
permeances of two pure gases, measured separately under the same pressure/temperature
conditions:
𝛼𝛼 𝑖𝑖 𝑖𝑖 =
𝑃𝑃 𝑃𝑃 𝑃𝑃 𝑃𝑃 𝑖𝑖 𝑃𝑃 𝑃𝑃 𝑃𝑃 𝑃𝑃 𝑗𝑗 (1)
where 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃
𝑖𝑖 and 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃
𝑖𝑖 are the permeance of the two pure gases, respectively, with i being the
most permeable gas.
For the MR-AR experiments, since the Co and Mo metal components of the fresh (as
received) catalyst, are in their oxidized form, they need to be sulfided prior to the initiation of the
reaction. The activation procedure involves the in-situ reduction of the metals using a gas mixture
containing H2, N2 and H2S (up to 5 mol% of H2S in H2/N2) using a temperature and pressure
protocol, as specified by the catalyst manufacturer [17]. In this study, we used a simulated coal-
derived syngas (purchased as a certified, pre-mixed gas mixture from Praxair, Specialty Gases &
Equipment), with composition: H2:CO:CO2:Ar:CH4:H2S = 0.51:1.00:0.36:2.28:0.1:0.0031, typical
of air-blown gasifier off-gas (we substitute N2 with Ar as to facilitate the analysis via the RGA,
since N2 interferes with the analysis of CO, but our studies have shown that both N2 and Ar act as
diluents and do not participate in the WGS reaction - Ar also serves in the dual capacity as an
internal standard gas).
27

For each experiment, the conversion of the MR and the AR sub-systems and the overall
combined MR-AR conversion were measured for different W/FCO values (where W is the weight
of the catalyst, and FCO is the feed molar flow rate (mol/h) in the MR), while employing a constant
H2O/CO ratio (equal to 2.8 in the experiments reported here). The MR-AR experiments are
performed at IGCC-relevant high-temperature and high-pressure (up to 25 bar) conditions. Before
initiating the combined MR-AR experiment, the MR conversion was first allowed to reach steady
state for at least 1 h. To “bench-mark” the MR performance, in the experiments we also measure
the corresponding conversion of the packed-bed reactor (PBR) under the same operating
conditions. This is accomplished by closing the exit on the permeation side of the MR, and again
allowing the system (now operating as a PBR) to reach steady state.
As shown in Figure 5 and Figure 6, the reject-side (retentate) stream (a CO2-rich stream,
that also includes some H2, unreacted CO, CH4, N2 and impurities like H2S) of the MR serves as
the feed for the AR. Though it is possible to operate the AR at different temperatures and pressures
than those in the MR, in the preliminary experiments reported here the operating temperature and
pressure for both reactors were kept the same. As noted above, one has the option to add additional
steam to the feed stream into the AR (to potentially compensate for H2O losses to the permeate
side of the MR) to maintain the desired H2O/CO ratio (for the experiments reported here, this ratio
for the AR is also maintained at 2.8). To initiate the operation of the combined MR-AR system,
once the MR operation reaches steady state, the MR’s reject-side stream is directed to the feed of
the AR via a 3-way valve to allow it to be first mixed with a sufficient flow of steam to maintain
the desired H2O/CO ratio. Prior to doing so, both AR’s are pressurized simultaneously to the
desired pressure and heated to the desired temperature with the aid of a 50 vol.% steam-50 vol.%
inert gas (Argon) stream. While the MR reject-side (intermixed with any additional steam, if so
desired) is directed into the first AR (AR I), the 50 vol.% steam/50 vol.% Argon stream continues
to be directed to second AR (AR II).
During the experiment, the AR I outlet gas is measured instantaneously using the RGA I.
Once the AR I system reaches the desired level of performance (e.g., a pre-determined hydrogen
product purity - in the experiments reported here we allow for complete adsorbent saturation in
order to validate catalyst and adsorbent stability), the MR reject-side stream (plus any added steam)
28

is then switched into the second AR (AR II), while the 50 vol.% steam/50 vol.% Argon stream is
directed into AR I to regenerate the adsorbent in the reactor at the desired temperature (a
temperature 400 ºC is employed here) and duration (for 30 min in the experiments reported here -
the regeneration conditions are a key optimization parameter, and means by which to optimally
select such parameters will be discussed in an upcoming publication). After completing the
evaluation of the performance of the AR I and AR II sub-systems, the MR’s reject-side stream is
switched back away from the feed of the AR, and the MR performance, including the gas
composition and flow rate of both the reject and the permeate sides, is studied again to verify if
the MR performance has remained stable. The operating cycle is then again repeated to gauge
system performance.
The total CO conversion of MR-AR overall system is defined as follows:
CO (%) =
𝐹𝐹 CO ,𝑓𝑓 𝑓𝑓𝑓𝑓𝑓𝑓
M
− ( 𝐹𝐹 CO , 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 A
+ 𝐹𝐹 CO , 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
)
𝐹𝐹 CO ,𝑓𝑓 𝑓𝑓𝑓𝑓𝑓𝑓
M
× 100% (2)
where 𝐹𝐹 C O, 𝑓𝑓 𝑃𝑃𝑃𝑃 𝑓𝑓 M
represents the molar flow rate of CO in the feed of the MR (mol/h), and 𝐹𝐹 C O, 𝑃𝑃 𝑒𝑒 𝑖𝑖 𝑒𝑒 A

and 𝐹𝐹 C O,𝑃𝑃 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
represent the molar flow rates of CO in the exit of the AR and the MR permeate-side,
respectively.
The CO conversion of the MR subsystem is calculated by Eq. (3)
CO (%) =
𝐹𝐹 CO ,𝑓𝑓 𝑓𝑓𝑓𝑓𝑓𝑓
M
−( 𝐹𝐹 CO , 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 R
+ 𝐹𝐹 CO , 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
)
𝐹𝐹 CO ,𝑓𝑓 𝑓𝑓𝑓𝑓𝑓𝑓
M
× 100% (3)
where 𝐹𝐹 C O,𝑃𝑃 𝑒𝑒 𝑖𝑖 𝑒𝑒 R
is the CO molar flow rate at the exit of the MR reject-side (mol/h).
The CO conversion of the PBR (when closing the MR permeate-side) is defined as follows
CO (%) =
𝐹𝐹 CO ,𝑓𝑓 𝑓𝑓𝑓𝑓𝑓𝑓
M
− 𝐹𝐹 CO , 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 R
𝐹𝐹 CO ,𝑓𝑓 𝑓𝑓𝑓𝑓𝑓𝑓
M
× 100% (4)
The CO conversion of the AR subsystem is defined as follows
29

CO (%) =
𝐹𝐹 CO , 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 R
− 𝐹𝐹 CO , 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 A
𝐹𝐹 CO , 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 R
× 100% (5)
where the CO molar flow (mol/h) at the exit of the MR reject-side functions as the feed of the AR.
The hydrogen purity (dry-basis) is calculated by Eq. (6)
𝑃𝑃𝑃𝑃 H
2
=
𝐹𝐹 H
2
, 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
𝐹𝐹 H
2
,𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
+ 𝐹𝐹 CO ,𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
+ 𝐹𝐹 CO
2
, 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
+ 𝐹𝐹 CH
4
, 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
+ 𝐹𝐹 A rg o n , 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
× 100% (6)
where 𝐹𝐹 H
2
,𝑃𝑃 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
, 𝐹𝐹 C O
2
, 𝑃𝑃 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
, 𝐹𝐹 C H
4
, 𝑃𝑃 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
and 𝐹𝐹 A rg o n, 𝑃𝑃 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
are the molar flow rates of hydrogen, CO2, CH4
and Argon at the exit of the MR permeate-side (mol/h), respectively.
The hydrogen recovery is calculated by Eq. (7)
𝑅𝑅 𝑃𝑃 H
2
=
𝐹𝐹 H
2
, 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 P
𝐹𝐹 H
2
,𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 R
+ 𝐹𝐹 H
2
, 𝑓𝑓 𝑒𝑒 𝑖𝑖 𝑒𝑒 P

× 100% (7)
where 𝐹𝐹 H
2
,𝑃𝑃 𝑒𝑒 𝑖𝑖 𝑒𝑒 R
is the hydrogen molar flow rate at the exit of the MR reject-side (mol/h).

2-3. Results and Discussion
2-3-1. Membrane Reactor Experiments
We employed in this study a CMS membrane prepared at M&PT. The membrane was loaded
into the MR module at the M&PT laboratories, and its permeation characteristics were tested using
two model inert gases (He and N2) prior to the module being shipped to USC. Upon receiving the
MR module from M&PT, the pure-gas permeances of N2 and He were again measured at the same
temperature and pressure conditions, and substantial differences were observed in the permeance
values measured at USC when compared to those measured at M&PT. A closer inspection of the
MR module indicated that the end-fittings appeared loose and were, then, re-tightened. Subsequent
30

measurement of the permeation properties indicated them to be more in line with those measured
at M&PT (see Table 2), with the He/N2 ideal selectivity (IS) measured at USC being, in fact, quite
higher (173 vs. 122, see Table 2) than the one measured at M&PT, mostly, due to a lower N2
permeance, indicative potentially of a leaking graphite sealing during the M&PT permeation test
(It should be noted that single-gas and mixed-gas permeances with these CMS membranes are
close to each other, and so are the ideal selectivity and separation factor. Single-gas permeances
are significantly easier, and also more reliable to measure than mixed-gas permeances, and are
thus utilized routinely here to monitor the state of the membrane during the MR-AR experiments.
Further details about the permeation properties and the mechanism of transport of these
membranes can be found elsewhere [17]). After the initial measurement of the membrane
properties, 10 g of WGS catalyst were intermixed with ~90 g of crushed quartz, and the mixture
was then carefully loaded into the MR module and activated according to the activation procedure
recommended by the catalyst manufacturer[17].

Table 2. Pure-gas permeances of He and N 2 and corresponding ideal selectivity at various temperatures
and pressures measured in the empty module.
Lab T (ºC) P (bar) Gas permeance
(m
3
(STP)/m
2
h bar)
Ideal selectivity
(IS)

He N2 He/N2
MP&T 250 2.38 1.1110 0.0091 122
USC 250 2.38 1.0454 0.0061 173
USC 250 25 0.9641 0.0068 142

In this study, we report experiments with the hybrid MR-AR system with the MR and AR
subsystems operating at the same temperature (250 °C) and pressure (15 and 25 bar), while varying
the W/FCO for the MR. As noted previously, prior to initiating these experiments with the combined
31

MR-AR system (i.e., switching the reject-side stream from the MR as feed into the AR) the MR
was allowed to reach steady state and its conversion, H2 purity (dry-basis) in the permeate stream
and recovery were all measured, and are presented in Figures 7, 8 and 9. We also measured the
conversion and hydrogen purity of the reactor functioning as a PBR under the same experimental
conditions (followed the method detailed in Section 2-2-3), and those are also reported on the same
Figures. (As noted in Section 2-2-2. the gas compositions in these experiments were measured
using a GC. For each experimental point, we withdrew and analyzed several samples until at least
three consecutive measurements were in the range of ~5% error).
As can be seen in Figure 7, the CO conversions for the MR are significantly higher than
those of the PBR under all conditions studied (i.e., different W/FCO and pressures). The most
important advantage, however, of using the MR is shown in Figure 8, which compares the
hydrogen purities of the processed syngas for the MR and PBR. Because the air-blown coal gasifier
off-gas contains a large fraction of N2, despite the relatively large conversions attained in the PBR,
the hydrogen purity of the resulting product is so poor that it is hardly appropriate for use in power
generation. On the other hand, the use of the MR that combines reaction and separation in one unit,
in addition to improving the CO conversion, also provides a hydrogen product stream with
sufficient purity to be directly usable in a hydrogen turbine for electricity generation (the calculated
equilibrium conversion under these conditions is quite high ~ 98%, and for the W/FCO and the
membrane area employed in this lab-scale study neither the PBR nor the MR approach this value).
The important role that the AR plays as an add-on 2nd stage to the MR can be seen by
studying the MR behavior in Figure 7-9 (Note that experimental results on conversion, H2 purity
and recovery in these Figures are consistent with the simulation findings employing a data-
validated model of the WGS reaction being carried out in a MR employing the same catalyst and
membranes developed by this team [17]. The experimental behavior of purity is in fact interesting,
as it does not show much dependence on MR feed-side pressure. A simple explanation for that, is
because the permeation of the various species is proportional to their partial pressure gradients,
which for these high (reject side/permeate side) total pressure ratios is approximately proportional
to the reject-side partial pressures. The ratio of fluxes of two given species (that determine the
32

purity of the permeate product) is thus substantially dependent on the reject-side composition and
not the total pressure).
The need to capture a large fraction of the carbon in the gasifier off-gas for storage and
sequestration (the present DOE target is >90%), dictates that the CO conversion in a single-stage
MR is quite high (for the off-gas composition used in these experiments, a CO conversion larger
than 85% is needed, but most likely significantly higher than that, given the unavoidable losses
during the CO2 separation and capture steps). Such high conversions are, definitely, attainable in
a MR, e.g., by increasing the amount of catalyst utilized (i.e., increasing the W/FCO, see Figure 7)
or the membrane area, but doing so implies substantial additional capital cost. In addition, as Figure
7-9 make amply clear, increasing CO conversion comes at a cost of diminished hydrogen product
purity, which also implies additional carbon losses. Another key challenge for the stand-alone MR
(as with the PBR), is meeting the purity requirements for the CO2 stream (current DOE
target >95%) for storage and sequestration. For meeting this requirement, the reject stream from
the MR (processing the air-blown gasifier syngas employed in this study) will have to undergo
significant additional processing (e.g., a two-stage Selexol process). The combined MR-AR
system readily meets the target without needing additional processing. It is superior, furthermore,
to a stand-alone AR which produces a low-purity hydrogen stream, not much unlike the PBR,
which requires substantial added treatment before it can be used for power generation (though the
stand-alone AR has an advantage over the PBR, in that it produces a substantially pure CO2 stream).
Another important objective of this preliminary lab-scale investigation is to validate the
ability of the materials utilized (catalysts, membranes and adsorbents) to function stably under
these harsh experimental conditions involving the processing of a simulated coal-gasifier off-gas
containing a large concentration of H2S with high-temperature and high-pressure steam. For that,
the catalytic activity of the catalyst was monitored continuously (by measuring the steady state
conversion in the MR, as well as the steady state conversion in the AR, after the adsorbent was
saturated - see discussion to follow) and was shown to remain virtually identical for the duration
of the experimental cycle lasting ~ two weeks. We monitored, in addition, the state of the CMS
membrane by measuring the pure-gas permeances of all syngas components (H2, N2, CO, CO2,
CH4), in addition to He and Argon, both before the MR-AR experiments were started, and also
33

after the MR-AR experiments were completed (as noted above, for these CMS membranes, single-
gas and mixed-gas permeances are close to each other [17], and the choice of measuring the single-
gas permeances is because they are much more convenient and accurate to measure). The measured
permeance along with the corresponding IS values are shown in Table 3. Within the experimental
accuracy of some of these measurements, the membrane properties remain quite invariant
(particularly of the slow non-adsorbing gases like Ar and N2, which are a much more sensitive
indicator of the membrane developing cracks and pinholes).
Table 3. Single-gas permeances and ideal selectivities before and after the MR-AR experiments at 250°C
and 25bar.

Gas permeance
(m
3
(STP)/m
2
h bar)
Ideal selectivity
(IS)
He H 2 Ar N 2 CO CO 2

CH 4 H 2/
Ar
H 2/
CO
H 2/
CO 2
H 2/
CH 4
Before 1.0090 1.2112 0.0130 0.0099 0.0143 0.0346 0.0088 93 85 35 138
After 1.0397 1.2925 0.0130 0.0107 0.0153 0.0369 0.0095 99 79 35 136

34

Figure 7. Conversion vs. W/F CO for the membrane reactor and packed-bed reactor at 250°C.

Figure 8. Hydrogen purity vs. W/F CO for the membrane reactor and packed-bed reactor at 250°C.
35

Figure 9. Hydrogen recovery vs. W/F CO for the membrane reactor at 250 °C.
2-3-2. Combined MR-AR Experiments
The performance of the two AR’s (AR I and AR II) during the combined MR-AR
experiments, in terms of the various exit gas molar flow rates in the AR (i.e., H2, CO, CO2), the
CO conversion in the AR (Eq. 5) as well as the total conversion in the MR-AR combined system
(Eq. 2) as a function of time, is presented in Figure 10-14 (the gas compositions and conversions
in these Figures are measured via the RGA instrument, which samples and measures the exit
composition at a frequency of 1 sample-point per second, so what is shown in Figure 10-14 are the
lines passing through this multitude of points). In Figure 10, we show the behavior for both AR’s
for an operating pressure (for both the MR and AR) of 25 bar, and a feed W/FCO (into the MR) of
33 g·h/mol during the first cycle of operation. Both reactors show typical AR behavior [29],
whereby the initial AR conversion begins at a high value (~100% in this case) and starts to decline
as the adsorbent gets saturated with CO2 leveling-off finally at the conversion that would be
attained if the reactor was operating as PBR (~64.26% for AR I and ~60.1% for AR II).
36

Figure 11 shows the behavior for both AR’s for a larger value of W/FCO (into the MR) of 77
g·h/mol. The experimental behavior is similar to that in Figure 10 (for the smaller value of W/FCO),
other than the fact that the two reactors show ~ 100% conversion for a more extended time period,
and the steady-state conversions in both ARs levels-off at a higher value of ~85%. Figure 12 and
13 show the behavior for the MR-AR experiments with the reactors operated at 250°C and at 15
bar of pressure. Though the qualitative behavior is similar to that at higher pressures, the reactor
appears to be less efficient at lower pressures never reaching 100% conversion. This is, likely, the
outcome of decreased residence times and adsorption rates at the lower pressures. In the present
study, during each MR-AR experiment, the ARs were allowed to run for ~1 h to approach steady-
state, so as to monitor the state of the catalyst and adsorbent. In field operations, of course, the AR
will be taken off-line, once the carbon loss (and/or hydrogen purity) would exceed (or fall below)
a certain threshold value dictated by the corresponding process design calculations.
After the adsorption/reaction part of the experimental cycle in a given AR was completed,
the temperature in this AR was raised to 400°C under a gas mixture (i.e., 50 vol.% steam - 50 vol.%
Argon) flow and kept there for an additional 30 min for the adsorbent to be regenerated. (In the
meantime, as noted in Section 2-2-2, the reject side from the MR was switched to the other AR for
the adsorption/reaction part of the cycle for this reactor to commence). Upon completion of the
regeneration the temperature of the AR was lowered under the same gas mixture flow to 250
o
C
for the second adsorption/reaction cycle to begin. Figure 14 shows the behavior of the MR-AR
system, operated at 250 °C and a pressure of 25 bar (with a MR feed W/FCO equal to 55 g·h/mol)
after two cycles of operation.

37

Figure 10. CO conversion in the AR and total MR-AR conversion, and molar flow rates of CO 2, H 2, CO
in the two AR’s. (Left) AR I, and (Right) AR II. Temperature of 250°C, pressure of 25 bar, H 2O/CO ratio
of 2.8, with the MR operated with a W/F CO of 33 g·h/mol.

The data in Figure 14 indicate quite a reproducible behavior, which is indicative of the fact that
the hydrotalcite adsorbent and WGS catalyst used in this study show good reversibility during the
sorption-desorption cycles and remarkable stability in the coal-gasifier off-gas atmosphere. That
these hydrotalcite materials are capable of functioning so well under these harsh conditions is in
line with cyclic adsorption/desorption studies by our Group under non-reactive conditions(, and
prior AR studies under more moderate conditions by other investigators [50]. And it points out the
promise that all materials (membranes, catalysts, and adsorbents) tested here show for the practical
application of the proposed technology.

Figure 11. CO conversion in the AR and total MR-AR conversion, and molar flow rates of CO 2, H 2, CO
in the two AR’s. (Left) AR I, and (Right) AR II. Temperature of 250°C, pressure of 25 bar, H 2O/CO ratio
of 2.8, with the MR operated with a W/F CO of 77 g·h/mol.

38

Figure 12. CO conversion in the AR and total MR-AR conversion, and molar flow rates of CO 2, H 2, CO
in the two AR’s. (Left) AR I, and (Right) AR II. Temperature of 250°C, pressure of 15 bar, H 2O/CO ratio
of 2.8, with the MR operated with a W/F CO of 33 g·h/mol.

Figure 13. CO conversion in the AR and total MR-AR conversion, and molar flow rates of CO 2, H 2, CO
in the two AR’s. (Left) AR I, and (Right) AR II. Temperature of 250°C, pressure of 15 bar, H 2O/CO ratio
of 2.8, with the MR operated with a W/F CO of 55 g·h/mol.

39

Figure 14. CO conversion in the AR and total MR-AR conversion, and molar flow rates of CO 2, H 2, CO
in the two AR’s. (Left Top) AR I, first cycle, (Right Top) AR II, first cycle, (Left Bottom) AR I, second
cycle, (Right Bottom) AR II, second cycle. Temperature of 250°C, pressure of 25 bar, H 2O/CO ratio of
2.8, with the MR operated with a W/F CO of 55 g·h/mol.
2-4. Conclusions
In this study, a hybrid MR-AR system, which consists of a membrane reactor followed by
two adsorptive reactors was experimentally evaluated for the WGS reaction in the context of the
IGCC power generation application. The stand-alone MR system showed good conversions of CO
(~80%) and hydrogen purity (>75%) under the IGCC-relevant operating pressure at 25 bar that
substantially exceed those of the PBR (<25%) under the same conditions. The combined MR-AR
lab-scale system was experimentally tested at different pressures and reactor space times, and
displayed superior performance to that of a conventional packed-bed reactor with near 100%
conversions attained while the AR’s are functional and with hydrogen purities which are
significantly better to those attained in the PBR. It was also shown to operate stably under the
40

experimental conditions studied. Specifically, our study has shown that the CMS membrane, sour
shift catalyst, and hydrotalcite adsorbent employed are all robust in the presence of a simulated
coal-gasifier off-gas, under IGCC-relevant high-temperature and high-pressure conditions. The
results of this initial feasibility-type study suggest that the proposed hybrid MR-AR system is a
promising technology for incorporation into IGCC power plants for simultaneous power
generation and pre-combustion CO2 capture. The additional parametric studies of the hybrid MR-
AR system to determine the optimal operating conditions to improve the process performance, and
long-term (>100 h) experimental evaluation to further validate materials stability and performance
will be detailed in the following chapter.

41

Chapter 3: MR-AR Experiments with Steam Sweep
3-1. Introduction
In the last Chapter we have presented a preliminary experimental study of a novel reactor
configuration, consisting of a membrane reactor (MR) followed by two adsorptive reactors (ARs)
in parallel, operating alternately, utilized for the production of high-purity hydrogen with
simultaneous CO2 capture during the water−gas shift reaction treating a coal gasifier off-gas. In
the study, we used a commercial sour-shift WGS catalyst (Co/Mo/Al2O3) in both the MR and the
AR. A carbon molecular sieve (CMS) membrane was used in the MR, and a hydrotalcite adsorbent
was used in the AR. The experimental results show that membrane, catalyst, and adsorbent all
operated stably under the integrated gasification combined cycle (IGCC)-relevant conditions. The
MR−AR reactor sequence displayed performance superior to that of a conventional packed-bed
reactor (PBR) with near 100% conversions attained while the ARs are functional (with an ultrapure
hydrogen stream exiting the AR and permeate-side hydrogen purities from the MR of ∼75−80%).
Thus, these findings manifest the ability of the hybrid MR−AR process configuration to operate
properly under the desired conditions and to intensify the efficiency of the WGS reaction, as well
as to validate its potential to function as a high-efficiency, ultra-compact process for incorporation
into IGCC power plants for environmentally-benign power generation with pre-combustion CO2
capture.
In the previous study, we operated the integrated MR-AR without steam sweep by
employing a specific CMS membrane with only one end open to flow (termed as the candle-filter
membrane configuration [17]). To compensate for the water losses into the permeate-side of the
MR, a separate steam-generating unit was employed to supply additional steam to the AR feed
stream; this however, increases the capital cost for the process due to the fact that high temperature
(250°C) and high pressure (up to 25 bar) steam is very costly. In order to improve the process
design for this integrated MR-AR system, we upgraded the integrated MR-AR lab-scale system so
that we can provide a steam sweep in the MR’s permeate side, and also to be able to increase the
permeate-side pressure above atmospheric.
42

The advantages of operating the MR-AR system with steam sweep include the following:
(i) Adding steam sweep during MR operation is able to reduce the water losses to the permeate-
side from the reject-side of the MR, thus enhancing the CO conversion and H2 recovery by
eliminating the loss of reactant H2O from the reject-side of the MR where the WGS reaction takes
place; (ii) the use of steam sweep on the permeate-side of the MR as well the increase in the MR’s
permeate-side pressure helps to maintain the H2O:CO ratio in the exit of the MR reject-side to a
sufficiently high level so that there is no need to supply any additional high-pressure and-
temperature steam to the AR feed stream. The latter requires installing a separate steam-generating
unit in between the MR and the AR system, as we did in the experiments reported in Chapter 2.
A key objective of this series of experiments, furthermore, was to investigate the feasibility of
running the integrated MR-AR system for long time periods (500 hr long, much longer than the run
times reported in Chapter 2). This is in preparation for a similar length, future bench-scale run under
field conditions employing real syngas. An important goal for the experiments was to investigate
catalyst/adsorbent stability and, in particular, the CMS membrane performance stability during
exposure to H2S and other syngas components for a period of >500 hr. Furthermore, during the over
500 hr long run, the plan is to carry-out additional parametric studies of the hybrid MR-AR system to
determine the optimal operating conditions in order to improve the process performance. The
optimization parameters we investigate include the operating conditions for the MR (e.g., W/FCO,
H 2O/CO ratio, steam sweep ratio, steam sweep pressure in the MR’s permeate side, etc.) and for the
AR (e.g., sorption-enhanced WGS reaction time, adsorbent regeneration time, regeneration pressure,
etc.). The experimental findings of this >500-hr experimental run are described below.
3-2. Experimental Section
3-2-1. Materials
For the experiments reported in this Chapter performed with the hybrid MR-2AR system
while utilizing steam as a sweep, a tubular CMS membrane with an inner diameter of 3.6 mm,
outer diameter of 5.6 mm, and a length of 254 mm was utilized in the MR. The CMS membrane
was prepared by our industrial collaborator in this project Media and Process Technology, Inc.
43

(M&PT). The membrane consists of a thin nano-porous CMS separation layer formed on the
outside surface of a M&PT commercial asymmetric mesoporous alumina ceramic tube. Unlike the
CMS membrane utilized in the experiments in Chapter 2, which had only one end open to flow
and was employed in the MR in the candle-filter membrane configuration [17, 53], the particular
CMS membrane utilized in the experiments reported here has both ends open to flow. During the
MR experiments, the membrane is sealed on both ends to the tubular SS reactor using graphite O’
rings and Swagelok compression fittings. Steam, when employed as a sweep, flows from the one
open end of the tube and exits from the other end. The steam sweep is generated by employing a
high-pressure water syringe pump and a steam-generating unit (evaporator). A commercial
Co/Mo/Al2O3 sour-shift catalyst provided by Clariant (USA) was utilized for both the MR and the
AR units. Its physical and chemical properties were reported in Chapter 2, as well as in a recent
paper [17] by our team. As noted in Chapter 2 and our recent paper [53], we have also used in the
AR unit a Mg-Al-CO3 double- layered hydroxide (with a Mg/Al molar ratio of 3:1) CO2 adsorbent
prepared by M&PT using conventional approaches [51, 52].
3-2-2. Experimental Set-up and Procedure
The schematic of the lab-scale MR-AR set-up equipped with a steam sweep system is
shown in Figure 16. The system was modified from the previous MR-AR system used in Chapter
2, by adding a high-pressure water syringe pump and evaporator on the MR permeate side, in order
to generate the steam that can then be used as sweep stream in the MR permeate side. In addition,
a back-pressure regulator (BPR) was installed at the line connected to the exit of the MR permeate
side prior to the water condenser, in order to be able to control the steam sweep pressure on the
MR’s permeate side (in the experiments in this Chapter, the pressure was varied in the range from
1-3 bar). 10 g of commercial Co/Mo/Al2O3 sour-shift catalyst intermixed with glass balls of
similar particle size (600~850 μm in diameter) were loaded into the annular space in the MR in
between the reactor body and the CMS membrane. The remaining parts of the experimental set-
up are detailed in Chapter 2 (see Section 2-2-2).

44

Figure 16. Experimental set-up used in the MR-AR experiments (red lines are heat-traced to prevent
water from condensing; MR: Membrane Reactor; AR: Adsorptive Reactor; MFC: Mass Flow
Controller; BFM: Bubble Flow-meter; RGA: Residue Gas Analyzer; GC: Gas Chromatography;
BPR: Back Pressure Regulator;).

To study whether the state of the CMS membrane is stable during the MR-AR experiments
with steam sweep, the single-gas permeances of the major syngas components (i.e., CO, CO2, CH4,
H2, N2) and of He (an inert fast gas which serves as a safe surrogate to monitor the H2 permeance
of the CMS membrane) were frequently measured at predetermined temperatures and pressures.
Specifically, to initiate these single-gas permeation tests, we adjusted the BPR to pressurize the
feed-side of MR up to its appropriate value, while the MR permeate-side pressure was, typically,
maintained at atmospheric conditions. A bubble flow-meter was employed to measure the MR’s
reject-side and permeate-side flow rates. The laboratory temperature and pressure were measured
with a digital thermometer and an analog barometer respectively in order to be able to convert the
volumetric flows to STP conditions (m
3
/h). We have previously validated experimentally [17] that
single-gas and mixed-gas permeances with these CMS membranes are very close to each other,
and so are the ideal selectivity and separation factor. Thus, in our studies of membrane stability
during the MR-AR experiments we have routinely utilized the single-gas permeances, which are
45

significantly easier and much more straightforward and reliable to measure, rather than mixed-gas
permeances and separation factor.
For the MR-AR experiments with steam sweep presented here, we have used fresh catalyst,
in which the Co and Mo metal components are in their oxidized form and, therefore, need to be
activated (sulfided) prior to the initiation of the reactor runs. The activation procedure, which is
described in detail in Chapter 2 (see Section 2-2-3) follows a temperature and pressure protocol
recommended by the catalyst manufacturer [17]. The MR-AR experiments presented here are
carried-out using a simulated coal-derived syngas feed mixture (H2: CO: CO2: Ar: CH4:
H2S=0.51:1:0.36:2.28:0.1:0.0031, typical of an air-blown gasifier off-gas). Note that we substitute
in the syngas composition N2 with Ar, so as to facilitate the analysis via the RGA measurements,
because we have found that N2 interferes with the analysis of CO. However, our studies have
shown that both of these gases act as diluents and inert gases and do not participate in the WGS
reaction. Ar also acts in the dual capacity as an internal standard gas during the experimental runs.
The long-term (>500 hr) experimental run began by testing the MR-AR system
components individually prior to initiating the feasibility testing of the combined MR-AR system.
Experiments have been run at various W/FCO values for the MR feed (where W is the weight of
undiluted catalyst (g), and FCO is the molar flow rate of CO (mol/hr) in the syngas), as well as
different H2O/CO ratios and steam sweep ratios (defined as the ratio of the sweep stream molar
flow rate to the feed molar flow rate) in the MR’s permeate side. For all the experiments reported
here, we employed a steam sweep at two different permeate-side pressures of 1 and 3 bar
(experiments with no sweep have also been carried-out). All the MR-AR experiments reported
here were performed at 250°C and a MR reject-side and AR pressure of 25 bar (experiments
employing a MR-AR pressure of 15 bar have also been carried out). As part of the testing of the
stability of the catalyst, adsorbent and membrane in the integrated MR-AR system, we have
performed multiple-cycle (typically 10 - 16 cycles) MR-AR experiments. During these
experiments, we have tested different MR feed pressures, W/FCO values, H2O/CO ratios, steam
sweep ratios and different steam sweep pressures in the MR permeate side and evaluated their
impact on the performance of the AR in the cyclic operation, the goal here being to study the
optimum conditions for the MR-AR system as a whole. Furthermore, during the multi-cycle run,
46

we have tested the system performance under different AR reaction times (25-58 min), different
regeneration pressure (15-25 bar), and different regeneration times at 400°C (10-70 min). Prior to
initiating the combined MR-AR experiments, the MR conversion was first allowed to reach steady
state for at least 2 hours. During this time period, the RGA was utilized to measure the gas
composition for both MR’s reject side and permeate side. To “benchmark” the MR performance
for these experiments, we also measured the corresponding conversion of PBR under the same
operating condition by closing the exit on the permeate side of the MR and allowing the system to
operate as a PBR under steady-state conditions.
As illustrated in Figure 16, the reject-side (retentate) stream (a CO2-rich stream, that also
includes some H2, unreacted CO, CH4, Ar, and impurities like H2S) from MR serves as the feed
for the AR. Similar with the MR-AR experiments without steam sweep detailed in Chapter 2 (see
Section 2-2-3), although we have the capability of operating the AR at different temperatures and
pressures than those in the MR, in the experiments presented here we have kept the operating
temperature and pressure for both reactors the same. As noted in Section 3-2-2 above, instead of
adding additional steam to the feed stream into the AR to potentially compensate for steam losses
into the permeate side of the MR (in an effort to maintain the desired H2O/CO ratio in the AR feed
stream), the installation of a separate steam-generating unit (high-pressure water syringe pump +
evaporator) on the MR permeate side allowed for steam to be generated and used as a sweep stream
in the MR permeate side. The use of steam sweep, and also increasing the permeate-side pressure
(up to 3 bar in the experiments reported in this Chapter) with the aid of the BPR installed at the
end of the MR’s permeate stream line, has allowed us to maintain the desired H2O/CO ratio in the
AR feed stream without needing to supply additional steam in between the MR and the AR units,
as we did in the experiments presented in Chapter 2.
To initiate the operation of the combined MR-AR system, the MR subsystem performance
(in terms of CO conversion, hydrogen recovery and purity, which are monitored via a GC that is
used to measure the composition of the reject side and that of the permeate side) is, typically,
allowed to stabilize before switching the MR reject stream as a feed into the AR(s). Meanwhile,
both ARs are simultaneously pressurized to the desired pressure (up to 25 bar) and heated to the
desired temperature using a 50 vol %/50 vol % steam/Argon gas mixture. Once the MR reject side
47

flow rate, gas composition and steam concentration are stable, we switch the MR reject-side stream
to the AR feed-side via a 3-way valve and begin the MR-AR multi-cycle experiments. While the
MR reject-side is directed into the first AR (AR I), the 50 vol %/50 vol % steam/Argon stream
continues to be directed into the second AR (AR II). During the experiment, the AR I outlet gas is
measured instantaneously via the RGA I; in most of the experiments reported here, we have
allowed for the sorption-enhanced water gas shift reaction time in the AR to be long enough (25-
58 min depending on the operating conditions selected) so that the adsorbent became saturated
with CO2, and we obtained the so-called “pseudo-steady state” for CO conversion in the AR prior
to being switched into the regeneration mode. Once the AR I system reached the “pseudo-steady
state”, the MR reject-side stream was then switched into the second AR (AR II), while the 50
vol %/50 vol % steam/Argon stream was directed into AR I to regenerate the adsorbent in the
reactor at the desired temperature (a temperature of 400°C was employed here) and duration (10-
70 min in these experiments for investigation of the optimization parameter). After completing the
evaluation of the performance for both the AR I and AR II subsystems, the MR’s reject-side stream
was switched back away from the feed of the AR, and the MR performance (including the gas
composition and flow rate of both reject and permeate sides) was studied again to verify if the MR
performance has remained stable. It should be noted that during the multiple-cycle (10-16 cycles)
run with the integrated MR-AR with steam sweep, the MR system performance and its robustness
to the simulated coal-derived syngas conditions was monitored before the multiple-cycle run, in
the middle of the multiple-cycle run and after the multiple-cycle run, which had been completed
by employing a GC to measure the MR reject-side and permeate-side stream gas composition and
a bubble flow meter (BFM) to measure the dry-gas flow rate.
3-3. Results and Discussion
3-3-1. Membrane Studies
As noted above, a CMS membrane prepared at M&PT was employed in this study. The
membrane was installed into the MR empty module at the M&PT laboratories, and prior to the
module being shipped to USC side, its permeation characteristics were tested using helium and
48

nitrogen (both of which serve as model inert gases, where He is indicative of fast gases like H2 and
the N2 is indicative of slow gases like CO). Upon receiving the MR module with the CMS
membrane from M&PT, the single-gas permeances of He and N2 were again measured at the same
temperature and pressure conditions at USC. The properties of the as received membrane (in the
empty module) measured at M&PT as well as at USC are shown in Table 4.
Table 4. Single- gas permeances of N 2 and He at various temperatures and pressures
measured in the empty module
Lab
T
(℃)
Pressure
(bar)
He
*
N2
*

SF
§

(He/N2)
M&PT 20 3.07 0.1477 0.00827 18
M&PT 250 3.07 1.2060 0.00675 179
USC 250 3.07 1.3265 0.00702 189
USC 250 25 1.2740 0.00824 155
*
Gas permeance [m
3
/m
2
h bar];
§
Separation factor.

There is a difference observed in the measurement of He permeances among the two
laboratories (See Table 4). The He permeance measured at USC is 9.1 % larger than the M&PT
measurement, however, there is only a slight difference observed in the measurement of N2, which
is below the 5% experimental error, which is typical with such measurements. Consequently, the
He/N2 separation factor measured at USC is about 5% (189 vs. 179) larger than its value measured
at M&PT. We have previously observed such differences (5-10%) in the He measurements among
the two laboratories and reported them in previous publications [17, 53].
Subsequently, 10 g of catalyst intermixed with glass balls of similar particle size (600~850
μm in diameter) were loaded into the reactor, and the membrane permeation characteristics were
measured again, and the values are shown in Table 5. There is a statistically significant decrease
in the He permeance, and a smaller (and, likely, statistically insignificant) decrease in the N2
49

permeance. That both permeances decrease, is indicative of the fact that the loading of the catalyst
caused no damage to the membrane, because when the membrane’s CMS layer is damaged the N2
permeance increases greatly. We ascribe the decrease in the He permeance to the catalyst/glass
balls potentially blocking/hindering access to parts of the surface of the CMS membrane.
Table 5. Single- gas permeances of N 2 and He at 250℃ and 25bar
measured in the module after loading the catalyst
T
(℃)
Pressure
(bar)
He
*
N2
*
SF
§

(He/N2)
250 25 1.0382 0.00807 129
*
Gas permeance [m
3
/m
2
.h.bar];
§
Separation factor.

The catalyst was then activated. The activation procedure has been described in detail
elsewhere [17, 53], and involves exposing the catalyst to 5% H2S in a N2/H2 mixture at high
temperatures for ~ 36 hr. After the catalyst activation, the membrane properties were measured
once more and are shown in Table 6 below.
Table 6. Single- gas permeances of He, N 2, H 2, N 2, A r, CO, CO 2 and methane at 250℃ and 25bar
after the catalyst activation
Test date He
*
N2
*
H2
*
CO
*
CO2
*
Ar
*
CH4
*

SF
§

(He/
N2)
SF
§

(H2/
CO)
After
catalyst
activation
1.0206 0.00857 0.9507 0.00981 - - - 119 97
After 2
days
- - - - 0.01375 0.00768 0.01038 - -
After 3
days
1.0192 0.00866 0.9679 0.00981 - - - 118 99
*
Gas permeance [m
3
/m
2
.h.bar];
§
Separation factor.

50

From Tables 5 and 6 above, one notices that a slight ~1.7% (statistically insignificant) decrease in
the He permeance and a small ~6.1% increase in the N2 permeance after activating the catalyst.
The H2/CO separation factor of around 98 - 99 is well above the DOE target value of 80.
One of the key objectives of the MR-AR with steam sweep experiments is to evaluate the
membrane properties stability under these harsh experimental conditions involving exposing the
membrane with large concentration of H2S and high-temperature and high-pressure steam during
the long period run (over 500 hr). During the subsequent four months of testing of the MR-AR
system, the membrane properties were shown to be very stable, as shown in Table 7 and in Figures
17 and 18. The only notable exception is the CO2 permeance for which the initial value measured
is ~11.7% less than its final value measured. We ascribe this difference to experimental error in
the measurement of the original value (since the value at the end of the 500 hr was measured
repeatedly), because there are only slight, if any, differences measured with the other slow gases
(e.g., the initial and final permeance values for CO and N2 are virtually identical), and there is no
valid scientific explanation why only the CO2 permeance would change significantly during the
long-period run (in fact, measurement of the permeance of condensable gases like CO2 in
microporous membranes, at high pressure, is notoriously difficult because of the long transients
observed).
Table 7 (and Figures 17 and 18) report the He, N2, H2 and CO permeances and the He/N2,
H2/CO selectivities as a function of the cumulative H2S/syngas exposure time. As the data
presented in the Table (and in the Figures) show, after 542 hr of exposure time to H2S/syngas, the
permeances for He, H2, N2 and CO were experimentally indistinguishable from those measured
before the run, and the H2/CO separation factor stayed constant at 98, which is significantly higher
than the project target value of 80. Thus, one may conclude that for this CMS membrane, the
membrane properties are very robust and stable under the cumulative 542 hours run of H2S
exposure and under the 250°C temperature and 25 bar pressure environments. After completing
all the experiment up to 25 bar pressure, we continued the experiment with 15 bar pressure. The
membrane still shows high He/N2 selectivities (~126) over a total of 742-hour MR-AR with steam
sweep run as shown in Table 9.
51

Table 7. Single- gas permeances of He, N 2, H 2, N 2, CO, at 250℃ and 25bar during the 500-hour run
Exposure to
Syngas/H 2S
Hr
He
*
N 2
*
H 2
*
CO
*

SF
§

(He/N 2)
SF
§

(H 2/CO)
Comments
0 1.0382 0.00807 - - 129 -
After
loading the
catalyst
36 1.0206 0.00857 0.9507 0.00981 119 97
After
catalyst
activation
36 1.0192 0.00866 0.9679 0.00981 118 99 -
50 0.97741 0.00810 - - 121 - -
66 1.0003 0.00804 - - 124 - -
84 0.99471 0.00817 - - 122 - -
108 0.995 0.008277 - - 120 - -
122 1.023 0.00857 - - 119 - -
186 1.004945 0.008252 - - 122 - -
244 1.005 0.008247 - - 122 - -
265 1.01894 0.00833 - - 122 - -
402 0.9943 0.008288 - - 120 - -
542 1.0061 0.00854 - - 118 - -
542 1.0087 - 0.9558 0.00980 - 98 -
*
Gas permeance [m
3
/m
2
h bar];
§
Separation factor.

Table 8. Single- gas permeances of A r, CO 2 and methane before and after the MR-AR experiment
Lab/Test
date
T
(℃)
Pressure
(bar)
CO2
*
Ar
*
CH4
*

USC,
09/29/2018
250 25 0.01375 0.00768 0.01038
USC,
01/18/2019
250 25 0.01558 0.00828 0.01027
*
Gas permeance [m
3
/m
2
.h.bar];
§
Separation factor.

Table 9. Single- gas permeances of He, N 2 at 250℃ and 15 bar after the experiment with a total 742-hour run
Exposure to
Syngas/H 2S
Hr
He
*
N 2
*
H 2
*
CO
*

SF
§

(He/N 2)
SF
§

(H 2/CO)
Comments
742 1.0203 0.008071 - - 126 -
After the MR-
AR
experiment
52

Figure 17. Single-gas permeances as a function of H 2S exposure time during the 500-hour run

Figure 18. He/N 2 and H 2/CO selectivities as a function of H 2S exposure time during the 500-hour run.
53

Another key focus of this research was also on the optimization operation parameter. Thus
we initiated the experiments with the hybrid MR-AR system with the MR and AR subsystems
operating at the same temperature (250°C) and pressure (15 and 25 bar), while varying the W/FCO,
steam sweep pressure, H2O/CO ratio, steam sweep ratio for the MR. Note that before initiating the
parametric study with the combined MR-AR with steam sweep system (i.e., switching the reject-
side stream from the MR as feed into the AR), the MR was allowed to operate until reaching steady
state and its conversion, H2 purity (dry-basis) in the permeate stream and recovery were all
measured and presented Figures 19 and 21-26. It should be noted that, since we employed the
steam sweep at the MR’s permeate side while not supplying any additional steam between the MR
and AR unit, we also monitored the MR’s reject-side H2O/CO ratio. For example, as shown in
Figure 20, under that experimental condition the MR’s reject-side H2O/CO ratio is below 1 when
the MR’s permeate-side steam sweep pressure is 1 bar. Thus, to prevent the catalyst in the AR unit
from getting “coked”, we employ the steam sweep pressure of 3 bar.

Figure 19. Conversion vs. W/F CO for H 2O/CO ratio of 3.22 with MR operated at SR=0.49, SP=3bar, 25
bar and 250°C
54

Figure 20. MR reject H 2O/CO ratio vs. sweep pressure with MR operated at W/F CO=66 g.h/mol, H 2O/CO
ratio of 3.22, SR=0.49, 25 bar and 250°C

Figure 21. H 2 purity vs. W/F CO for H 2O/CO ratio of 3.22 with MR operated at SR=0.49, SP=3bar, 25 bar
and 250°C
55

Figure 22. H 2 recovery vs. W/F CO for H 2O/CO ratio of 3.22 with MR operated at SR=0.49, SP=3 bar, 25
bar and 250°C

Figure 23. H 2 purity and H 2 recovery vs. H 2O/CO ratio with MR operated at W/F CO=66 g.h/mol,
SR=0.49, SP=3 bar, 25 bar and 250°C
56

Figure 24. H 2 purity and H 2 recovery vs. pressure for H 2O/CO ratio of 4.3 with MR operated at
W/F CO=66 g.h/mol, SR=0.49, SP=3 bar, 250°C

Figure 25. H 2 purity and H 2 recovery vs. sweep ratio for H 2O/CO ratio of 4.3 with MR operated at
W/F CO=66 g.h/mol, SP=3 bar, 15 bar and 250°C
57

Figure 26. Conversion vs. MR H 2O/CO ratio with MR operated at W/F CO=66 g.h/mol, SR=0.49, SP=3
bar, 25 bar and 250°C
In addition, we also measured the conversion and hydrogen purity of the reactor functioning as a
PBR under the same experimental conditions (using the approach detailed in Section 3-2-2), and
these are also presented in the same figures.
As can be seen in Figure 19, for the experiment performed at H2O/CO=3.22, with steam
sweep ratio=0.49, steam sweep pressure=3 bar, the difference between the MR CO conversion and
PBR CO conversion under different W/F CO is relatively small (~3 % for W/F CO =44, ~6 % for W/F CO
=66). This is due to the fact that when employing the steam sweep pressure up to 3 bar, the
difference between the MR’s reject side and permeate side pressure decreases, and so is the
difference with respect to CO conversion between the MR and the PBR. Also note that in Figure
26, the difference between the MR and the PBR for CO conversion decreases when increasing the
MR’s feed side H2O/CO ratio (~3 % for H2O/CO ratio=4.3, ~6 % for H2O/CO ratio=3.22), which
suggests that it is less efficient from the MR’s performance standpoint to operate at very high
H2O/CO ratio (also the operating cost will increase when increasing the H2O/CO ratio since the
high-temperature and high-pressure steam is very costly).
58

The key advantage of using the MR is shown in Figure 21, which compares the H2 purities
of the processed syngas for the MR and the PBR. Since the air-blown coal gasifier off-gas contains
a large fraction of N2, the hydrogen purity of the PBR is so poor that it is hardly appropriate for
use in power generation. However, the use of the MR that combines reaction and separation in one
unit, in addition to enhancing the CO conversion, also provides a H2 product stream with superior
purity which can be directly usable in a hydrogen turbine for power generation. It should be noted
that the calculated equilibrium conversion under these conditions is quite high (over 98 %), and
for the W/FCO and the membrane area employed here, neither the PBR nor the MR reach this value.
Also, as seen in Figure 21, the H2 purity decreases when increasing the W/FCO for MR while the
situation is reversed for the PBR, with the explanation for this behavior found in our recent
publication [17, 21, 22, 24]. However, increasing W/FCO can significantly enhance the H2 recovery,
thus, the investigation of an optimal operating W/FCO condition is required, if one needs to meet
both the H2 purity and recovery requirements.
From Figures 23 and 25, one may note that the change in both the H2 purity and recovery
is insignificant when increasing either the H2O/CO ratio or the steam sweep ratio. This
experimental result validates the modeling finding by our team and the UCLA team [17, 54] which
collaborates with us on this project. One would recommend using a H2O/CO ratio, typically,
between 2 to 3 and steam sweep ratio ~0.1 to reduce the capital cost and the energy penalty of
generating the high-temperature and/or high-pressure steam. Furthermore, it can be seen in Figure
24, when increasing the system pressure (from 15 bar to 25 bar), there is only a slight change for
the H2 purity, but significant enhancement for the H2 recovery, thus it would be much more
efficient to operate the system under high pressure.
3-3-2. MR-AR Multiple-Cycle Run
To assure that upstream of the AR (i.e., the MR reject side which serves as the feed stream
for AR), the state of the membrane and catalyst in the MR is stable, during each multi-cycle
experiment, we continue to monitor the MR subsystem performance with respect to MR CO
conversion, H2 purity and recovery in order to verify both membrane and catalyst stability. Figure
27 shows these MR properties during a multi-cycle MR-2AR run as a function of time of stream
59

(expressed in this Figure as syngas/H2S total exposure time). At the beginning of this run, the
catalyst and membrane had been exposed to syngas/H2S for a total of 405 hr and by the time the
final MR performance measurement, shown on this Figure, was made the catalyst and membrane
had undergone an additional ~70 hr of exposure. It is clear from the Figure, that both and
membrane and catalyst employed in the MR show quite stable performance.

Figure 27. The MR properties during a multi-cycle run. Experimental conditions: T=250°C, feed
pressure=25 bar, permeate-side pressure 3 bar, with steam sweep ratio=0.49, H 2O/CO ratio of 4.3,
W/F CO= 66 g·h/mol, air-blown gasifier model syngas (CMS#23).
In the MR-AR with steam sweep multiple-cycle run experiments, to properly characterize
the AR behavior [21, 22, 24, 53], we have defined the “AR effect time” to be two different criteria:
(i) the difference between the H2 emergence time (to account for any system dead times) and the
time when the CO2 exit composition reaches 5 %; (ii) the difference between the H2 emergence
time (to account for any system dead times) and the time when the AR’s CO conversion decreases
60

below 95 %. During the aforementioned time period, the AR works in the sorption-enhanced water
gas shift reaction mode, whereby the adsorbent is still effective, and only a small amount of CO2,
specifically below 5%, exits the reactor, and the conversion in the AR is >95%, which also means
that the conversion for the combined MR-AR system is >>95% as well (as the exit stream from
the MR is the feed-stream into the AR).. Figures 28-29 show the “AR effect time” during a multi-
cycle MR-AR run (for the first 10 cycles we employed the MR-2AR configuration, with the
remaining 6 cycles being run in the MR-AR configuration). From this Figure, one notices that the
“AR effect time” takes a few cycles to settle to the eventual “steady-state” value of ~400 sec.

Figure 28. The AR effect time w.r.t 95 % CO conversion during multi-cycle run. Experimental
conditions: T=250°C, feed pressure=25 bar, permeate-side pressure 3 bar, with steam sweep, W/F CO= 66
g·h/mol, air-blown gasifier model syngas (CMS#23).
61

Figure 29. The AR effect time w.r.t 5 % CO 2 composition during multi-cycle run. Experimental
conditions: T=250°C, feed pressure=25 bar, permeate-side pressure 3 bar, with steam sweep, W/F CO= 66
g·h/mol, air-blown gasifier model syngas (CMS#23).

62

Figure 30. The AR steady-state CO conversion during a multi-cycle run. Experimental conditions:
T=250 °C, feed-side pressure=25 bar, permeate-side pressure 3 bar, with steam sweep, W/F CO= 66
g·h/mol, air-blown gasifier model syngas (CMS#23).
In addition, for the experiments shown in Figure 28-30, we allowed for the sorption-enhanced
water gas shift reaction time to be long enough (58 min in this case) so that the adsorbent becomes
saturated with CO2. Figure 30 reports the “pseudo-steady state” CO conversion in the AR prior to
being switched into the regeneration mode. It takes again a few cycles before the conversion settles
down to its eventual steady state value and the catalyst activity and adsorbent performance become
stable.
Furthermore, we have also previously investigated catalyst robustness in the AR during
adsorbent regeneration. We have performed a continuous 18-cycle experiment during which we
employed various regeneration temperatures (i.e., 350, 400, 450°C) and varied the regeneration
times (i.e., 10, 30, 60 min) for each selected temperature, with the results shown in Figure 31. (For
these experiments the AR sorption-enhanced water gas shift reaction run-time is selected long
enough so that the AR reaches its pseudo-steady state CO conversion). As can be seen in Figure
31, the CO steady state conversion remained quite stable, which validated the fact that the catalyst
is very robust during the various regeneration treatments.
63

Figure 31. AR pseudo-steady conversion after adsorbent saturation for various regeneration protocols, as
shown on the Figure. Temp=250°C, pressure=5 bar, Wc/F CO=121 g·h/mol, WAd/Wc= 6.9:1.
Because temperature is also a key operating parameter during the study, during the
experiments, the temperatures profiles during the sorption-enhanced water gas shift reaction mode
and the adsorbent regeneration mode at different bed points of the AR were also recorded via three
two-point thermocouples installed in the AR. A typical temperature profile is shown in Figure 32
(note that Inlet-0’’, Middle-0’’ and Outlet-0’’ signify three different equidistant axial positions in
the reactor, specifically 1.375 in, 2.75 in, and 4.125 in from the entrance of the bed, while Inlet-
0.5’’, Middle-0.5’’ and Outlet-0.5’’ signify three radial positions 0.5” away from the bed axis). As
Figure 32 indicates, the reactor is fairly isothermal with radial and axial profiles being, typically,
less than 2-5°C. The bed pressure drop is also very small (~0.1 psi) under all conditions studied
during the MR-AR experiments. The lack of significant pressure drops and temperature gradients
make it a more straightforward task to model the experimental data, and thus to validate the reactor
models to be used in process design and optimization and in the TEA studies.
64

Figure 32. Temperature profiles of AR operated at 250°C and 25 bar using W/F CO = 66 and H 2O/CO
=3.22, steam sweep ratio=0.5, permeate side pressure=3 bar.
3-4. Conclusions
In this study, an integrated MR-AR system, which consists of a MR followed by two ARs,
was modified from the previous MR-AR system used in Chapter 2, in order to have the capability
of employing steam sweep in the MR permeate side and controlling the steam sweep pressure in a
variable range (1-3 bar). This integrated MR-AR system with steam sweep was then
experimentally evaluated for the WGS reaction in the context of the IGCC power generation
application. The CMS membrane employed in the study has demonstrated to exhibit very robust
and stable performance during the long-term run (over 500 hr run of H2S exposure and under the
temperature of 250°C and up to 25 bar pressure environments) and maintained high He/N2
selectivity (~126) over a total of 742-hour of operation of the MR-AR run. The combined MR-AR
lab-scale system was experimentally tested during numerous multiple-cycle runs and displayed
65

superior performance to that of a conventional PBR with high purities for the hydrogen product
which can be directly usable in a hydrogen turbine for power generation.
Also, another key focus of this research was on the optimization of operation parameters,
for which we have performed the experiments with the hybrid MR-AR system with the MR and
AR subsystems operating at the same temperature (250°C) and pressure (15 and 25 bar), while
varying the W/FCO, steam sweep pressure, H2O/CO ratio, steam sweep ratio for the MR. This
parametric study shows that:
(i) By employing the steam sweep and increasing the sweep pressure, the MR’s reject-side
H2O/CO ratio managed to stay at a sufficiently high level (above 1), thus significantly enhancing
the CO conversion and H2 recovery by eliminating the loss of reactant H2O from the reject-side of
the MR where the WGS reaction takes place, and also reduce the energy penalty during the MR-
AR operation since there is no need to supply any additional high-pressure and -temperature steam
to the AR feed stream to prevent the catalyst in the AR unit from getting “coked”.
(ii) The change in both the H2 purity and recovery is insignificant when increasing either the
H2O/CO ratio or the steam sweep ratio. Thus, since it will be less efficient from the MR’s
performance standpoint to operate at very high H2O/CO ratio and/or steam sweep ratio, one would
recommend using H2O/CO ratio, typically, between 2 to 3 and steam sweep ~0.1 to reduce the
capital cost and the energy penalty of generating the high-temperature and/or high-pressure steam.
(iii) Increasing W/FCO can significantly enhance the H2 recovery, however, the H2 purity decreases
with increasing W/FCO. Thus, the investigation of an optimal operating W/FCO condition is
required if one needs to meet both the H2 purity and recovery requirements.
(iv) There is only a slight change for the H2 purity, but significant enhancement for H2 recovery,
thus it would be much efficient to operate the system under high pressure.
During the MR-AR with steam sweep multiple-cycle run, we continued to monitor MR
subsystem performance with respect to MR CO conversion, H2 purity and recovery in order to
verify both membrane and catalyst stability. The membrane and catalyst in the MR both displayed
66

stable performance during the long-period run. To properly characterize and monitor the AR
behavior, we have defined the “AR effect time” in two different ways, and during the MR-AR
multiple-cycle run, we have found that the “AR effect time” takes a few cycles to settle to the
eventual “steady-state”. We have also monitored the “pseudo-steady state” CO conversion in the
AR prior to being switched into the regeneration mode by allowing for the sorption-enhanced
water gas shift reaction time to be long enough so that the adsorbent becomes saturated with CO2.
We have also found it takes again a few cycles before the conversion settles down to its eventual
steady state value and the catalyst activity and adsorbent performance become stable. Further, the
catalyst in the AR demonstrated very robust and stable performance during the continuous 18-
cycle experiment under various regeneration treatment. Thus, one may conclude that the membrane,
catalyst and adsorbent are very robust and stable under the large concentration H 2S, high-temperature
and high-pressure environment during the long-period MR-AR multiple-cycle run.
67

Chapter 4: MR-AR Studies Using Oxygen-blown Coal
Gasifier Off-gas
4-1. Introduction
In Chapter 2, we have presented a preliminary MR-AR study with a membrane reactor
(MR) with no steam sweep and the permeate side operating at atmospheric pressure, followed by
two adsorptive reactors (ARs) in parallel, operating alternately, with maximum 2 cycles operation
for ARs. In Chapter 3, we have upgraded the system by employing steam sweep in the MR
permeate side and controlling the steam sweep pressure in a variable range (1-3 bar). This
integrated MR-AR system was experimentally evaluated for the WGS reaction in the context of
IGCC power generation application for a long-period operation. We have also experimentally
tested the combined MR-AR system during numerous multiple-cycle (up to 18-cycle operation for
ARs) runs and performed parametric study under different operating conditions for process
optimization. The membrane and catalyst in the MR, in addition, both demonstrated very stable
performance during the long-period operation, and so did the catalyst and adsorbent in the AR
during the multiple-cycle (up to 18 cycles) run under various regeneration treatments. During the
MR-AR multiple-cycle run, for the AR unit we have found that both the “AR effect time” and
“pseudo steady state” CO conversion take a few cycles to settle down to the eventual “steady-state”
where the catalyst activity and adsorbent performance become stable. Thus, from the last two
Chapters we have concluded that the membrane, catalyst and adsorbent are very stable and robust
under the IGCC environment (large concentration H2S, high-temperature and high-pressure)
during the long-period MR-AR multiple-cycle run.
The MR-AR experiments presented in Chapters 2 and 3 are carried-out using a simulated
coal-derived syngas feed mixture with composition H2: CO: CO2: N2: CH4: H2S=0.51: 1: 0.36:
2.28: 0.1: 0.0031, typical of an air-blown coal gasifier off-gas (it should be noted that N2
constitutes a large portion ~56% in this gas mixture). In this Chapter, we have investigated the
feasibility of running the MR-AR experiment using a different simulated coal-derived syngas feed
68

mixture with composition H2: CO: CO2: N2: CH4: NH3: H2S=0.91: 1: 0.59: 0.035: 0.081: 0.0086:
0.0038, typical of an oxygen-blown coal gasifier off-gas (this gas mixture by comparison
aforementioned air-blown gasifier off gas in Chapters 2 and 3 contains minute amounts of N2,
~3 %, and larger portions of hydrogen, ~34 %, and carbon dioxide, ~21%). The large fraction of
H2 and CO2 and small amount of nitrogen suggest potential differences in behavior of the MR-AR
when employing these two different gas mixtures. For example, one expects the flow of the reject-
side gas of the MR to be substantially smaller as compared with the air-blown gasifier off-gas case,
since most of the hydrogen as fast gas will permeate through the membrane.
In addition, in the studies in this Chapter, we have investigated using a lower regeneration
temperature for the AR unit, i.e., 350°C versus 400°C employed for the most part in Chapters 2
and 3. The advantages of operating the MR-AR system utilizing lower regeneration temperature
for AR includes decreasing the total time required for adsorbent regeneration and the power load
for operating the oven at such high temperature; this then reduces the operating cost and energy
penalty for the MR-AR process. A lower regeneration time also brings the desorption period much
closer to the reaction/adsorption time, so one may be able to effectively utilize two beds for
operation of the MR-AR system versus the four or greater number of units that are presently used
to accommodate the large imbalance between reaction/adsorption and desorption times
Similarly with the last two Chapters, one of the key objectives for this series of MR-AR
experiments, was to investigate the feasibility of running the integrated MR-AR system using
oxygen-blown coal gasifier off-gas for long-time periods (over 300 hr to date – the study is
presently ongoing), which serves as preparation for the future planned bench-scale field test
employing real oxygen-blown gasifier off-gas under field conditions. A primary focus for the
experiments was to evaluate the catalyst/adsorbent stability and, in particular, the CMS membrane
performance stability over a long-period run (>300 hr). To improve the process performance when
carrying out MR-AR experiments employing this type of syngas, a systematic parametric study
was performed to determine the optimal operating conditions. The optimization parameters we
investigated here include the operating conditions for the MR (e.g., W/FCO, H2O/CO, steam sweep
pressure and steam sweep ratio in the MR’s permeate side, etc.) and for the AR (i.e., the sorption-
69

enhanced WGS reaction temperature). The detailed experimental findings from running the MR-
AR system employing the oxygen-blown coal gasifier off-gas are described below.
4-2. Experimental Section
4-2-1. Materials
For the experiments in this Chapter, we utilized the hybrid MR-AR system described in
Chapter 3. However, the CMS membrane that we employed here was a different membrane (CMS-
MPT-25) than the one used in Chapter 3 (CMS-MPT-23), also prepared by our industrial
collaborator for this project Media and Process Technology, Inc. It is again a tubular CMS
membrane with an inner diameter of 3.6 mm, outer diameter of 5.6 mm, and a length of 254 mm
with a thin nano-porous CMS separation layer formed on the outside surface of a M&PT
commercial asymmetric mesoporous alumina ceramic tube. Similarly, since we added steam
sweep in the MR’s permeate side, we employed membrane configuration open on both ends rather
than the candle-filter membrane configuration we utilized in the Chapter 2 [17, 53]. During the
MR experiments, the membrane was sealed on the both ends to the tubular stainless-steel reactor
using Swagelok compression fittings and graphite O’ rings. We utilized the same catalyst
(commercial Co/Mo/Al2O3 sour-shift catalyst provided by Clariant (USA)) from the previous
studies for both the MR and the AR units. We have also used the same AR units with those in
Chapter 3.
4-2-2. Experimental Set-up and Procedure
The schematic of the lab-scale MR-AR set-up is shown in Figure 33. Since this is initial
phase of the study for carrying-out the MR-AR experiments using the oxygen-blown coal gasifier
off-gas, in the results reported in this Chapter we only operated the system with the MR being
followed by only one AR rather than two ARs employed in the multiple-cycle runs in Chapters 2
and 3. For the experiments, we have upgraded the system by re-working the heating tapes utilized
in the system and the heat insulation, to make sure that all the piping including the back-pressure
regulators (BPRs) are heat-traced and well-insulated, thus to prevent the steam from condensing
70

within the system and remedy potential oscillation and pressure shocks caused by the “pooling
effect” of water condensing and then suddenly vaporizing in the system. Again, in the experiments
we used 10 g of commercial Co/Mo/Al2O3 sour-shift catalyst intermixed with glass balls of similar
particle size (600~850 μm in diameter). The rest of the parts of the experimental set-up are detailed
in Chapter 2 (see Section 2-2-2).

Figure 33. Experimental set-up used in the MR-AR experiments (red lines are heat-traced to prevent
water from condensing; MR: Membrane Reactor; AR: Adsorptive Reactor; MFC: Mass Flow
Controller; BFM: Bubble Flow-meter; RGA: Residue Gas Analyzer; GC: Gas Chromatography;
BPR: Back Pressure Regulator;).

To make sure the state of the CMS membrane remains stable during the MR-AR
experiments using the oxygen-blown gasifier off-gas, the single-gas permeance of N2, which is
indicative of the slow gases like CO and of He, which is an inert fast gas to serve as a safe surrogate
to H2, of the CMS membrane were frequently measured at pre-determined temperatures and
pressures (we also use the He/N2 ideal selectivity to track the H2/CO ideal selectivity during the
long-term run). Before the initiation of the MR-AR experiments, the single-gas permeances of the
major syngas components (i.e., CO, CO2, CH4, H2, N2) and of He were measured using the
experimental procedure detailed in Chapter 3 (note that we still substitute N2 in the syngas
composition with Ar to facilitate the analysis via the RGA measurements, since we have found
71

that N2 interferes with the analysis of CO). The previous studies have already shown that both of
these gases act as diluents and inert gases and do not participate in the WGS reaction.
For the experiments using the oxygen-blown gasifier off-gas presented here, we have used
the same catalyst employed in the previous study in Chapter 3. But since the catalyst was exposed
to air during the loading procedure, the Co and Mo metal components are potentially in their
oxidized form; therefore, the catalyst needed to be re-activated (sulfided) prior to the initiation of
these MR-AR runs. The activation procedure followed has been described in detail in Chapter 2
(see Section 2-2-3) and involves employing a temperature and pressure protocol recommended by
the catalyst manufacturer [17].
The MR-AR experiments using oxygen-blown gasifier off-gas were initiated by testing the
MR-AR system components individually prior to carrying out the feasibility testing of the
combined MR-AR system. Specifically, the experimental run began by testing the MR unit at
various W/FCO (i.e., 44, 55, 66) values for the MR feed (where W (g) is the weight of undiluted
catalyst, and FCO is the molar flow rate of CO (mol/hr) in the syngas), as well as different H2O/CO
ratios (i.e., 3, 4, 5), steam sweep ratios (i.e., 0.1, 0.3, 0.5, 0.7 defined as the ratio of the sweep
stream molar flow rate to the feed molar flow rate) in the MR’s permeate side, and various sweep
pressures in the MR’s permeate side (i.e., 1, 3, 5 bar). In all the MR-AR experiments reported here,
the MR was operated at 250°C (the AR, however, was operated at various temperatures, i.e., 225°C,
250°C and 275°C for AR unit). Both the MR reject-side and the AR operated at a pressure of 25
bar).
As part of the evaluation of the stability of the catalyst, adsorbent and membrane in the
integrated MR-AR system using the oxygen-blown gasifier off-gas, we have performed 10-cycle
MR-AR runs. During the multi-cycle run, we fixed the regeneration temperature as 350°C for 8
min, and the AR reaction time as 42 min. Prior to initiating the combined MR-AR experiments,
the MR conversion was first allowed to reach steady state for at least 2 hours. During this time
period, the RGA was utilized to measure the gas composition for both the MR’s reject side and
permeate side. To “benchmark” the MR performance for these experiments, we also measured the
72

corresponding conversion of PBR under the same operating condition by closing the exit on the
permeate side of the MR and allowing the system to operate as a PBR under steady-state conditions.
As illustrated in Figure 33, the reject-side (retentate) stream (a CO2-rich stream, that also
includes some H2, unreacted CO, CH4, Ar, and impurities like H2S) from the MR serves as the
feed for the AR. For this study, as noted above, we have kept the operating pressure for both the
MR and the AR the same, while operating the AR at different temperatures (i.e., 225°C, 250°C
and 275°C). The use of steam sweep, and also increasing the permeate-side pressure (up to 5 bar
in the experiments reported in this Chapter) with the aid of the BPR installed at the end of the
MR’s permeate stream line, has allowed us to maintain the desired H2O/CO ratio in the AR feed
stream without needing to supply additional steam in between the MR and the AR units, as we did
in the experiments presented in Chapter 2.
To initiate the operation of the combined MR-AR system, the MR subsystem performance
(in terms of CO conversion, hydrogen recovery and purity, which are monitored via a GC that is
used to measure the composition of the reject side and that of the permeate side) is, typically,
allowed to stabilize before switching the MR reject stream as a feed into the AR(s). Meanwhile,
the AR unit is simultaneously pressurized to 25 bar and heated to the desired temperature using a
50 vol %/50 vol % steam/Argon gas mixture. Once the MR reject side flow rate, gas composition
and steam concentration are stable, we switch the MR reject-side stream to the AR feed-side via a
3-way valve and begin the MR-AR multi-cycle experiments (the AR is in “reaction mode” during
this period of time, for a fixed period of 42 min). While the MR reject-side is directed into the AR
unit, the 50 vol %/50 vol % steam/Argon stream is directed to the vent. During the experiment, the
AR outlet gas is measured instantaneously via the RGA. Fixing the AR reaction mode time as 42
min, allows the adsorbent to become saturated with CO2, and to obtain the so-called “pseudo-
steady state” for the CO conversion in the AR prior to being switched into the regeneration mode.
Upon completion of the sorption-enhanced WGS reaction for 42 min and the AR system reaching
the “pseudo-steady state”, the MR reject-side stream is then switched into the vent, while the 50
vol %/50 vol % steam/Argon stream is directed into the AR unit to regenerate the adsorbent in the
reactor at 350°C and duration of 8 min for optimized process design.
73

After completing the evaluation of the performance for MR-AR multiple-cycle run, the
MR’s reject-side stream was switched back away from the feed of the AR, and the MR
performance (including the gas composition and flow rate of both the reject and permeate sides)
was studied again to verify whether the MR performance has remained stable, i.e., to validate its
robustness to this simulated coal-derived syngas conditions. For that we employ a GC to measure
the MR reject-side and permeate-side stream gas composition and a bubble flow meter (BFM) to
measure the dry-gas flow rate.
4-3. Results and Discussion
4-3-1. Membrane Studies
In this Chapter, a CMS membrane (CMS-MPT-25) prepared by M&PT was employed in
the study. The membrane was installed into the MR empty module at the M&PT laboratories, and
prior to the module being shipped to USC, its permeation characteristics were tested using He and
N2 (both of which serve as model inert gases, whereby He is indicative of the behavior of fast
gases like H2 and the N2 is representative of slow gases like CO). Upon receiving the MR module
with the CMS membrane from M&PT, the single-gas permeances of He and N2 were again
measured at the same temperature and pressure conditions at USC. The properties of the as received
membrane (in the empty module) measured at M&PT as well as those measured at USC are shown in
Table 10.
Table 10. Single- gas permeances of N 2 and He at various temperatures and pressures
measured in the empty module.
Lab
T
(℃)
Pressure
(bar)
He
*
N2
*

SF
§

(He/N2)
M&PT 17 3.07 0.3194 0.00308 104
M&PT 250 3.07 1.1618 0.0060 193
USC 250 3.07 1.1292 0.0047 240
74

USC 250 25 1.0548 0.0057 185
*
Gas permeance [m
3
/m
2
h bar];
§
Separation factor.

There is a significant difference observed in the measurement of N2 permeances among the
two laboratories (See Table 10). The N2 permeance measured at M&PT is 21.7 % larger than the
USC measurement. The main reason why M&PT has higher N2 permeance measurement can be
explained as the loose end-fittings in the MR’s permeate side which were re-tightened after
receiving the MR module at USC. However, there is only a slight difference observed in the
measurement of He (~2.8%), which is below the 5% experimental error, which is typical with such
measurements. Consequently, the He/N2 separation factor measured at USC is, in fact, quite higher
(240 vs. 193 see Table 10) than the one measured at M&PT, mostly, due to a lower N2 permeance,
indicative potentially of a leaking graphite sealing during the M&PT permeation test.
Subsequently, 10 g of catalyst intermixed with glass balls of similar particle size (600~850
μm in diameter) were loaded into the reactor, and the membrane permeation characteristics were
measured again, and the values are shown in Table 11. There is a slight difference for He
permeance before and after loading the catalyst, which is below the 5% experimental error and
typical with such measurements. However, there is a statistically significant decrease in the N2
permeance (~11%). The slight difference in the He permeance and a larger decrease in the N2
permeance, is indicative of the fact that the loading of the catalyst (separation factor at 94 vs. 83
see Table 11) is quite successful and caused no damage to the membrane, since if the membrane’s
CMS layer is damaged the N2 permeance increases significantly. We ascribe the decrease in the
N2 permeance to the catalyst/glass balls potentially blocking/hindering access to parts of the
surface of the CMS membrane.
Table 11. Single- gas permeances of N 2 and He at 23℃ and 25bar
measured in the module before and after loading the catalyst.
Test date T
(℃)
Pressure
(bar)
He
*
N2
*
SF
§

(He/N2)
75

Before
loading the
catalyst
23 25 0.2507 0.00303 83
After
loading the
catalyst
23 25 0.2525 0.0027 94
*
Gas permeance [m
3
/m
2
.h.bar];
§
Separation factor.

The catalyst was then re-activated. The activation procedure has been described in detail
elsewhere [17, 53]. After the catalyst activation, the membrane properties were measured once
more and are shown in Table 12 below.
Table 12. Single- gas permeances of He, N 2, H 2, N 2, A r, CO, CO 2 and methane at 250℃ and 25bar
after the catalyst activation.
Test date He
*
N2
*
H2
*
CO
*
CO2
*
Ar
*
CH4
*

SF
§

(He/N2)
SF
§

(H2/CO)
Before
catalyst
activation
1.0548 0.0057 - - - - - 185 -
After
catalyst
activation
1.0644 0.0061 - - - - - 174 -
After 2
days
1.0542 0.0061 - - - 0.0081 - 173 -
After 3
days
- - - 0.0097 0.0255 - - - -
After 4
days
- - - - 0.0258 - - - -
After 5
days
1.0403 - 1.2932 - - - - - 133
*
Gas permeance [m
3
/m
2
.h.bar];
§
Separation factor.

From Table 12 above, one notices that a slight <1% (statistically insignificant) increase in the He
permeance and a small ~6.6% increase in the N2 permeance after activating the catalyst. The
H2/CO separation factor of around 133 is well above the DOE target value of 80.
76

One of the key objectives of the MR-AR experiments using oxygen-blown gasifier off-gas
is to evaluate the membrane properties stability under these harsh experimental conditions
involving exposing the membrane with large concentration of H2S and high-temperature (250°C)
and high-pressure steam during the long-period run (over 300 hr). During the subsequent 2 months
of testing of the MR-AR system, the membrane properties were shown to be very stable, as shown
in Table 13 and in Figures 34 and 35.
Table 13 (and Figures 34 and 35) report the He, N2, H2 and CO permeances and the He/N2,
H2/CO selectivities as a function of the cumulative H2S/syngas exposure time. As the data
presented in the Table (and in the Figures) show, after 344 hr of exposure time to H2S/syngas, the
permeances for He, N2 were experimentally indistinguishable from those measured before the run,
which suggests the membrane still maintains satisfactory performance, and one may conclude that
for this CMS membrane, the membrane properties are very robust and stable under the cumulative
344 hours run of H2S exposure and under the 250°C temperature and 25 bar pressure environments..
It should be noted that the MR-AR experiments are still on-going, and we will measure the H2/CO
separation factor at the end of the operation.
Table 13. Single- gas permeances of He, N 2, H 2, N 2, CO, at 250℃ and 25bar during the long-term run.
Lab/Test
date
Exposure
Hours
He
*
N2
*
H2
*
CO
*
CO2
*
SF
§

(He/N2)
SF
§

(H2/CO)
03/28/2019
a
0 1.0548 0.0057 - - - 185 -
04/07/2019 24 1.0206 0.0061 - - - 174 -
04/09/2019
b
24 1.0542 0.0061 - - - 173 -
04/10/2019 24 - - - 0.0097 0.0255 - -
04/11/2019 24 - - - - 0.0258 - -
04/12/2019 24 1.0403 - 1.2932 - - - 133
05/14/2019 281 1.0363 0.0061 - - - 169 -
05/17/2019 296 1.0331 0.0061 - - - 170 -
05/24/2019 344 1.0421 0.0062 - - - 167 -
*
Gas permeance [m
3
/m
2
h bar];
§
Separation factor.
a
After loading catalyst
b
After steam and argon treatment
77

Figure 34. Single-gas permeances as a function of H 2S exposure time during the long-period run.

Figure 35. He/N 2 and H 2/CO selectivities as a function of H 2S exposure time during the long-period run.
78

Another key focus of this research was also on the optimization of operation parameters
when using the oxygen-blown gasifier off-gas. Thus, we initiated experiments with the MR alone
(the temperature was kept as 250°C and pressure at 25 bar) by varying the W/FCO, steam sweep
pressure, H2O/CO ratio, steam sweep ratio for the MR. Subsequently, we carried-out experiments
with the hybrid MR-AR system with the MR and AR subsystems operating at the same pressure
(25 bar), while varying the AR temperature. Note that before initiating the parametric study with
the combined MR-AR with steam sweep system (i.e., switching the reject-side stream from the
MR as feed into the AR), the MR was allowed to operate until reaching steady state and its
conversion, H2 purity (dry-basis) in the permeate stream and recovery were all measured and
presented on Figures 37-46. Since we employed the steam sweep at the MR’s permeate side while
not supplying any additional steam in between the MR and AR unit, we also monitored the MR’s
reject-side H2O/CO ratio. For example, as shown in Figure 36, under these experimental conditions
the MR’s reject-side H2O/CO ratio is below 1 when the MR’s permeate-side steam sweep pressure
is below 5 bar. Thus, to prevent the catalyst in the AR unit from getting “coked”, we employed the
steam sweep pressure of 5 bar during the MR-AR experiments.

Figure 36. MR reject H 2O/CO ratio vs. sweep pressure with MR operated at W/F CO=55 g.h/mol, H 2O/CO
ratio of 4, SR=0.5, 25 bar and 250°C.
79

Figure 37. Conversion vs. W/F CO for H 2O/CO ratio of 3 with MR operated at SR=0.3, SP=3bar, 25 bar
and 250°C.

Figure 38. H 2 Purity vs. W/F CO for H 2O/CO ratio of 3 with MR operated at SR=0.3, SP=3 bar, 25 bar and
250°C.
80

Figure 39. H 2 Recovery vs. W/F CO for H 2O/CO ratio of 3 with MR operated at SR=0.3, SP=3 bar, 25 bar
and 250°C.

Figure 40. Conversion vs. MR sweep ratio for H 2O/CO ratio of 3 with MR operated at W/F CO=66
g.h/mol, SP=3 bar, 25 bar and 250°C.
81

Figure 41. H 2 purity and H 2 recovery vs. MR sweep ratio for H 2O/CO ratio of 3 with MR operated at
W/F CO=66 g.h/mol, SP=3 bar, 25 bar and 250°C.

Figure 42. Conversion vs. MR H 2O/CO ratio with MR operated at W/F CO=55 g.h/mol, SR=0.5, SP=3 bar,
25 bar and 250°C.
82

Figure 43. H 2 purity vs. H 2O/CO ratio with MR operated at W/F CO=55 g.h/mol, SR=0.5, SP=3 bar, 25 bar
and 250°C.

Figure 44. H 2 Recovery vs. H 2O/CO ratio with MR operated at W/F CO=55 g.h/mol, SR=0.5, SP=3 bar, 25
bar and 250°C.
83

Figure 45. Conversion vs. MR sweep pressure for H 2O/CO ratio of 4 with MR operated at W/F CO=55
g.h/mol, SR=0.5, 25 bar and 250°C.

Figure 46. H 2 purity and H 2 recovery vs. MR Sweep pressure for H 2O/CO ratio of 4 with MR operated at
W/F CO=55 g.h/mol, SR=0.5, 25 bar and 250°C.
84

We also measured the conversion and hydrogen purity of the reactor functioning as a PBR under
the same operating conditions (using the approach detailed in Section 3-2-2), and these are also
presented in the same figures.
As can be seen in Figure 37, for the experiment performed at H2O/CO=3, with a steam
sweep ratio=0.3, steam sweep pressure=3 bar, the MR’s CO conversion increases significantly
when increasing the W/F CO, the difference between the MR CO conversion and PBR CO
conversion under different W/F CO, however, is relatively small (~7.31 % for W/F CO =44, ~8.46 %
for W/F CO =66). Nevertheless, the difference between the MR CO conversion and PBR CO
conversion under different W/F CO using oxygen-blown gasifier off-gas here is larger than the case
of using air-blown gasifier off-gas in Chapter 3 (~3 % for W/F CO =44, ~6 % for W/F CO =66, note
that the sweep ratio for these runs was higher equal to 0.49 here). We ascribe the small difference
between the MR CO conversion and PBR CO conversion to employing the steam sweep pressure
of 3 bar, which lowers the difference between the MR’s reject side and permeate side pressure,
and so is the difference with respect to CO conversion between the MR and the PBR. Figure 45
shows how the difference between the MR CO conversion and PBR CO conversion decreases
when increasing the sweep pressure over the MR’s permeate side (~7.4% for sweep pressure=1,
~4.4 % for sweep pressure =5 when operating the MR at W/FCO=55 g.h/mol, SR=0.5, H2O/CO=4,
25 bar and 250°C), which indicates that the MR becomes less efficient when increasing the sweep
pressure. However, it should be noted that, as shown in Figure 36, the MR’s reject-side H2O/CO
ratio is below 1 when the MR’s permeate-side steam sweep pressure is below 5 bar at same
experimental conditions (the catalyst in the MR is likely to get “coked” under this condition),
which dictates the optimization study for process design. Also note that in Figure 42, increasing
the H2O/CO ratio can significantly increase the MR CO conversion, however, the difference
between the MR and the PBR for CO conversion shows a minimum at H2O/CO ratio=4 when
varying the MR’s feed side H2O/CO ratio (~8.76 % for H2O/CO ratio=3, ~6.03 % for H2O/CO
ratio=4 and ~8.07 % for H2O/CO ratio=5). Furthermore, as shown in Figure 40, the difference
between the MR and the PBR for CO conversion increases significantly when increasing the sweep
ratio in the MR’s permeate side (~7.46 % for sweep ratio=0.1, ~13.2 % for sweep ratio=0.7).
However, the operating cost will significantly increase when increasing the H2O/CO ratio and
85

sweep ratio when operating the MR since the high-temperature and high-pressure steam is very
costly. Thus, one needs in addition to performing a parametric study to also carry-out a Technical
& Economic Analysis (TEA) study to determine the optimal operating conditions for determining
the design of both the field-scale and eventually of the commercial size system.
The most important advantage, however, of using the MR is shown in Figure 38 and Figure
43, which compares the H2 purities of the processed syngas for the MR and the PBR. the hydrogen
purity of the PBR is very poor (<50 %) that it is hardly appropriate for use in power generation.
However, the use of the MR that combines reaction and separation in one unit, in addition to
enhancing the CO conversion, also provides a H2 product stream with superior purity which can
be directly usable in a hydrogen turbine for power generation. It should be noted that, in Figure
38, both the H2 purity and the difference between the MR and the PBR for H2 purity decrease when
increasing the W/FCO, however, there is a statistically insignificant change for both the H2 purity
and the difference between the MR and the PBR for H2 purity when increasing the H2O/CO ratio
as shown in Figure 43. Also, as seen in Figure 39 and 44, the H2 recovery increases when
increasing the W/FCO for MR while the situation is reversed when increasing the H2O/CO ratio
(note that for Chapter 3 when using air-blown gasifier off-gas the H2 recovery increases when
increasing the H2O/CO ratio). The explanation can be found in some of our group’s publications
[17, 21, 22, 24, 53]. Thus, an investigation of the optimal conditions for W/FCO and H2O/CO ratio
is required, if one needs to meet both the H2 purity and recovery requirements.
From Figure 41, one notices that the change in both the H2 purity and recovery is
insignificant when increasing the steam sweep ratio (similar behavior can be found in Chapter 3
when using air-blown gasifier off-gas). However, as seen in Figure 46, the H2 recovery decreases
when increasing the sweep pressure for MR while the situation is reversed for H2 purity. These
experimental findings validate the modeling findings by our team and the UCLA team [17, 54]
which collaborates with us on this project. For process design, one would recommend using a
smaller H2O/CO ratio (e.g., 2 or 3) and steam sweep ratio~0.1 to reduce the capital cost and the
energy penalty of generating the high-temperature and/or high-pressure steam. However, from our
parametric study and modeling investigation, one would recommend the H2O/CO ratio above 4
and steam sweep above 0.5 while keeping the sweep pressure up to 5 bar to assure the H2O/CO
86

ratio in the MR’s reject side above 1, thus to prevent the catalyst in the MR from getting “coked”
during the operation.
4-3-2. MR-AR Multiple-Cycle Run
To make sure the upstream of the AR (i.e., the MR reject side, which serves as the feed
stream for AR), the state of the membrane and catalyst in the MR is stable, prior to the initiation
of the MR-AR experiments using oxygen-blown gasifier off-gas, and during each multi-cycle run,
we continue to monitor the MR subsystem performance with respect to MR CO conversion, H2
recovery and purity in order to verify both membrane and catalyst stability. Figure 47 shows these
MR properties during a multi-cycle MR-AR run as a function of time on stream (expressed in this
Figure as syngas/H2S total exposure time). At the beginning of this run, the catalyst and membrane
had been exposed to syngas/H2S for a total of 186 hr and by the time the final MR performance
measurement, shown on this Figure, was made the catalyst and membrane had undergone an
additional ~158 hr of exposure. It is clear from the Figure, that both and membrane and catalyst
employed in the MR show quite stable performance.

87

Figure 47. The MR properties during a multi-cycle run. Experimental conditions: T=250°C, feed
pressure=25 bar, permeate-side pressure 5 bar, with steam sweep ratio=0.5, H 2O/CO ratio of 4, W/F CO=
55 g·h/mol, oxygen-blown gasifier model syngas (CMS-MPT-25).
In Chapter 3, in the MR-AR experiment multiple-cycle run experiments using air-blown
gasifier off-gas, we have defined two different criteria of “AR effect time” to properly characterize
the AR behavior. For the study in this Chapter, we use “AR effect time” defined as the difference
between the H2 emergence time (to account for any system dead times) and the time when the
AR’s CO conversion decreases below 95 % for simplicity. During the aforementioned time period,
the AR works in the sorption-enhanced WGS reaction mode, whereby the adsorbent is effective
and the CO conversion in the AR is at or near 100 %, subsequently as the adsorbent in the AR
gradually becomes saturated with CO2, the performance of AR starts to deteriorate as CO
conversion decreases below 95 % and eventually AR reaches its “pseudo-steady state” CO
conversion, by which time the AR is switched into the regeneration mode. Figure 48-50 show the
“AR effect time” during a multiple-cycle MR-AR run with various AR temperatures (note that the
MR is operated at same conditions). As seen in Figure 48, the AR operated at 225°C is less efficient
with large oscillation for AR effect time with very short AR effect time for some cycles, and the
AR effect time is still not able to reach its eventual “steady-state” after 10 cycles. From Figure 49
and 50, one may notice that the “AR effect time” takes a few cycles to settle to the eventual
“steady-state” (e.g., the steady-state AR effect time for AR operating at 250°C is approximately
~354 sec, the steady-state AR effect time for AR operating at 275°C is approximately ~250 sec).
One may conclude that the optimal operating temperature for AR is 250°C.
88

Figure 48. The AR effect time w.r.t 95 % CO conversion during multi-cycle run. Experimental
conditions: AR is operated at 225°C and 25 bar with MR operated at T=250°C, feed pressure=25 bar,
permeate-side pressure 5 bar, with steam sweep, W/F CO= 55 g·h/mol, Oxygen-blown gasifier simulated
syngas (CMS-MPT-25).

89

Figure 49. The AR effect time w.r.t 95 % CO conversion during multi-cycle run. Experimental
conditions: AR is operated at 250°C and 25 bar with MR operated at T=250°C, feed pressure=25 bar,
permeate-side pressure 5 bar, with steam sweep, W/F CO= 55 g·h/mol, Oxygen-blown gasifier model
syngas (CMS-MPT-25).

Figure 50. The AR effect time w.r.t 95 % CO conversion during multi-cycle run. Experimental
conditions: AR is operated at 275°C and 25 bar with MR operated at T=250°C, feed pressure=25 bar,
permeate-side pressure 5 bar, with steam sweep, W/F CO= 55 g·h/mol, Oxygen-blown gasifier model
syngas (CMS-MPT-25).
90

Figure 51. The AR steady state CO conversion during multi-cycle run. Experimental conditions: AR is
operated at 225°C and 25 bar with MR operated at T=250°C, feed pressure=25 bar, permeate-side
pressure 5 bar, with steam sweep, W/F CO= 55 g·h/mol, Oxygen-blown gasifier model syngas (CMS-MPT-
25).

91

Figure 52. The AR steady state CO conversion during multi-cycle run. Experimental conditions: AR is
operated at 250°C and 25 bar with MR operated at T=250°C, feed pressure=25 bar, permeate-side
pressure 5 bar, with steam sweep, W/F CO= 55 g·h/mol, Oxygen-blown gasifier model syngas (CMS-MPT-
25).

Figure 53. The AR steady state CO conversion during multi-cycle run. Experimental conditions: AR is
operated at 275°C and 25 bar with MR operated at T=250°C, feed pressure=25 bar, permeate-side
pressure 5 bar, with steam sweep, W/F CO= 55 g·h/mol, Oxygen-blown gasifier model syngas (CMS-MPT-
25).
In addition, for the experiments shown in Figure 51-53, we allowed for the sorption-enhanced
WGS reaction time to be long enough (42 min in this case) so that the adsorbent becomes saturated
with CO2. We investigated the AR performance with respect to “pseudo-steady state” CO
conversion in the AR prior to being switched into the regeneration mode. It takes again a few
cycles before the CO conversion settles down to its eventual steady-state value and catalyst activity
and adsorbent performance become stable. Also, as seen in Figure 51, the stability of AR
performance with respect to “pseudo-steady state” CO conversion is very poor at 225°C compared
with the AR operated at 250°C or/and 275°C. Though AR operating at 275°C will increase the
92

capital cost and energy penalty, operating at 275°C can also significantly decrease the temperature
difference in AR between reaction mode and regeneration mode, and thereby decrease the AR
regeneration time and the AR reaction time since they are linked with each other. Thus, one needs
to determine the optimal operating temperature for AR to improve the process performance and at
the same time to reduce the capital cost and energy penalty.
The dynamic cyclic performance of the AR during the combined MR-AR experiments
using the oxygen-blown gasifier off-gas (here, we include the cycle 5-7 for AR at 250°C with MR
operating at W/FCO= 55 g·h/mol, sweep pressure of 5 bar) , in terms of the various exit gas molar
flow rates in the AR (i.e., H2, CO, CO2), and the CO conversion in the AR as a function of time,
is presented in Figure 54-56 (the gas compositions and conversion in these Figures are measured
via the RGA instrument, which samples and measures the exit composition at a frequency of 1
sample point per 3 seconds, so what is shown in Figures 54-56 are the lines passing through this
multitude of measuring points). As seen in these Figures, the AR subsystem shows typical AR
behavior (the details have been described in Chapter 2 and Chapter 3), whereby the initial AR CO
conversion begins at 100% for an extended time (~250~400 sec) and starts to decline as the
adsorbent gets saturated with CO2, leveling off eventually at the “pseudo steady state” conversion
that would be attained if the reactor was operating as PBR (~65~70 %). Furthermore, these 3 cycles
dynamic profiles demonstrate a quite reproducible behavior (the AR effect time is ~350~450 sec
considering the 50 sec experimental error), which is indicative the fact that the hydrotalcite
adsorbent and WGS catalyst used in this study show good reversibility during the sorption-
desorption cycles (reaction-regeneration modes) and remarkable stability in the oxygen-blown
gasifier off-gas environment.
93

Figure 54. CO conversion, and molar flow rates of CO 2, H 2, CO in the AR cycle 5 (temperature of
250°C, pressure of 25 bar), with the MR operated at T=250°C, feed pressure=25 bar, permeate-side
pressure 5 bar, with steam sweep, W/F CO= 55 g·h/mol, Oxygen-blown gasifier simulated syngas (CMS-
MPT-25).

94

Figure 55. CO conversion, and molar flow rates of CO 2, H 2, CO in the AR cycle 6 (temperature of
250°C, pressure of 25 bar), with the MR operated with at T=250°C, feed pressure=25 bar, permeate-side
pressure 5 bar, with steam sweep, W/F CO= 55 g·h/mol, Oxygen-blown gasifier simulated syngas (CMS-
MPT-25).

Figure 56. CO conversion, and molar flow rates of CO 2, H 2, CO in the AR cycle 7 (temperature of
250°C, pressure of 25 bar), with the MR operated with at T=250°C, feed pressure=25 bar, permeate-side
pressure 5 bar, with steam sweep, W/F CO= 55 g·h/mol, Oxygen-blown gasifier simulated syngas (CMS-
MPT-25).
95

Figure 57. Temperature profiles of AR cycle 5 operated at 250°C and 25 bar using W/F CO = 55 and
H 2O/CO =4, steam sweep ratio=0.5, permeate side pressure=5 bar.
Because temperature is also a key operating parameter in the study, during the experiments,
the temperatures profiles during the sorption-enhanced water gas shift reaction mode and the
adsorbent regeneration mode at different bed points of the AR were also recorded via three two-
point thermocouples installed in the AR. A typical temperature profile is shown in Figure 57 (note
that Inlet-0’’, Middle-0’’ and Outlet-0’’ signify three different equidistant axial positions in the
reactor, specifically 1.375 in, 2.75 in, and 4.125 in from the entrance of the bed, while Inlet-0.5’’,
Middle-0.5’’ and Outlet-0.5’’ signify three radial positions 0.5” away from the bed axis). As Figure
57 indicates, the reactor is fairly isothermal with radial and axial profiles being, typically, less than
2-3°C. The bed pressure drop is also very small (~0.1 psi) under all conditions studied during the
MR-AR experiments. The lack of significant pressure drops, and temperature gradients make it a
more straightforward task to model the experimental data, and thus to validate the reactor models
to be used in process design and optimization and in the TEA studies.
96

4-4. Conclusions
In this study, we have investigated the feasibility of running the integrated MR-AR system
with steam sweep using oxygen-blown coal gasifier off-gas which contains small amount of N2
(~3 %) and larger portions of hydrogen (~34 %) and carbon dioxides (~21%) compared with the
air-blown gasifier off gas used for MR-AR experiments in previous Chapters. Note that for this
preliminary study for this type of syngas, we only operated the MR followed by one AR unit. In
order to improve the process design (i.e., to reduce the capital cost and energy penalty for
regeneration and to shorten the time required for adsorbent regeneration) for the integrated MR-
AR, we utilized 350°C rather than 400°C as regeneration temperature and 8 min as regeneration
time, which allowed us to reduce the AR reaction mode time down to 42 min (note that the AR
regeneration time is linked with reaction time in the AR). The CMS membrane employed in the
study has demonstrated very robust and stable performance during the long-term run (~342 hr run
of H2S exposure and under the temperature of 250°C and 25 bar pressure environments) and
maintained quite high He/N2 selectivity (~170). The combined MR-AR system using the oxygen-
blown gasifier off-gas was experimentally tested during numerous multiple-cycle runs and
displayed superior performance to that of a conventional PBR with high purities for the H2 product
which can be directly usable in a hydrogen turbine for power generation.
We have also carried-out the parametric study for the optimization of operation conditions,
for which we have tested the MR unit at various W/FCO for the MR feed, as well as different
H2O/CO ratios, steam sweep ratios in the MR’s permeate side, and various sweep pressures in the
MR’s permeate side. All the MR-AR experiments reported here, the MR was operated at 250°C
(we have employed various temperatures, i.e., 225°C, 250°C and 275°C for AR unit) and a MR
reject-side and AR pressure of 25 bar. The parametric study shows that:
(i) The MR’s reject-side H2O/CO ratio increases when employing the steam sweep and increasing
the sweep pressure, since by adding steam sweep and increasing the sweep pressure can
significantly raise the partial pressure of water in the MR’s permeate side, and thus eliminate the
loss of reactant H2O from the reject-side of the MR where the WGS reaction takes place. However,
in the parametric study for MR using oxygen-blown gasifier off-gas, the CO conversion and H2
97

recovery of MR decreases when increasing the sweep pressure, and so does the difference between
the MR CO conversion and PBR CO conversion, which indicates that the MR is less efficient with
increasing sweep pressure. However, to maintain the MR’s reject-side H2O/CO ratio above 1, and
thus to prevent the catalyst in the MR and AR units (the MR’s reject stream serves as the feed to
the AR) from potentially getting “coked”, one needs to operate the MR with sweep pressure ~5
bar.
(ii) Increasing the W/FCO can significantly increase the MR’s CO conversion, the difference
between the MR CO conversion and PBR CO conversion under different W/F CO, however, is
relatively small due to the high sweep pressure we employed in the study. Nevertheless, the
difference between the MR CO conversion and PBR CO conversion under different W/F CO using
oxygen-blown gasifier off-gas is larger than the case of using air-blown gasifier off-gas. Also,
increasing the H2O/CO ratio can significantly increase the MR CO conversion, however, the
difference between the MR and the PBR for CO conversion shows a minimum when operating
under H2O/CO ratio =4. Furthermore, the difference between the MR and the PBR for CO
conversion increases significantly when increasing the sweep ratio in the MR’s permeate side.
However, one needs to take into account that the operating cost will significantly increase when
increasing the H2O/CO ratio and sweep ratio in the MR, which dictates that a TEA study be
performed to determine the optimal operating conditions for process design.
(iii) Both the H2 purity and the difference between the MR and the PBR for H2 purity decrease
when increasing the W/FCO. Also, the H2 recovery increases when increasing the W/FCO for MR
while the situation is reversed when increasing the H2O/CO ratio.
(iv) There is a statistically insignificant change for both the H2 purity and the difference between
the MR and the PBR for H2 purity when increasing the H2O/CO ratio. Also, there is only a slight
change in both the H2 purity and recovery when increasing the steam sweep ratio. In addition, the
H2 recovery decreases when increasing the sweep pressure for MR while the situation is reversed
for H2 purity.
98

During the MR-AR multiple-cycle run, we continued to monitor the MR subsystem
performance with respect to MR CO conversion, H2 purity and recovery to assure both membrane
and catalyst in the MR are stable. The membrane and catalyst in the MR both displayed stable
performance during the long-period run. For this study, to properly characterize and monitor the
AR behavior, we use the “AR effect time” defined with respect to 95% CO conversion threshold.
We have investigated the AR performance with respect to “pseudo-steady state” CO conversion
and “AR effect time” during the multiple-cycle run with various AR temperature while the MR is
operated at same conditions. We found that it took a few cycles before both the “AR effect time”
and “pseudo-steady state” CO conversion settle down to its eventual steady state value and the
catalyst activity and adsorbent performance become stable. Further, the AR operated at 225°C is
less efficient with respect to “AR effect time”, and the stability of AR performance with respect to
“pseudo-steady state” CO conversion is very poor at 225°C compared with the AR operated at
250°C or 275°C. We found that the optimal operating temperature for AR is 250°C and we will
continue to use this temperature for future experiments. Furthermore, the AR dynamic profiles in
the cyclic operation demonstrate a quite reproducible behavior, which is indicative of the fact that
the hydrotalcite adsorbent and WGS catalyst used in this study show good performance during the
sorption-desorption cycles (reaction-regeneration modes) and remarkable stability in the oxygen-
blown gasifier off-gas environment. Thus, one may conclude that the membrane, catalyst and
adsorbent are very robust and stable under the large concentration of H2S, high-temperature and
high-pressure oxygen-blown gasifier off-gas environment during the long-period MR-AR
multiple-cycle run.

99

Chapter 5: Future Work
In the above Chapters, we have experimentally validated the feasibility of running the
integrated MR-AR system under the IGCC-relevant conditions. Our future work involves several
topics for further investigation:
(i) We need to study in greater detail the kinetics of the commercial sour-shift WGS catalyst
(Co/Mo/Al2O3) for different simulated coal-derived syngas (both air-blown gasifier off-gas and
oxygen-blown gasifier off-gas) under IGCC relevant conditions. In order to do so, we will study
the effect of temperature, pressure and H2S concentration, and we will investigate three different
microkinetic models to determine which one fits best the kinetics of the catalyst we employed in
our research.
(ii) We need to investigate the hydrotalcite adsorbent kinetics by studying the CO2 adsorption
isotherms and dynamics. In order to do so, we need to analyze the previous experimental data
performed in the static CO2 adsorption system and employ a model that we have alreday developed
to further validate the experimental results.
(iii) We need to continue working on the integrated MR-AR system under lab-scale conditions
with simulated oxygen-blown gasifier off-gas for even longer periods of operation and carry-out a
much longer (30-cycle) MR-AR run to evaluate the MR-AR system performance with respect to
the stability of catalyst, membrane and adsorbent during this long-period and numerous-cycle run.
As part of this effort, we will develop mathematical models for both the MR and AR systems and
use them to validate our experimental findings.
(iv) We need to study the integrated MR-AR system under field conditions with a real syngas from
a coal-gasifier. AS part of this effort, we will first address the scale-up aspects and complete an
initial technical and economic analysis. We will build a bench-scale MR-AR system and install it
in the field and we will carry-out the “bench-scale” experiments with real coal gasifier off-gas. We
will evaluate the stability of the catalyst, membrane and adsorbent in the field-study during the
long-period run. We will also perform parametric study for optimization operation.
100

References:
[1] J. Conti, P. Holtberg, J. Diefenderfer, A. LaRose, J.T. Turnure, L. Westfall, International
energy outlook 2016 with projections to 2040, USDOE Energy Information Administration (EIA),
Washington, DC (United States …, 2016.
[2] S. Bilgen, Structure and environmental impact of global energy consumption, Renewable and
Sustainable Energy Reviews 38 (2014) 890-902.
[3] R.B. Gupta, Hydrogen fuel: production, transport, and storage, Crc Press2008.
[4] L. Barreto, A. Makihira, K. Riahi, The hydrogen economy in the 21st century: a sustainable
development scenario, International Journal of Hydrogen Energy 28 (2003) 267-284.
[5] I. Jain, Hydrogen the fuel for 21st century, International journal of hydrogen energy 34 (2009)
7368-7378.
[6] D. Stolten, B. Emonts, Hydrogen Science and Engineering, 2 Volume Set: Materials, Processes,
Systems, and Technology, John Wiley & Sons2016.
[7] J.R. Elliott, C.T. Lira, Introductory chemical engineering thermodynamics, Prentice Hall PTR
Upper Saddle River, NJ1999.
[8] K. Gillingham, J.H. Stock, The Cost of Reducing Greenhouse Gas Emissions, Journal of
Economic Perspectives 32 (2018) 53-72.
[9] S.A. Rackley, Carbon capture and storage, Butterworth-Heinemann2017.
[10] C.-H. Yu, C.-H. Huang, C.-S. Tan, A review of CO2 capture by absorption and adsorption,
Aerosol Air Qual. Res 12 (2012) 745-769.
[11] K. Babita, S. Sridhar, K. Raghavan, Membrane reactors for fuel cell quality hydrogen through
WGSR–Review of their status, challenges and opportunities, international journal of hydrogen
energy 36 (2011) 6671-6688.
[12] R. Smith, M. Loganathan, M.S. Shantha, A review of the water gas shift reaction kinetics,
International Journal of Chemical Reactor Engineering 8 (2010) 4.
[13] D. Mendes, A. Mendes, L. Madeira, A. Iulianelli, J. Sousa, A. Basile, The water ‐gas shift
reaction: from conventional catalytic systems to Pd ‐based membrane reactors—a review, Asia ‐
Pacific Journal of Chemical Engineering 5 (2010) 111-137.
101

[14] M. Abdollahi, J. Yu, H.T. Hwang, P.K. Liu, R. Ciora, M. Sahimi, T.T. Tsotsis, Process
intensification in hydrogen production from biomass-derived syngas, Industrial & Engineering
Chemistry Research 49 (2010) 10986-10993.
[15] M. Abdollahi, J. Yu, P.K. Liu, R. Ciora, M. Sahimi, T.T. Tsotsis, Hydrogen production from
coal-derived syngas using a catalytic membrane reactor based process, Journal of Membrane
Science 363 (2010) 160-169.
[16] P.K. Liu, M. Sahimi, T.T. Tsotsis, Process intensification in hydrogen production from coal
and biomass via the use of membrane-based reactive separations, Current Opinion in Chemical
Engineering 1 (2012) 342-351.
[17] A. Garshasbi, H. Chen, M. Cao, S. Karagöz, R.J. Ciora Jr, P.K. Liu, V.I. Manousiouthakis,
T.T. Tsotsis, Membrane-based reactive separations for process intensification during power
generation, Catalysis Today (2017).
[18] S. Dabir, M. Cao, R. Prosser, T. Tsotsis, Feasibility Study of Biogas Reforming To Improve
Energy Efficiency and To Reduce Nitrogen Oxide Emissions, Industrial & Engineering Chemistry
Research 56 (2017) 1186-1200.
[19] G.J. Stiegel, M. Ramezan, Hydrogen from coal gasification: An economical pathway to a
sustainable energy future, International Journal of Coal Geology 65 (2006) 173-190.
[20] J. Yu, M. Tan, P.K. Liu, M. Sahimi, T.T. Tsotsis, Hydrogen production from biomass-derived
syngas using a membrane reactor based process, Industrial & Engineering Chemistry Research 53
(2013) 819-827.
[21] A. Harale, H.T. Hwang, P.K. Liu, M. Sahimi, T.T. Tsotsis, Experimental studies of a hybrid
adsorbent-membrane reactor (HAMR) system for hydrogen production, Chemical engineering
science 62 (2007) 4126-4137.
[22] A. Harale, H.T. Hwang, P.K. Liu, M. Sahimi, T.T. Tsotsis, Design aspects of the cyclic hybrid
adsorbent-membrane reactor (HAMR) system for hydrogen production, Chemical Engineering
Science 65 (2010) 427-435.
[23] S.Y. Lim, B. Park, F. Hung, M. Sahimi, T.T. Tsotsis, Design issues of pervaporation
membrane reactors for esterification, Chemical Engineering Science 57 (2002) 4933-4946.
[24] B. Fayyaz, A. Harale, B.-G. Park, P.K. Liu, M. Sahimi, T.T. Tsotsis, Design aspects of hybrid
adsorbent− membrane reactors for hydrogen production, Industrial & engineering chemistry
research 44 (2005) 9398-9408.
102

[25] B.-G. Park, T.T. Tsotsis, Models and experiments with pervaporation membrane reactors
integrated with an adsorbent system, Chemical Engineering and Processing: Process
Intensification 43 (2004) 1171-1180.
[26] B.M. Esteves, C.S. Rodrigues, L.M. Madeira, Wastewater Treatment by Heterogeneous
Fenton-Like Processes in Continuous Reactors, Applications of Advanced Oxidation Processes
(AOPs) in Drinking Water Treatment, Springer2017, pp. 211-255.
[27] M. Soria, S. Tosti, A. Mendes, L.M. Madeira, Enhancing the low temperature water–gas shift
reaction through a hybrid sorption-enhanced membrane reactor for high-purity hydrogen
production, Fuel 159 (2015) 854-863.
[28] F. García-García, M. León, S. Ordóñez, K. Li, Studies on water–gas-shift enhanced by
adsorption and membrane permeation, Catalysis Today 236 (2014) 57-63.
[29] D.P. Harrison, Sorption-enhanced hydrogen production: a review, Industrial & engineering
chemistry research 47 (2008) 6486-6501.
[30] E. Van Selow, P. Cobden, P. Verbraeken, J. Hufton, R. Van den Brink, Carbon capture by
sorption-enhanced water− gas shift reaction process using hydrotalcite-based material, Industrial
& Engineering Chemistry Research 48 (2009) 4184-4193.
[31] E. Van Selow, P. Cobden, R. Van den Brink, J. Hufton, A. Wright, Performance of sorption-
enhanced water-gas shift as a pre-combustion CO2 capture technology, Energy Procedia 1 (2009)
689-696.
[32] P. Cobden, P. Van Beurden, H.T.J. Reijers, G. Elzinga, S. Kluiters, J. Dijkstra, D. Jansen, R.
Van Den Brink, Sorption-enhanced hydrogen production for pre-combustion CO2 capture:
Thermodynamic analysis and experimental results, International journal of greenhouse gas control
1 (2007) 170-179.
[33] A.M. Ribeiro, C.A. Grande, F.V. Lopes, J.M. Loureiro, A.E. Rodrigues, A parametric study
of layered bed PSA for hydrogen purification, Chemical Engineering Science 63 (2008) 5258-
5273.
[34] B. Dou, Y. Song, C. Wang, H. Chen, Y. Xu, Hydrogen production from catalytic steam
reforming of biodiesel byproduct glycerol: Issues and challenges, Renewable and Sustainable
Energy Reviews 30 (2014) 950-960.
[35] A. Iulianelli, P. Ribeirinha, A. Mendes, A. Basile, Methanol steam reforming for hydrogen
generation via conventional and membrane reactors: a review, Renewable and Sustainable Energy
Reviews 29 (2014) 355-368.
103

[36] Y.-J. Wu, P. Li, J.-G. Yu, A.F. Cunha, A.E. Rodrigues, Sorption-enhanced steam reforming
of ethanol for continuous high-purity hydrogen production: 2D adsorptive reactor dynamics and
process design, Chemical Engineering Science 118 (2014) 83-93.
[37] K. Johnsen, H. Ryu, J. Grace, C. Lim, Sorption-enhanced steam reforming of methane in a
fluidized bed reactor with dolomite as CO2-acceptor, Chemical Engineering Science 61 (2006)
1195-1202.
[38] H.T.J. Reijers, S.E. Valster-Schiermeier, P.D. Cobden, R.W. van den Brink, Hydrotalcite as
CO2 sorbent for sorption-enhanced steam reforming of methane, Industrial & engineering
chemistry research 45 (2006) 2522-2530.
[39] Y.-J. Wu, P. Li, J.-G. Yu, A. Cunha, A. Rodrigues, Sorption-enhanced steam reforming of
ethanol on NiMgAl multifunctional materials: Experimental and numerical investigation,
Chemical engineering journal 231 (2013) 36-48.
[40] H.M. Jang, K.B. Lee, H.S. Caram, S. Sircar, High-purity hydrogen production through
sorption enhanced water gas shift reaction using K2CO3-promoted hydrotalcite, Chemical
engineering science 73 (2012) 431-438.
[41] R.W. Stevens Jr, A. Shamsi, S. Carpenter, R. Siriwardane, Sorption-enhanced water gas shift
reaction by sodium-promoted calcium oxides, Fuel 89 (2010) 1280-1286.
[42] B. Najmi, O. Bolland, K.E. Colombo, A systematic approach to the modeling and simulation
of a Sorption Enhanced Water Gas Shift (SEWGS) process for CO2 capture, Separation and
Purification Technology 157 (2016) 80-92.
[43] J. Boon, P. Cobden, H. Van Dijk, M. van Sint Annaland, High-temperature pressure swing
adsorption cycle design for sorption-enhanced water–gas shift, Chemical Engineering Science 122
(2015) 219-231.
[44] T.M. Murphy, G.T. Offord, D.R. Paul, Fundamentals of membrane gas separation, Membrane
operations: innovative separations and transformations (2009) 63-82.
[45] D. Parsley, R.J. Ciora Jr, D.L. Flowers, J. Laukaitaus, A. Chen, P.K. Liu, J. Yu, M. Sahimi,
A. Bonsu, T.T. Tsotsis, Field evaluation of carbon molecular sieve membranes for the separation
and purification of hydrogen from coal-and biomass-derived syngas, Journal of membrane science
450 (2014) 81-92.
[46] M. Dadwhal, T.W. Kim, M. Sahimi, T.T. Tsotsis, Study of CO2 diffusion and adsorption on
calcined layered double hydroxides: The effect of particle size, Industrial & Engineering
Chemistry Research 47 (2008) 6150-6157.
104

[47] Y. Ding, E. Alpay, High temperature recovery of CO2 from flue gases using hydrotalcite
adsorbent, Process Safety and Environmental Protection 79 (2001) 45-51.
[48] C. Li, M. Wei, D.G. Evans, X. Duan, Layered Double Hydroxide ‐based Nanomaterials as
Highly Efficient Catalysts and Adsorbents, Small 10 (2014) 4469-4486.
[49] H. Van Dijk, K. Damen, M. Makkee, C. Trapp, Water-gas shift (WGS) operation of pre-
combustion CO2 capture pilot plant at the Buggenum IGCC, Energy Procedia 63 (2014) 2008-
2015.
[50] H. Van Dijk, S. Walspurger, P. Cobden, R. Van den Brink, F. De Vos, Testing of hydrotalcite-
based sorbents for CO2 and H2S capture for use in sorption enhanced water gas shift, International
Journal of Greenhouse Gas Control 5 (2011) 505-511.
[51] Y. Kim, W. Yang, P.K. Liu, M. Sahimi, T.T. Tsotsis, Thermal evolution of the structure of a
Mg− Al− CO3 layered double hydroxide: Sorption reversibility aspects, Industrial & engineering
chemistry research 43 (2004) 4559-4570.
[52] W. Yang, Y. Kim, P.K. Liu, M. Sahimi, T.T. Tsotsis, A study by in situ techniques of the
thermal evolution of the structure of a Mg–Al–CO3 layered double hydroxide, Chemical
Engineering Science 57 (2002) 2945-2953.
[53] H. Chen, M. Cao, L. Zhao, R.J. Ciora Jr, P.K. Liu, V.I. Manousiouthakis, T.T. Tsotsis,
Experimental Study of an Intensified Water–Gas Shift Reaction Process Using a Membrane
Reactor/Adsorptive Reactor Sequence, Industrial & Engineering Chemistry Research 57 (2018)
13650-13660.
[54] S. Karagöz, T.T. Tsotsis, I. Vasilios, Multi-scale Modeling and Simulation of a Novel
Membrane Reactor (MR)/Adsorptive Reactor (AR) Process, Chemical Engineering and
Processing-Process Intensification (2019).
[55] S. Karagöz, F.E. da Cruz, T.T. Tsotsis, V.I. Manousiouthakis, Multi-scale membrane reactor
(MR) modeling and simulation for the water gas shift reaction, Chemical Engineering and
Processing-Process Intensification 133 (2018) 245-262.

Abstract (if available)

Abstract The overarching objective of this lab-scale study is to prove the technical feasibility of the membrane- and adsorption-enhanced water gas shift (WGS) process that employs a carbon molecular sieve (CMS) membrane reactor (MR) followed by two adsorption reactors (ARs) in parallel, operating alternately, utilized for the production of high-purity H₂ with simultaneous CO₂ capture during WGS reaction treating a coal gasifier off-gas. In the study whose results are presented here, a commercial sour-shift WGS catalyst (Co/Mo/Al₂O₃) was employed in both the MR and the AR. A CMS membrane was used in the MR, and a hydrotalcite adsorbent was used in the AR. In a preliminary experimental study, we first investigated the integrated MR-AR system without steam sweep in the MR’s permeate side. The experimental results show that membrane, catalyst, and adsorbent all operated stably under the integrated gasification combined cycle (IGCC)-relevant conditions. The stand-alone MR system displayed superior performance in CO conversion and hydrogen purity compared with the conventional packed-bed reactor (PBR) and the MR-AR reactor sequence demonstrated high performance superior to that of a PBR with near 100% conversions attained while the ARs are functional (with an ultrapure hydrogen stream exiting the AR and permeate-side hydrogen purities from the MR of ~75-80%) under the IGCC-relevant operating condition with pressure up to 25 bar and temperature of 250℃. These results have experimentally validated the ability of the hybrid MR-AR process configuration to operate stably and properly under the desired conditions and to intensify the efficiency of the WGS reaction. ❧ Further, we have also investigated the feasibility of performing the lab-scale experiments with the integrated MR-AR system employing steam sweep in the MR’s permeate side. The system was modified and upgraded in order to have the capability of utilizing steam sweep in the MR permeate side and controlling the steam sweep pressure in a variable range (1-3 bar). This integrated MR-AR system with steam sweep was then experimentally evaluated for the WGS reaction under IGCC power generation conditions. The CMS membrane for the MR in the study has displayed very robust and stable performance during the long-term run (over 500 hr run of H₂S exposure and under the 250℃ temperature and up to 25 bar pressure environments) and maintained high He/N₂ selectivity (~126) over a total of 742 hours of operation during the MR-AR experiments. We have experimentally tested the combined MR-AR with steam sweep system in multiple-cycle runs (10-16 cycles), and the system has demonstrated superior performance to that of a conventional PBR with high purities for the hydrogen product which can be directly usable in a hydrogen turbine for power generation. In addition, we have carried-out parametric studies for optimization of the operation of the integrated MR-AR system by investigating various operating conditions for both the MR and the AR. Also, during the MR-AR with steam sweep multiple-cycle run, the state of the membrane, catalyst in both the MR and the AR manifested stable performance. We found that the AR performance (with respect to the catalyst and adsorbent performance in the AR) takes a few cycles to settle to the eventual “steady-state” value. In general, the membrane, catalyst and adsorbent are very robust and stable under the IGCC power generation conditions (large concentration H₂S, high-temperature and high-pressure) environment during the long-period MR-AR multiple-cycle run. ❧ We have also investigated the feasibility of performing the lab-scale experiments with the integrated MR-AR system employing oxygen-blown gasifier off-gas as the MR’s feed gas. The CMS membrane for the MR in this study has demonstrated very robust and stable performance during the long-term run (over 344 hr of H₂S exposure and under the 250℃ temperature and 25 bar pressure environments) and maintained high He/N₂ selectivity (~170) during the MR-AR experiments. We have also performed multiple-cycle runs for the integrated MR-AR system and carried-out parametric studies for optimization of the operation of the integrated MR-AR system. The membrane, catalyst and adsorbent are also very robust and stable during the MR-AR experiments in multiple-cycle runs under the oxygen-blown gasifier off-gas environment.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Process intensification in hydrogen production via membrane-based reactive separations

PDF

The use of carbon molecule sieve and Pd membranes for conventional and reactive applications

PDF

A hybrid adsorbent-membrane reactor (HAMR) system for hydrogen production

PDF

Methanol synthesis in the membrane reactor

PDF

On the use of membrane reactors in biomass utilization

PDF

A flow-through membrane reactor for destruction of a chemical warfare simulant

PDF

An integrated 'one-box' process for hydrogen production

PDF

Methanol synthesis in a membrane reactor

PDF

A study of the application of membrane-based reactive separation to the carbon dioxide methanation

PDF

Methanol synthesis in a contactor-type membrane reactor

PDF

Mesoscale SOFC-based power generator system: modeling and experiments

PDF

The purification of contaminated air streams via recative separation techniques

PDF

Novel methods for landfill gas and biogas clean-up

PDF

Fabrication of silicon carbide sintered supports and silicon carbide membranes

PDF

The study of CO₂ mass transfer in brine and in brine-saturated Mt. Simon sandstone and the CO₂/brine induced evolution of its transport and mechanical properties

PDF

Dynamics of direct hydrocarbon polymer electrolyte membrane fuel cells

PDF

Optimizing biomembrane reactor systems for water reclamation and reuse applications

PDF

High-throughput nanoparticle fabrication and nano-biomembrane interactions

PDF

Continuum and pore netwok modeling of preparation of silicon-carbide membranes by chemical-vapor deposition and chemical-vapor infiltration

PDF

Investigation of gas sorption and mass transfer in gas shales

Asset Metadata

Creator Cao, Mingyuan (author)

Core Title A high efficiency, ultra-compact process for pre-combustion CO₂ capture

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Chemical Engineering

Publication Date 07/26/2020

Defense Date 03/22/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag adsorptive reactor,CO₂ capture,membrane reactor,OAI-PMH Harvest,WGS reaction

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Tsotsis, Theodore (committee chair), Egolfopoulos, Fokion (committee member), Jessen, Kristian (committee member)

Creator Email high007spy@hotmail.com,usccaomingyuan@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-199319

Unique identifier UC11662680

Identifier etd-CaoMingyua-7673.pdf (filename),usctheses-c89-199319 (legacy record id)

Legacy Identifier etd-CaoMingyua-7673.pdf

Dmrecord 199319

Document Type Dissertation

Format application/pdf (imt)

Rights Cao, Mingyuan

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA