BAHAGIAN A – Pengesahan Kerjasama*
Adalah disahkan bahawa projek penyelidikan tesis ini telah dilaksanakan melalui
kerjasama antara
dengan
Disahkan Oleh
Tandatangan :
…………………………… Tarikh ;
Nama
:
……………………………
Jawatan
:
……………………………
…………...
(Cop Rasmi)
* Jika penyediaan tesis/projek melibatkan kerjasama.
BAHAGIAN B – Untuk Kegunaan Pejabat Sekolah Pengajian Siswazah
Tesis ini telah diperiksa dan diperakui oleh:
Nama dan Alamat
Pemeriksa Luar
:
Professor Nik Meriam Binti Nik Sulaiman
Jabatan Kejuruteraan Kimia,
Fakulti Kejuruteraan,
Universiti Malaya,
50603, Kuala Lumpur
Nama dan Alamat
Pemeriksa Dalam I
:
Professor Madya Adnan Bin Ripin
Fakulti Kejuruteraan Kimia & Sumber Asli,
Universiti Teknologi Malaysia,
81310 Skudai, Johor
Disahkan Oleh Penolong Pendaftar di SPS:
Tandatangan :
………………………….
Nama
GANESAN A/L ANDIMUTHU
:
Tarikh :
……………..
DEVELOPMENT OF ADSORPTION SELECTIVE CARBON MEMBRANE
USING CELLULOSE ACETATE FOR SEPARATION OF O2/N2 AND
C1 – C4 HYDROCARBONS/N2
ABDUL RAHIM BIN JALIL
A thesis submitted in fulfilment of the
requirements for the award of the degree of
Master of Engineering (Chemical)
Faculty of Chemical and Natural Resources Engineering
Universiti Teknologi Malaysia
FEBRUARY 2006
ii
I declare that this thesis entitled “Development of Adsorption Selective Carbon
Membrane Using Cellulose Acetate for Separation of O2/N2 and C1 – C4
Hydrocarbons/N2” is the result of my own research except as cited in the references.
The thesis has not been accepted for any degree and is not concurrently submitted in
candidature of any other degree.
Signature
:
……………………….
Name
:
Abdul Rahim Bin Jalil
Date
:
…………………………
iii
“To my beloved mother, father, my family who gave me encouragement towards this
study and to my wife who gave me inspiration and encouragement towards the
success of this study, may our dream come true”
iv
ACKNOWLEDGMENT
I would like to express my gratefulness to Allah S.W.T for giving me strength
and wisdom in my research work. In preparing this thesis, I was in contact with many
people, researchers, academicians, technicians and practitioners. They all have
contributed to my understanding and valuable thoughts during my research.
First and foremost, I would like to express my special thanks to my
supervisor, Associate Professor Dr. Mohd Ghazali Bin Mohd Nawawi, for his
encouragement, guidance, ideas which enlighten my curiosity, suggestion, advice
and friendship. I am gratefully expressing my thanks to my whole family who
understand me and gave me the spirit and continuing support to finish this study.
I am grateful to Universiti Teknologi Malaysia for granting me generous
financial support under Industrial and Technology Development fellowship award
that enabling this research to be done successfully.
My fellow collogues who also should be recognized for their moral support.
Their view and tips are useful indeed, but it is not possible to list them all in this
limited space.
v
ABSTRACT
The objective of this study is to develop a new kind of carbon membrane
which could separate gas based on the adsorption concept. This study was subjected
to encounter challenges imposed by general trade off between permeability and
selectivity of membrane. Membrane was prepared from a thermosetting polymer that
acts as a carbon precursor in this type of membrane. The membrane is formed by
pyrolysis of cellulose acetate supported over a microporous ceramic membrane used
for microfiltration at 3000C, 3250C, 3500C, 4000C, 4500C, and 5000C under N2
flowrate equal to 200 ml/min. The membrane then was further subjected to an
oxidative treatment at temperature between 1500C to 4000C with an interval of 500C.
The pyrolysis temperature was found plays an important role in changing the
morphology of the carbon membrane been developed. Increasing the pyrolysis
temperature produces more pores with smaller diameter, thus reducing the
permeability of the penetrates. The optimum pyrolysis temperature for O2/N2
separation is at 4000C which give the value of the selectivity about 3.92. This value
has exceeded an excellent value that is selectivity above 3.0 as suggested by
Kulprathinja (1988). A hierarchal way to developed the adsorption selective carbon
membrane has been done. The prepared membrane shows high permeabilities and
selectivity towards separation of gas mixtures formed by hydrocarbons and N2.
Membrane prepared at 4000C was further subjected to an air oxidation at 3000C and
gave the value for single gas experiment, C2H6/N2; 2.52, C3H8/N2; 2.44, n-C4H10/N2;
2.35. For binary gas experiment, the selectivity for C2H6/N2; 3.3, C3H6/N2; 14.4, nC4H10/N2; 26.05. A selective and high permeability carbon membrane based on
cellulose acetate could be developed.
vi
ABSTRAK
Objektif utama penyelidikan ini dijalankan adalah untuk menghasilkan
sejenis membran karbon yang mampu memisahkan gas berdasarkan konsep
penjerapan. Penyelidikan ini dijalankan untuk mengatasi dan mengetahui hubungan
timbal balik antara kebolehtelapan dan kemimilihan membran karbon. Membran
disediakan daripada bahan polimer (suhu terkawal) iaitu cellulose acetate. Membran
karbon dihasilkan daripada proses pirolisis satu lapisan nipis polimer ini pada
penyokong seramik pada suhu 3000C, 3250C, 3500C, 4000C, 4500C dan 5000C
didalam aliran nitrogen pada kadar 200ml/min. Membran ini kemudian melalui
proses pengoksidaan proses pengoksidaan antara 1500C hingga 4000C dengan beda
suhu 500C. Suhu pemanasan tanpa udara memainkan peranan yang paling penting
dalam menghasilkan membran karbon ini. Suhu pirolisis yang terlalu tinggi akan
menghasilkan liang-liang rongga dengan diameter yang lebih kecil dan dengan ini
akan menyebabkan penurunan kepada nilai kebolehtelapan membran karbon yang
dihasilkan. Suhu optimum bagi proses pemisahan O2/N2 adalah pada 4000C dimana
nilai kemimilihannya ialah 3.92. Nilai ini telah melebihi nilai yang dicadangkan
untuk pemisahan O2/N2 yang optimum iaitu 3.0 seperti dicadangkan oleh
Kulprathinja (1988).Kaedah yang bersistematik telah dilakukan bagi mendapatkan
suhu optimum pengoksidaan. Membran yang disediakan pada suhu pirolisis 4000C
dan suhu pengoksidaan 3000C memberikan nisbah pemisahan gas bagi ujikaji gas
tulen, C2H6/N2; 2.52, C3H8/N2; 2.44, n-C4H10/N2; 2.35. Ujikaji bagi campuran gas
hidrokarbon dan N2, kemimilihan C2H6/N2; 3.3, C3H6/N2; 14.4, n-C4H10/N2; 26.05.
Membran karbon yang memiliki nilai ketelapan dan kemimilihan yang tinggi dapat
dihasilkan.
vii
TABLE OF CONTENTS
CHAPTER
1
TITLE
INTRODUCTION
1.1
2
PAGE
Membrane-Based Gas Separation Process
1
1.1.1
Historical and Current Status
1
1.1.2
Problem Statement
2
1.2
Objective of Work
5
1.3
Scope of Work
5
LITERATURE REVIEW
2.1
Development of Adsorption Selective
7
Carbon Membrane for Gas Separation
2.1.1 Introduction
7
2.1.2 Fundamentals of Membrane Technology
2.1.2.1 Advantages of Membrane
11
Technology
2.1.2.2 Fundamentals of Gas Permeation
2.1.3 Basic Principle of Adsorption Selective
13
19
Carbon Membrane
2.1.4 Evolution and Development
21
2.1.5 Carbon Membrane
24
2.2
Ceramic Asymmetric Membrane
28
2.3
Parameters Effecting Gas Separation
32
Performance
viii
3
2.3.1 Pyrolysis Parameter
32
2.3.2 Coating Procedure
33
2.3.3 Oxidation Time and Temperature
35
2.3.4 Pressure and Temperature Difference
35
RESEARCH METHODOLOGY
3.1
Materials
37
3.2
Experimental Methods
38
3.3
Preparation Of Carbon Membrane
38
3.3.1 Preparation of Carbon Membrane
38
Support
3.3.2 Preparation of Carbon Precursor
38
3.3.3 Preparation of Adsorption Selective
39
Carbon Membrane
3.4
Design and Fabrication of Gas separation
45
Test Rig
3.5
Gas Permeation Measurement
48
3.6
Characterization of Prepared Carbon membrane
51
3.6.1 General Overview
51
3.6.2 Performance Study of the
51
Membrane
3.6.2.1 Morphologies of Carbon
52
Membrane
3.6.2.2 Effect of Separation Pressure
52
and Temperature on Gas
Separation Performance.
3.6.2.3 Effect of Oxidation Temperature
on Gas Separation Performance.
52
ix
4
RESULT AND DISCUSSION
4.1
Introduction
53
4.2
Membrane Morphology
53
4.2.1 Effect of Pyrolysis Temperature on the
59
Membrane Developed.
4.3
Permeability and Selectivity of Unmodified
64
Ceramic Membrane
4.4
Permeability and Selectivity Properties of
65
CA Carbon Membrane.
4.4.1 Permeability and selectivity of Oxygen
66
and Nitrogen.
4.4.2 Effect of oxidative treatment on the
76
permeability and selectivity of
hydrocarbon, oxygen and nitrogen.
5
REFERENCES
CONCLUSIONS AND RECCOMMENDATIONS
5.1
Conclusions
85
5.2
Recommendations
87
89
x
LIST OF TABLES
TABLE NO.
TITLE
PAGE
2.1
Membrane Separation Process
13
2.2
General hierarchy of permeability of common gas
17
2.3
Advantages and Disadvantages of Ceramic Membranes
31
3.1
Carbonization condition used to prepare cellulose-based
41
carbon membrane.
3.2
Oxidation condition
43
4.1
Permeability and selectivity of oxygen and nitrogen (1 bar)
67
and different separation temperature.
4.2
Permeability and selectivity of oxygen and nitrogen (2 bar)
68
and different separation temperature.
4.3
Permeability and selectivity of oxygen and nitrogen (3 bar)
69
and different separation temperature.
4.4
Permeability and Selectivity of Modified Membrane
77
(Oxidized at 1500C)
4.5
Permeability and selectivity of methane and nitrogen (1 bar)
at different separation temperature.
80
xi
4.6
Separation of single gas ( Temperature:270C; Pressure : 1 bar)
81
4.7
Separation of binary gas ( Temperature:270C; Pressure : 1 bar)
81
xii
LIST OF FIGURES
FIGURE NO.
TITLE
PAGE
2.1
Schematic representation of a chemical process
10
2.2
Schematic representation of mass transport phenomena
16
occurring in solution diffusion membrane
2.3
Structure of inorganic membrane
21
2.4
Schematic representation of different membrane morphologies
22
2.5
Schematic representation of an asymmetric membrane
29
2.6
Pore size range of ceramic membranes and related application fields
29
3.1
Preparation procedure for adsorption selective carbon membrane
44
3.2
Schematic representation of gas separation apparatus
46
3.3
Schematic representation of gas permeation cell
47
3.4
Schematic design of gas permeation cell
47
4.1
Surface View of unmodified ceramic membrane
54
4.2
Cross section view of unmodified ceramic membrane
54
4.3
Surface view of polymer membrane
55
4.4
Cross Section View of Membrane Prepared at Pyrolysis
56
0
Temperatures 400 C (Magnification 200X).
4.5
Cross Section View of Membrane Prepared at Pyrolysis
56
Temperature 4000C (Magnification 500X).
4.6
Surface View of Membrane Prepared at Pyrolysis
Temperature 4000C (Magnification 500X).
57
xiii
4.7
Cross Section View of Membrane Prepared at Pyrolysis
57
Temperature 4000C; Oxidation Temperature 3000C.
(Magnification 200X).
4.8
Cross Section View of Membrane Prepared at Pyrolysis
0
58
0
Temperature 400 C; Oxidation Temperature 300 C.
(Magnification 500X).
4.9
Surface View of Membrane Prepared at Pyrolysis
58
Temperature 4000C; Oxidation Temperature 3000C.
(Magnification 4000X).
4.10
Cross Section View of Membrane Prepared at Pyrolysis; 4000C,
60
Heating rate 50C/min (Magnification 200X).
4.11
Surface View of Membrane Prepared at Pyrolysis; 3000C,
60
0
Heating rate 2 C/min (Magnification 200X).
4.12
Cross Section View of Membrane Prepared at Pyrolysis; 3000C,
61
0
Heating rate 2 C/min (Magnification 200X).
4.13
Surface View of Membrane Prepared at Pyrolysis; 3250C,
61
0
Heating rate 2 C/min (Magnification 500X).
4.14
Cross Section View of Membrane Prepared at Pyrolysis; 3250C,
62
Heating rate 20C/min (Magnification 500X).
4.15
Surface View of Membrane Prepared at Pyrolysis; 3500C,
62
Heating rate 20C/min (Magnification 500X).
4.16
Cross Section View of Membrane Prepared at Pyrolysis; 4500C,
0
Heating rate 2 C/min (Magnification 500X).
63
xiv
4.17
Cross Section View of Membrane Prepared at Pyrolysis; 4500C,
63
Heating rate 20C/min (Magnification 500X).
4.18
Cross Section View of Membrane Prepared at Pyrolysis; 5000C,
64
Heating rate 20C/min (Magnification 4000X).
4.19
Separation properties of unmodified ceramic membrane at 1 bar
65
4.20
Separation Properties of CA based Carbon Membrane at 1 bar
72
(a) 27 degree Celsius; (b) 55 degree Celsius; (c) 100 degree Celsius
4.21
Separation Properties of CA based Carbon Membrane at 2 bar
73
(a) 27 degree Celsius; (b) 55 degree Celsius; (c) 100 degree Celsius
4.22
Separation Properties of CA based Carbon Membrane at 3 bar
74
(a) 27 degree Celsius; (b) 55 degree Celsius; (c) 100 degree Celsius
4.23
Permselectivity as function of separation temperature for oxygen
and nitrogen separation.
75
4.24
Separation properties of C8 at 1 bar.
78
4.25
Permselectivity of methane and nitrogen versus separation
79
temperature.
4.26
Modification of gas permeability through cellulose acetate
83
derived carbon membrane with temperature; single gas.
4.27
Modification of gas permeability through cellulose acetate
derived carbon membrane with temperature; binary gas mixtures.
83
xv
LIST OF SYMBOLS
A
-
Membrane area (cm2)
c
-
Concentration (cm3 (STP)/cm3)
d
-
Kinetic diameter
D
-
Diffusion coefficient (cm2/s)
J
-
Diffusion flux (cm3 (STP)/cm2.s)
MW
-
Molecular weight (g/mol)
Q
-
Volumetric flow rate (cm3/s)
S
-
Solubility coefficient (cm3(STP) /cm3.s.cmHg)
T
-
Temperature ( Kelvin)
Ji
-
Flux of component i
l
-
Membrane thickness (cm)
dc/dx -
Concentration gradient
Di/Dj -
Diffusivity
Kv
-
Geometric constant for viscous or Poiseuille flow through porous
media (dimensionless)
Permi -
Permeability of component i
p
-
Pressure (cm Hg)
pus
-
Upstream pressure (cm Hg)
p ds
-
Downstream pressure (cm Hg)
¨p
-
Pressure gradient (cm Hg)
¨pc
-
Pressure drop by capillary action (cm Hg)
¨pi
-
Difference of partial pressure of component i
p
-
Average pressure (bar)
Qi
-
Volumetric flow rate of component i (cm3/s)
Å
-
Angstrom ( 10-10)
Į
-
Selectivity (dimensionless)
xvi
Ș
-
Viscosity (poise)
P
l
-
Pressure normalized flux (cm3(STP) /cm2.s.cmHg)
Si/Sj
-
Selectivity solubility
v
-
Mean molecular velocity (cm/s)
xvii
LIST OF APPENDICES
APPENDIX
TITLE
PAGE
A1
Permeability and Selectivity of Unmodified Membrane
101
A2
Permeability and Selectivity of Hydrocarbon and
102
Nitrogen gas at 1 bar (single gas)
A3
Permeability and Selectivity of Hydrocarbon and
103
Nitrogen gas at 1 bar (binary gas)
A4
Sample of Gas Permeability Calculation
104
B1
Dimension of Permeation Cell
105
B2
Assembly Component of Permeation Cell
106
B3
Gas Separation Apparatus
107
CHAPTER 1
INTRODUCTION
1.1
Membrane-Based Gas Separation Process
1.1.1 Historical and Current Status
Membranes are increasingly playing on significant role in chemical
technology and being used in variety of applications in our daily life. At the present
time, there is a growing interest in the development of gas separation membranes
based on material providing in terms of chemical and mechanical stability. For the
past several decades, membrane process has gone from laboratory curiosity to
commercial reality. The key functionality of the membranes actually is the ability to
control the permeation of chemical species. In separation applications, the goal is to
allow one component of a mixture to permeate the membrane while hindering other
components (Shiflett, 2002). Mitchell (1831) reported the first scientific observation
for gas separation process. He observed that balloons made of India rubber (natural
rubber) put in gas atmosphere were blown up with different velocity is depending of
the nature of gases. At about the same time in 1855, Fick performed his classical
study, “Uber diffusion” which then was formulated as Fick’s first law for diffusion in
membrane (Fick, 1995).
2
The most remarkable contribution to gas separation using membrane was
attributed by Sir Thomas Graham in 1860 that proposed the formation of Graham’s
law and postulated solution diffusion mechanism (Graham, 1866). He discussed in
modern terms and demonstrated experimentally that mixtures of gas can be separated
via membrane. Successful commercialization of gas separation system is a major
breakthrough in research and development of membrane technology. Recently, a
considerable progress has been made in commercial use of membranes for gas
separations, covering many existing and emerging applications.
1.1.2 Problem Statement
Conventional processes for the separation of certain gases from gaseous
mixture are based on the physical properties of various constituent that we want to
separate. An example is the removal of hydrogen sulfide from natural gas. This
process leaves sulfides as a waste, thus adding the complexity of the whole process.
A real application of the complexity of this process is in the refinery itself where a
complex unit is design and operated in order to remove the sulfur produce in a safely
manner.
An alternative ways is needed in order to improve the competitiveness of the
process. Gas separation membranes seem the solution to overcome this problem. At
present, the major interest in membrane technology is focused in findings inorganic
material that resistant to thermal and chemical influences, withstand harsh
environment and also could produce a higher permeability and selectivity compared
to polymeric membrane.
As the new competitive edge, a major demand emerged in the field of gas
separations using membrane. Polymeric membrane has been used extensively
produced by various researchers and surprisingly result has been achieved. There is
a lack in terms of stability of polymeric membrane that cannot be encountered by the
conventional polymeric membranes but can be done by using carbon membrane.
Carbon membrane could withstand a harsh operating environment such as high
3
operating pressure and temperature without loose of the performance. Carbon
membrane technology has been focused in the gas separation process. It is an effort
to developed carbon membrane who could give a higher permeability and selectivity
of each process before it could apply to the industry broadly. The main problem that
must be overcome before it could be applied to the industry is the fabrication of this
material in a manner that it is reproducible and scalable for manufacturing (Fauzi,
2003).
It has reported that the major barrier in the developing carbon membrane
compromise of many aspects. It involves many aspects such as producing the carbon
membrane itself and also the module involved to attach them for carrying out the
separation analysis. In terms of producing the carbon membrane itself, the material
used as the carbon precursor itself has the significant impact on the overall cost for
producing the carbon membrane. An investigation need to be carried out to find a
more suitable carbon precursor other than polyimide, which is mainly used by other
researcher. This is why this study try to find out the performance of carbon
membrane developed from cellulose acetate. It is not a simple task, but by
contributing a new material that can improve the separation performance without
losing the economically processibility could be a breakthrough in the fields of gas
separations using carbon membrane.
There are many ways in developing carbon membrane but mostly it come
back to the objective of the development of the membrane itself, the pore of the
membrane need to be controlled in a reproducible and tailored fashion. By tailoring
the pore of the membrane, specific applications could be identified. It has been
postulated that the presence of adsorbed molecules forms a barrier to the diffusion of
non-adsorbed molecules and hence hinder their transport (Yang et al, 1999). Based
on this postulated statement, research work has been carried out in order to find the
compatibility of the carbon membrane based from cellulose acetate to separate
absorbable and non-absorbable gases. Thus the motive of this research is subjected
to the development of a new kind of carbon membrane that is adsorption selective
carbon membrane instead of molecular sieve carbon membrane. It is capable of
separating gas based on their adsorption characteristics of the gas molecule and the
membrane itself.
4
One of the major problem encounters in the application of carbon membrane
is the hydrophobic problem. One of the best solutions to overcome this problem is
by coating the carbon membrane with suitable barrier without greatly inhabiting the
flux of other permeating species. This can be accomplished by developing carbon
composite membranes. Jones and Koros (1995) used Teflon Af1600 and Teflon
Af2400 as hydrophobic element. They also suggest that the coating solution can be
made by dissolving the polymeric material in an appropriate solvent so that the
polymer concentration is between 0.5 and 2.0% by weight. They proposed some
coating material such as poly (4-methyl-1-pentene), PMP. Compared to the work
they have done, they found out that The Teflon AF material far less restrictive to the
flux of O2 and N2 than using PMP as hydrophobic element.
Verma and Walker (1992) have proposed a simpler method where they treat
the carbon surface with various agents such as H2 and Cl2 to make it more
hydrophobic but the surface modification would likely change the molecular sieving
properties of the membranes. The trends that the water permeance and O2 permeance
show are typical of capillary condensation. As the relative humidity the water film
thickness increases and menisci will began to form inside the pore and the pore will
began to fill with water blocking the flow of the oxygen (Cooper and Lin, 2002).
This problem was encountered in this study by adding a drying compartment
using silica gel. The gas was pre-dried before pass through the selective carbon
membrane. When we deal with the capability or usefulness of the membrane, we
always refer to the efficiency of the membrane. The efficiency of membrane
separation process does not only depend on the membrane alone. It is also depends
on the way the membrane is installed in the form of membrane module. Many
researchers have proposed membrane modules depends on their applications.
In the field of carbon membrane, the most challenging part to be considered
is the poor mechanical stability of the carbon membrane. This study used a tubular
ceramic membrane as a supporting module in terms to encounter this problem. A gas
separation apparatus were also designed in order to attached the carbon membrane
been developed
5
1.2
Objective of Work
Based on the background of this study, objectives of this study are
categorized as following;
(i)
To develop a new type of adsorption selective carbon membranes for
gas separation using cellulose acetate as a carbon precursor.
(ii)
To determine an optimum preparation condition of the new type of
adsorption selective carbon membrane for membrane separation
process.
(iii)
To analyze the membrane been developed in terms of selectivity and
permeability using oxygen, nitrogen and C1-C4 hydrocarbon gas.
1.3
Scopes of Work
In order to achieve the objective mentioned in 1.3, below are the steps in
order to accomplish this experiment. The scopes of work will be carried out
(i)
To prepare carbon membrane using cellulose acetate as a carbon
precursor using dip coating technique. Asymmetric ceramic
membrane was used as the supporting material.
(ii)
To design and fabricate gas separation unit in order to carry out gas
separation analysis of single and binary mixtures (50/50 by volume)
of gas. Types of gases used were oxygen, nitrogen methane, ethane,
propane and n-butane.
(iii)
To determine the optimum pyrolysis condition for the carbon
membrane been developed in the range of carbonization temperature
at 3000C, 3250C, 3500C, 4000C, 4500C and 5000C.
(iv)
To determine an optimum oxidation temperature between 1500C to
4000C with an interval of 500C.
6
(v)
To study the effect of permeation temperature (270C, 550C and 1000C)
and feed pressure (1 bar, 2 bar and 3 bar).
(vi)
To analyze the component exist in the permeate stream using Hawlett
Packard Agilent 6890N.
(vii)
To characterize and determine the structure and morphology of
modified membrane using Nikon Microscopes and PHILIPS XL-40
Scanning Electron Microscopy (SEM).
CHAPTER 2
LITERATURE REVIEW
2.1
Development of Adsorption Selective Carbon Membrane for Gas
Separation
2.1.1 Introduction
Before any membrane can be used in commercial scale, it must post ideal
criteria that are has a good selectivity and also high flux. In general, there is a classic
tradeoff between these two values. Usually it is preferable to choose membranes
with high selectivity rather than high flux because membrane with low flux can be
compensate by increasing the membrane area or minimizing the membrane
thickness. Membranes used for gas separation also did not excluded by these
criteria.
Ideally, membrane for gas separation has higher selectivity and higher
permeability, is the most economical gas separation process. However the most
early membrane-based gas separation was limited for commercial applications due to
lack of productivity. It happens because membrane has to be relatively thick and
dense to avoid irregularities on membrane surface that cause dramatic loss in
selectivity (Geankoplis, 1993).
8
Gas flow process through microporous material is important to many
industrial applications using membrane gas separations. In particular, recent effort
(oil industry, natural gas processing, etc) concentrate on the exploitation of carbon
membrane. Carbon membrane is also superior to other methods available nowadays
such as distillation, adsorption and absorption that based on energy consumptions.
This separation has been recognized as a key technology for use by the
petrochemical industries (Katsaros et al, 1997).
Gas separation by means of microporous carbon is based on interaction
between components of gas mixtures with respect to the carbon membrane. When
the size of the micpore in the carbon membranes is in the range of 3-5 Å, gas
molecules (< 4 Å) show significant difference in gas diffusivity and mixtures of these
gases can effectively separated according to the molecular sieving mechanism. In
this condition, the gas transport rate through the membrane depends on the effective
size of gas molecules instead of adsorption effects. Membrane with this
characteristic identified as Molecular Sieve Carbon Membrane (MSCM). It is well
known that precursors of MSCM include synthetic polymers such as cellulose
acetate, polyacronitrilles, phenolic resins and many other thermosetting polymers
(Fuertas, 2000). These polymers are initially crosslinked or become crosslinked
during pyrolysis process.
An enlargement of pore size of MSCM from 3-5 Å to 5-7 Å, will
dramatically change the gas separation mechanism. This enlargement can be done
by air oxidation above 1000C to produce different characteristics of MSCM. Thus
gas separation through the membrane with enlarge micropores is governed by
adsorption strength instead of molecular sieve mechanism. Micropores enlargement
cause the lost in selectivity in mixtures formed by permanent gases like O2/N2, He/N2
and CO2/N2. This characteristic is not suitable for separating gas mixtures formed by
permanent gases. In addition, an enlargement of the porosity allows the permeation
of molecules sizes around 4-5 Å, such as hydrocarbon. Because hydrocarbon shows a
high affinity to adsorb on carbon surface, these gases are easily adsorbed on the
enlarge micropores of carbon membrane. Under these circumstances, the gas
transport of adsorbed species takes place according to the surface diffusion as
described by Barrer et al. (1976).
9
When a carbon membrane formed by enlarged micropores comes into contact
with gas mixtures containing gases with different affinities towards carbon (i.e
hydrocarbons/hydrogen, hydrocarbons/nitrogen, etc) it is the more strongly
condensable components that are preferentially adsorbed onto the micropores of the
membrane. This will reduce the open porosity and consequently will limit the
diffusion of less absorbable gases in the micropores.
As a result, the more strongly adsorbed components permeate preferentially
through the membrane (permeate side) whereas the less adsorbed components of the
feed gas mixtures are recovered at the high-pressure side (retentate stream). This
kind of membrane would be effective to separate non-adsorbable or weakly
adsorbing adsorbable gases (i.e He, H2, air, O2, CH4, etc) from adsorbable gases such
as hydrocarbons (C2+), NH3, SO2, H2S, VOCs, CFCs, etc. From now on this type of
membrane will be known as Adsorption Selective Carbon Membrane (ASCM)
(Fuertas, 2001).
2.1.2 Fundamentals of Membrane Technology
The growing significance of membrane and membrane process as efficient
tools for laboratory and industrial scale mass separations is based on the several
properties, characteristics off all membrane separation process, which make them
superior to many conventional mass separation methods (Ghazali, 1997). An
appreciable energy saving offered by membrane separation process regarding for
replacing conventional technique process like distillation, cryogenic distillation, ion
exchange and many more chemical treatment systems. Membrane separation system
also offered greater flexibility in designing the systems itself while still produces
high quality products. The application of the membrane itself (separation process) is
part of in the daily life such as in chemical process.
Chemical process, in general can be regarded, as a sequence of pretreatment
step, a reaction step and a separation step, transforming the incoming raw materials
10
(input) into the desired products (output) to fulfill the whole process. Engineering of
chemical process is carried out in the framework of minimization of energy
consumption and waste disposal. This can be seen as in Figure 2.1
energy
Boundry
recycle
input
pretreatment
reaction
output
separation
waste
Figure 2.1: Schematic representation of chemical process
Every process involved energy consumption. This energy required for
processing raw material, pretreatment process, reaction process, separation process,
and product itself. Besides the desired products, often the undesired or waste
products will also be produced. To fulfill the tasks of diminishing energy
consumption and minimization of waste been produced, a modular, energy efficient
and highly selective separation technique is required such as membrane process
technology.
Membrane come from Latin word that is membrane, membrane in a general
definition is a selective layer between two phases. The membrane can be defined
essentially as a barrier, which separate two phases and restricts transport of various
substances in a selective manner. A membrane can be heterogeneous or
homogenous, symmetric or asymmetric in structure, solid or liquid; can carry a
11
positive or negative charge or be neutral or sometimes bipolar. A membrane
separation system separates an influent stream into two effluent stream known as the
permeate stream (low pressure side) and the retentate stream (high pressure side).
The permeate is the portion of component has pass through the semi-permeable
membrane whereas the retentate stream contains the component that have been
rejected by the membrane.
In a membrane process, a membrane acts as a selective interphase between
two bulk phases. By means of a driving force, some of the species from a multicomponent mixture are transported through the membrane into the other bulk phase
while the membranes retain other components. Selective mass transport has
occurred. Often a membrane, which has a sufficiently high selectivity, is
accompanied by a low transmembrane flux and vice versa making highly selective
membrane process too expensive. Membrane separation process enjoys numerous
industrial applications such as environmentally benign and it also an appreciable
energy savings technique.
Industrial process have a different mode of operation depends on the
complexity of the process itself. The complexity of the process required the most
advance separation tools. Membrane separation process may have different mode of
operation, different structures used as separating barrier and different driving force
used to transport the different chemical species but they actually posses several
features in common which makes them very attractive as a separation tools.
Criteria for selecting membranes for a given application are complex;
durability, mechanical integrity at the operating conditions, productivity and
separation efficiency are important stipulation that must be balance in all cases
(Koros, 1994). Of all these characteristics for a given membrane, selectivity or
separation efficiency and permeation rate (productivity) is clearly the most basic.
The higher the selectivity the more efficient the process, the lower driving force
(pressure ratio) required to achieve a given separation and therefore, the lower the
operating cost of the membrane system. The higher the flux the smaller the required
membrane area and therefore the lower the capital cost of the membrane system
(Koros, 2000).
12
2.1.2.1 Advantages of Membrane Technology
Membrane and membrane separation process have been developed and
optimized more than two decades even for industrial applications (Bruschke, 1995).
Separation of various mixtures, especially organic liquid is a very necessary unit
operation in a chemical industry. A large number of conventional techniques are
available such as adsorption, cryogenic process, distillation, solvent extraction and
fractional crystallization. A new method that can exceed this conventional method is
offered by the application of the membrane itself
By using membrane, conventional separation process can be done in ambient
temperature. This is an excellent criterion that can be meet by the application of
membrane. Therefore solution that is sensitive to temperature can be treated without
damaging the chemical and physical structure of the solution. This is an important
criterion in food industry and biotechnology for processing product that is sensitive
to temperature (Weber and Waren, 1986).
Lately, membrane separation process has been widely used to replace
common separation process that used high technology and also high in capital cost.
Membrane systems offer a low capital cost investment; ease of operation, low energy
consumption and moreover is space efficiency. Membrane process offers a wide
range of application ranging from industrial applications such as gas separations to
dairy products. The demand and the driving force for the wide application of
membrane itself offer the potential of this research to be done. Membrane separation
process can be described as in Table 2.1
13
Table 2.1: Membrane Separation Process
Membrane
Physical State
Driving
Separation
Applications
Process
Feed /
Force
Mechanism
(Separation of)
Permeate
Microfiltration
Liquid / liquid
Pressure
Sieving
Suspended materials
Sieving
Macromolecular
(10-100 kPa)
Ultrafiltration
Liquid / liquid
Pressure
(0.1-1 MPa)
Reverse
Liquid / liquid
Osmosis
Dialysis
Liquid / liquid
solution
Pressure
Solution -
Microsolutes and salts
(1-10 Mpa)
Diffusion
from solutions
Concentration
Diffusion
Low molecular species
Difference
from macromolecular
solutions
Electrodialysis
Gas Separation
Pervaporation
Liquid / liquid
Gas / gas
Liquid / gas
Electric
Selective
Desalination of water
Potential
ion transport or process streams.
Pressure
Solution -
Gases from gas
(0.1- 10 Mpa)
Diffusion
mixtures
Partial
Solution -
Solvent and azeotropic
Pressure
Diffusion
mixture
Difference (0100 KPa)
2.1.2.2 Fundamentals of Gas Permeation.
Thomas Graham (1866) performed the first recorded experiments on the
transport of gases and vapors in polymeric membranes when he observed that a wet
pig bladder inflated to the bursting point when placed in an atmosphere of carbon
dioxide. In fact Thomas Graham become the pioneer of membrane science and
14
technology not only by devising and testing a permeability rate-measuring device
using flat membranes with a vacuum on one side displacing a mercury column but he
also postulating a mechanism for the permeation process. This mechanism viewed
the permeation process as the solution of gases in the upstream surface of the
membrane, diffusion across the membrane, and then evaporation across the
membrane surface (Kesting and Fritzsche, 1993). This is the basis for the so-called
solution diffusion model, which is used in various forms and modifications in the
handling of most membrane problem today (Stannett, 1968).
There are many types of membrane used as separation tools for gas
separation means nowadays. Membrane used in gas separation can be classified as
porous or non-porous (dense) membrane. Transport gases through non-porous
membrane are a complex process that may consist of a sequence of steps stated as
following (Kesting and Fritzsche, 1993):
(i)
Adsorption of gases at upon the upstream boundary
(ii)
Activated diffusion through the membrane
(iii)
Dissolutions or evaporation from the downstream boundary
Gas transport through nonporous membrane is determined predominantly by
solution-diffusion mechanism (Pinnau and Koros,1992). In essence, solution
diffusion mechanism describing a gas transport through nonporous membranes
involving a combination of Fick’s Law of diffusion and Henry’s Law of solubility.
Gas diffusion through nonporous membranes can be described as Fick’s law, which
stated as
J
D
dc
dx
(2.1)
Where J is the diffusion flux or penetrate gas or amount of penetrate gas
permeate through membrane area in unit time, D is diffusion coefficient and dc/dx is
local concentration gradient of sorbed penetrate gas at given position in time (Zoland
and Flemming, 1992). On the other hand, according to Henry’s law, concentration,
15
c and pressure, p of penetrate gas at membrane interface can be related by solubility
coefficient, S ;
c
(2.2)
Sp
where solubility coefficient as reciprocal of Henry’s law constant. Equations (2.1)
and equation (1.2) are then combined and integrated to give
Q
DS
'pA
l
(2.3)
where Q is a volumetric flow rate or permeation rate of penetrate gas, A and l is
membrane effective thickness and area 'p is pressure difference across the
membrane which is given by 'p
p us p ds where pus and p ds are upstream and
downstream pressure respectively. Hence permeability can be defined by this
equation, which is therefore given by
P
DS
(2.4)
In a solution diffusion membrane, the active layer in this type of membrane is
dense in structure. Since they are dense, they do not separate species on the basis of
ordinary sieving mechanism. The separation mechanism for dense membranes is so
called solution-diffusion mechanism (Wessling, 1993). At the feed side, molecules
of component i dissolve in the polymer phase and thermodynamic equilibrium exists
between the penentrants sorbed in the membrane phase and the penentrants in the
feed or permeate side compartment. The chemical potential of component i at the
downstream side lower than that at the feed side, which means that the concentration
of i at the downstream side is also lower. This driving force causes a continuous
diffusional mass transport of the species i through the membrane.
The capability of solution-diffusion membrane to separate multi component
mixtures is based on specific thermodynamic interactions between membrane
material with the different components in the mixture, and furthermore on the
16
selective, diffusive mass transport through a dense homogenous layer of the
membrane itself. A schematic representation of mass transport phenomena occurring
in a solution-diffusion membrane can be shown in Figure 2.2.
Figure 2.2: Schematic representation of mass transport phenomena occurring in
solution-diffusion membrane (Wessling, 1993)
The selective diffusion of the penentrant molecules through a layer of a dense
polymer is mainly influenced by the molecular structure of the polymer itself. Short
range of motions polymer chains, like chain bending, bond rotation and phenyl ring
flips which of course depend on the molecular structure, allow penentrant molecules
to proceed into the direction of the driving force.
The actual thickness of skin layer of asymmetric membrane is usually not
measurable and cannot be determined explicitly using presently available methods.
Instead total gas permeation rate can be expressed as;
§P·
¨ ¸
©l ¹
Q
A'p
(2.5)
17
where
P l is defined as pressure normalized flux or permeability coefficient
divided by effective skin thickness, A is surface area, Q is volumetric flow rate and
'p pressure difference across the membrane. The common unit of pressurenormalized flux is GPU (1 GPU = 1 x 10-6 cm3 (STP)/cm2.s). Different gases has a
different rate of permeability depends on the size and characteristics of the gas itself.
A relative permeability of common gases can be seen in Table 2.2
Table 2.2: General hierarchy of permeabilities of common gases
Relative
Permeability
Gases
Fast
Medium
Slow
H 2O
NH3
H2
He
H2S
CO2
Ne
O2
C2H2
Ar
Xe
CO
CH4
N2
C2H6
C3H8
Selectivity of membrane for mixtures of penetrate gases is a degree of
separation or recovery. Selectivity is dimensionless parameter. When downstream
pressure is negligible compared to upstream pressure (or the absolute downstream
pressure is close to zero), selectivity Į is determined by relative permeability of
component i and j,
D ij
Pi
Pj
Di S i
Dj S j
(2.6)
Intrinsic selectivity is a standard selectivity attributed by dense film of
membrane material corresponding to particular gas pair (Zoland and Fleming, 1992).
Meanwhile the apparent selectivity is determined by experiments. Ratios of
D
i
D j and S i S j are called mobility (or diffusivity) selectivity and the solubility
18
selectivity respectively. The mobility selectivity is governed by the dimensions of
the penetrating gases and the packing formation formed in membrane matrix to
function as a selective media. The solubility selectivity (thermodynamic factor) is
determined by interactions between gas molecules with membrane materials (Odani
and Masuda, 1992).
Gas transport through porous membrane can occur by viscous or Poiseuille
flow, transition flow and Knudsen flow (Claussi et al, 1999). If mean free path of
gas molecules is small compared to pore dimension, only gas-gas collision is
significant and gas transport take place by viscous mechanism or Poiseuille flow. In
this regime, permeance can be expressed as;
§P·
¨ ¸
©l ¹
Kv
K
p
(2.7)
where K v is a geometric constant dependent on porous morphology, Ș is gas
viscosity and p is average pressure. Consequently, selectivity of gas transport
through porous membrane by Poiseuille flow is simply the inverse ratio of gas
viscosities,
D ij
Kj
Ki
(2.8)
If mean free path of gas molecule is large enough relative to pore dimension,
gas-wall collision is significant and gas transport becomes dominated by Knudsen
flow. In this regime, permeance is given by;
§P·
¨ ¸
©l ¹
Kk v
(2.9)
where Kk is a geometric constant dependent on porous morphology and v is mean
velocity of gas which is function of molecular weight. Therefore the selectivity of
19
gas transport through porous membrane by Knudsen flow is estimated by square root
of inverse ratio of gas molecular weights,
D ij
MW j
MWi
(2.10)
2.1.3 Basic Principle of Adsorption Selective Carbon Membrane
Traditional polymer membranes that used for gas separation applications
have been achieved an upper limit. This information has not surpassed in many
years. Robeson (1991) has described the limit of separation performance exhibited
by conventional processabble polymer membranes. The explanation was using a
graph relating gas permeability to selectivity. It is appears that to exceed this limit,
new materials need to be developed.
Concern on this issue, a new type of membrane need to be produced. Carbon
membrane seems to be the alternative to exceed the upper limit achieved for
traditional polymer membranes. There are a bundle of carbon materials that can be
used as a starting material for developing Adsorption Selective Carbon Membrane
(ASCM). Experimental results shown that carbon material has the ability to
distinguish penetrants on a molecular scale and selectivity separate gas pair with
similar sizes (Singh, 1997). The concept of carbon membrane for gas separation can
be found in the early 1970’s. Molecular sieve carbon membrane can be obtained by
pyrolysis of many thermosetting polymers such as polyacronitrille (PAN), poly
(furfuryl alcohol) (PFA), poly (vinyl chloride) (PVDC), and also from oil palm shell.
It also has been proved that molecular sieves membranes to be effective for gas
separations in adsorption applications (Koresh and Soffer, 1980). They also
described that the pore dimensions of the membrane is depend on the carbon
precursor and also processing conditions when developing the carbon membrane.
20
Performance of membrane-based separation process strongly depends on
permeability and selectivity of membrane. Permeability indicates flux or permeation
rate of gases through membrane while selectivity indicates degree of separation or
recovery. Membrane that has high value of permeability leads to a higher
productivity while higher selectivity leads to higher recovery and lower power costs.
Membrane which posses high value of both permeability and selectivity will lead to
the most economical gas separation process but unfortunately there will be some
trade off between this two membrane characteristics. Both of these parameters tend
to exhibit an inverse relation representing a major problem in production and
application of commercial gas separation membranes (Koros and Mahajan, 2000).
Thus, the effectiveness of the membrane in a certain application depends on
the detailed morphology and microstructure of the membrane system, in addition to
the performance above mentioned is related with physical chemistry mechanisms.
The synthesis process critically determines them and this is why a details preparation
of the procedures is so important. Eventually, there are four different mechanisms
for separation of gas mixtures through a porous membrane; Knudsen diffusion;
partial condensation/capillary condensation; surface diffusion/selective adsorption
and molecular sieving (Rao and Sircar, 1993). Carbon molecular sieve membranes
have been identified as very promising candidates for gas separations both in term of
separation properties and stability.
Adsorption Selective Carbon Membrane (ASCM) work by means of
adsorption effects of gases to carbon membrane itself. By enlargement of porosity of
the membrane allows the permeation of the molecules of around 4.5 Å such as
hydrocarbons. This can be done by an oxidative treatment of the carbonized
membrane. From a structural point of view, adsorption selective carbon membrane
is constituted by carbon film with microporous wider than molecular sieve carbon
membrane probably in the range of 5 - 7 Å (Fuertas, 2001).
21
2.1.4 Evolution and Development.
As early as 1831, Mitchell reported that India rubber membranes possed
carbon dioxide substantially faster than hydrogen under equivalent conditions. It has
been known his work is first reported work of gas permselectivity of a given
membrane. Most of the early membrane was limited to commercial application due
to lack of productivity because of the relative thickness and dense to avoid
irregularities on membrane surface that cause dramatic loses in selectivity.
Consequently the greatest limitations using membrane for gas separation process is
the thickness of the membrane itself. (Koros, 1994).
The development of porous inorganic membrane has started long before the
development of today synthetic membranes. The first inorganic membranes were
developed for separation of uranium isotopes, therefore there were mainly used for
military purposes or nuclear applications (Soria, 1995). Nowadays inorganic
membrane becomes an important tool for beverage productions, water purification
and separation of dairy products. (Keizer andVerweij, 1996). Inorganic membrane
can be divided into two classes that are non-porous (dense) membrane and porous
membrane as shown in Figure 2.3
Structure of inorganic membrane
x
x
x
Dense
Nickel
Solid Electrolytes
(Zirconia)
Metal (palladium,
silver and their
alloys)
x
x
x
x
x
Porous
Oxides(Alumina,
titania)
Carbon
Glass (silica)
Metal
Zeolite
Asymmetric
Symmetric
Figure 2.3: Structure of inorganic membrane (Ismail, 2001)
22
Dense (non-porous) inorganic is membranes such as nickel, solid electrolytes
and certain metal such as palladium, silver and their alloys while porous inorganic
membrane such as carbon, silica, metal and also zeolite. Dense inorganic membrane
have limited industrial applications because of their low permeability value
compared to porous organic membrane, therefore nowadays-commercial industrial
membrane market is dominated by porous membrane (Soria, 1995). Porous
inorganic membrane can be also being separated into two different groups that are
asymmetric and symmetric in structure. Figure 2.4 show a schematic diagram of
different membrane morphologies.
Figure 2.4: Schematic representation of different membrane morphologies
(Mulder, 1991)
23
Symmetric membranes can be dense, or can have straight or sponge-like
pores. Homogeneous membranes are merely dense structures. They often used in
research work to characterize the membrane properties, but rarely used in
commercially due to the impractical for large scale industrial separation processes
because the flux of these membranes are generally low.
Asymmetric membranes are made up of a very thin dense layer on top of a
much more porous support sublayer. The resistance to mass transfer is determined
largely by the thin top layer, and the porous support provides the membranes with
sufficient mechanical strength. Another example of an asymmetric membrane is a
composite membrane. Composite membranes also consist of a very thin dense
polymer layer on a microporous support with small resistance to mass transport. The
composite membranes are difference from asymmetric membranes where in the
composite membranes the dense and sublayer layers are made from different
polymeric materials, whereas asymmetric membranes are typically prepared from a
single polymer material. This means that polymers with poor mechanical strength
but good selectivity can be utilized as the top thin dense selective barrier for
composite membranes. Kurdi and Krumbley (1999) state that the desirable structural
characteristics of an asymmetric membrane for gas separations are as follow;
1. The number of dead ended pores should be minimized.
2. The number of selective pores in which surface diffusion of permeate gas
is predominant should be as large as possible.
3. The membrane should have a well-interconnected porous network
structure to avoid high flow resistance.
4. The skin and overall cast film should be thin to obtain high permeance.
5. Macrovoid-free structure to avoid the formation of weak layer under the
skin or avoid the presence of nonselective pores.
6. High mechanical strength to avoid compaction under high-pressure
operation.
24
At present time, there is a growing interest in the development of gas
separation membranes based on materials providing in terms of chemical stability,
thermal stability and most important thing is improving the selectivity. Among nonpolymeric material, molecular sieving materials such as zeolites and carbon have a
potential to push the upper boundary of the permeability versus selectivity tradeoff
relationship. Carbon molecular sieves showed attractive characteristic among
molecular sieving materials such as excellent shape selectivity for planar molecule, it
offers high hydrophobicity, heat resistance and high corrosion resistance (Kyotani,
2000). It is well known that the pyrolysis of certain type of substance (natural or
polymeric) leads to carbon material with very narrow micropore distribution below 1
nm which make it possible to separate gases with similar molecular dimensions.
Eventually, the criteria that been mentioned above can be fulfilled with the
development of carbon membrane.
2.1.5 Carbon Membrane
Gas separations by membranes has acquired a significant role in industry due
to their economic competitiveness compared to the existing separation process
(Centeno and Fuertas, 2002). At preset, a major interest in membrane technology is
finding suitable inorganic material that resistant to thermal and chemical influences
and has higher permeabilities and selectivities compared to polymeric membranes.
The concept of carbon membrane or film gas separation has started in the early 1970.
During that time Ash et al compressed non-porous graphite carbon into a plug, called
a carbon membrane but they meet shrinkage problems, which lead to cracking and
deformation of the membrane. Hence they failed to obtain continuous membrane.
Actually, there is multiple ways to developed membrane for gas separation
application. One approach to develop separation membranes suitable for gaseous
systems is to prepare composite membranes by depositing a polymer films from feed
solution in a porous support. The carbon membrane is obtained by subsequent
carbonization process under controlled conditions. The support material can be as
thick as possible as long as it can provides an adequate mechanical strength and is
25
highly permeable so as not to reduce the permeation rate of gases through the
membrane.
In practice, carbon membrane has been prepared in two main configurations:
(a) unsupported membranes, and (b) supported membranes (Fuertas, 1998). Both of
these types of carbon membranes present some drawbacks. The brittleness of
unsupported carbon membranes create some difficulties for practical use while the
supported membrane itself require that the cycle of precursor depositioncarbonization must be repeated several times in order to obtain an almost crack-free
membrane. Almost all polymeric membrane used in gas separation is of asymmetric
membrane type (Kesting and Fritzsche, 1993), they are constituted of two structully
distinct layers, one of which is thin, dense selective layer and the other a thick
microporous layer who function is provide a physical support to the dense skin.
Taking into account that the facts that cracks in the carbon molecular sieves
usually result from defects from existing on the surface of the microporous support,
the presence of sponge like structure will lessen the effects upon dense carbon film.
The structure of the microporous carbon support is important in order to obtain a
crack free thin film of carbon molecular sieve membranes. In fact when carbon
support without layer is coated, the polymeric solution partially slipped in the
substrate and defect the final membrane. In order to prevent this from happening, a
thin layer (thickness around 10 µm) formed by graphite particles (mean diameter; 3
µm was deposited on the carbon support (Fuertas and Centeno, 1999). Regarding to
their work, they suggest that the existence of the intermediate layer in one casting
step is not sufficient to obtain good carbon molecular sieve membranes. However if
an asymmetric membrane were used, it will drastically reduce the process, the reason
is that probably that the fact that the sponge like structure reduces the defects of the
supports on the thin molecular sieve film.
In the same year, they proposed new material as an intermediate layer in
order to improve the support surface and achieved defect free membrane. A paste
formed by fine graphited particles (Timrex, KS6, TIMCAL GT) with a mean
diameter of 3 µm blended with a polyamide-imide resin (Rhodeftal 311 provided by
Cibo-Geigy) was carefully deposited over a polished surfaced of carbon supported by
26
means of a knife. The support with the intermediate layer was cured (1000C) and
carbonized under vacuum at temperature between 5500C and 7000C (heating rate, 0.5
O
C/min). Ideally the selective membrane material may be directly placed over the
structural support material but it requires a defect free support. Otherwise the defects
of the substrate may be translated to carbon membrane film originating small
pinholes that destroy the molecular sieve properties required for gas separations.
There has been a considerable growth in carbon membrane fabrication for the
separation of gas mixtures for the last few decades. Numerous novel membranes has
been synthesized and evaluated. As an old, yet new material, carbon membrane has
gained great interest regarding to their advantages (Liang et al, 1999);
1. Carbon membrane has higher permselectivity than any known
polymer membranes.
2. Carbon membrane has superior stability in the presence of high
temperature, organic vapor or solvents, and non-oxidizing acids or
bases.
3. The pore dimension of the carbon membrane can be finely adjusted by
simple thermochemical treatment to meet different separation needs.
4. Carbon membrane has superior adsorptive properties for some
specific gases, which can enhance its gas separation capacity.
Eventually, carbon membrane may be developed by means of pyrolyzing of
many thermosetting polymers. Many works has been done in a way to developed
carbon membrane with optimum selectivity and permeability. As noted before,
ASCM can be developed by means of pyrolyzing any thermosetting polymers either
in an inert pyrolysis or vacuum pyrolysis and air oxidation above 1000C. The
characteristics of the membrane developed are governed by several factors such as
carbon precursor used, pyrolyzing time and temperature, oxidation time and many
more. A number of variables will affect the process and a protocol must be
optimized for specific polymer precursor and for specific applications (Jones and
Koros, 1995).
27
Carbon membrane has been prepared from numerous polymeric precursors as
stated before. Rao and Sircar in 1992 obtained a carbon membrane by pyrolysis of
polyvinylidene chloride-acyrlate terpolymer latex coated on a porous graphite
support. The resulting result led to the separation of H2/hydrocarbon mixtures by
selective adsorption and surface diffusion of the larger components. Polyimides also
have been used extensively as a carbon precursor by many researchers. Hayashi et al
in 1995 used a PMDA-ODA polyimide coated on the outer surface of porous
alumina tube. They modified the resulting CMSM by chemical vapor deposition
(CVD) method using propylene as carbon source. This treatment increased the
permselectivity of different pairs (O2/N2, CO2/N2) and He/N2 but a loss of permeance
were observed.
There is bundle of technique to deposit carbon precursor depends on the
physical state of the carbon precursor. The advantages of liquid phase precursor
came to the fore when considering a new means to deliver the polymer to the support
surface in controlled and reproducible fashion (Foley and Acharya, 1999).
Reproducibility proved to be problematic, however rapid research development to
overcome this problem been done. Foley and Acharya used spray-coating technique
by means of depositing carbon precursor on the stainless steel porous support to
overcome this problem. The entire synthesis process is been done in standard fume
hood with no added precautions required for particulate removal. Subsequently in
2000, Foley and Shiflett have proposed new method by means of depositing the
carbon precursor. They use ultrasonic deposition method whereby this method
provides a greater degree of control over deposition step.
As stated before, adsorption selective carbon membrane separate mixture of
gases depends on their adsorption capability of each gas to the carbon material on the
membrane instead of molecular sieving mechanism in molecular sieve carbon
membrane. This is a clear differentiation between molecular sieve carbon membrane
and adsorption selective carbon membrane. In other words, gas that has high affinity
towards carbon is most likely to be adsorbed on the surface of ASCM compared to
other gases that have low affinity towards carbon.
28
There are many ways in developing carbon membrane but it offers some
bottlenecks prior to preparation of it. Practical problem prevent the development of
carbon membrane technology mainly related to the mechanical instability of the
carbon membranes. If the carbon membrane is not grown on a porous substrate, the
most obvious problem encounter by carbon membrane is its very brittle. Preparation
of carbon composite membranes formed by a CMS film supported on a microporous
substrate seems to be a solution to overcome this drawback. This is the reason why
the development of carbon composite membrane supported on microporous substrate
need to be studied in order to enhance the efficiency of the carbon composite itself.
2.2
Ceramic Asymmetric Membrane
The asymmetric membrane system can be shown as in Figure 2.5. It consists
of a porous support with few millimeters in thickness, with pores in the range of 1-10
µm, a porous intermediate layer of 10-100 µm thickness, with pores of 50-500 nm,
and a top layer (the proper separation layer) with a thickness of 1 µm (or smaller) –
10 µm with pores of 2 – 50 nm. The intermediate layer must prevent the penetration
of the precursor of the top layer into the pores of the support during the synthesis and
collapse of the thin finished top layer into the large pores of the support. In all cases
the top layer must be defect free (no crack or pinholes) and have properly a narrow
pore size distribution. This is an ideal criteria need for an optimum separation to
occur. In order it to happen, these sets require demands on the quality of the
intermediate layer and of the support. It may also require development of special
technologies to overcome inferior qualities of the support system.
Ceramic membrane has been used widely and its application can be
simplified in Fig 2.6. Application of ceramic membranes in a wide range need the
most advanced technologies but the most frequently used principle to meet this
requirements is the formation of a layer consisting of a packing of well ordered,
uniform-sized particles. The size and shape of the particles determine the minimum
obtainable mean size and pore size of the ceramic membrane that will be developed.
These parameters as well as the porosity can be changed by further heat treatment
subjected to the ceramic membrane.
29
3
2
1
1. Porous Support
2. Intermediate Layer
3. Separation Layer
Figure 2.5: Schematic representation of an asymmetric membrane
Application fields
Gas (vapor)
separation
Application
field
Reverse osmosis
ultrafiltration
microfiltration
hyperfiltration
0.1
Moving
particles
ions
atoms
1
10
Macromol.
colloidal
100
1000
nm
Microscopic
particles
Molecules, particles
Figure 2.6: Pore size range of ceramic membranes and related application fields
(Mulder, 1991)
30
The need of ceramic membrane as a porous substrate supporting material is
clearly the solutions to overcome the major problem encounter by carbon membrane.
It is difficult to produce an unsupported carbon membrane that continuous and free
of crack and voids. During carbonization cycle molecules will decompose, escape
and the membrane will shrink. When the shrinkage is too much or uneven, the
resulting carbon membrane will crack. Furthermore carbon membrane itself fragile,
insufficient strength and can rupture easily. In order to encounter this problem, a
porous substrate is needed. A carbon membrane with carbon molecular sieve
functions can be prepared on a porous substrate by a pyrolysis process or membrane
deposition process. The porous substrate can provide the mechanical strength
required by carbon molecular sieve (Hong, 2004).
Ceramic tubular membrane has been chose as the membrane support in this
research. Given their unique mechanical strength, thermal stabilities and organic
solvent resistance, ceramic membranes offers an excellent potential for gas
separations in process industries where operating conditions are rather severe (Yang
et al, 1999). Ceramic materials can withstand high operating pressure and
temperature. Schumacher reported the first tubular membrane in 1860 when he
dipped a test tube into cellulose nitrate (collodion) solutions. The quality of any
support used in membrane-based gas separation system is especially critical if the
formation of the top layers in mainly determined by the capillary action on the
support. Besides a narrow pore size distribution the wettability of the support system
plays a role. In many cases an intermediate layer with pore size and thickness lie
between those of the main support and top layer. This intermediate layer can be used
to improve the quality of the support system. If large capillary pressure is used to
form such intermediate layer, defects (pinholes) in the support will be “transferred”
to this layer. This can be avoided by decreasing the acting capillary pressure or even
by eliminating them. This can be done in several ways (Bhave, 1991).
A common method to slip-cast ceramic membranes is to start with a colloidal
suspension or polymeric solution as noted before. This is called a “slip”. The porous
support system is dipped in the slip and the dispersion medium is forced into the
pores of the support by a pressure drop 'Pc created by capillary action of the
31
microporous support. At the interface the solid particles are retained and
concentrated at the entrance of the pores to form a gel layer as in the case of sol-gel
process. It is important that the formation of the gel layers starts immediately and
that the solid particles do not penetrates the pores of the support system. This means
that the solid concentration in the slip must not be close to its gelling state, the
particle (or agglomerate) size must not be too small compared with the pore size of
support unless agglomerates are formed in the pore entrance immediately at the start
of the process. Nowadays many high technology ceramic membranes have been
developed and their came is the same principles as stated above. The advantages and
disadvantages of ceramic membranes can be best shown in Table 2.3
Table 2.3: Advantages and Disadvantages of Ceramic Membranes
Advantages
1. High temperature stability
2. Mechanical stability under large pressure gradients (noncompressibble, no
creep)
3. Chemical stability (especially in organic solvents)
4. No ageing, long lifetime
5. Rigorous cleaning operation allowable (steam sterilization, high backflush
capability)
6. High throughput volume and diminished fouling
7. Good control of pore dimension and pore size distribution
Disadvantages
1.
2.
3.
4.
Brittle character needs special configurations and supporting system
Relatively high capital installation cost
Relatively high modifications costs in case of defects
Sealing technology for high-temperature applications may be complicated
32
2.3
Parameter Effecting Gas Separation Performance
According to the solution-diffusion model, the transport phenomena in gas
separation using ASCM are strongly depending on the solubility and diffusivity of
the permeating components in the gas mixtures itself. Parameters that effect gas
separation performance mainly focused on the processing stage for developing on
Adsorption Selective Carbon Membrane. The parameter involved such as pyrolysis
parameter, coating procedure, oxidation time and temperature and pressure
difference.
2.3.1 Pyrolysis Parameter.
Pyrolysis parameter usually differs and depends on the carbon precursor used.
This parameter includes type and flow rate of inert gas used and pyrolysis
temperature. Changing the pyrolysis parameters will alters the structure of carbon
material and changes the transport properties of specific carbon material been used
(Koros and Steel, 2003). One such parameter is pyrolysis temperature. As the
pyrolysis temperature is increased, the permeability tends to be decreased with the
increase in selectivity, presumably due to the associated change in the pore size
distribution of the material been used. To tailor the separation performance of
carbon membranes, the pyrolysis temperature can be varied in accordance with the
type of precursor been used. It is desirable to keep the processing temperature low
enough to prevent graphitization, especially for coke-forming precursor materials.
For carbon materials, processing temperature are typically in the range 5000C10000C, and carbon molecular sieve membrane synthesis temperature fall within this
range (Geizler and Koros, 1996).
Pyrolysis of carbon material can be done either in an inert pyrolysis condition
or vacuum pyrolysis condition. Both of this condition offered a trade-off between
selectivity and permeability. When pyrolyzed in a vacuum, the polyimide probably
degraded via unimolecular degradation mechanism. The membrane developed will
33
be more selective in separating gas mixture with decreasing value of flux of
permeating species. When an inert gas was used, the degradation process was
“enhanced” presumably due to increased gas phase heat and mass transfer. By
accelerating the carbonization reaction, the inert gas molecules appeared to produce
more “open” porous matrix in the CMS membrane resulting to a higher permeability
and less selective pore structures.
The firing temperature can be altered in a way to improve the membrane
performance. This can be done either by stages heating or directly heating to the
targeted temperature. Heating rate plays an important factor also because an
excessive heating rate may lead to an excessive change in molecular orientation and
may cause the layer to crack easily during the carbonization stage. Some researcher
suggested that the optimum pyrolysis temperature were 7000C for some
thermosetting polymers, but mainly the best optimum pyrolysis temperature depends
on the microstructure and morphology of the carbon precursor itself.
2.3.2 Coating procedure.
An important consideration in designing gas separation membrane is the
thickness of the membrane itself since the permeation rate is maximized when the
thickness is minimized (Fuertas and Centeno, 1999). Coating with suitable carbon
precursor can be done in many ways. In 1999, Foley and Acharya use spray coating
technique to deposit poly (furfuryl) alcohol on a stainless steel disks in a
reproducible manner. The entire process of the coating step was been done in
standard fume hood, with no added precautions required for particulate removal.
Hayashi et al (1996) coated a polyamic acid film synthesis from 3,3’4,4’ biphenyltetracarboxylic dianhydride (BPDA) and 4,4’– oxydialine (ODA) on the
outer surface of a porous alumina support tube.
Coating materials is a critical factor as it provides interactions between the
permeating species and non-permeating species. Foley and Shiflet (2000) coat a
porous stainless steel using ultrasonic deposition with poly (furfuryl alcohol)(PFA)
34
as a coating material. They found that the membranes with the highest selectivity’s
are clustered about a carbon mass to surface area value between 3.4 to 3.6 mg cm-2.
This finding supports the idea that there must be a critical thickness for the
membrane to improve gas separation properties. The coating temperature also gives
a significant impact on the gas separation properties of the membrane itself. They
found that initial coating at higher temperatures might produce high fluxes. It is
believe that by initial coating at higher temperature may have produced a more
porous bridge between the final CMS layer and microporous metal support.
The basic idea behind the asymmetric structure is to minimized the overall
hydraulic resistance of the permeate flow through the membrane structure. The
permeate flux through a given layer is inversely proportional to the layer thickness
and is under simplified assumptions proportional to the same power of the pore size
of the porous layer. It is desirable to have a separative layer (membrane) as thin as
possible and yet possessing defect-free physical integrity and one or more layers
support, which provide an adequate mechanical strength with negligible hydraulic
resistance. This also helps to reduce the pressure required for back flushing in
microfiltration operation. In cases where the precursor particles in the membrane
layer is to small in size compared to the pore size of the bulk of the support, the
membrane particles will significantly penetrates the support pores and the resulting
permeability of the support composite will deteriorate. A practical solution to
adverse this problem is to add one or more intermediate layer having pore size
between those of membrane layer the bulk support.
Additives such as zeolite, titanium oxide, poly(ethylene)glycol can be added
to increase flux rate of the membrane. These additives may have some
characteristics that may enhance either selectivity or permeability rate of the
membrane itself. Many kind of depositing technique of these additives may be found
in literature such as spray coating technique as mentioned above, dip coating,
chemical vapor deposition method and many more method that has been establish by
many researcher.
35
2.3.3 Oxidation Time and Temperature.
It is known that transport properties of gas molecules in CMS by molecular
sieve mechanism. By oxidizing the membrane in air at temperature above 1000C the
will be an enlargement of the membrane micropores. Thus the gas permeation
through the membrane with enlarged micropores is governed by adsorption strength
instead of molecular size. Micropores enlargement is the cause of an important loss
in separation selectivity in mixtures formed by permanent gases like O2/N2, CO2/N2
etc. Thus O2/N2 separation selectivity can change from values in the range of 10-15,
characteristics of CMSM to values of around 1-2. Evidently a membrane with this
characteristic is not suitable for separating gas mixtures formed by permanent gas.
An enlargement of porosity allows the permeation of molecules with sizes of around
4 - 5 Å such as hydrocarbons (Fuertas, 2001).
As the oxidation time increased, permeances abruptly diminished as the
molecular sizes increases indicating that gas transport take place according to the
molecular sieving mechanism instead of adsorption effects. Air oxidation produces
drastic changes on the permeation of pure gas through the carbon membrane with
respect to the non-oxidized sample. It has been shown that as a consequence of
oxidation the membrane are now permeable to hydrocarbons( dk > 4 Å), which
permeate at higher comparable rates than those permanent gases. Minimum
permeances for nitrogen are also observed for all oxidation temperature. As the
oxidation temperature rise, there is an increase in gas permeance for all gases and
diminution in gas permselectivity of permanent gas pairs such as O2/N2 or CO2/N2.
(Fuertas, 2001).
2.3.4 Pressure and Temperature Difference.
Pressure difference will act, as a driving force is needed in terms of
increasing the permeability and selectivity of the membrane that being developed.
A pressure differential is maintained in between the upstream and downstream sides
36
providing the driving force for permeation. The downstream side can maintained as
a vacuum, or at any pressure below the upstream pressure (Koros, 2003). The bigger
the driving force the more permeance will be achieved for certain gases theoretically
but this is governed by the certain membrane itself. As stated before, certain
membrane has a different morphology, different mechanical and different chemical
integrity.
The higher the driving force could diminish the morphology of the membrane
itself due to the lack of mechanical strength of the membrane. It is true that pressure
has the most significant effect on the permeability of gases. Wang et al (1995) has
shown in his study that increasing pressure will alters the separation factors of
helium and oxygen with respect to nitrogen when the pressure is varied. Their study
also includes the transport behavior of fast and slow gas. The transport behavior of
fast gases and slow gases in asymmetric membranes are dramatically different. It is
not only depends on how large is the pressure difference but it also depend on the
relative contribution of the resistance of the porous medium and non-porous medium
varied with other parameters as well.
The permeability and selectivity of asymmetric membrane could be increased
or decreased by raising the temperature at a relatively low pressure depending on the
membrane structure parameters (Wang et al, 1995). When an asymmetric membrane
has a small fraction of defects, there appears an optimum temperature at which the
separation factor is optimum. Under such conditions, high permeability and
selectivity values could be obtained by suitably adjusting the operating pressure and
temperature.
CHAPTER 3
RESEARCH METHODOLOGY
3.1
Materials
Carbon precursor used in this experiment is cellulose acetate. It been selected
as membrane material because of it commercial availability, low cost and ease of
processing. It has been used as a membrane material for many years in reverse
osmosis since it has a high salt rejection and it is relatively inexpensive (Reid and
Breton, 1959). In fact it was one of few polymers to be successfully fabricated into
an asymmetric membrane. Cellulose acetate is also one of few polymers currently
being used in commercial gas separations (Soffer at al, 1987; Stern 1994). Numbers
of researcher has found that the gas permeabilities of asymmetric cellulose acetate
membranes are very high (Gantzel and Mertin, 1970; Stern at all, 1974). Cellulose
acetate is an organic ester that can also dissolve in many organic solvent and widely
used in plastic and coating application.
N,N-dimetilacetamide(DMAc)(boiling point;165.2OC) was chosen as a
solvent in this research in order to prepare the carbon precursor solution. For
cleansing purpose, nitric acid (HNO3)(boiling point; 1220C) and deionized water
were used. All chemicals were used as received in reagent grade purities.
38
3.2
Experimental Methods
The experimental work of this study were divided into three major parts,
which include the development and preparation of carbon-based membranes,
engineering aspects; that is design and fabrication of gas separation test rig and
analysis of the prepared membranes.
3.3
Preparation of Carbon Membrane
Adsorption selective carbon membrane been developed in a laboratory scale
experiments. A method of making adsorption selective carbon membrane
compromise of certain steps such as introducing the carbon precursor in the porous
substrate, heating the porous substrate containing the precursor under certain
condition sufficient enough to convert the precursor into porous adsorptive material
and cooling to ambient temperature (Rao et al,1992).
3.3.1 Preparation of Carbon Membrane Support
Ceramic tubular membranes for microfiltration process have been selected for
the membrane support. The dimensions of the support are 8 mm inside diameter, 12
mm outside diameter and 75 mm in length. For cleaning purposes, the ceramic
membrane was immersed in 30 wt% of HNO3 for 24 hour and then immersed in
deionized water for 2 hour. The membrane supports then were dried at temperature
above 1000C to remove any excess of water and nitric acid. The weights of the dried
membrane were taken as Wo.
The membrane developed is in tubular form whereas ceramic asymmetric
tubular membrane used for microfiltration process was used as the membrane
support. The average pore size was declared by the manufacture to be 1 micron.
The need of this support is clearly to overcome the disadvantages of carbon
membrane mechanical instability if it stands alone. It will provide a sufficient
strength required by the carbon membrane.
39
3.3.2 Preparation of Carbon Precursor
As been mentioned before the carbon precursor that has been selected in this
study is cellulose acetate. Cellulose acetate is one of membrane material applied in
the separation of CO2 from natural gases. It is inexpensive and highly qualified for
the preparation of membranes. This chemical was supplied from Fluka Chemika
with Mf ~ 37000. Solutions were prepared gravimetrically by adding cellulose
acetate with preset concentrations with the desired amount of solvent. The solvent
that were used in this procedure is N,N-Dimethyl acetamide (DMAc)
[CH3CON(CH3)2] (purum;>98.0%) that were also supplied by Fluka Chemica.
The solutions then were stirred for 72 hours at ambient air in order to ensure
homogenization of the solution. The precursor then were filtered and allowed to
stand for degassing before it was deposited on the porous ceramic membrane. By
using liquid phase precursor offer some advantages when considering a new means
to deliver the polymer to the support surface in a controlled and reproducible fashion
(Foley and Achrya, 1999).
Dip coating technique were chosen as a method to deposit the carbon
precursor. Coating the precursor was selected at the inner side of the ceramic
support. This was selected in order to ease for handling purposes and also to avoid
any crack that may happen during handling the develop carbon membrane. For the
coating steps of carbon precursor, the weight of uncoated membrane will be Wco.
The time for the coating steps were set to 15 seconds for each coating step. The
coated membrane then were taken out and cured at controlled temperature using
Carbolite furnace (CWF 1100) with heating rate 20C/min to targeted temperature
1200C and maintain for an isothermal period about 2 hour before allowing the
furnace to cool down. The weight of the coated ceramic then weighed as Wcx where
x carries for number of coating step.
40
3.3.3 Preparation of Adsorption Selective Carbon Membrane
Carbon membrane can be developed either by depositing carbon particulate
with specific pore or by carbonization of polymeric precursor. It is known that
carbon membrane can easily be produced by pyrolyzing any polymeric material.
A number of variables can affect the pyrolysis process and the protocols must be
optimized for specific applications (Jones and Koros, 1995).
During this research certain parameters has been fix and some parameters are
to be manipulated. Parameters that has been fixed such as type of module the
membrane been develop that is tubular form, the concentration of the precursor were
set at 20 wt% Cellulose Acetate (CA) in N,N-dimetilacetamide(DMAc) for all
sample except sample C1 that will be the unmodified membrane. Some of
researchers in the field of carbon membrane used different kind of polymer with
different type of concentrations giving different result in terms of the performance of
the develop membrane. The concentration used was based on the viscosity of the
precursor itself that is 30 cp as suggested by Centeno et all (2004). The type and
flow of inert gas been used that is nitrogen. For pyrolysis temperature the membrane
were carbonized at 3000C, 3250C, 3500C, 4000C, 4500C and 5000C. For oxidation
temperature, the membrane was oxidized in the range of 1500C to 4000C with 500C
interval for 0.5 hour. Certain precautions were taken into consideration in order to
maintain the reproducibility of the membrane been developed.
Prior to the pyrolysis process, the quartz tube were cleaned with potassium
hydroxide (KOH)(0.5 M) before each experiment was done. The cleaning was done
in order to ensure that any particulate that may evolve from precursor from previous
run that may trap in the tube side did not deposit back to the prepared membrane for
the next set of pyrolysis. The quartz tube then was purged with nitrogen for 15
minutes to ensure that all the air had been removed before firing. The flow rate of
nitrogen was set to 200 ml/min during the entire pyrolysis procedure.
41
The firing temperature was set to preset temperature with preset heating rate
and maintain for certain isothermal period at targeted temperature before allowing it
to cool down under the flowing of nitrogen gas. The targeted temperature for the
pyrolysis procedure can be shown in Table 3.1.
Table 3.1: Carbonization condition used to prepare cellulose-based carbon
membrane
Sample
Carbonization
Heating Rate
o
Temperature ( C)
SoakingTime
o
( C/min)
(min)
Carbonization
Atmosphere
C1
0
0
0
-
C2
300
2
30
Nitrogen
C3
325
2
30
Nitrogen
C4
350
2
30
Nitrogen
C5
400
2
30
Nitrogen
C6
450
2
30
Nitrogen
C7
500
2
30
Nitrogen
C8
400
2
30
Nitrogen
C9
400
2
30
Nitrogen
C10
400
2
30
Nitrogen
C11
400
2
30
Nitrogen
C12
400
2
30
Nitrogen
C13
400
2
30
Nitrogen
C14
400
5
30
Nitrogen
In between runs, the furnace was baked in air at 1500C to ensure removal of
any deposited materials that may affect consecutive runs. This standardized protocol
was maintained to ensure reproducibility of the pyrolysis process. There are two
alternatives to develop carbon membrane from polymeric precursor either using
vacuum atmosphere or inert atmosphere condition. Inert pyrolysis conditions were
chosen because this protocol has been proved by previous researcher such as Fuertas
(2001), Geizler (1996) and Koros (1994) to produce high value flux of permeating
species compared to the vacuum pyrolysis process.
42
As far as the separation mechanism is concerned, there are two types of
carbon membrane have been developed that is Molecular Sieve Carbon Membrane
(MSCM) and Adsorption Selective Carbon Membrane (ASCM). The former contains
pores that approach the molecular dimensions of gases (<4 Å) and exhibit selectivity
accordingly to the size and shape of the molecules. The MSCM separate effectively
gas molecules with similar size (Hayashi et all, 1995). On the other hand ASCM
perform on the selective adsorption of certain components of the gas mixtures on the
pores surface followed by surface diffusion of the adsorbed molecules across the
pore. They present micropores in the range of 5 – 7 Å and separate non-adsorable or
weakly adsorable gases (H2, N2, etc) from adsorable gases (hydrocarbon, CFC, etc).
(Rao and Sircar, 1993).
Molecular Sieve Carbon Membrane (MSCM) can be transformed into
Adsorption Selective Carbon Membrane (ASCM) by means of air oxidation at
temperature between 1500C and 4500C. An enlargement of the micropore in the
carbon membrane from 3 to 5 Å for Molecular Sieve Carbon membranes to value in
the range of 5 – 7 Å produces a change in separation mechanism (Fuertas, 2001).
Based on this fact, an oxidative treatment need to be done in order to developed a
new kind of carbon membrane that is adsorption selective carbon membrane
(ASCM).
All of the samples were tested using oxygen and nitrogen to check their
performance in terms of permeability and selectivity. Based on the data collected the
sample who gave the highest value of permselectivity further subjected to oxidation
treatment in order to developed adsorption selective carbon membrane. Oxidation
plays an important role in tailoring the develop ASCM. As an example Fuertas
(2000) and Singh (1997) manipulated these variables in the range of 0.5 hour to 8
hour. In order to minimize these variables, this research has fixed this value. Heating
rate for oxidation steps were set to a fixed value that is 20C/min and were oxidized at
certain temperature with an isothermal period for about 30 minutes and then allowed
to cool down at ambient temperature. This exersice were done for sample C8 to C14.
The difference in air oxidation temperature and time of oxidation will be carried out
as shown in Table 3.2 for the developing the new type of ASCM.
43
Table 3.2: Oxidation Condition
Oxidation temperature (0C)
150
200
250
300
350
400
Oxidation period (hour)
0.5
0.5
0.5
0.5
0.5
0.5
All of the prepared ASCM were mounted in the gas separation rig develop.
The membrane then was feed with a high purity gas from compressed cylinder. The
permeation protocol will be done in a different temperature. Different pressure of
feed will be applied to the modified membrane. The permeation cell will be flushed
with nitrogen priory before each set of experiment been done. All of this procedure
for developing a new kind of adsorption selective carbon membrane can be
simplified as shown in Figure 3.1
44
Dissolved CA in DMAc (wt%)
ambient temperature
Stirred for 72 hour, filtered
Coating Procedure
- Tubular Ceramic Support
- Immersed = 30 wt% HNO3
(24 hr) and Deionized
Water (2 hr)
Drying the coated ceramic at
controlled temperature
Polymer Membrane
Pyrolysis Process under
N2 flow
1. 3000C
2. 3250C
3. 4000C
4. 4500C
5. 5000C
Permeation Test
NO
Does Į O2/N2 > 2.99
YES
Adsorption Selective Carbon
Membrane
Permeation Test Using
CH4 and N2
Oxidation 1500C-4000C
for 0.5 hour
Air Oxidation
Process
Compare the result
which gives
the highest value
Permeation Test Using CH4, C2H6, C3H8, n-C4H10 and N2
Figure 3.1: Preparation procedure for adsorption selective carbon membrane
45
3.4
Design and Fabrication of Gas separation Test Rig
Gas separation rig were developed and fabricated in order to determine the
compatibility of the modified membrane. In the present study the membranes that
produced were characterized using a laboratory scale of gas separation unit. The
schematic representation of the gas separation apparatus can be seen in Figure 3.2
consists of the following apparatus:
(i)
Gas cylinder
(ii)
Mixture compartment
(iii)
Demoisturizer
(iv)
Flowmeter
(v)
Membrane cell
(vi)
Pressure regulator
(vii)
Throttle Valve
(viii)
Vacuum pump
The main component of this gas separation rig is the membrane cell. The end
cap of the tubular membrane cell is fabricated using 316 stainless steel while the
tubular compartments were using grade 1 aluminum. A schematic diagram and an
schematic design of the permeation cell can be seen in Figure 3.3 and 3.4 respectively.
The details of the tubular permeation cell are illustrated in Appendix B1. The
permeation cell was attached horizontally in the gas separation rig with one opening
for each end cap for inlet and outlet of the feed gas. The permeate gas left the
permeation cell through a bottom side opening that connected to a wet test meter
(GCA / Precision Scientific).
The three portions of the cell were clamped and tightly sealed using rubber
O –ring. The effective membrane cell was 19.60 cm2. The inlet feed gas was
controlled via pressure regulator. Different type of gas was injected into the inner side
of tubular ceramic membrane where the selective carbon surfaces were attached. The
Fotek (TC 48-DD-A) digital controller was used to control the permeation cell
temperature by using heated rod and a thermocouple. Furthermore Swagelok tube and
fittings were used to connect equipments in the gas separation rig been developed.
46
2
2
11 1
5
3
8
7
7
6
4
9
13
10
11
12
Figure 3.2: Schematic representation of gas separation apparatus
1; Mixture compartment, 2; silica gel compartment, 3,7,9; flowmeter, ,
4,5,10; three way valve, 6; permeation cell, 8; collector, 11; pressure gouge
12; throttle valve, 13;Gas analyzer
47
Feed
Rerentate
Permeate
Figure 3.3: Schematic representation of gas permeation cell
Figure 3.4: Schematic design of gas permeation cell
48
3.5
Gas Permeation Measurement
Gas permeation test will be performed with a rig setup as shown in Figure
3.3. Each sample also was treated priory to an elevated temperature at 1000C for 10
minutes to ensure all H2O vapor trapped at carbon surface were completely
vaporized. Carbonized membrane and oxidized membrane then mounted in a
stainless steel cylindrical permeation cell and tightened with a rubber O-ring.
Effective permeation area was 19.60 cm2. Prior to testing, care was taken to check
for any leaking that might happen using soap bubbles. The gas separation apparatus
then were purge with nitrogen for 15 minutes priory before any runs were
established. After this protocol has been done, a high pressure-high purity gases
either single gas or binary gas can be introduced into the inner space of the modified
ceramic tubular membrane.
Permeation of gas either single or binary was set to 9 conditions as mention
in section 3.6.2.3. A pressure differential were maintained between upstream and
downstream sides which in this case is the inner side of ceramic and the outer side of
the ceramic support, providing the driving force for permeation. The downstream
side can be maintained at vacuum or any pressure below the upstream pressure
(Koros et al, 2003). All the permeability data for each runs were taken after 1 minute
assuming that constant flow rate for permeates side has been achieved.
Permeability and selectivity of this asymmetric membrane were determined
by constant pressure-variable volume method. In adsorption selective carbon
membrane, the skin layer is the most important one. It can separate different gases
depending on the adsorption effects of gases to the carbon surfaces of the membrane.
It is assume that the skin layer of the carbon membrane formed on the tubular
ceramic membrane that acts as a support has a small fraction of defects.
The membrane performance is characterized by the flux of gas component
across the membrane. The volumetric gas flow rate, Q is a total volume that passes
through the carbon membrane in certain time was calculated by using the equation
below:
49
Q=
V
t
(3.1)
Where,
Q = volumetric flow rate, cm3/s
V = volume of permeable gas, cm3
t = time, second (s)
Permeability can be obtained from the calculation of the volumetric flow rate
for the gas. The compositions of permeate and retentate gas were determined by
using gas analyzer. Permeance of component i, Permi was calculated using
Permi
Ji
'p i
Qi
A'p i
(3.2)
Where,
P = Permeability constant, cm3 (STP) / cm2.s.cmHg
J i = flux of component i
Qi = volumetric flow rate of component i at STP, cm3/s
A = surface area of carbonized membrane, cm2
'p i = difference of partial pressure of component i between the feed
side and the permeate one, cm Hg
The pressure difference of feed side and permeate side were almost keep at
certain value depending on the preset value. The partial pressure of component i at
the permeate side was assumed to be 0 kPa because the amount of permeate gas was
smaller than the feed gas. The common unit of permeability coefficient is Barrer (1
Barrer =10-10 cm3(STP).cm/cm2.s.cmHg).
The actual thickness of membrane usually is not measurable and cannot be
determined explicitly using available methods. Thus, the absolute value of the
permeability coefficient remains unknown. Instead the total gas permeation rate was
determined as;
50
§P·
¨ ¸
© l ¹i
Qi
A'p i
(3.3)
Where Pi l is defined as pressure normalized flux or permeability of gas i
(permeability coefficient divided by effective skin thickness). Each set of data are
determined as an average of three replicates for consistency of the data. The
common unit for measuring the pressure normalized flux of gas applied in membrane
research area in GPU unit. The equation was;
GPU = 1 x 10-6
cm 3 ( STP )
cm 3 .s.cmHg
(3.4)
The permeability of component i relative to component j is defined as
selectivity and can be shown as below;
D
i
j
Permi
Perm j
(3.5)
Gas permeation rate was measured by using pure and mixture of gasses from
compressed cylindrical at various pressure gradients. Penetrate gas was introduced
without dilution into the feed side (Hayashi et al, 1996). As for that, the feed
pressure was controlled at 1, 2 and 3 bar while permeate side were open to vacuum.
Experiments were carried out at different temperature (270C, 550C and 1000C) with
the lower temperature experiment were carried out priory before high temperature
experiment. This procedure will prevent the material from being subjected to undue
thermal cycling and reduces history dependent behavior from the previous runs.
Permeate stream were collected for a period of time. The analysis of the
component exist in permeate stream were carried out using Hawlett Packard Agilent
6890 N. These systems were equipped with thermal conductivity detector (TCD)
and four series column. The peaks detected were identified by matching their
retention time with Scott Gas Standard (P/N 5080 -8755).
51
3.6
Characterization of Prepared Carbon Membrane
3.6.1 General Overview
The prepared adsorption selective carbon membranes were characterized in
order to obtain information about:
(i)
Morphologies of the adsorption selective carbon membrane.
(ii)
Gas separation performances in term of permeation flux and
selectivity for the unmodified ceramic membrane and modified
ceramic membrane.
(iii)
Factors effecting gas separation performance.
3.6.2 Performance Study of the Membrane
The weight change during each heat treatment of the polymeric precursor will
be evaluated by means of weight change of coated ceramic membrane. In order to
obtain information about the textural characteristics, a sample of the precursor was
carbonized under the same conditions used to prepare the carbon membranes. This
study includes the morphologies of the membrane being developed, the effect of feed
pressure, the effect of separation temperature and the effect of oxidation temperature
on membrane that being developed.
Some limitation was encountered in the development of adsorption selective
carbon membrane using cellulose acetate as the carbon precursor. Each of the
samples was only develop once at certain pyrolysis and oxidative treatment. This is
due to the limitation of the asymmetric ceramic membrane that acts as supporting
module of the carbon membrane. Reproducibility of the develop membrane that give
an optimum value for separation could not be done.
52
3.6.2.1 Morphologies of Carbon Membrane.
The structural morphologies of the adsorption selective carbon membranes
were determined using a PHILIPS XL-40 Scanning Electron Microscope (SEM) at
an accelerating voltage of 20 kV. The specimens were sputter-coated with gold prior
to macroscopic observation. When the specimens were ready, they were put into the
SEM apparatus for observation of the surface view and cross section view of the
membrane structures.
3.6.2.2 Effect of Separation Pressure and Temperature on Gas Separation
Performance
In the development of carbon membrane for industrial scale applications
require a significant reduction in overall cost. An alternative way is to alter the best
operating condition for the separation process to occur.
Single and binary gas mixture was prepared with the equal percentage. The
gas separation experiments were carried out at 1, 2 and 3 bar. All of the preset
pressure was controlled using pressure regulator that was attached at compressed gas
cylinder. The separation temperature was set to three different temperatures that is
270C, 550C and 1000C. The permeance of the membrane will be measured in terms
of permeability of gases.
3.6.2.3 Effect of Oxidation Temperature on Gas Separation Performance
It is known that the pore size of the carbon membrane produced can be
altered by using oxidation process. The effect of oxidation temperature on gas
separation process will be studied. The oxidation temperature will be set for six
different environments. The holding time for oxidation period were maintains for
0.5 hour. The temperature of oxidation will be set from 1500C to 4000C with an
interval of 500C as shown in Table 3.2.
CHAPTER 4
RESULTS AND DISCUSSION
4.1
Introduction
In order to investigate the behavior and performance of carbon membrane been
developed, membrane characterization has been carried by means to investigate in
terms of:
i) morphology of the developed membrane that been pyrolyzed at different
temperature
ii) Permeability and selectivity of unmodified membrane.
iii) Permeability and selectivity of carbon membrane been developed for
O2/N2 separation using single gas.
iv) Effect of oxidative treatment on the carbonized membrane for O2/N2,
CH4/N2, C2H6/N2, C3H8/N2, n-C4H10/N2 separation using single and
binary gases at feed pressure equal to 1 bar.
4.2
Membrane Morphology
The membrane morphology was studied using Nikon Microscopes for
preliminary view of unmodified and polymer membrane (Figure 4.1 to Figure 4.3).
For the carbonized and oxidized membrane, the morphology was studied using
54
PHILIPS XL-40 Scanning Electron Microscope (SEM) at an accelerating voltage of
20 kV. (Figure 4.4 to Figure 4.18)
Figure 4.1: Surface layer view of unmodified ceramic membrane
Figure 4.2: Cross-section view of unmodified ceramic membrane
55
Figure 4.3: Surface view of polymer membrane
The SEM photograph of the surface layer and cross-section of the adsorption
selective carbon membrane can be seen shown in Figure 4.4 to Figure 4.18. From the
photograph we can note that there is a difference layer between the carbon membrane
and the ceramic support. Two different parts can be distinguish, the carbon layer and
the ceramic support. It can be also clearly seen that the selective layer is finely
adhered to the ceramic support.
Figure 4.4 to figure 4.6 show the carbonized membrane at 400 degree Celsius
while Figure 4.7 to Figure 4.9 shows the carbonized membrane at 400 degree Celsius
and oxidized membrane at 300 degree Celsius. This is an optimum condition for the
preparation of the carbon membrane been developed. Note that the absence of pores
at a very high level of magnification in the adsorption selective carbon membrane
surface layer and cross-section indicating the dense structure of the homogenous
carbon membrane.
56
Figure 4.4: Cross Section View of Membrane Prepared at Pyrolysis
Temperature 4000C (Magnification 200X)
Figure 4.5: Cross Section View of Membrane Prepared at Pyrolysis
Temperature 4000C (Magnification 500X)
57
Figure 4.6: Surface View of Membrane Prepared at Pyrolysis Temperature
4000C (Magnification 500X)
Figure 4.7: Cross Section View of Membrane Prepared at Pyrolysis; 4000C,
Oxidation Temperature; 3000C (Magnification 200X)
58
Figure 4.8: Cross Section View of Membrane Prepared at Pyrolysis; 4000C,
Oxidation Temperature; 3000C (Magnification 500X)
Figure 4.9: Surface View of Membrane Prepared at Pyrolysis; 4000C, Oxidation
Temperature; 3000C (Magnification 4000X)
59
4.2.1 Effect of Pyrolysis Temperature on the Membrane Developed
It has been mention in section 3.3, preparation of adsorption selective carbon
membrane requires two important steps that is pyrolysis temperature and oxidation
temperature. Pyrolysis temperature plays an important role in the pore formation in
carbon membrane. Many parameters are involved during this process such as targeted
temperature, heating rate, soaking time and also the type of inert gas used. To
minimize all of this parameter, certain parameter has been fix such as soaking time,
type and the flow rate of the flowing inert gas been used as been mention earlier in
section 3.3.3.
Upon pyrolysis, the volatilities will be evolved from the precursor membrane
and this will create voids or pores in the carbon matrix. The heating rate and targeted
pyrolysis temperature will determine the rate of volatilities evolution and subsequently
will control the amount and size of the pores present in carbon membrane. It is the
key factor in the preparation of the developed membrane. If the heating rate were too
high, it may lead to cracking due to the contraction of the material and or thermal
stress (Fuertas and Centeno, 2004). This statement was also observed by this study.
We could observe from SEM photograph (Figure 4.10), there was some deformation
at a higher heating rate. It has been experimentally proved that at a heating rate
50C/min to the carbonization temperature 4000C produce some cracks in the
developed carbon membrane.
Furthermore, increasing the pyrolysis temperature will change the gas
permeation activity. These mechanisms depend hardly on the size of the pore in
carbon membrane developed. As been lay out in Figure 4.11 to Figure 4.18 it can be
see clearly that the impact of manipulating the pyrolysis temperature of the develop
membrane. As the pyrolysis temperature been increased, we could see clearly there is
some deformation of the carbon layer. Thus this lead to the lower permeation of
permeating gas due to the collapse of the pore network develops at higher pyrolysis
temperature. The data can be seen clearly as tabulated in Table 4.2 to Table 4.3 at
different operating temperature (270C,500C,1000C) and operating pressure (1 bar, 2
bar, 3 bar)
60
Figure 4.10: Cross Section View of Membrane Prepared at Pyrolysis; 4000C,
Heating rate 50C/min (Magnification 200X).
Figure 4.11: Surface View of Membrane Prepared at Pyrolysis; 3000C, Heating
rate 20C/min (Magnification 200X).
61
Figure 4.12: Cross Section View of Membrane Prepared at Pyrolysis; 3000C,
Heating rate 20C/min (Magnification 200X).
Figure 4.13: Surface View of Membrane Prepared at Pyrolysis; 3250C, Heating
rate 20C/min (Magnification 500X).
62
Figure 4.14: Cross Section View of Membrane Prepared at Pyrolysis; 3250C,
Heating rate 20C/min (Magnification 500X).
Figure 4.15: Surface View of Membrane Prepared at Pyrolysis; 3500C, Heating
rate 20C/min (Magnification 500X).
63
Figure 4.16: Cross Section View of Membrane Prepared at Pyrolysis; 4500C,
Heating rate 20C/min (Magnification 500X).
Figure 4.17: Cross Section View of Membrane Prepared at Pyrolysis; 4500C,
Heating rate 50C/min (Magnification 50X).
64
Figure 4.18: Surface View of Membrane Prepared at Pyrolysis; 5000C, Heating
rate 20C/min (Magnification 4000X).
4.3
Permeability and Selectivity of Unmodified Ceramic Membrane
In order to evaluate the performance of carbon membrane been developed,
certain types of gas were tested for gas separations that is nitrogen, oxygen, carbon
dioxide, methane, ethane, propane and n-butane. The data for the permeation of gases
in unmodified ceramic membrane will be a basis of comparison for the carbon
membrane been developed. The permeability data for unmodified membrane can be
found in Appendix A1 for oxygen and nitrogen separation. The trend of the separation
could be seen clearly in figure 4.19 for the unmodified membrane at operating
pressure equal to 1 bar. The changes in gas permeances and selectivity will show
what is the mechanism of gas separation occurred during the separation process.
65
Permeability N2
Selectivity
635
630
625
620
615
610
605
600
595
590
585
1.02
1.015
1.01
1.005
1
0.995
0.99
0.985
0.98
0.975
0.97
27
55
Selectivity
Permeability
Permeability O2
100
Separation Temperature
Figure 4.19: Separation Properties of Unmodified Ceramic Membrane at 1 bar
Calculations of the permeability of the gases were carried out three times in
order for constituency of the data. As we can see from Figure 4.19, the trend shows
that the separation is enhanced as the separation temperature is increased. This trend
proves that the separation occur are dominantly effected by the presence of heat in the
system but we will see this trend will not be the same for the separation of
hydrocarbon gas and nitrogen for the develop carbon membrane.
4.4
Permeability and Selectivity Properties of CA Carbon Membrane
In order to evaluate the performance of carbon membrane been developed, two
types of gas were tested for gas separations that is oxygen and nitrogen. The changes
in gas permeance and selectivity are accordingly related to the pyrolysis temperature
and oxidative treatment been done. Permeability and selectivity data of all the carbon
membrane developed at certain pressure and temperature are represented hereinafter.
66
Each of the samples was purged with nitrogen priory before each type of gas
been introduced. It can be clearly shown and discussed in depth by altering the
preparation parameter significantly change the permeability and selectivity of the
prepared membrane. The membrane was tested primarily for oxygen and nitrogen
separation before oxidative treatment was done. An optimum condition based on the
oxygen and nitrogen separation will be chosen for oxidative steps and will be discuss
in the next subchapter.
4.4.1 Permeability and Selectivity of Oxygen and Nitrogen
It is widely known that applications of both of this gas in our daily life play an
important role for the need of pure gas. Pure oxygen is widely used for medical
purposes while pure nitrogen used especially for blanketing perishable fruit and also
shipment of flammable liquids. During this study, the range of pyrolysis temperature
was set between 3000C to 5000C. Experimental results shown in Table 4.1 to Table
4.3, clearly showed that by increasing separation temperature significantly increased
the permeability of the flux for the carbon membrane prepared at certain carbonization
temperature. All of the morphology of the membrane at certain pyrolysis temperature
can be seen from Figure 4.1 to Figure 4.18. The resultant morphology of carbon
membrane had varied with different pyrolysis temperature. Before carbonization were
done, it can be clearly shown from Figure 4.3, there is no pore could be observed on
the surface of the membrane. However, the probabilities of pore network still exist in
the membrane but maybe not connected through the surface.
Pyrolysis Temperature (OC)
300
325
350
400
450
500
Sample
C2
C3
C4
C5
C6
C7
95.08
100
87.05
27
89.55
105.05
100
55
115.05
113.98
27
55
183.56
100
178.65
27
180.06
185.6
100
55
180.56
179.05
27
55
200.85
100
182.72
27
184.35
219.65
100
55
195.45
189.65
27
55
O2 Permeability(GPU)
Permeation Temperature (OC)
74.06
65.05
50.05
50.85
49.32
40.95
60.53
54.79
45.55
89.26
72.00
65.75
144.96
113.85
107.19
200.16
144.15
133.99
N2 Permeability(GPU)
1.28
1.38
1.74
2.06
2.33
2.78
3.03
3.28
3.92
2.07
2.51
2.72
1.39
1.62
1.70
1.09
1.36
1.42
Selectivity
Table 4.1: Permeability and selectivity of oxygen and nitrogen (1bar) at different separation temperature
67
Pyrolysis Temperature (OC)
300
325
350
400
450
500
Sample
C2
C3
C4
C5
C6
C7
358.6
100
299.05
27
350.03
422.56
100
55
412.65
405.14
27
55
529.53
100
507.68
27
510.05
632.55
100
55
600.58
555.05
27
55
584.88
100
555.32
27
575.86
598.64
100
55
568.05
555.53
27
55
O2 Permeability(GPU)
Permeation Temperature (OC)
254.79
226.37
180.26
250.00
210.05
200.58
273.65
220.65
205.00
423.15
393.41
324.91
433.05
411.65
387.88
459.66
430.67
414.55
N2 Permeability(GPU)
1.41
1.55
1.67
1.69
1.96
2.02
1.93
2.31
2.47
1.49
1.53
1.71
1.35
1.39
1.43
1.30
1.32
1.34
Selectivity
Table 4.2: Permeability and selectivity of oxygen and nitrogen (2 bar) at different separation temperature
68
Pyrolysis Temperature (OC)
300
325
350
400
450
500
Sample
C2
C3
C4
C5
C6
C7
1026.00
100
956.50
27
958.00
1050.05
100
55
988.85
968.95
27
55
1285.66
100
1040.55
27
1108.00
1068.28
100
55
1065.08
1058.55
27
55
1213.50
100
1068.5
27
1105.00
1096.08
100
55
1085.05
1066.86
27
55
O2 Permeability(GPU)
Permeation Temperature (OC)
894.35
815.74
613.95
758.55
650.06
584.5
947.05
805.19
704.75
951.00
877.52
805.65
1109.08
963.05
894.65
1033.68
1008.86
908.5
N2 Permeability(GPU)
1.15
1.17
1.56
1.38
1.52
1.66
1.36
1.37
1.47
1.12
1.21
1.31
1.09
1.15
1.19
1.06
1.07
1.17
Selectivity
Table 4.3: Permeability and selectivity of oxygen and nitrogen (3 bar) at different separation temperature
69
70
As we can see from the data tabulated in Table 4.2 to Table 4.4 and plotted in
Figure 4.19 to Figure 4.21, as pyrolysis temperature were increased from 3000C to
5000C, the permeance for both gas were decreased. This is because of pore
formation similar to the size of oxygen and nitrogen molecules. The changes in
permeability value are associated with the modification of textural characteristic
(mean micropore size and micropore volume) of the carbon membrane itself parallel
to the pyrolysis temperature.
It was found that the higher the temperature used for the pyrolysis process,
the smaller were the pores of the product, and thus smaller the molecule could
permeates through such membrane (Soffer et al, 1987). We could observe clearly
that from Figure 4.3 to Figure 4.18 there is some trend of uneven carbon layer
formed at the surface of the supported module (ceramic ultrafiltration membrane). As
the pyrolysis temperature were increased, we could see clearly that the surface of the
carbon membrane develop were totally collapsed thus hindering most of the
permeating species. As a prove for comparison, we could see from Scanning
Electron Micrograph picture that a fine layer (Figure 4.9: Surface View of Membrane
prepared at pyrolysis temperature; 4000C, Magnification; 4000X) change to a
collapsed structure (Figure 4.18: Surface View of Membrane prepared at pyrolysis
temperature; 5000C, Magnification; 4000X).
In the absence of defects, the selectivity is a function of material properties at
the operating conditions (Koros, 2000). Based on the experimental data collected,
the optimum separation temperature for O2/N2 separation achieved at separation
temperature 300C at feed pressure equal to 1 bar. The value achieve were 3.92. This
value almost reach the value for an attractive oxygen separation ranging from 4 to 6
with oxygen permeability 250 Barrer suggested by Puri (1996). This value also a
slight higher for the same polymer developed for gas separations develop by
Kulprathipanja (1988) who reach oxygen/nitrogen selectivity 2.99 and he stated that
selectivity above 3.0 is considered conductive to an excellent separation. Indeed
Kammermeyer (1960) achieve oxygen/nitrogen selectivity around 1.8. The value for
oxygen/nitrogen selectivity 2.99 were selected as a target in finding an optimum
value for the develop carbon membrane before an oxidative treatment were done.
71
The effect of feed pressure does not significantly affect the permselectivity
properties of the carbon membrane. From the data tabulated in Table 4.1 to Table
4.3 and plotted in Figure 4.20 to Figure 4.22, we can see clearly that the permeability
was increased due to increasing value of volumetric feed rate. This study found
some contradictions to the other experimental result by previous researcher as at
higher pressure level the total permeance should be reduced. This is suspected due to
the experimental set up which was difficult to maintain a constant permeate pressure.
These findings were also observed by Rauntenbach on his study regarding to the
impact of operating pressure on hollow fiber membrane used for gas separations.
The questions always arise either the permeance depends on the feed pressure or the
pressure difference across the membrane contributed to the total permeability of the
gas. Based on this findings, it can be concluded that the pressure difference play an
important role and not the absolute pressure level for the asymmetric permeation
properties.
72
Oxygen
Nitrogen
Permselectivity
3.5
3
2.5
2
1.5
1
0.5
0
300
325
350
(a)
400
450
500
Pyrolysis Tem perature
Oxygen
Nitrogen
Permselectivity
250
3.5
3
200
2.5
150
2
100
1.5
1
50
0.5
0
0
300
325
(b)
350
400
450
500
Pyrolysis Tem perature
Oxygen
Nitrogen
Permselectivity
Permeability (GPU)
250
3.5
3
200
2.5
150
2
100
1.5
1
50
0.5
0
(c)
Permselectivity O2/N2
Permeability (GPU)
Permselectivity O2/N2
4.5
4
Permselectivity O2/N2
Permeability (GPU)
200
180
160
140
120
100
80
60
40
20
0
0
300
325
350
400
450
500
Pyrolysis Tem perature
Figure 4.20: Separation Properties of CA based Carbon Membrane at 1 bar
(a) 27 degree Celsius; (b) 55 degree Celsius; (c) 100 degree Celsius
73
Nitrogen
Permselectivity
600
3
500
2.5
400
2
300
1.5
200
1
100
0.5
0
Permselectivity O2/N2
Permeability (GPU)
Oxygen
0
300
(a)
325
350
400
450
500
Pyrolysis Tem perature
Oxygen
Nitrogen
Permselectivity
2.5
700
Permeability (GPU)
2
500
400
1.5
300
1
200
0.5
Permselectivity O2/N2
600
100
0
(b)
0
300
325
350
400
450
500
Pyrolysis Tem perature
Oxygen
Nitrogen
Permselectivity
700
2.5
2
Permeability
500
400
1.5
300
1
200
0.5
Permselectivity O2/N2
600
100
0
(c)
0
300
325
350
400
450
500
Pyrolysis Tem perature
Figure 4.21: Separation Properties of CA based Carbon Membrane at 2 bar
(a) 27 degree Celsius; (b) 55 degree Celsius; (c) 100 degree Celsius
74
Oxygen
Nitrogen
Permselectivity
1200
1.8
Permeability (GPU)
1.4
800
1.2
1
600
0.8
400
0.6
0.4
200
0.2
0
0
300
(a)
Permselectivity O2/N2
1.6
1000
325
350
400
450
500
Pyrolysis Tem perature
Oxygen
Nitrogen
Permselectivity
1.6
1200
Permeability (GPU)
1.2
800
1
600
0.8
0.6
400
0.4
200
0.2
0
0
300
(b)
Permselectivity O2/N2
1.4
1000
325
350
400
450
500
Pyrolysis Tem perature
Nitrogen
Permeability
1.6
1200
1.4
1.2
1000
1
800
0.8
600
0.6
400
0.4
200
0.2
0
(c)
Permselectivity O2/N2
Permeability (GPU)
Oxygen
1400
0
300
325
350
400
450
500
Pyrolysis Tem perature
Figure 4.22: Separation Properties of CA based Carbon Membrane at 3 bar
(a) 27 degree Celsius; (b) 55 degree Celsius; (c) 100 degree Celsius
75
For the membrane prepared at pyrolysis temperature 4500C and 5000C,
experimental data shown that there is a loss in term of permselectivity compared to
all of the carbon membrane prepared which shown some increment. This can be
seen clearly in Figure 4.18. This occurrence happen due to an excessive
carbonization temperature which lead to the collapse of pore network structure and
then hindered the permeation properties of gas in a selective manner. An excessive
carbonization temperature reduces the permeability of both oxygen and nitrogen.
Consequently, work done by Soffer (1987) have shown that the higher the
temperature used for the pyrolysis of the precursor, the smaller were the pores of the
product, and thus the smaller the molecules which could permeate through such
membranes. Mainly these characteristics were used to define the membrane
developed. It is well known that in carbon molecular sieve membrane, the pore is
controlled by carbonization process. Controlling this process can be varies and
mainly it depends on the precursor itself, heating pattern and also the soaking time of
the carbonization process. In fact if an excessive carbonization occurs to the carbon
precursor, the permeation of both gas are selectively eliminated as a result of
shrinkage of carbon structure. (Kasubake at al, 1998).
Permselectivity
Permselectivity versus Temperature at 1 bar
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
20
40
60
80
100
120
Tem perature
C1= Unmodified Membrane
C3 = Pyrolysis Temperature; 325 Degree Celcius
C5 = Pyrolysis Temperature; 400 Degree Celcius
C7 = Pyrolysis Temperature; 500 Degree Celcius
C2 = Pyrolysis Temperature; 300 Degree Celcius
C4 = Pyrolysis Temperature; 350 Degree Celcius
C6 = Pyrolysis Temperature; 450 Degree Celcius
Figure 4.23: Permselectivity as function of separation temperature for oxygen
and nitrogen separation
76
The effect of separation temperature on permselectivity is illustrated in
Figure 4.23 where all of the samples are plotted. As we can see the optimum
separation temperature were observed to be at 27 degree Celsius. Additionally, from
the change of permselectivity, we could see that the values are all higher than those
predicted for Knudsen separation mechanism which gives the value for oxygen and
nitrogen separation equal to 0.94. The gas permeance was increased with separation
temperature indicating that the transport was an activated process as expected for
molecular sieve mechanism. All of this observation can be concluded that the carbon
membrane developed separate gas based on their molecular size or in other words the
gas transport through the membrane occurs to molecular sieving mechanism instead
of Knudsen diffusion. This observation is also observed in zeolites and other
microporous solid (Karger, 1992).
A membrane permeability and selectivity are material properties of the
membrane itself, and thus this property are ideally constant with feed pressure, flow
rate and other process conditions (Koros, 2003). However permeability and
selectivity are both temperature dependent. Permselectivity were increased
significantly by decreasing temperature but compared by increasing pressure this
value was declined proportionally.
4.4.2 Effect of oxidative treatment on the permeability and selectivity of
hydrocarbon, oxygen and nitrogen
Depending on the separation mechanism, there are two types of carbon
membrane that is molecular sieve carbon membrane (MSCM) and adsorption
selective carbon membrane (ASCM). Since MSCM have micropores close to the
size of molecule gas, the diffusivity of these gases through the membrane changes
abruptly with the molecular size and shape (Fuertas, 2000). This kind of membrane
has shown effective to separate gas mixtures such as oxygen and nitrogen but if an
oxidative treatment were done, there is some loss in terms of selectivity of the gas
been separated. This is due to the pore enlargement of the carbon membrane
develop.
77
Based on the data collected for oxygen and nitrogen separation and plotted in
Figure 4.23 showed that membrane produced at carbonization temperature at 400
degree Celsius give the highest selectivity for all the carbon membrane been
developed. This temperature was selected as an optimum carbonization temperature.
Hereinafter the development for adsorption selective carbon membrane will be
carried out by using the sample prepared at this oxidation temperature.
The resulting carbon membrane was subjected to an oxidative treatment as in
Table 3.3. Due to the enlargement of the pore size of the carbon membrane, it is
expected that the permeability will increase and the selectivity will deteriorate for
oxygen and nitrogen separation. It has been proven experimentally when oxidative
treatment was done at 1500C. As we can see in Table 4.4, there is a slight increment
in terms of permeability and some loss in term of selectivity for oxygen and nitrogen
separation compared to the value achieved at carbonization 4000C only. Figure 4.24
show the impact of operating separation temperature for sample C8 at feed pressure
equal to 1 bar. As we can see the value for permeability of both gas increased with
temperature significantly but the permselectivity value drop proportionally with the
increasing separation temperature.
Table 4.4: Permeability and Selectivity of Modified Membrane (Oxidized at
1500C)
Pressure
Permeation
(bar)
Temperature (OC)
1
2
3
Permeability O2
Permeability N2
Selectivity
27
190.55
169.55
1.12
55
195.58
178.56
1.09
100
200.55
198.55
1.01
27
523.55
514.52
1.02
55
536.58
520.05
1.03
100
538.85
521.55
1.04
27
1153.55
1025.55
1.12
55
1208.88
1094.05
1.10
100
1294.55
1198.75
1.09
78
Nitrogen
Permselectivity
205
1.14
200
1.12
195
1.1
190
1.08
185
1.06
180
1.04
175
1.02
170
1
165
160
0.98
155
0.96
150
Permselectivity
Permeability (GPU)
Oxygen
0.94
27
55
100
Separation Tem perature
Figure 4.24: Separation properties of C 8 at 1 bar
The main objective of this study is to show the compatibility of the carbon
membrane developed in term of separation of hydrocarbon and nitrogen. The data
for all of hydrocarbon analysis are presented in Table 4.5 to Table 4.7, Figure 4.24 to
Figure 4.26. Analysis of the hydrocarbon gas was carried out at pressure equal to 1
bar. In order to determine the optimum oxidation temperature in the range between
1500C to 4000C, a systematical pathway was used. Permeation experiments using a
high purity nitrogen and methane were carried out and used as a benchmark property
to determine the optimum oxidation temperature in the range mention before. The
result for methane and nitrogen permeation are tabulated in Table 4.5
Figure 4.24 show the effect of separation temperature on the permselectivity
of each oxidized sample for methane and nitrogen experiment. As expected, the
value of permselectivity increased as separation temperature decreased. Previous
work done by Fuertas (2001) shown that a rise in separation temperature from 250C
to 1500C, produces a strong reduction in the permselectivity value as a consequence
of the desorption of hydrocarbon molecules from the carbon membrane surface.
79
3
Permselectivity
2.5
C8
2
C9
C10
1.5
C11
C12
1
C13
0.5
0
27
55
Separation Temperature
100
Figure 4.25: Permselectivity of Methane and Nitrogen versus Separation
Temperature
Oxidation Temperature(OC)
150
200
250
300
350
400
Sample
C8
C9
C10
C11
C12
C13
574.61
100
649.12
27
602.10
573.21
100
55
603.07
661.44
27
55
517.85
100
626.22
27
521.16
401.77
100
55
397.32
426.425
27
55
269.45
100
290.65
27
273.73
208.47
100
55
214.26
220.42
27
55
CH4 Permeability(GPU)
Permeation Temperature (OC)
310.60
305.50
288.50
289.50
280.50
275.60
258.50
250.56
248.50
243.5
236.5
230.5
215.56
210.56
200.45
198.55
178.55
169.55
N2 Permeability(GPU)
1.85
1.97
2.25
1.98
2.15
2.40
2.00
2.08
2.52
1.65
1.68
1.85
1.25
1.30
1.45
1.05
1.20
1.30
Selectivity
Table 4.5: Permeability and selectivity of methane and nitrogen (1 bar) at different separation temperature
80
81
The value for permselectivity was increase at lower temperature as a
consequence of increase in the potential barrier of nitrogen due to increasing of
hydrocarbon adsorption to the membrane surface. Based on this experimental result,
we can see in the range of oxidation temperature from 1500C to 4000C, the highest
value for selectivity were achieved at oxidation temperature equal to 3000C which
gave the value for methane and nitrogen permselectivity equal to 2.52. This result
will be the basis for the development of adsorption selective carbon membrane using
cellulose acetate as the carbon precursor. Analysis was further carried out for the
separation of other hydrocarbon gases that is ethane, propane and n-butane at feed
pressure equal to 1 bar. The sample used for all the analysis is sample C 11. The
result for analysis was presented in Table 4.6 for permeation of single gas and table
4.7 for permeation of binary mixtures with equal percentage.
Table 4.6: Separation of single gas (Temperature; 270C, Pressure; 1bar)
Permeability (GPU)
Hydrocarbon
CH4 :
C2H6 :
C3H8 :
n-C4H10:
Permselectivity (Hydrocarbon/N2)
N2
626.22
606.5
585.6
405.85
248.50
248.50
248.50
248.50
2.52
2.44
2.35
1.63
Table 4.7: Separation of binary gas (Temperature; 270C, Pressure; 1bar)
Permeability (GPU)
Hydrocarbon
C2H6 :
C3H8 :
n-C4H10:
208.74
167.31
115.95
Selectivity (Hydrocarbon/N2)
N2
63.25
11.62
4.45
3.30
14.40
26.05
82
We can see from both table there is some different in terms of value of
permselectivity and selectivity of the gas been separated. For the permeation of single
gas, the value of hydrocarbon slightly higher than for binary gases experiments. On
the other hand the value for nitrogen decrease as the molecular weight of hydrocarbon
increased in the binary gas experiment compared to single gas experiments which are
constant. In consequence, the measured (hydrocarbon/N2) selectivity is larger than the
estimated from pure gas experiments. In addition, the selectivity is increased with the
increase in molecular weight of hydrocarbon and that means the membranes
selectivity increases with the hydrocarbon condensability. This proves that the
molecules of the hydrocarbons occupying the pore exist in the membrane surface and
partially inhabiting the diffusion on non-absorbable species (N2).
This fact support the statement postulated by Yang in 1999, which postulated
that the presence of adsorbed molecules form a barrier to the diffusion of nonadsorbed molecules, and hence hinder the transport across the membrane. Ash et al
(1973) also observed that the permeances of weakly adsorbed components are
drastically reduced in the presence of strongly adsorbed components. This study also
reaches the same postulate found by other researcher. As a prove we could see that
from the experimental result the degree of hindrance of N2 diffusion increases from
methane to butane because the more condensable hydrocarbon is more strongly
adsorbed. On the other hand, the adsorption of hydrocarbons itself on the membrane
surface effectively reduces the amount of open void space and hence alter the
permeation of nitrogen through the void space exist.
The impact of operating temperature was further carried out for the separation
of single and binary gases. The data are tabulated in Appendix A. The data are best
represented in Figure 4.18 and Figure 4.19. As we can see from the figure, we could
see that there is a trend of dependency of permeability to the operating temperature.
As the operating temperature were raised the permeability of hydrocarbon gases were
declined while the permeability for nitrogen gas were increased.
83
700
260
256
C3H8
500
400
258
CH4
CH4
C2H6
C3H8
n-C4H10
254
252
C2H6
N2
250
300
n-C4H10
N2
n-C4H10
248
200
Nitrogen Permeability (GPU)
Hydrocarbon permeability (GPU)
N2
CH4
C2H6
C3H8
600
246
100
244
0
242
27
55
100
Separation Temperature
Figure 4.26: Modification of gas permeability through cellulose acetate derivedcarbon membrane with temperature; single gas
250
80
70
C2H6
200
60
C3H8
50
150
40
n-C4H10
100
30
Mixtures C3H8-N2
N2 Permeability (GPU)
Hydrocarbon Permeability (GPU)
Mixtures C2H4-N2
20
50
Mixtures n-C4H10-N2
10
0
0
27
55
100
Separation Temperature
Figure 4.27: Modification of gas permeability through cellulose acetate derivedcarbon membrane with temperature; binary gas mixtures
84
This important effect of the adsorption process and the hindrance effect of
hydrocarbon to nitrogen gas on the carbon performance were clearly seen now. As we
could see from Table 4.6, Table 4.7, Figure 4.26 and Figure 4.27, the experimental
result shown that by simply using binary gas, the value of selectivity was increased.
The increment of the selectivity for certain pair of gases were drastically change were
the separation factor for each hydrocarbon separation such as for C3H8 to N2
separation for binary gas were improved about 5 times and for n-C4H10 to N2
separation from the single gas experiment to binary gas experiment were also
improved about 15.95 times. This prove that in the presence of hydrocarbon gas, the
component that is more preferred to adsorbed on the carbon surface will hindered the
presence of non adsorbing gas, in this case is nitrogen. The decrease of nitrogen
permeability was mainly because of decrease of gas diffusivity of nitrogen itself.
Based on the experimental data collected, the optimum value was achieved for
the operating temperature of the hydrocarbon gas and nitrogen is best operated on the
lower temperature. We could see that for the non-absorbable gas (N2), the value for
permeability was increased parallel with the increasing of separation temperature
while for hydrocarbon gases, this value were declined gradually. This pattern were
shown in both test either single or binary gas. Consequence to this behavior, the value
for separation factor was decreased as the operating temperature increased. This
phenomenon was also observed by other researchers.
Based on the result for the experimental work that has been done, we can
conclude that cellulose acetate is a promising candidate for the separation of
hydrocarbon gases from nitrogen gas. I t has been proved experimentally by this work
that this material is effective for the recovery of hydrocarbons from gas mixture
formed by hydrocarbons and nitrogen due to the selective adsorption of the
hydrocarbons molecules onto the pore of the membrane and followed by the surface
diffusion of the adsorbed molecules across the pore. A separation process was highly
selective and a hydrocarbon-rich gas was obtained as the permeate stream.
CHAPTER 5
CONCLUSIONS
5.1 Conclusions
Adsorption selective carbon membrane can be achieved by controlled
atmosphere of pyrolyzing and oxidation of cellulose acetate as the carbon precursor.
A carbon membrane with a high value of permeability and selectivity for
hydrocarbon gases and considerable permselectivity for permanent gas was
achieved. A new kind of carbon membrane based on the adsorption capability were
produced and so called adsorption selective carbon membrane. Many parameters
were involved in developing a new type of carbon membrane. It ranges from
selection of the polymeric precursor, carbonization condition and also oxidation
condition. They are identified as a dominant fabrication parameters that determine
the performance of the adsorption selective carbon membrane been developed.
From physical point of view, a dense separation layer was observed on the
carbon membrane developed. This can be seen clearly as shown before in Figure 4.3
to Figure 4.10. This is also proves that a single cycle coating layer that has been
done in this research is enough to produce a fine layer of selective film adhered to
the ceramic support. As we know, temperature plays an important role in the making
of this adsorption selective carbon membrane. These variables consist of heating rate
and also targeted temperature for the carbonization and also oxidation. Increasing the
heating rate to the targeted carbonization may lead to the shrinkage of the carbon
86
precursor itself. As prove we can see in Figure 4.10 where the selective layer has
been completely crack thus diminished the selective layer. While increasing the
carbonization temperature itself may also lead to the collapse of the pore network of
carbon membrane been developed.
In order to know the performance of the carbon membrane develop,
permeation test need to be carried out. As we can see from the result been layout in
Chapter 4, we can conclude that the changes of gas permeances with pyrolysis
temperature itself is related to the modification of the textural characteristics (mean
micro pore size and mean micro pore volume) of the carbon membrane develop.
We always heard that one of the major problem encountered by carbon
membrane is the hydrophobic problem where as. Some of researches such as Jones
and Kores (1995) used Teflon AF 1600 and Teflon as hydrophobic element to solve
this problem. One of the major concerns in this issue is by adding more material in
the carbon precursor itself may lead to the risk of contributing of restricting the flow
of flux of the permeating species. A simple approach were identified by this research
by doing simple pretreatment before introducing the hydrocarbon and N2 gases to the
selective film of the carbon membrane. A canister contain humidity controller using
silica get will act as moisture dryer in the system. The gases were dried before
introduce to the selective carbon layer.
As the research been carried out, optimum value for the oxygen and nitrogen
separation were identified. The selectivity value were 3.92 was achieved for carbon
membrane developed at pyrolysis temperature equal to 4000C with an operating
temperature at 270C and operating pressure equal to 1 bar. This value is slight higher
for the same polymer develops by Kulprathipanja (1988) that reach O2 / N2
selectivity equal to 2.99. He also stated that selectivity above 3.0 is considered
conductive for excellent separation.
The develop membrane then were subjected to further heat treatment at
controlled oxidation environment. As we can see from the result tabulated in Chapter
4, we can conclude based on the graph that the value of selectivity for the
hydrocarbon separation increased as the molecular weight of the hydrocarbons
87
increased. The degree of N2 hindrance can also be observed at this condition. This is
believe because of the more adsorable compound that is hydrocarbon is preferred to
occupied the pores exist in the carbon membrane thus restrict and hindering the flow
of the N2 itself.
The value of this hydrocarbon and N2 selectivity for each pair of hydrocarbon
and N2 being separated were mainly increased with the decreasing value of operating
separation temperature. Operating pressure has been identified does not significantly
impact the selectivity of the carbon membrane been developed.
5.2
Recommendations
Based on the result and conclusion of this study, some recommendations for
future work compromise of certain aspect. A comprehends study on other fabrication
parameter such as type of inert gas used and also the flow rate of the gas during
pyrolysis process.
In the perspective of polymer membrane preparation wise, two or three layer
of coating been done during polymer deposition to the support on order to check
either it can improve or not the selectivity of the carbon membrane produced. A
detailed study on the carbon precursor concentration could also be done to know the
effect of the precursor concentration itself. Other thermosetting polymer should be
studied also to know their performance as long as their hold an asymmetric
configuration in structure.
Heating also plays an important role of the improvement of the developed
membrane. It is suggested that stages heating should be done instead of direct
heating that has been done in this research. Soaking time also could be manipulated
to know the impact of the develop membrane. Vacuum pyrolysis should be done to
compare the result of this research that used inert pyrolysis as one of the protocol to
carbonized the polymer membrane itself.
88
In order to understand more on the degree of hindrance of N2 to the presence
of hydrocarbon in the system, the concentration of the permeating gas should be
varied. This can be done by manipulating the concentration of hydrocarbon and
nitrogen.
Reproducibility is an issue in this research, due to the lack of ceramic
asymmetric membrane that acts as the supporting module. The developed membrane
was prepared once for each of preparation condition as been laid out in Chapter 3.
The morphology test were problematic due to this test were once through test where
the develop carbon membrane could not be used again. The samples need to be
produced more than one to encounter this problem.
REFERENCES
Ahmad Fauzi Ismail. and David, L.I.B (2001).A review on the latest development of
carbon membranes for gas separation. Journal of Membrane Science.193: 1-18.
Ahmad Fauzi Ismail. and David, L.I.B (2003).Future direction of R&D in carbon
membranes for gas separations. Membrane Technology. April. :4-8.
Auriol, A and Tritten, D. (1973). A process for the manufacture of porous filter
supports. French Patent, 2,463,636”.
Ash.R, Barrer.M.R, Lowson.R.T (1973). Transport of single gases and binary
mixtures in microporous carbon membrane. Journal of Membrane Science. 1: 17
Barrer, M.R. (1939). Permeation, Solution and Diffusion of Gases in Organic
Polymers. Trans. Faraday Society. 35: 628
Barrer, M.R., Ash.R and Sharma, P. (1976). Sorption and flow of carbon dioxide
and some hydrocarbons in microporous carbon membrane. Journal of Chemical
Society Faraday Transional. 69: 2166-2172
Bhave, R. R.(1991) Inorganic Membranes, Synthesis, Characteristics and
Applications. Van Nostrand Reinhold.
Boddeker, K.W (1995).Commentary: Tracing membrane science. Journal of
Membrane Science. 100: 65-68.
Bottino, A., Cannapelli, G. and Munari, S. (1986). Factors effecting the structure
and properties of asymmetric polymer membranes. In: Membrane and
Membrane Process. Plenum Press. 163-178
90
Bruschke, H.E.A (1995). Industrial Application of Membrane Separation Process.
Pure and Applied Chemistry. 67: 993-1002.
Breck, D.W. (1974). Zeolite Molecular Sieves. New York. John Wiley
Centeno, T. A. and Fuertas, A. B. (1999).Supported carbon molecular sieve
membranes based on phenolic resin. Journal of Membrane Science.160: 201-211.
Centeno, T. A. and Fuertas, A. B. (2000). Carbon molecular sieve gas separation
membranes based on poly(vinyl chloride-co-vinyl chloride)Carbon 38:10671073.
Centeno, T. A., Vilas, J. L. and Fuertas, A. B. (2004). Effects of phenolic resin
pyrolysis conditions on carbon membrane performance for gas separation.
Journal of Membrane Science 228: 45-54.
Changhai Liang. , Guangyan Cha and Shucai Gua. (1999).Carbon membrane for gas
separation derived from coal tar pitch. Carbon.37: 1391-1397
Clausi, D.T., Mckelvey, S.A and Koros, W. J. (1999). Characterization of
Substructure Resistance in Asymmetric Gas Separation Membranes. Journal of
Membrane Science.160:51-64
Cooper, C.A. and Lin, Y. S. (2002). Microstructural and gas separation properties of
CVD modified mesoporous Ȗ- alumina membranes. Journal of Membrane
Science.195:31-50
Cowper, C. J. and DeRose, A. J. “The Analysis Of Gases By Chromatography”.
Pergamon Series in Analytical Series Volume 7.
Deraquin, B.V and Levi, S. M. (1964). The Physical Chemistry of Coating Thin
Films on a Moving Support. The Focal Press London
Doshi, J.K (1987). Enhanced Gas Separation Process (U.S Patent 4,690,695)
91
Fick, A.(1995). On Liquid Diffusion. Journal of Membrane Science.100:33-38.
Foley, H. C., Lafyatis, D.S. and Jeannie Tung. (1991).Poly (furfuryl alcohol)Derived Carbon Molecular Sieve: Dependence of Adsorptive Properties on
Carbonization Temperature, Time and Poly (ethylene glycol) Additives.
Industrial Engineering and Chemical Research. 30: 865-873.
Foley, H.C. and Acharya, M. (1999). Spray coating of nanoporous carbon
membrane for air separation. Journal of Membrane Science.161:1-5.
Foley, H. C. and Shiflett, M.B (2000). On the preparation of supported carbon
membranes. Journal of Membrane Science.179: 275-282.
Fuertas, A. B. (1998). Preparation of Supported Asymmetric Carbon Molecular
Sieve Membranes. Journal of Membrane Science.144: 105-111.
Fuertas, A. B. and Centeno, T. A. (1999). Preparation of supported carbon molecular
sieve membranes. Carbon.37: 679-684
Fuertas, A. B. (2000). Adsorption selective carbon membrane for gas separation.
Journal of Membrane Science.177: 9-16.
Fuertas, A. B. (2001). Effects of air oxidation on gas separation properties of
adsorption selective carbon membranes. Carbon.39: 697-706.
Fuertas, A. B. (2001). Preparation and characterization of adsorption selective
carbon membrane for gas separation. Adsorption. 7: 117-129.
Gantzel, P. K. and Mertin, U (1970). Gas Separations with High Flux Cellulose
Acetate Membranes. Ind. Eng. Chem. Process Des. Devolop. Vol 2: 331-332
Geankoplis, C. J. (1993). Transport Process and Unit Operations. 3rd Ed.
Singapore: Prentice Hall International Editions.
92
Geizler, V. C. and Koros, W. J (1996). Effect of Polyimide Pyrolysis Conditions on
Carbon Molecular Sieve membrane Properties. Industrial Engineering and
Chemical Research. 37: 2999-3003.
Ghazali, M.N.W. (1997). Pervaporation Dehydration of Isopropanol-Water Systems
Using Chitosan Membranes:University of Waterloo. Ph.D. Thesis.
Graham, T. (1866). On the Absorption and Dialytic Separation of Gases by Colloid
Septa. Part I. Action of a Septum of Caoutchoue. Philos. Mag. 32: 401-420.
Graham, T. (1995). On the law of diffusion of gases. Journal of Membrane
Science.100: 17-21.
Halil Kalircilar and Ali Culfaz. (2002) “Role of water content of clear synthesis
solutions on the thickness of silicate layers grown on porous Į-alumina tube”
Microporous and Mesaporous Material. 52. 39-54.
Hatori, H., Yamada, H., Shiraishi, H., Nakama, H. and Yoshitomo, S. (1992). Carbon
molecular sieve from polyimide. Carbon. 30: 719-720
Hatori, H., Yamada, H. and Shiraishi, H. (1992). Preparation of microporous carbon
membrane from phase inversion membrane. Journal of Applied Polymer Science.
57: 871-876
Hayashi,J. Mizuta, Yamamoto.M, Kasubake.K and Morooka.S. (1996). Separation
of ethane/ethylene and propane/propylene systems with carbonized BPDA-pp’
ODA polyimide membranes. Industrial Engineering Chemical Research. 35:
4176-4181
Hensema, E.R. and C.A Smolders. (1990). Latest developments in membrane
technology for gas and vapour separation. Process Technology Proceedings,
Volume 8. Elsevier Science Publishers.
93
Hong, N.C.F. (2004). Preparation of carbon molecular sieves membranes on porous
substrate. (U.S Patent 6,730,364)
Jintong Li, Shichang Wang, Kazukiyo Nagai, Tsutomu Nakagawa and Albert W-H
Mau (1998).Effect of polyethyleneglycol (PEG) on gas permeabilities in cellulose
acetate (CA) blend membranes. Journal of Membrane Science.138: 143-152.
Jones,C. W. and Koros,W. J. (1995).Characterization of Ultramicroporous Carbon
Membranes with Humidified Feeds. Industrial Engineering and Chemical
Research. 34: 158-163.
Jones,C. W. and Koros,W. J. (1995). Carbon composite membranes; A Solution to
Adverse Humidity Effects. Industrial Engineering and Chemical Research 34:
164-167.
Jyh-Jeng Shieh, Tai-Sung Chung, Rong Wang, Srivinasan, M.P. and Paul, D.R.
(2001).Gas separation performance of poly (4-vinylpyridine) /polyethermide
composite hollow fibers. Journal of Membrane Science. 182: 111-123.
Katsaros, F. K, Steriotis, T. A., Stubos, A. K., Mitropoulos, A., Kanellopoulos, N.K.
and Tennison, S. (1997). High pressure gas permeability of microporous carbon
membranes. Microporous Material. 8: 171-176.
Keizer, K. and Verweij, H. (1996). “Progress in inorganic membranes” Chemical
Technology 37
Kesting, R.E. and Fritzche, A. K. (1993). Polymeric Gas Separations Membranes.
New York. John Wiley & Sons Inc.
Koresh, J.E and Soffer, A. (1980). Study of molecular sieve carbons: Part 1, Gradual
pore opening and mechanism of molecular sieving. Journal Chemical Society
Faraday Trans. 76: 2457-2469.
94
Koros, W.J and Chern, R.T.(1987). Separation of Gases Mixtures Using Polymer
Membranes. In: Handbook of Separation Process Technology. New York WileyInterscience.
Koros, W. J. (1994). Composite Carbon Fluid Separation Membranes (U.S Patent
5,288,304)
Koros,W.J and Mahajan, R. (2000). Pushing the limits on possibilities on large scale
gas separation: which strategies. Journal of Membrane Science 175: 181-196.
Koros, W.J. (2003). Carbon Molecular Sieves and Methods for Making the Same
(U.S Patent 6,562,110)
Kurdi, J. and Trembly, A. Y.(1999). Preparation of Defect Free Asymmetric
Membranes for Gas Separations. Journal of Applied Polymer Science. 73: 14711482.
Kulprathipanja et al (1988). Separation of Fluids by Means of Mixed Matrix
Membranes. (U.S Patent 4,740,219)
Kyotani, T. (2000). Control of pore structure in carbon. Carbon .38: 269-286
Lafyatis, D. S., Tung, J and Foley, H. C.(1991).Poly(furfuryl alcohol)-Derived
Carbon Molecular Sieve: Dependence of Adsorptive properties on Carbonization
Temperature, Time and poly(ethylene glycol) Additives. Industrial Engineering
and Chemical Research. 37: 2999-3003
Kusabake, K., Yamamato, M and Marooka, S (1998). Gas permeation and
micropore structure of carbon molecular sieving membranes modified by
oxidation. Journal of Membrane Science.149: 59-67.
Mattson, J. S. and Mark H. B.(1971). Activated Carbon New York: Marcel Dekker,
95
Menendez, I. and Fuertas, A. B(2001). Aging of carbon membrane under different
environment. Carbon .39: 733-740.
Mitchell, J.K (1831). On the Penetrative of Fluids. Royal Institution Journal. 2.
101-118.
Misra, R., Barker, A. J. and East, J.(2002). Controlled drying to enhance properties
of technical ceramics. Chemical Engineering Journal .86: 111-116
Mulder, M (1991). Basic Principle of Membrane Technology. Netherlands. Kluwer
Academic Publisher.
Murphy, M.K., Beaver, E.R. and A.W. Rice. (1989). Post-treatment of asymmetric
membranes for gas application. Membrane Separations in Chemical
Engineering. 85: 34-40.
Odani, H. and Masuda, T. (1992). Design of Polymers Membranes for Gas
Separation. In: Toshima, N. Polymers for Gas Separation. New York: VCH
Publishers Inc. 107-116.
Ogawa, M. and Nakano, Y. (2000). Separation of CO2/CH4 through carbonized
membrane prepared by gel modification. Journal of Membrane Science. 173:
123-132
Patel, N., Keijiokabe and Oya, A.(2002). Designing carbon materials with unique
shapes using polymer blending and coating technique. Carbon .40: 315-320
Pinnau, I. and Koros, W.J. (1992). Gas Permeation Properties of Asymmetric
Polycarbonate, Polyestercarbonate and Flourinated Polyimide Membranes
Prepared by the Generalized Dry-Wet Phase Inversion Process. Journal of
Applied Polymer Science. 71: 81-96.
Puri,P.S. (1996). Gas separation membranes current status. La Chemica e Industria.
78: 815
96
Raman, N. K. and Brinker, C. J. (1995) “Organic template approach to molecular
sieving silica membranes”. Journal of Membrane Science.105: 273-279.
Rauntenbach.R, Struck.A, Melin.T, Rooks.M.F.M. Impact of operating pressure on
the permeance of hollow fiber gas separation membranes. Journal of Membrane
Science.169: 217-223.
Rao, M.B (1992). Gas Separation by Adsorbent Membranes. (U.S Patent 5,104,425)
Rao, M.B. and Sircar, S. (1993). Nanoporous carbon membrane for separation of
gaseous mixture by selective surface flow. Journal of Membrane Science. 85:
253-264
Rao, M.B. and Sircar, S. (1996). Performance and pore characterization of
nanoporous carbon membranes for separation of gas separation. Journal of
Membrane Science. 110: 109-118
Reid, R.C.. and Breton, E. (1959). Water and ion flow across cellulosic membrane.
Journal of Applied Polymer Science. 1. 133-143.
Reid, R.C, Praunitz, J.M. and Sherwood, T.K. (1997). The Properties of Gases and
Liquids. New York. McGraw-Hill.
Robeson, L.M (1991). Correlation Separation factor Versus Permeability for
Polymeric Membranes. Journal of Membrane Science. 62:165-185
Robeson, L.M (1999). Polymer membranes for gas separation. Current Opinion in
Solid State and Materials Science for Gas Separation. 549-552
Shiflett, M. B. and Foley, H. C. (1999). Ultrasonic deposition of high selective
nanoporous carbon membranes. Science.285: 1902-1905.
97
Shiflett, M. B. and Foley, H. C. (2000). On the preparation of supported nanoporous
carbon membranes. Journal of Membrane Science.179: 275-282.
Shiflett, M. B (2002). Synthesis, Characterization and Application of Nanoporous
Carbon Membrane. University of Delware: Ph.D. Thesis.
Singh, A. (1997). Membrane Material with Enhanced Selectivity; an entropic
interpretation. University of Texas, Austin: Ph.D. Thesis.
Soffer, A., Koresh, J.E. and Saggy, S. (1987). Separation Device. (U.S Patent
4,685,940)
Soria, R. (1995).Overview on industrial membrane. Catalyst Today .25: 285.
Stannet, V. (1955). The permeability of polymer films to gases – A simple
relationship. Journal of Polymer Science. 16: 89-91
Stannet, V. (1968). Simple Gas. In: Crank, J. Diffusion in Polymers. New York.
Academic Press.
Steel, K. M. and Koros. W.J. (2003). Investigation of porosity of carbon materials
and related effects on gas separation properties. Carbon. 41: 253-266
Stern. S. A (1972). Gas Permeation Process. In: Lacey, R.E. Industrial Processing
with Membranes. Canada: John Wiley & Sons Inc. 279-303.
Stern. S. A, Sen.S.K and Rao.A.K (1974). The permeations of gases through
Symmetric and Asymmetric (Leob-type) Cellulose Acetate Membranes. Journal
Macromolecul Science-Physics .B10: 507-528.
Stern. S. A and Frisch. H.L (1983). Diffusion of small molecules in polymers. In:
CRC Critical Reviews in Solid State and Materials Science Vol 11 No 2. 123187
98
Stern. S. A (1994). Polymers for Gas Separations: The Next Decade. Journal of
Membrane Science.94: 1-65.
Trick, K. A. and Sabila, T. E. (1995). Mechanisms of the pyrolysis of phenolic resin
in a carbon/phenolic composite. Carbon. 33: 719-720
Ulhorn, R. J. R. and Huis, M.B.H.J in Veld, T., Keizer, K. and A.J. Burggraaf.
(1990). Transport and separation of condensable gases with microporous ceramic
membranes. Process Technology Proceedings. 8: 491-497.
Verma, S. K. and Walker, P. L. Jr (1992). Carbon molecular sieve with stable
hydrophobic surface. Carbon 30: 837-844.
Verweij, H. and Renate, M. (1998). Improved Performance of Silica Membranes for
Gas Separation. Journal of Membrane Science.143: 37-51.
Vos, R. M., Maier, W. F. and Verweij, H.(1999). Hydrophobic silica membranes for
gas separations. Journal of Membrane Science.158: 277-288
Wang, D., Li, K., and Teo, W. K. (1995). Effects of temperature and pressure on gas
permselection properties in asymmetric membranes. Journal of Membrane
Science.105: 89-101
Weber,W. F. and Waren, B (1986). Membranes Replacing Other Separation
Technologies. Chemical Engineering Progress.11: 23-24
Winston, W.S and Sirkar, K. K. (1992). Overview. In: Winston, W.S Membrane
Handbook. New York. Van Nostrand Reinhold. 3 - 15.
Weh, K., Noack, M., Sieber, I. and Caro, J. (2002). Permeation of single gases and
gas mixtures through faujasite-type molecular sieve membranes. Microporous
and Mesaporous Material .54: 27-36.
99
Wessling. M (1993). Relaxation Phenomena in Dense Gas Separation Membranes.
University of Waterloo. Ph.D Thesis.
Xianshe Feng, Chuen Y. Pan, Ivory, J. and Ghosh, D. (1998). Integrated
membrane/adsorption process for gas separation. Chemical Engineering Science.
53: 181-196.
Yang, R.T. (1987). Gas Separation by Adsorption Process. London. Imperial
College Press.
Yang, M., Crittenden, B. D., Perera, S. P., Moueddeb, H. and Dalmon, J. A. (1999).
The hindering effect of adsorbed components on the permeation of non-adsorbing
component through a microporous silicate membrane: the potential barrier theory.
Journal of Membrane Science. 156: 1-9.
Yong, C., Fauda, A. E., and Matsura, T. (1989). Effect of drying condition on the
performance and quality of synthetic membranes used for gas separations.
Membrane Separations in Chemical Engineering. Volume 85. 18-33.
Zhau, W., Yoshino, M., Kita, H. and Okamoto, K. (2003). Preparation and gas
permeation properties of carbon molecular sieve membranes based on sulfonated
phenolic resin. Journal of Membrane Science. 217: 55-67.
Zoland, R. R. and Fleming, G.K. (1992). Definitions. In: Winston, W.S. and Sirkar,
K.R. Membrane Handbook. New York: Van Nostrand Reinhold. 19-24.
Zoland, R. R. and Fleming, G. K. (1992). Theory. In: Winston W.S and Sirkar, K.R.
Membrane Handbook. New York: Van Nostrand Reinhold. 25-53
Zoland, R. R. and Fleming, G. K. (1992). Design of Gas Permeation Systems. In:
Winston, W.S and Sirkar, K. R. Membrane Handbook. New York: Van Nostrand
Reinhold. 54-77.
100
Zoland, R. R. and Fleming, G. K. (1992). Applications. In: Winston, W.S and
Sirkar, K. R. Membrane Handbook. New York: Van Nostrand Reinhold. 78-94.
APPENDIX A
APPENDIX A.1: Permeability and Selectivity of Unmodified Membrane
Pressure
(bar)
1
2
3
Permeation
Permeability O2 Permeability N2
Selectivity
O
Temperature ( C)
27
599.8
607.3
0.987
55
604.4
609.5
0.991
100
629.0
618.3
1.017
27
916.8
942.7
0.972
55
937.3
963.6
0.972
100
998.9
982.8
1.016
27
1385.5
1307.6
1.06
55
1395.3
1326.6
1.051
100
1499.5
1456.5
1.029
102
APPENDIX A.2: Permeability and Selectivity of Hydrocarbon and Nitrogen
gas at 1 bar (single gas)
Hydrocarbon
gas
CH4
C2H6
C3H8
n-C4H10
Permeation
Hydrocarbon
Nitrogen
Temperature ( C)
Permeability
Permeability
27
626.22
248.5
2.52
55
521.16
250.56
2.08
100
517.85
258.5
2.00
27
606.5
248.5
2.44
55
463.53
250.56
1.85
100
369.65
258.5
1.43
27
585.6
248.5
2.36
55
551.32
250.56
2.20
100
465.3
258.5
1.80
27
405.85
248.5
1.63
55
263.08
250.56
1.05
100
245.57
258.5
0.95
0
Selectivity
103
APPENDIX A.3: Permeability and Selectivity of Hydrocarbon and Nitrogen
gas at 1 bar (binary gas)
Hydrocarbon
Permeation
Hydrocarbon
Nitrogen
gas
Temperature (OC)
Permeability
Permeability
27
208.74
63.25
3.3
55
158.45
69.32
2.28
100
102.6
73.04
1.40
27
167.31
11.62
14.39
55
118.5
20.45
5.79
100
65.5
29.5
2.22
27
115.95
4.45
26.05
55
101.5
8.43
12.04
100
60.5
15.43
3.92
C2H6
C3H8
n-C4H10
Selectivity
104
APPENDIX A.4: Sample of Gas Permeability Calculation
Pressure
(Bar)
1.0
Test
Gas
Permeability Permeability Permeability Average
1
2
3
Permeability Selectivity
(GPU)
O2
607.4
590.5
601.5
599.8
N2
605.5
610.8
605.6
607.3
0.987
The permeability and selectivity of unmodified ceramic membrane was
calculated as followed:
For nitrogen, N2, at pressure 1.0 bar (76 cmHg)
Diameter, D
= 0.8 cm
Length, L
= 7.8 cm
Quantity, n
= pieces
Volume, V
= cm3
Area of carbon membrane,
A = n( 2ʌjL )
= 19.6035 cm2
The calculation for permeation rate for oxygen gas,
p
l
Q
A'p
P1
100cm 3
1GPU .cm 2 s.cmHg
X
19.6035cm 2 (76cmHg )(110.5s ) 1X 10 6 cm 3 ( STP )
P1
607.4GPU
P2
590.5GPU
P3
601.5GPU
Paverage
P1 P2 P3
3
P = 599.8 GPU
APPENDIX B
2
1
7
5
6
4
3
Dimensions of permeation cell.
Reference
Length (cm)
Number
1
10.2
2
13.0
3
7.8
4
1.0
5
7.2
6
9.5
7
3.5
Appendix B.1: Dimensions of permeation cell
Aluminum Cell
Appendix B.2: Assembly component of permeation cell
Gasket
Stainless Steel End Cap
106
107
Figure B.3: Gas separation Apparatus