fabrication of carbon membranes for gas separation

8/12/2019 Fabrication of Carbon Membranes for Gas Separation

1/19

Review article

Fabrication of carbon membranes for gas separationa review

S.M. Saufi, A.F. Ismail *

Membrane Research Unit, Faculty of Chemical Engineering and Natural Resources Engineering, Universiti Teknologi Malaysia,

81310 UTM, Skudai, Johor Bahru, Malaysia

Received 14 April 2003; accepted 15 October 2003

Abstract

Carbon membrane materials are becoming more important in the new era of membrane technology for gas separation due to

their higher selectivity, permeability and stability in corrosive and high temperature operations. Carbon membranes can be pro-

duced by pyrolysis of a suitable polymeric precursor under controlled conditions. This paper reviews the fabrication aspects of

carbon membranes, which can be divided into six steps: precursor selection, polymeric membrane preparation, pretreatment of the

precursor, pyrolysis process, post-treatment of pyrolyzed membranes and module construction. The manipulation of the pre-

treatment variables, pyrolysis process parameters and post-treatment conditions were shown to provide an opportunity to enhance

the separation performance of carbon membranes in the future. By understanding the available methods, one can choose and

optimize the best technique during the fabrication of carbon membranes. Furthermore, areas of future potential in carbon mem-

brane research for gas separation were also briefly identified.

2003 Elsevier Ltd. All rights reserved.

Keywords: A. Porous carbon; Carbon precursor; B. Pyrolysis; Heat treatment

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242

2. General conditions for the fabrication of carbon membranes . . . . . . . . . . . . . . . . . . . . . . . . . . 242

3. Precursor selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242

3.1. Polyacrylonitrile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

3.2. Polyimide and derivatives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

3.3. Phenolic resin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245

3.4. Polyfurfuryl alcohol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246

4. Polymeric membrane preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246

5. Pretreatments of precursor membranes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

5.1. Oxidation pretreatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2475.2. Stretching. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248

5.3. Chemical treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248

6. Pyrolysis process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

7. Post-treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

7.1. Post-oxidation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

7.2. Chemical vapor deposition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

* Corresponding author. Tel.: +60-7-5535592; fax: +60-7-5581463.

E-mail address: [email protected] (A.F. Ismail).

0008-6223/$ - see front matter 2003 Elsevier Ltd. All rights reserved.

doi:10.1016/j.carbon.2003.10.022

Carbon 42 (2004) 241259

www.elsevier.com/locate/carbon
http://mail%20to:%[email protected]/http://mail%20to:%[email protected]/


2/19

7.3. Post-pyrolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

7.4. Coating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

8. Module construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

9. Future directions in carbon membrane research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253

Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

1. Introduction

The development of membrane processes for the

separation of gas mixtures has exhibited a remarkable

progress during the last two decades. Strong interest has

existed in the synthesis of gas separation membranes

based on materials that provide better selectivity, ther-mal stability and chemical stability than those that

already exist (i.e. polymeric membranes) [1,6]. Non-

polymeric and inorganic membranes have attracted

considerable attention in recent years in studies aimed

at improving the performance of membrane materials

for gas separation.

Among non-polymeric membranes, molecular sieving

materials [6] such as silica, zeolites and carbon have the

potential to push the upper boundary of the perme-

ability vs. selectivity tradeoff relationship. Carbon

molecular sieves showed attractive characteristic among

molecular sieving materials such as excellent shapeselectivity for planar molecule, high hydrophobicity,

heat resistance and high corrosion resistance [14]. In

addition, it is more feasible to form carbon molecular

sieve membranes (CMSM) [8,11].

Recently, Ismail and David [17] provided a review on

the latest developments in the field of carbon mem-

branes for gas separation. In the present paper, we re-

view the general fabrication aspects to produce carbon

membranes including pre- and post-treatment methods.

The manipulation of the pretreatment variables, pyro-

lysis process parameters and post-treatment conditions

were seen to generate opportunities to enhance the

separation performance of carbon membranes in thefuture.

2. General conditions for the fabrication of carbon

membranes

Producing a high performance carbon membrane is

not an easy task, since it involves many steps that must

be controlled and optimized. In general, the fabrication

of carbon membranes involves six important steps, as

shown in Fig. 1. At the same time, there are several

factors that need to be considered in order to ensure the

success of each step.

Among these steps, the pyrolysis process is the most

important step and can be regarded as the heart of the

carbon membrane production process. During this

stage, the pore structure of the carbon membrane is

formed, and this determines the ability of a carbon

membrane to separate gases [20]. In the next section,

various methods and parameters that are involved in

each step will be discussed and compared. The choice of

the most suitable method depends on the intended

application of the carbon membrane.

3. Precursor selection

The choice of the polymeric precursor is the first

important factor since pyrolysis of different precursors

may bring about different kinds of carbon membranes

[21]. Carbon membranes can be produced through the

carbonization or pyrolysis process of suitable carbon-

containing materials such as thermosetting resin,

graphite, coal, pitch and plants, under inert atmosphere

or vacuum [22]. Lately, numerous synthetic precursors

have been used to form carbon membranes, such as

polyimide and derivatives, polyacrylonitrile (PAN),

Step 1:

PrecursorSelection

HighPerformance

CarbonMembrane

Step 2:PolymericMembranePreparation

Step 6:Module

Construction

Step 3:

Pretreatment

Step 5:

Post Treatment

Step 4:Pyrolysis/

Carbonization

Fig. 1. Carbon membrane fabrication process.

242 S.M. Saufi, A.F. Ismail / Carbon 42 (2004) 241259


3/19

phenolic resin, polyfurfuryl alcohol (PFA), polyviny-

lidene chlorideacrylate terpolymer (PVDCAC), phe-

nol formaldehyde, cellulose and others.

A thermosetting polymer can often withstand high

temperatures [24] and neither liquefies nor softens dur-

ing any stage of pyrolysis. Suitable precursor materials

for carbon membrane production will not cause anypore-holes or cracks to appear after the pyrolysis step

[2]. Based on experience with activated carbon produc-

tion, polymer precursors are preferred, especially when a

carbon with few inorganic impurities is needed [25].

Therefore, primarily thermosetting polymeric precursors

have been used in previous research to manufacture

carbon membranes, as shown in Table 1. Detailed

descriptions for each precursor are given in the follow-

ing section.

3.1. Polyacrylonitrile

In the area of carbon fiber production, PAN fibers

have been recognized as the most important and

promising precursor for producing high performance

carbon fiber. It dominates nearly 90% of all worldwide

sales of carbon fibers [109]. There are numerous

advantages of PAN fibers, including a high degree of

molecular orientation, higher melting point (PAN fiber

tends to decompose below its melting point, Tm of 317

330 C) and a greater yield of the carbon fiber [3].

PAN fibers form a thermally stable, highly oriented

molecular structure when subjected to a low tempera-

ture heat treatment, which is not significantly disrupted

during the carbonization treatment at higher tempera-tures. This means that, the resulting carbon fibers have

good mechanical properties [3]. One of the main prob-

lems of unsupported carbon membrane (i.e. hollow

fiber) is brittleness. Therefore, by applying the PAN

which are widely used in the production of high strength

carbon fiber as carbon membranes, this problem can be

minimized [18].

Schindler and Maier [20] were the first researchers

involved in the preparation of PAN carbon membrane

based on their US patent published in 1990. They had

prepared porous capillary carbon membranes by nitro-

gen gas pyrolysis system at pyrolysis temperatures

ranging from 800 to 1600 C. Before preoxidation and

carbonization, the precursor membranes were subjected

to pretreatment with hydrazine solution. Later in 1992,

Yoneyama and Nishihiro [59] from Mitshubishi Rayon,

Japan patented their invention on the preparation of

porous hollow fiber carbon membranes.

During the period of 19941995, Linkov et al. did as

much effort in the research related to the PAN-based

carbon membranes. They had prepared PAN carbon

hollow fiber membranes by blending with methyl

methacrylate [30], polyethylene glycol (PEG) [31] and

polyvinylpyrrolidone (PVP) [31] in order to alter the

final pore size distribution of the carbon membranes.

They had also coated their PAN hollow fiber carbon

membranes with zeolites [50], polyimide [55] and per-

fluoro-sulfonated ionomer (PSI) [66] to produce com-

posite membranes. After this period, there was not much

research involving on the use of PAN as a precursor for

carbon membranes until late 2002, when David andIsmail [5] prepared their carbon hollow fiber membrane

from PAN precursor.

3.2. Polyimide and derivatives

Polyimides are among the most stable classes of

polymers. They can be used at temperatures up to 300

C and higher, and will usually decompose before

reaching their melting point. Because they do not go

through a melting phase transition and lose their shape,

polyimides are good precursors for glassy carbon [32].

Many researchers have used polyimides as their pre-

cursor for making carbon membranes (Table 1). Jones

and Koros [33] reported that the best carbon mem-

branes, in terms of both separation and mechanical

properties, were produced from the pyrolysis of aro-

matic polyimides. Moreover, Hatori et al. [38] found

that polyimide blended with PEG was also a good pre-

cursor for porous carbon membranes.

Polyimides are rigid, high-melting point, high glass

transition temperature (Tg), thermally stable polymers

formed by the condensation reactions of dianhydrides

with diamines (see Fig. 2) [34]. Normally, polyimide was

synthesized in lab scale with different type of dianhy-

drides and diamines in order to tailor its separationproperties when being used as membrane material as

shown in Fig. 2. A typical synthesis of polyimide

might involve separately suspending powder 3,30,4,40-

biphenyltetracarboxylic dianhydride (BPDA) and

4,40-oxydianiline (ODA) in a solvent such as N,N0-

dimethylacetamide (DMAC) and adding each suspen-

sion together drop wise under an inert environment to

produce a homogeneous polyamic acid solution [75].

Then, the polyamic acid is subjected to an imidization

process, which transforms it into polyimide polymer.

Further details about the preparation of polyimides can

be found in these papers [6,3537].

One of common commercial polyimide used for pre-

paring carbon membrane is Kapton (DuPont trade-

mark) polyimide. Hatori et al. [83] had successfully

prepared carbon molecular sieves film with homoge-

neous fine pores and without cracks or large pores, by

carbonizing Kapton film at 800 C. Kapton polyimide

was also successfully carbonized in the form of a capil-

lary membrane which can retain high gas permselectivity

even at high temperature [54]. The membrane exhibited

a permselectivity of 2000 for He/N2 at 0 C and 170 at

250 C. Recent study on Kapton polyimide-based car-

bon membrane showed that the permselectivities for

S.M. Saufi, A.F. Ismail / Carbon 42 (2004) 241259 243


4/19

Table 1

Representive examples of carbon membrane precursors and pyrolysis conditions

Precursor Configuration Pyrolysis condition (temperature;

heating rates; soak times; atmosphere)

Reference

Acrylonitrile Hollow fiber 8001600 C; ; 560 min; N2 [20]

Acrylonitrile Hollow fiber 6001200 C; ; 10 min; N2 [59,79]

PAN Hollow fiber 250 C800 C, 9 C/min; 10180 min; N2 [5,80]

PAN Hollow fiber 600 and 950 C; 1 C/min; ; N2 [30]PAN Hollow fiber 900 C; 5 C/min ; N2 [31]

Cellulose Hollow fiber or

supported film

400900 C; 110 C/min; ; Ar [2]

Cellulose Hollow fiber or

supported film

500800 C; ; 12 h; Ar [68]

Cellulose Hollow fiber 120400 C; 0.10.6 C/min; ; Ar [61,81]

Coal tar pitch Plate 600900 C; 1 C/min; ; N2 [22]

Condensed polynuclear aromatic

(COPNA)

Supported film 4001000 C; 5 C/min; 0 h; N2 [82]

Kapton and matrimid Supported film 450700 C; 0.5 C/min; 1 h; vacuum [40]

Kapton and matrimid Film 500800 C; 1 C/min or 4 C/min; 2h or

8 hr; vacuum

[41]

Kapton polyimide Film 800 C; 3 C/min; ; [8385]

Kapton polyimide Film 1000 C; 10 C/min; 2 h; vacuum [86,13]

Kapton polyimide Film 6001000 C; 10 C/min 2 h; vacuum or Ar [21]

Kapton polyimide Capillary 950 C; 5 C/min; 120 min; [54]

Phenol formaldehyde Flat 800950 C; ; 120180 min; N2 [88]

Phenol formaldehyde Tubular 800 C; 50 C/h; ; N2 [89]

Phenol formaldehyde Supported film 800C; ; ; vacuum [87]

Phenol formaldehyde Supported film 900C; 0.5 C/min; 60 min; Ar [19]

Phenol resin Supported film 700850 C; ; 3 h; N2 or CO2 [90]

Phenolic resin Supported film 900C; 25 C/h; 1 h; N2 [91]

Phenolic resin Supported film 5001000C; 5 C/min; ; vacuum [39]

Phenolic resin Supported film 700C; 0.5 C/min; ;vacuum [44]

Phenolic resin Supported film 700C; ; ; vacuum [70,92]

Phenolic resin Supported film 700C; ;; vacuum [93]

Phenolic resin Supported film 800C; 50 C/h; ; N2 [74]

Phenolic resin Supported film 700C; ; ; vacuum [63]

Poly (dimethylsilane) Supported film 300950 C; ; ; Ar [94]

Poly (para/phenylene pyromellitimide)

blend with PEG

Film 600 C; 3 C/min; 1 h; Ar [38]

PVDCAC Supported film 600 C; 1 C/min; 3 h; N2 [95]




PVDC-VC Supported film 5001000 C; 1 C/min; ; vacuum [99]

PEI Supported film 800 C; 0.5 C/min; 1 h; vacuum [16]

PEI Supported film 350 C; 1 C/min; 30 min; Ar and followed

by 600 C; 4 h

[26,100]

PFA Supported film 500800 C; 10 C/min; 2 h; He or N2 [43]

PFA Supported film 600 C; 5 C/min; 12 h; He [52]

PFA Supported film 300 C; 1.5 C/min; 2 h N2 and followed

by 500 C; 6h

[10]

PFA Supported film 150600C; 5 C/min; 02 h; He or N2 [101]

PFA Supported film 450 C; 5 C/min; 120 min; He [102]



PFA Supported film 600 C; 5 C/min; 12 h; He [103]


PFA Supported film 450600 C; 1 C/min; 60 min; [105]

Polyimide Supported film 550700 C; 0.5 C/min; ; vacuum [15]

Polyimide Supported film 550 C; 0.5 C/min; 1 h; vacuum [6]

Polyimide Supported film 800 C; 0.25 C/min; 2 h; vacuum [64]

Polyimide Supported film 500900 C; 5 C/min; 0 h; N2 [37,106]

Polyimide Supported film 700800 C; 5 C/min; 0 h; Ar [8,107]

Polyimide Supported film 600900 C; 5 C/min; 0 h; N2 [69]

Polyimide Supported film 500700 C; 5 C/min; 0 h; N2 [36]

Polyimide Hollow fiber 500550 C; 0.2513.3 C/min; 2 h; Vac uum [11,33,27]



5/19

O2/N2, CO2/CH4 and CO2/N2 are 4, 16 and 9, respec-

tively, at 25 C [40].

Another economical polyimide used to form carbon

membrane is polyetherimide (PEI), which is manufac-

tured by General Electric under trade name of Ultem

1000. This material has been used to prepare polymeric

membranes due to its superior strength and chemical

resistance [16,26]. An early work regarding the use of

PEI as a carbon membrane precursor was disclosed by

Fuertes and Centeno [16]. They had prepared supported

PEI carbon film by casting in only one step, which gave

the permselectivity values of 7.4, 121, 25 and 15 for O2/

N2, He/N2, CO2/CH4and CO2/N2, respectively, at 25 C

[16].

Sedigh et al. [26] dip-coated a PEI film inside a

ceramic tube and then carbonized it at 600 C under

flowing argon for 4 h. They had investigated the trans-

port characteristics of their PEI carbon membrane usingsingle gases (CH4, H2, CO2), binary mixtures (CO2/CH4,

H2/CH4, CO2/H2), and also ternary mixtures of CO2/H2/

CH4. CO2/CH4 separation factors as high as 145 for the

equimolar binary and 155 for the ternary mixture were

obtained with this PEI-based CMSM. More recently,

Countinho et al. [48] had prepared PEI carbon hollow

fiber membrane for use as a catalyst supports in a

membrane reactor. They had investigated pyrolysis

process parameters as well as stabilization parameters to

find an optimum condition for preparing PEI based

carbon membranes.

3.3. Phenolic resin

Phenolic resins are very popular and inexpensive

polymers, which are employed in a wide range of

applications, spanning from commodity construction to

high technology materials [39]. Phenolic resins present

suitable features to be applied as carbon membrane

precursors, such as:

(1) The pyrolysis of these materials provides carbon

films with molecular sieve properties [44].

(2) They provide a high carbon yield after carboniza-

tion [39].

(3) They are low cost polymers with numerous applica-

tions in different fields [44].

Fuertes and coworkers [39,44,63,70,92] investigated

intensively the preparation of carbon membranes from

phenolic resin precursors. They coated phenolic resin

film on a porous carbon disk by a spin coating tech-

nique, cured the supported film in air at 150 C for 2 h,

and then carried out a carbonization step under vac-

uum. The phenolic resin carbonized at 700 C showed

a O2/N2 selectivity around 10, and separation factors of

Table 1 (continued)

Precursor Configuration Pyrolysis condition (temperature;

heating rates; soak times; atmosphere)

Reference

Polyimide Hollow fiber 500550 C; 0.2513.3 C/min; 2 h;

vacuum and inert gas

[62,32]

Polyimide Hollow fiber 550 C and 800 C; 0.25 C/min; 2 h;

vacuum or He

[34,47]

Polyimide Hollow fiber 600900 C; 1 C/min; 1 h; N2 [65]Polyimide Hollow fiber 750 C; ; 60 min; vacuum [108]

Polyimide Hollow fiber 6001000 C; ; 3.6 min; N2 [57]

Polyimide Hollow fiber 500900 C; ; 0.5 s20 min; N2 [110]

Polyimide Hollow fiber 750 C; 2.6 C/min; 3 h; N2 or 850 C;

2.6 C/min; 3 h; vacuum

[35,111]

Polyimide Hollow fiber 500700 C; ; 5 C/min; N2 [112]

Polyimide Hollow fiber 500900 C; ; 0.130 min; N2 or He or Ar [58]

Polypyrrolone Film 300800 C; 58 C/min; 110 h; N2 [46]

Sulfonated phenolic resin Supported film 250800C; 5 C/min; 1.5 h; N2 [45]

() Data not available/not presented.

Fig. 2. Basic molecular structure of polyimide [36].



6/19

160 and 45 for CO2/CH4 and CO2/N2, respectively, in

gas mixture separation experiments [39]. The phenolic

resin was also used to prepare adsorption-selective car-

bon membranes (ASCM) [70], which can separate non-

adsorbable or weakly adsorbable gases (i.e. He, H2, air,

O2, CH4, CO2, etc.) from adsorbable gases such as

hydrocarbons (C2

), NH3, SO2, H2S, etc.Zhou et al. [45] introduced sulfonic acid groups in a

thermosetting phenolic resin to make a carbon mem-

brane precursor. The sulfonic acid groups evolve upon

heating as small molecular gases or fragments such as

sulfur dioxide and water, and leave void spaces in the

thermoset matrix during the pyrolysis step. The

decomposition of sulfonic acid groups takes place before

substantial carbonization takes place. A detailed

description of the preparation method was provided in

Zhou et al.s paper [45].

3.4. Polyfurfuryl alcohol

Polyfurfuryl alcohol (PFA) is an amorphous polymer

with a non-graphitizable structure, and potentially a

good precursor to prepare CMSM [29]. Because of its

simple molecular structure and formation mechanism, it

is an appropriate material for fundamental experimental

and atomistic simulation studies. On the other hand,

Acharya [43] chose PFA based on the knowledge that

PFA derived carbon could have desirable properties

such as a narrow pore size distribution and chemical

stability. However, PFA does not appear to have the

best mechanical and elastic properties required forforming a thin film on a rigid support [43]. Since PFA is

in liquid state at room temperature, it can be used in

supported films only [10].

Foley and coworkers [9,51,52,101104] used PFA

extensively as a precursor during the preparation of

nanoporous carbon (NPC) membrane for gas separa-

tion. An early work had uncovered the used of a spray-

coating technique in order to produce a thin layer of

NPC on the porous surface of a stainless steel disk

support [43,52]. A selectivity for O2/N2 separation of

up to 4 was achieved for the supported membrane

carbonized at 600 C in flowing helium [52]. Later, Foley

and coworkers [9,51,103] had improved their prepara-

tion technique by using an ultrasonic deposition sys-

tem which can distribute uniformly the polymeric

solution onto the support. They successfully prepared

reproducible NPC membranes which improved O2/

N2 selectivity within the range of 230, and provide

large separation factors for H2/CH4, CO2/CH4, N2O/

N2, and H2/CO2 of 600, 45, 17, and 14 respectively [9].

Other research studies involving the use of PFA pre-

cursor for carbon membrane synthesis was done by

Chen and Yang [10], Wang et al. [105], and Sedigh et al.

[29].

4. Polymeric membrane preparation

Polymeric membranes or precursors for fabricating

carbon membranes must be prepared at optimum con-

ditions in order to produce a carbon membrane of high

quality. In the case of poor quality polymeric mem-

branes, a pyrolysis process cannot be expected to satis-factorily produce superior carbon membrane properties.

Therefore, precursor membranes must be prepared in

defect free form in order to minimize problems in sub-

sequent processing during the manufacture of carbon

membranes.

A polymeric membrane can be produced in two main

configurations as a precursor for carbon membranes,

namely unsupported membranes and supported mem-

branes. Unsupported membranes have three different

configurations: flat (film), hollow fiber, and capillary.

Supported membranes can adopt two configurations:

flat and tube. Detailed descriptions of these two cate-

gories can be found in Ismail and Davids review [17]. In

most cases, supported polymeric membranes are pro-

duced, because of the poor mechanical stability (i.e.

brittleness) of unsupported carbon membranes.

For making the supported carbon membranes, vari-

ous options are available for coating the supports with

thin polymeric films, such as ultrasonic deposition

[9,51], dip coating [8], vapor deposition [105], spin

coating [16] and spray coating [52]. Ultrasonic deposi-

tion (UD) provides nearly zero spray velocities, droplet

sizes that are often narrowly distributed in size from 10

to 102 lm, accurate precursor delivery rates, and more

localized coating compared to spray and misting [9].Acharya in his thesis listed the requirements for a

method to be acceptable when coating with a polymeric

solution [43], namely:

(1) It should be able to control the amount of material

being deposited on the support;

(2) It should produce a uniform distribution of material

on the support;

(3) It should work for a number of different supports;

and

(4) It should not destroy or alter the support in any

way.

Because of the shrinkage of the polymer material

during pyrolysis, the coating procedure has to be re-

peated until a defect free carbon molecular sieve is ob-

tained [53]. Defects of the surface of the substrate may

be translated to the carbon membrane film, thereby

originating small pinholes that destroy the molecular

sieve properties required for gas separation. The coating

of the support surface with an intermediate layer re-

duces the number of defects existing on the original

substrate [15]. Moreover, the rapid coagulation of

polymer prevents the infiltration of solution into the



7/19

porous carbon substrate, and an excellent polymeric film

is achieved [16]. Chen and Yang [10] noted that further

work was needed to understand the effect of different

supports on the quality of the membrane produced. The

choice of a support material is based on:

Economics (i.e. cost, availability, morphology); Durability;

Heat transfer characteristics;

Chemical reactivity; and

Compatibility with the carbon.

For unsupported polymeric membranes (i.e. hollow

fibers), spinning parameters are crucial factors that must

be controlled during the preparation of the membranes.

These parameters include the amount and types of

polymer, solvents, and additives mixed into the spinning

dope solution, the dope and bore fluid flow rates, the

fiber take-up velocity, the air gap distance (unless wet

spinning is used), and the coagulant bath temperature

[32,42]. The main challenges in hollow fiber spinning,

as proposed by Puri [49], are:

Synthesizing polymeric melts and solutions having

appropriate thermodynamic and rheological proper-

ties.

Determining the kinetic and thermodynamic condi-

tions involved in transforming the solution/melt to

a solid.

Annealing and crystallizing (and orienting) the

formed bodies.

Post-treating to alter surface properties and stabilizethe products.

As a conclusion, during the preparation of precursor

membranes, the process parameters involved must be

optimized to form a good precursor for pyrolysis.

Moreover, the choices of supported or unsupported

carbon membranes depend on the application for which

the carbon membranes will be used. Normally, most of

todays gas separation membranes are formed into

hollow-fiber modules [56]. Hollow fiber membranes offer

a greater area per specific module volume than flat film

membranes, by a factor of 140 [5].

5. Pretreatments of precursor membranes

Polymeric membranes are often subjected to pre-

treatments before they undergo the pyrolysis process.

This step can ensure the stability of the polymeric pre-

cursor and the preservation of its structure during

pyrolysis. In fact, carbon membrane of good quality, in

terms of stability and separation performance, can be

produced using specific pretreatments for a given pre-

cursor.

Pretreatment methods can be divided into physical

and chemical methods. Physical pretreatments consist

of stretching or drawing hollow fiber membranes prior

to pyrolysis. In contrast, chemical pretreatments involve

some chemical reagents, which are applied to the poly-

meric precursor. Sometimes the precursor is subjected

to more than one pretreatment method to achieve thedesired properties in a carbon membrane. The most

important and popular pretreatment method employed

by previous researchers has been the oxidation treat-

ment. In fact, it can be stated that, in virtually every

fabrication process, a heat-assisted oxidation treat-

ment in oxygen, air or an oxidative atmosphere is in-

volved.

5.1. Oxidation pretreatment

Kusuki et al. [57] reported that certain precursors,

when not preoxidized, underwent softening during

the pyrolysis step, and the resulting carbon membrane

had a low membrane performance. The stabilization

or oxidation step is also intended to prevent the melting

or fusion of the fibers (or membranes), and to avoid

excessive volatilization of elemental carbon in the sub-

sequent carbonization step, and thereby maximizing

the final carbon yield from the precursor. Thus, the

oxidation treatment is considered very important

and can have a substantial influence on the resulting

performance of a carbon membrane. Oxidation pre-

treatments can be applied at very different ranges of

thermal soak times, depending largely on the precursoruses. In all cases, the aim is the same, that is, to con-

tribute to the stabilization of the asymmetric structure

of the precursor and provide sufficient dimensional

stability to withstand the high temperatures of the

pyrolysis steps. Table 2 lists selected oxidation pre-

treatments applied to different precursors by various

researchers.

Centeno and Fuertes [44] observed that air oxidation

of a phenolic resin film, prior to the carbonization step,

at temperatures ranging from 150 to 300 C (for 2 h)

improved the gas permeation rate of the resulting car-

bon membrane significantly. This observation was taken

to imply that the oxidation prior to carbonization leads

to an enlargement of carbon film pores [44]. However, in

the case of CMSM made from poly (vinylidene chloride-

co-vinyl chloride) (PVDC-VC), they found that oxida-

tive pretreatment in air at 200 C for 6 h produced a less

permeable membrane than the untreated membrane, but

a much more selective membrane (i.e. selectivity of O2/

N2 increases from 7.7 to 13.8) [99]. Therefore, it is

necessary to optimize the preoxidation conditions to

gain the best cost-effectiveness due to the permeability

and selectivity of carbon membrane being inversely re-

lated to each other [22].



8/19

5.2. Stretching

Stretching or drawing is categorized as a physical

pretreatment and usually involves the hollow fiber pre-

cursor only. This technique is adapted from the fabri-

cation of carbon fiber and is sometimes referred to as a

post-spinning treatment. An ideal post-spinning modi-

fication scheme would allow the removal of surface de-

fects, attenuation of variations in filament diameter, and

an enhanced retention of molecular orientation prior toheat treatment in order to obtain fibers with a good

balance of stiffness and strength [28].

Draw ratios are preferably as high as possible with-

in the range in which the structure of fiber is not rup-

tured. Typical upper limits of the total draw ratio

are around 80% of the draw ratio at which fiber break

occurs [59]. Drawing can also take place during the

spinning process to no more than a few times the ori-

ginal length of the fibers. This offers an advantage in

that the pore systems obtained after the spinning pro-

cess can be preserved to a larger extent in the carbon

membrane and thus ensure greater dimensional stability[20].

Stretching PAN precursor fibers in the preoxidation

stabilization stage results in an increase in the Youngs

modulus of the final carbon fiber. Since the dipoledi-

pole interactions among the nitrile groups obstruct the

molecular chains from becoming fully oriented during

stretching, a reduction of these interactions can make

the drawing process more effective. In particular, the

introduction of solvent molecules or heat has been cited

as an effective means of decreasing these interactions in

PAN molecules [60].

5.3. Chemical treatment

Pretreatment of a membrane with certain chemicals

can provide enhanced uniformity of the pore system

formed during pyrolysis. Among the chemicals com-

monly used for chemical pretreatment are hydrazine,

dimethylformamide (DMF), hydrochloric acid and

ammonium chloride [2].

During the manufacture of porous carbon mem-

branes by Schindler and Maier [20], an acrylic precursorwas subjected to a pretreatment with an aqueous solu-

tion of hydrazine. It had been found that this pretreat-

ment improved the dimensional stability of the

membrane during the subsequent processing steps and,

in particular, that the tar formation and clogging of the

pores could be prevented during these steps. Very good

results were obtained for acrylic using a pretreatment

with 80% hydrazine hydrate for 30 min at a solution

temperature of 90 C [20].

During the chemical pretreatment, the membrane is

fully immersed in the appropriate solution. After that,

the membrane is washed and dried before it is fed to thefirst heat treatment station. In certain cases, it has also

been proven that it may be advantageous to evacuate

the pores of the membrane by applying a low air pres-

sure, and subsequently fill them with nitrogen gas at

normal pressure prior to pretreatment with an aqueous

solution. This way, one can obtain membranes with

higher carbon content than the ones made from pre-

cursor whose pores were still filled with air during the

pretreatment [20].

Another type of chemical pretreatment involves using

catalysts such as mineral acids and acidic salts such as

Table 2

Examples of preoxidation conditions of selected polymeric precursors for carbon membranes

Precursor Configuration Temperature and time Reference

Acrylonitrile Hollow fiber 200300 C; 3 h; air [79]

Acrylonitrile Hollow fiber 180350 C; 120 min; air [20]

Polyacrylonitrile Hollow fiber 250 C; 30 min; air or O2 [5,80]

Polyacrylonitrile Hollow fiber 265 C; 30 min; air or N2 [30]

Polyacrylonitrile Hollow fiber 270 C; 30 min; air [31]Coal tar pitch Plate 200260 C; 120 min; O2 [22]

Phenol resin Supported film 300 C; 1 h; air [90]

Phenolic resin Supported film 150300C; 2 h; air [44]

Phenolic resin Supported film 150C; 2 h; air [39,63,70,92,93]

Poly(dimethylsilane) Supported film 200 C; air; 1 h [94]

PVDC-VC Supported film 150200 C; air; 6 h [99]

Polyfurfuryl alcohol Supported film 90 C; 3 h; air [10]

Polyimide Hollow fiber 400 C; 30 min; air [112]

Polyimide Hollow fiber 300C; 1 h; air [65]




Polyimide Hollow fiber 250495 C; 0.1100 h; air [58]


Polyimide Supported film 400500 C; 01 h; air [36]



9/19

phosphoric acid and diammonium hydrogen phosphates

before carbonization. However, the preparation of car-

bon hollow fiber membranes with carbonization cata-

lysts causes certain problems. In carbon hollow fibers,

carbonization must take place uniformly both inside and

outside of the fiber, and pitting must be avoided because

the selectivity of the membrane depends strongly on theuniformity of the pores produced during carbonization.

Pitting occurs immediately if the catalyst is not uni-

formly distributed throughout the fiber, due to locally

catalyzed oxidation on the surface [61].

Another method of catalytic carbonization involves

applying a gaseous catalyst such as hydrogen chloride

(HCl), or ammonium chloride (NH4Cl) in a stream of

inert gas. This method of application leads to two useful

results [31]:

(1) The catalyst can be more uniformly distributed

throughout the fiber bundle, thereby avoiding hot

spots; and

(2) The inert gas acts as a purging gas, removing tars

which are formed during pyrolysis and could other-

wise impair membrane properties by causing unde-

sirable occlusions therein.

However, all the water must be removed before

exposing the fiber to the gaseous catalyst in order to

prevent local dissolution of the fiber surface, which may

lead to the fusion or cementation of adjacent fibers. This

may in turn lead to the fracture of some of the cemented

fibers due to non-uniform contraction during pyrolysis,

and due to local pitting in the area attacked by thewater-soluble catalyst. Accordingly, it is critical to apply

the catalyst to the fibers only after the last traces of

water have been removed [61].

6. Pyrolysis process

Pyrolysis (sometimes referred to as carbonization) is a

process in which a suitable carbon precursor is heated in

a controlled atmosphere (vacuum or inert) to the pyro-

lysis temperature at a specific heating rate for a suffi-

ciently long thermal soak time. The pyrolysis process isconventionally used for the production of porous car-

bon fibers, and causes the product to have a certain

microporosity of molecular dimensions that is respon-

sible for the molecular sieve properties of the carbon

[20].

During pyrolysis of a polymer, byproducts of differ-

ent volatilities are produced and generally cause a large

weight loss [41]. Typical volatile byproducts may include

ammonia (NH3), hydrogen cyanide (HCN), methane

(CH4), hydrogen (H2), nitrogen (N2), carbon monoxide

(CO), carbon dioxide (CO2) and others, depending on

the polymer [62]. The polymer precursors are initially

cross-linked or become cross-linked during pyrolysis.

This prevents the formation of large graphite-like crys-

tals during carbonization, and leads to the formation of

disordered structures (non-graphiting carbons) with a

very narrow porosity [15]. Pyrolysis creates an amor-

phous carbon material that exhibits a distribution ofmicropore dimensions with only short-range order of

specific pore sizes, and also pores larger than the ultra-

micropores required to exhibit molecular sieving prop-

erties. These larger pores connect the ultramicropores

that perform the molecular sieving process, and there-

fore allow high gas separation productivities to be at-

tained [34].

Generally, the pore system of carbon membranes is

non-homogeneous, as it is comprised of relatively wide

openings with a few constrictions [23]. The pores vary in

size, shape and degree of connectivity depending on the

morphology of the organic precursor and the chemistry

of its pyrolysis. An idealized structure of a pore in a

carbon material is shown in Fig. 3. The pore mouth d

often referred to as an ultramicropore (


10/19

temperature of the carbonaceous precursor and its

graphitization temperature (generally about 3000 C).

Usually, pyrolysis is carried out in the range between

500 and about 1000 C [21,62]. Depending on the

experimental conditions chosen, the pyrolysis process

removes most of the heteroatoms that are originally

present in the polymeric macromolecules, while leavinga cross-linked and stiff carbon matrix behind [26]. Table

1 lists typical pyrolysis temperatures used by previous

researchers for the production of carbon membranes.

Pyrolysis temperature has a strong influence on the

carbon membranes properties in terms of the membrane

structure, separation performance (permeability and

selectivity) and the transport mechanism for gas sepa-

ration. The optimum temperature depends very much

on the type of the precursor used, and generally an in-

crease in the pyrolysis temperature will result in a de-

crease in gas permeability and an increase in selectivity

[21,34,62]. The increase in pyrolysis temperature will

also lead to a carbon membrane with higher compact-

ness, a more turbostratic structure, higher crystallinity

and density, and smaller average interplanar spacing

between the graphite layers of the carbon [4].

The thermal soak time can be different depending on

the final pyrolysis temperature [20]. This parameter may

be used to fine-tune the transport properties of a carbon

membrane using a particular final pyrolysis temperature

[80]. Previous studies [21,41,54,64,80] showed that

increments in thermal soak time would increase the

selectivity of carbon membranes. It is believed that only

microstructural rearrangement occurs during the ther-

mal soak time, thus affecting the pore size distributionand average porosity of carbon membranes [34].

Heating rates will determine the evolution rate of

volatile component from a polymeric membrane during

its pyrolysis, and consequently affect the formation of

pores in a carbon membranes [2]. Widely different

heating rates can be used, ranging from 1 to >10 C/min

[21]. However, lower heating rates are preferable in

order to produce small pores, to increase of carbon

crystallinity, and consequently to produce carbon

membranes of higher selectivity [21,54]. Higher heating

rates may lead to the formation of pinholes, microscopic

crack, blisters and distortions, which in extreme cases

may render membranes useless (i.e. of low selectivity)

for gas separation [2].

The pyrolysis atmosphere must be controlled in order

to prevent undesired burn off and chemical damage of

the membrane precursor during pyrolysis. Therefore,

the pyrolysis can be carried out either in vacuum or inert

atmosphere. Vacuum pyrolysis was reported to yield

more selective but less permeable carbon membranes

(from a polyimide precursor) than an inert gas pyrolysis

system [34,62]. When dealing with the inert gas pyrolysis

system, one must consider how the inert gas flow rate

that will affect the performance of the resulting carbon

membranes. Generally, an increase in gas flow rate will

improve the permeability of carbon membranes without

interfering with their selectivity very much [62,80].

In principle, one can determine which pyrolysis

parameters are important and contribute most signifi-

cantly to the structural changes of the material. In that

case, it would be possible to predict the trends oftransport properties for a given carbon material more

effectively. Recently, some researchers [26,34,43,64,65]

pyrolyzed their precursor in a step-by-step fashion (i.e.

after reaching a certain temperature, holding the mem-

brane at that temperature for a period of time and then

heating it again) until the final pyrolysis temperature is

reached.

7. Post-treatment

As a result of the pyrolysis process, polymeric mem-

branes are transformed into carbon membranes with

varying degrees of porosity, structure and separation

properties that depend to an extent on the carbonization

conditions employed. In most cases, it is found to be an

advantage that the pore dimensions and distribution in a

carbon membrane can be finely adjusted by a simple

thermochemical treatment to meet different separation

needs and objectives [22]. Therefore, various post

treatment methods have been applied to meet the de-

sired pore structure and separation properties of carbon

membranes, including post-oxidation, chemical vapor

deposition (CVD), post-pyrolysis and coating. At the

same time, these post-treatments can also repair thedefects and cracks that exist in the carbon membrane.

7.1. Post-oxidation

Post-oxidation or activation is the favorite post-

treatment used by researchers to alter the pore structure

of carbon membranes. Typically, when a membrane is

exposed to an oxidizing atmosphere after the pyrolysis

step, the ensuing oxidation increases the average pore

size [2,20,67,68].

The oxidation of carbon membranes can be per-

formed using pure oxygen, oxygen admixed with other

gases, air, or other oxidizing agents such as steam, car-

bon dioxide, nitrogen oxides and chlorine oxides or

solutions of oxidizing agents such nitric acid, mixtures

of nitric and sulfuric acids, chromic acid and peroxide

solutions at elevated temperatures [68].

Kusakabe et al. [69] showed in their study that the

post-oxidation of a polyimide carbon membrane in a

mixture of O2N2 increased its permeability to CO2 and

O2 without damaging its permselectivity greatly. Due to

the oxidation process, the micropore volume increased

significantly but the pore size distribution was not

broadened. Different activation temperatures and dwell



11/19

times have been applied to obtain desired pore struc-tures in different materials, as illustrated in Table 3.

As the oxidation temperature is raised, there is an

increase in gas permeance for all gases (permanent and

hydrocarbon) and a reduction in permselectivity of

permanent gas pairs such as O2/N2 or CO2/N2 [70].

However, hydrocarbons gases (i.e. n-C4H10, C2H4,

C3H6) exhibit higher permeances than permanent gases

(i.e. He, N2), which is a consequence of enhanced

transport via hydrocarbon adsorption in the micropores

[70]. It is believed that the activation process increases

the permeability by expelling carbon atoms from the

pore walls, thus enlarging them [67]. The variation ob-served in permselectivities suggests that for permanent

gases (i.e. O2, N2, etc.), the transport mechanism

through the membrane changes as oxidation progresses,

from a molecular sieving mechanism to surface diffusion

for highly oxidized samples [70].

7.2. Chemical vapor deposition

The selectivity of a carbon membrane may be in-

creased through the introduction of organic species into

the pore system of the carbon membrane and their

pyrolytic decomposition (i.e. chemical vapor deposition,

CVD) [37,67,68,71]. Generally, to manufacture carbon

molecular sieves, the inherent pore structure of the car-

bonaceous precursor is initially tailored into a suitable

pore size range by controlling the thermal pretreatment,

followed by a final adjustment of the pore apertures byCVD [72]. Moreover, it is possible to produce asym-

metric membranes by CVD, using homogeneous (sym-

metric) carbon membranes as the starting materials [67].

The organic species used for cracking and the con-

ditions of their pyrolysis should be carefully chosen to

produce sufficient and selective deposition of carbon at

the pore apertures [73]. Among the organic molecules

which may be used for this purpose are ethane, propane,

ethylene, benzene, and other hydrocarbons [68]. Verma

and Walker [72] selected propylene as the preferred or-

ganic source of pyrolytic carbon since its heat of

cracking (DH) is low, and the gas is easy to handle underambient temperature. A few additional organic sources

for CVD and their application conditions are summa-

rized in Table 4.

The CVD of carbon onto a carbon membrane may

bring about three distinct results, namely homogeneous

deposition, adlayer deposition and in-layer deposition as

depicted in Fig. 4(a)(c) [67]. The preferred mode is

homogeneous deposition but, in practice, the three

modes need not necessarily be completely distinct. For

instance, there may be an adlayer on top of the in-layer

or some deposition in depth in addition to the in-layer.

7.3. Post-pyrolysis

Aside from CVD, post-pyrolysis is another treatment

that can be used to decrease membrane pore size.

Table 3

Examples of post-oxidation treatments of selected carbon membranes

Precursor Configuration post-oxidation conditions Reference

Cellulose Hollow fiber or supported film 400 C; air; 15 min [2]

Phenol formalde hyde Film supported on stainle ss ste el plate 800 C; CO2; 60 min [87]

Phenol formaldehyde Flat 800950 C; 8090 ml/min

0.52.0% O2N2; 3060 min

[88]

Phenol formaldehyde Film supported on outer faceof tubular membrane

800 C; CO2; [89]

Phenolic resin Film supported on inner face

of ceramic tubular membrane

100475 C; 30 min6 h; air [70]



300400 C; 30 min; air [92]



75350 C; 30 min; air [93]

Phenolic resin Film supported on outer face

of Novalak resin tube

800 C; ; CO2 [74]

Polyimide Film supported on outer face

of porous a-alumina tube

300 C; 3 h; mixture O2N2 or O2 [69]


Table 4

Chemical vapor deposition (CVD) of carbon membrane

Precursor CVD conditions Reference

Cellulose 1-1-1-trichloroethane saturated in Ar; 620 C; 5 min [67]

Phenolic resin N2 saturated with trychloroethane (TCE); 500 C; 110 min [93]

Polyfurfuryl alcohol 20% propylene in He; 600C; 2 h [52]

Polyimide 0.05 mole fraction propylene in carrier gas; 650C; 1 h [8]



12/19

However, this treatment is rarely used, because the first

pyrolysis step at high temperature produces small pores

efficiently due to the shrinkage of the carbon structure.Typically, post-pyrolysis is applied after post-oxidation

in order to recover from an excessive pore enlargement.

Sometimes post-oxidation and post-pyrolysis are re-

peated several times until the desired pore size distri-

bution is achieved.

The process of sintering is one way to decrease in the

pore volume. Sintering involves pore collapse, which

proceeds more readily on the smaller pores. Heating the

carbon membranes in an inert or reducing atmosphere

to above 400 C (typically, 800 C) is one way to induce

pore sintering. The more facile collapse of smaller pores

is due to the fact that the surface energy of high cur-

vature (small pore size) surfaces is higher. Thus, the

contractive surface forces acting to diminish the pore

wall surface areas and to close the pores are greater for

smaller pores. Therefore, the smaller pores are the first

to become closed by sintering [67].

On the other hand, Menendez and Fuertes [63] ap-

plied post-pyrolysis treatments to recover an aged car-

bon membrane used in their research. They found that

the heat treatments, at low temperatures (120 C, under

vacuum) or at high temperatures (600 C, under vac-

uum), can only partially restore the original permeance

properties of the membrane.

7.4. Coating

Some researchers have also applied the coating

technique to repair defects in their carbon membranes.

Petersen et al. [54] treated a carbonized membrane by

coating a thin layer of polydimethylsiloxane (PDMS) to

prevent gas flow through defect structures. The mem-brane was immersed in a 2% solution of PDMS

(WACKER E41) in heptane, and then withdrawn to

harden the coated layer [54]. On the other hand, a

method entailing surface coating with an alcohol solu-

tion of 60% B rank phenolformaldehyde resin (PFR)

and dispersant, followed by recarbonization (700 C, 60

min), was used to prepare carboncarbon composite

membranes with improved separation performance by

Liang et al. [22]. On the carboncarbon composite

membrane coated with PFR alcohol solution and added

dispersant, the gas permeation rates decreased by a

factor of 10, and the ideal separation factors increased,

probably due to the PFRs dispersing effect [22].

Jones and Koros [27,76] coated their carbon mem-

brane to counteract the decrease in carbon membrane

performance when exposed to water vapor. They used

poly(4-methyl-1-pentene) (PMP) from Scientific Poly-

mer Products, and DuPonts Teflon AF 1600 and AF

2400, as coating agents. They managed to provide a

successful protection barrier, which significantly limited

the permeation of water vapor or other impurities such

as hydrocarbons without significantly inhibiting the

permeation of the faster fluid component or lowering its

selectivity. Accordingly, they have patented [27] com-

posite membranes that retain good fluid separationproperties and are resistant to the adverse effects on

membrane performance commonly observed in envi-

ronments having high humidity.

8. Module construction

To ensure the efficiency of a membrane process in a

given application, the geometry of the membrane and the

way it is installed in a suitable device (i.e. a module) is

also important [77]. The selection of a membrane module

is mainly determined by economic considerations. These

economic considerations must take into account all the

cost factors and not just the cost of the module. The

cheapest configuration is not always the best choice be-

cause the type of application, and thus the functionality

of a module, is also an important factor [79].

For commercial applications of membranes, it is

preferable to fabricate a module with an asymmetric

structure and capillary or hollow fiber configurations in

order to increase the rate of permeation of the products

[54]. In general, the characteristics of modules that must

be considered in all system designs are summarized in

Table 5.

Fig. 4. Mechanism of a carbon deposition on the pore system of

carbon membrane: (a) Homogeneous carbon deposition on membrane

pore walls; (b) in-layer carbon deposition on membrane pore wall

entrances; and (c) adlayer carbon deposition outside membrane pores

[67].



13/19

Many researchers have proposed a wide variety of

modules for carbon membrane production. The most

challenging task that must be considered while fabri-

cating these modules is improving the poor mechanical

stability of carbon membranes. This is more crucial in

hollow fiber carbon membranes, since they are self-

supporting. Selected examples of modules fabricated by

various researchers are shown in Figs. 5 and 6.

9. Future directions in carbon membrane research

There is a growing interest in the development of gas

separation membranes based on materials that provide

better selectivity, thermal stability and chemical stability

than those that already exist (i.e. polymeric membranes)

[6]. Carbon membranes offer the best candidates for thedevelopment of new membrane technologies, because of

their stability and molecular sieving capabilities. The

most notable advantages of carbon membranes have

been recently reviewed [17] in comparison to those of

polymeric membranes, in order to highlight the factors

that make carbon membranes very attractive and useful

as separation tools. Table 6 summarizes the perfor-

mance of carbon membranes for the separation of

mixtures of permanent gases as reported in the litera-

ture.

Up to now, the separation of propene/propane gas

mixtures has been investigated mainly by using poly-

meric membranes. However, their drawbacks include a

relatively lower permselectivity and the lack of thermal

and chemical stabilities of the polymeric membranes.

These drawbacks have led to further trials to develop

thermally and chemically stable non-polymeric mem-

branes that exhibit better separation performance, spe-

cifically for propene/propane separation. One of the

choices is carbon molecular sieve membranes [86].

Table 5

Important considerations for the preparation of a membrane module

[50,77,78]

Category Characteristic

Production High surface area (i.e. along with high

packing density and the ability to form

stable membranes)

The ease of formation of a given configurationand structure

Nature of polymer/membrane

Cost Low operating cost

Low investment cost

Maintenance Low fouling tendency

Ease to clean

Membrane replacement is possible

Efficiency The extent of use

The nature of the desired separation

Structural strength

Fig. 5. Module constructed by Ismail and coworkers [5,18], Chen [28] and David [80].

Fig. 6. Module constructed by Koros and coworkers [34].



14/19

Table 6

Performance of carbon membranes for the separation of mixtures of permanent gases

Precursor Pyrolysis P (O2) P (CO2) P (H2) a (O2/N2) a(CO2/N2) a (CO2/C

6-FDA/BPDA-DAM

550 C, 2 h 2530 GPU 7385

Matrimid 5218 550 C, 2 h 1113 GPU 6983

6FDA/BPDA-

TMPD

500 C, 120 min 2050 GPU 8.511.5 5556

550 C, 120 min 1440 GPU 1114 140190

BPDA- ODA/DAT 600 C, 0 min 180 Barrer 60

BPDA pp0-ODA 600 C 330 Barrer 30 80

700 C 160 Barrer 55 60

800 C 7.8 Barrer 30 100

BPDA-pp0 ODA 700 C, 0 min 9.7 47

CVD 14 73

BPDA-pp0 ODA 700 C, 0 min 90222 Barrer 1440

BPDA-pp0 ODA 700 C, 0 min 35

BPDA-pPDA 550700 C 1.6850.1 Barrer 2.5271.4 Barrer 1.314 1.821 1.4

BPDA-pPDA 550 C 18 Barrer 5.5 37.4

BTDA-TDI/MDI 700 C, 60 min 6 2229

Cellulose 800 C, Ar, 300 Barrer

600 C 98 Barrer 12

400 C 56 Barrer 11

Coal tar pitch 900 C

COPNA resin 4001000 C 3545

Kapton 800 C 1.2 Barrer 13 Barrer 4. 62 50

Kapton 950 C, 120 min 0.95 GPU

Kapton 600 C, 120 min 300 GPU 1200 GPU 1500 GPU 3 12

800 C, 120 min 60 GPU 200 GPU 790 GPU 6 20

Kapton 600 C, 2 h 383 Barrer 1820 Barrer 1600 Barrer 4.7 22.2 800 C, 2 h 34.8 Barrer 128 Barrer 669 Barrer 11.5 42.2

1000 C, 2 h 0.96 Barrer 4.15 Barrer 59.4 Barrer 23.4 101

Matrimid and

Kapton

475650 C 17 Barrer 212 Barrer 2.95.9 5.432.9

PAN 500 C, 10180

min

6.7671.66 GPu 1.283.74

PDMS 600 C 16.2 Barrer

PEI 600 C 91

PEI 700 C, 60 min 7.4 15 25

PFA 600 C 1.27.2 2.73.7


15/19

PFA 300 C, 120 min 0.064 Barrer 0.828 Barrer 13.5

450 C, 120 min 0.334 Barrer 3.63 Barrer 30.4

600 C, 120 min 0.762 Barrer 8.136 Barrer 2.80

PFA 150600 C 0.0681.6 Barrer 0.83245.4 Barrer 2.130.4

450 C, He 1.67 Barrer 10.5 Barrer 9.9

450 C, H2 55.26 Barrer 4.1

PFA 450 C, 120 min 1.4522.742 Barrer 9.320.34 2.26.4

PFA 600 C 4.655.07 Barrer 34.9216.02 Barrer 15336.24 Barrer 10.612.7 9282

Phenol formalde-

hyde

800950 C,

120180 min

2300 Barrer 5100 Barrer 10.65

Phenol formalde-

hyde

900 C, 60 min 0.351 GPU 0.597 GPU 4.95 GPU 5.24 8.91 20.58

Phenolic resin 700 C, 1 coat 13.4 Barrer 46.1 Barrer 10.7 36.9 85.3

700 C, 3 coat 4.5 Barrer 16.1 Barrer 11.0 39.5 115.7

Phenolic resin 700 C 11.7 Barrer 10 45

Phenolic resin 700 C 11.4 38 120

Phenolic resin 700 C 13.2 39 101

Phenolic resin 700 C 1.59.2 2.439

Phenolic resin 700 C 22.2295.2 Barrer 127.2960 Barrer 3.39 1731 1367

Polyamic acid 750 C, 3 h 200

Polyamic acid 750 C, N2 189.6 Barrer 42

Polyimide 550 C, vacuum 2550 GPU 372473 GPU 7.49.0

550 C, Ar 1284 GPU 451713 GPU 2.86.1

550 C, He 73140 GPU 428676 GPU 4.76.1 550 C, CO2 75306 GPU 400654 GPU 2.66.1

800 C, vacuum 1.64.2 GPU 6.810.3

Polyimide 535 C 952 Barrer 4.6

550 C 239 Barrer 9.9

800 C 23 Barrer 12.3

Polyimide 700 C, 3.6 min 1000 GPU

850 C, 3.6 min 180 GPU

Polyimide 600 C 280 GPU 590 GPU 3.2 6.8

700 C 170 GPU 420 GPU 3.7 10

Polyimide 700 C, 3.6 min 420 GPU

Polyimide 500900 C 30800 GPU

Polypyrrolone 500 C, 1 h 674 Barrer 2970 Barrer 6580 Barrer 9.1 40 120

700 C, 1 h 68.1 Barrer 250 Barrer 1720 Barrer 11 39 180

800 C, 1 h 0.183 Barrer 0.178 Barrer 25.4 Barrer 7.8 7.6 7.1

PVDCPVC 700 C 14 25 60

Sulfonated

phenolic resin

500 C, 90 min 0.13240 GPU 0.29800 GPU 10.51950 GPU 1.85.2 427


16/19

Hayashi et al. [106,107] reported that a special carbon-

ized membrane is capable of recognizing size differences

between alkane and alkene molecules. It is also possible

to employ carbon membranes to separate isomers of

hydrocarbons into normal and branched fractions [7].

A carbon membrane reactor constitutes one of the

most promising applications of carbon membranes. Theperformance of a carbon membrane for gas separation

and for the dehydrogenation of cyclohexane to benzene

was examined by Itoh and Haraya [108]. They con-

cluded that the performance of their carbon membrane

reactor for dehydrogenation was fairly good compared

with that of a normal reactor, i.e., functioning at equi-

librium [108]. On the other hand, Lapkin [12] used a

macroporous phenolic resin carbon membrane as a

contactor for the hydration of propane in a catalytic

reactor. He found that the use of this porous contactor-

type reactor for this high-pressure catalytic reaction is

practicable.

However, the production of carbon membranes cur-

rently involves a very high cost. The cost of a carbon

membrane per unit of membrane area is reported to be

between one and three orders of magnitude greater than

that of a typical polymeric membrane [18]. Therefore,

carbon membranes must achieve a superior performance

in order to compensate for their higher cost. Optimiza-

tion of fabrication parameters during the pyrolysis

process is arguably the best way to achieve this goal.

Since there are a large number of parameters involved,

using computer programming to optimize the pyrolysis

process makes the task more accurate and satisfactory.

In addition, the effect of pretreatments and post-treatments during the membrane fabrication process

should also be examined. These steps provide a great

opportunity for tailoring the separation properties of a

given carbon membrane. As an example, post-oxidation

treatments can be very useful to tailor the properties of a

molecular sieving carbon membrane in order to make an

adsorption-selective carbon membrane, which is suitable

for the separation of hydrocarbon mixtures based on

adsorption differences. In fact, there are several pre-

treatment methods that can be used to control the uni-

formity of pore formation during pyrolysis. Hopefully,

through manipulation of the available pretreatment and

post-treatment methods, a stable carbon membrane

exhibiting high separation properties can be produced.

The most popular precursor currently used for man-

ufacturing carbon membranes is polyimide. Although

polyimide offers the best separation potential for carbon

membranes, its cost is too high and its commercial

availability is sometimes very limited. Therefore, in or-

der to reduce the overall cost and production time

during carbon membrane fabrication, an alternative

polymer must be selected. Our group and others have

proposed the use of polyacrylonitrile (PAN) as a carbon

membrane precursor, since PAN offers lots of advan-

tages (i.e. high carbon yield, mechanical and chemical

stability, ready availability, etc.). However, we are still

progressing towards increasing their separation proper-

ties. We believe that satisfactory results may be achieved

via optimization of the pyrolysis process and combina-

tion of suitable pretreatment and post treatment meth-

ods [18].Once, a carbon membrane of high performance is

produced, the effect of exposure to water vapor must be

considered. It has been reported that the selectivity of

a typical membrane decreases as the amount of sorbed

water increases [11]. Not many studies have taken into

account this phenomenon very seriously, though it be-

comes quite important when a carbon membrane is

commercialized. This is primarily due to the fact that

water vapor can be found in a large number of process

streams [11]. To date, only Jones and Koros [27,76] have

studied this phenomenon in detail, and they proposed

the use of composite membranes (coated with DuPonts

Teflon AF1600 and AF2400) as a solution to this

drawback. On the other hand, the humidity found in the

ambient atmosphere can also have an adverse effect on

carbon membrane performance. Therefore, the study of

storage conditions for carbon membranes is also an

important consideration for carbon membrane research

in the future.

Acknowledgements

The authors gratefully acknowledge the financialsupport under National Science Fellowship (NSF) from

the Ministry of Science, Technology and Environment

of Malaysia (MOSTE).

References

[1] Stern SA. Review: polymers for gas separations: the next decade.

J Membr Sci 1994;94:165.

[2] Soffer A, Rosen D, Saguee S, Koresh J. Carbon membranes. GB

patent 2207666, 1989.

[3] Donnet JB, Bansal RC. Carbon fiber. New York: Marcel

Dekker; 1984.[4] Tanihara N, Shimazaki H, Hirayama Y, Nakanishi S, Yoshinaga

T, Kusuki Y. Gas permeation properties of asymmetric carbon

hollow fiber membranes prepared from asymmetric hollow fiber.

J Membr Sci 1999;160:17986.

[5] David LIB, Ismail AF. Influence of the thermostabilization

process and soak time during pyrolysis process on the polyac-

rylonitrile carbon membranes for O2/N2separation. J Membr Sci

2003;213:28591.

[6] Fuertes AB, Centeno TA. Preparation of supported asymmetric

carbon molecular sieve membranes. J Membr Sci 1998;144:105

11.

[7] Soffer A, Gilron J, Cohen H. Separation of linear from branched

hydrocarbons using a carbon membrane. US patent 5914434,

1999.



17/19

[8] Hayashi J, Mizuta H, Yamamoto M, Kusakabe K, Morooka S.

Pore size control of carbonized BPTA-pp0ODA polyimide

membrane by chemical vapor deposition of carbon. J Membr

Sci 1997;124:24351.

[9] Shiflett MB, Foley HC. On the preparation of supported

nanoporous carbon membranes. J Membr Sci 2000;179:27582.

[10] Chen YD, Yang RT. Preparation of carbon molecular sieve

membrane and diffusion of binary mixtures in the membrane. Ind

Eng Chem Res 1994;33:314653.

[11] Jones CW, Koros WJ. Characterization of ultramicroporous

carbon membranes with humidified feeds. Ind Eng Chem Res

1995;34:15863.

[12] Lapkin A. Hydration of propene using a porous carbon

membrane contactor. Membr Tech 1999;116:59.

[13] Suda H, Haraya K. J Chem Soc Chem Commun 1995:117980.

[14] Kyotani T. Control pore in carbon. Carbon 2000;38:26986.

[15] Fuertes AB, Centeno TA. Preparation of supported carbon

molecular sieve membrane. Carbon 1999;37:67984.

[16] Fuertes AB, Centeno TA. Carbon molecular sieve membranes

from polyetherimide. Microporous Mesoporous Mater 1998;26:

236.

[17] Ismail AF, David LIB. A review on the latest development of

carbon membranes for gas separation. J Membr Sci 2001;193:118.

[18] Saufi SM, Ismail AF. Development and characterization of

polyacrylonitrile (PAN) based carbon hollow fiber membrane.

Songklanakarin J Sci Technol 2002;24(Suppl):84354.

[19] Wei W, Hu H, You L, Chen G. Preparation of carbon molecular

sieve membrane from phenolformaldehyde novolac resin. Car-

bon 2002;40:4657.

[20] Schindler E, Maier F. Manufacture of porous carbon mem-

branes.US patent 4919860, 1990.

[21] Suda H, Haraya K. Gas permeation through micropores of

carbon molecular sieve membranes derived from kapton polyi-

mide. J Phys Chem B 1997;101:398894.

[22] Liang C, Sha G, Guo S. Carbon membrane for gas separation

derived from coal tar pitch. Carbon 1999;37:13917.

[23] Koresh JE, Soffer A. Study of molecular sieve carbons, part 1pore structure, gradual pore opening and mechanism of molec-

ular sieving. J Chem Soc Faraday Trans I 1980;76:245771.

[24] Morthon-Jones DH. Polymer processing. London: Chapman

and Hall; 1984 [Chapter 2].

[25] Laszlo K, Bota A, Nagy LG. Comparative adsorption study on

carbons from polymer precursors. Carbon 2000;38:196576.

[26] Sedigh MG, Xu L, Tsotsis TT, Sahimi M. Transport and

morphological characteristics of polyetherimide-based carbon

molecular sieve membranes. Ind Eng Chem Res 1999;38:3367

80.

[27] Koros WJ, Jones CW. Composite carbon fluid membranes.

US patent 5288304, 1994.

[28] Chen JC. Modification of polyacrylonitrile (PAN) precursor

fiber via post-spinning plasticization and stretching. The Penn-

sylvania State University, PhD thesis, 1998.[29] Sedigh MG, Onstot WJ, Xu L, Peng WL, Tsotsis TT, Sahimi M.

Experiments and simulation of transport and separation of gas

mixtures in carbon molecular sieves membranes. J Phys Chem A

1998;102:85809.

[30] Linkov VM, Sanderson RD, Jacobs EP. Highly asymmetrical

carbon membranes. J Membr Sci 1994;95:939.

[31] Linkov VM, Sanderson RD, Jacobs EP. Carbon membranes

from precursors containing low-carbon residual polymers. Polym

Inter 1994;35:23942.

[32] Geiszler VC. Polyimide precursors for carbon membranes.

University of Texas, PhD thesis, 1997.

[33] Jones CW, Koros WJ. Carbon molecular sieve gas separation

membranes Ipreparation and characterization based on polyi-

mide precursors. Carbon 1994;32:141925.

[34] Vu DQ, Koros WJ, Miller SJ. High pressure CO2/CH4 separa-

tion using carbon molecular sieve hollow fiber membranes. Ind

Eng Chem Res 2002;41:36780.

[35] Ogawa M, Nakano Y. Gas permeation through carbonized

hollow fiber membranes prepared by gel modification of polya-

mic acid. J Membr Sci 1999;162:18998.

[36] Yamamoto M, Kusakabe K, Hayashi J, Morooka S. Carbon

molecular sieve membrane formed by oxidative carbonization of

a copolyimide film coated on a porous support tube. J Membr Sci

1997;133:195205.

[37] Hayashi J, Yamamoto M, Kusakabe K, Morooka S. Simulta-

neous improvement of permeance and permselectivity of 3,304,

40-biphenyltetracarboxylic dianhydride-4,40-oxydianiline polyi-

mide membrane by carbonization. Ind Eng Chem Res 1995;34:

436470.

[38] Hatori H, Kobayashi T, Hanzawa Y, Yamada Y, Iimura Y,

Kimura T, et al. Mesoporous carbon membranes from polyimide

blended with poly(ethyleneglycol). J Appl Polym Sci 2001;

79:83641.

[39] Centeno T, Fuertes AB. Supported carbon molecular sieve

membranes based on phenolic resin. J Membr Sci 1999;160:201

11.

[40] Fuertes AB, Nevskaia DM, Centeno TA. Carbon compositemembranes from matrimid and kapton polyimide. Microporous

Mesoporous Mater 1999;33:11525.

[41] Steel KM. Carbon membranes for challenging gas separations,

University of Texas, PhD thesis, 2000.

[42] Clausi DT, Koros WJ. Formation of defect-free polyimide

hollow fiber membranes for gas separations. J Membr Sci 2000;

167:7989.

[43] Acharya M. Engineering design and theoretical analysis of

nanoporous carbon membranes for gas separation, University of

Delaware, PhD thesis, 1999.

[44] Centeno TA, Fuertes AB. Carbon molecular sieve membranes

derived from a phenolic resin supported on porous ceramic

tubes. Sep Purif Tech 2001;25:37984.

[45] Zhou W, Yoshino M, Kita H, Okamoto K. Carbon molecular

sieve membranes derived from phenolic resin with a pendantsulfonic acid group. Ind Eng Chem Res 2001;40:48017.

[46] Kita H, Yoshino M, Tanaka K, Okamoto K. Gas permselectivity

of carbonized polypyrrolone membrane. J Chem Soc Chem

Commun 1997:10512.

[47] Koros WJ, Vu, DQ. High carbon filamentary membrane and

method making the same. US patent application 2002033096,

2002.

[48] Coutinho MB, Salim VMM, Borgees CP. Preparation of carbon

hollow fiber membranes by pyrolysis of polyetherimide. Carbon

2003;41:170714.

[49] Puri PS. Fabrication of hollow fiber gas separation membranes.

Gas Sep Purif 1990;4:2936.

[50] Smith SPJ, Linkov VM, Sanderson RD, Petrik LF, OConnor

CT, Keiser K. Preparation of hollow-fibre composite carbon

zeolite membranes. Micro Mat 1995;4:38590.[51] Shiflett MB, Foley HC. Ultrasonic deposition of high-selectivity

nanoporous carbon membranes. Science 1999;285:19025.

[52] Acharya M, Foley HC. Spray-coating of nanoporous carbon

membranes for air separation. J Membr Sci 1999;161:15.

[53] Linkov VM, Sanderson RD, Lapidus AL, Krylova AJ. Carbon

membrane-based catalysts for hydrogenation of CO. Catal Lett

1994;27:97101.

[54] Petersen J, Matsuda M, Haraya K. Capillary carbon molecular

sieve membranes derived from kapton for high temperature gas

separation. J Membr Sci 1997;131:8594.

[55] Linkov VM, Sanderson RD, Rychkov BA. Mater Lett 1994;20:

43.

[56] Baker RW. Future directions of membrane gas separation

technology. Ind Eng Chem Res 2002;41:1393411.



18/19

[57] Kusuki Y, Shimazaki H, Tanihara N, Nakanishi S, Yoshinaga T.

Gas permeation properties and characterization of asymmetric

carbon membranes prepared by pyrolyzing asymmetric polyi-

mide hollow fiber membrane. J Membr Sci 1997;134:24553.

[58] Tanihara N, Kusuki Y. Membrane and process for its prepara-

tion. EP patent 1034836, 2000.

[59] Yoneyama H, Nishihara Y. Porous hollow carbon fiber film and

method of manufacturing the same. EP patent 0394449, 1990.

[60] Gupta AK, Paliwal DK, Baja P. JMS-Rev Macromol Chem

Phys 1991;C31:1.

[61] Soffer A, Gilron J, Saguee S, Hed-Ofek R, Cohen H. Process for

the production of hollow carbon fiber membranes. European

patent 0671202, 1995.

[62] Geiszler VC, Koros WJ. Effect of polyimide pyrolysis conditions

on carbon molecular sieve membrane properties. Ind Eng Chem

Res 1996;35:29993003.

[63] Menendez I, Fuertes AB. Aging of carbon membranes under

different environments. Carbon 2001;39:73340.

[64] Ghosal AS, Koros WJ. Air separation properties of flat sheet

homogeneous pyrolytic carbon membranes. J Membr Sci

2000;174:17788.

[65] Barsema JN, van der Vegt NFA, Koops GH, Wessling M.

Carbon molecular sieve membranes prepared from porous fiberprecursor. J Membr Sci 2002;205:23946.

[66] Linkov VM, Bobrova LP, Tomifeev SV, Sanderson RD.

Composites membranes from perfluoro-sulfonated ionomer on

carbon and ceramic supports. Mater Lett 1995;24:14751.

[67] Soffer A, Azariah A, Amar A, Cohem H, Golub D, Saguee S, et

al. Method of improving the selectivity of carbon membranes by

chemical vapor deposition. US patent 5695818, 1997.

[68] Soffer A, Koresh J, Saggy S. Separation device. US patent

4685940, 1987.

[69] Kusakabe K, Yamamoto M, Morooka S. Gas permeation and

micropore structure of carbon molecular sieving membranes

modified by oxidation. J Membr Sci 1998;149:5967.

[70] Fuertes AB. Effect of air oxidation on gas separation properties

of adsorption-selective carbon membranes. Carbon 2001;39:

697706.[71] Cabrera AL, Zehner JE, Coe CG, Gaffney TR, Farris TS.

Preparation of carbon molecular sieves I. Two-step hydrocarbon

deposition with a single hydrocarbon. Carbon 1993;31:96976.

[72] Verma SK, Walker Jr PL. Preparation of carbon molecular

sieves by propylene pyrolysis over microporous carbons. Carbon

1992;30:82936.

[73] Chihara K, Suzuki M. Carbon 1979;17:339.

[74] Katsaros FK, Steriotis TA, Stubos AK, Mitropoulos A, Kanel-

lopoulos NK, Tennison S. High pressure gas permeability of

microporous carbon membranes. Microporous Mater 1997;8:

1716.

[75] Shiflett MB. Synthesis, characterization and application of

nanoporous carbon membranes, University of Delaware, PhD

thesis, 2002.

[76] Jones CW, Koros WJ. Carbon composite membranes:a solutionto adverse humidity effects. Ind Eng Chem Res 1995;34:1647.

[77] Strathmann H. Membrane processes for sustainable industrial

growth. Membr Tech 1999;113:911.

[78] Baker R. Future directions of membrane gas-separation tech-

nology. Membr Tech 2001;138:510.

[79] Yoneyama H, Nishihara Y. Carbon based porous hollow fiber

membrane and method for producing same. US patent 5089135,

1992.

[80] David LIB. Development of asymmetric polyacrylonitrile carbon

hollow fiber membrane for oxygen/nitrogen gas separation,

Universiti Teknologi Malaysia, MSc Thesis, 2001.

[81] Gilron J, Soffer A. Knudsen diffusion in microporous carbon

membranes with molecular sieving character. J Membr Sci 2002;

209:33952.

[82] Kusakabe K, Gohgi S, Morooka S. Carbon molecular sieving

membranes derived from condensed polynuclear aromatic (COP-

NA) resins for gas separation. Ind Eng Chem Res 1998; 37:42626.

[83] Hatori H, Shiraishi M, Nakata H, Yoshitomi S. Carbon

molecular sieve films from polyimide. Carbon 1992;30:71920.

[84] Hatori H, Yamada Y, Shiraishi M. Preparation of macroporous

carbon films from polyimide by phase inversion method. Carbon

1992;30:3034.

[85] Hatori H, Shiraishi M, Nakata H, Yoshitomi S. Carbon

molecular sieve films from polyimide. Carbon 1992;30:3056.

[86] Suda H, Haraya K. Alkene/alkane permselectivities of a carbon

molecular sieve membrane. J Chem Soc Chem Commun 1997;

1:934.

[87] Katsaros FK, Steriotis TA, Stefanopoulos KL, Kanellopoulos

NK, Mitropoulos ACh, Meissner M, et al. Neutron diffraction

study of adsorbed CO2 on a carbon membrane. Physica B 2000;

276278:9012.

[88] Wang S, Zeng M, Wang Z. Asymmetric molecular sieve carbon

membranes. J Membr Sci 1996;109:26770.

[89] Steriotis TA, Beltsios K, Mitropoulos ACh, Kanellopoulos N,

Tennison S, Wiedenman A, et al. On the structure of an

asymmetric carbon membrane with a novolac resin precursor.

J Appl Polym Sci 1997;64:232345.[90] Sakata Y, Muto A, Uddin MDA, Suga H. Preparation of porous

carbon membrane plates for pervaporation separation applica-

tions. Sep Purif Tech 1999;17:97100.

[91] Clint JH, Lear AM, Oliver LF, Tennison SR. Membranes. EP

patent 0474424, 1992.

[92] Fuertes AB. Adsorption-selective carbon membrane for gas

separation. J Membr Sci 2000;177:916.

[93] Fuertes AB, Menendez I. Separation of hydrocarbon gas

mixtures using phenolic resin-based carbon membranes. Sep

Purif Tech 2002;28:2941.

[94] Lee LL, Tsai DS. Synthesis and permeation properties of silicon

carbon-based inorganic membrane for gas separation. Ind Eng

Chem Res 2001;40:6126.

[95] Rao MB, Sircar S. Nanoporous carbon membranes for separa-

tion of gas mixtures by selective surface flow. J Membr Sci1993;85:25364.

[96] Rao MB, Sircar S, Golden TC. Gas separation by adsorbent

membranes. US patent 5104425, 1992.

[97] Rao MB, Sircar S. Performance and pore characterization of

nanoporous carbon membrane for gas separation. J Membr Sci

1996;110:10918.

[98] Thaeron C, Parrillo DJ, Sircar S, Clarke PF, Paranjape M,

Pruden BB. Separation of hydrogen sulfide-methane mixtures by

selective surface flow membrane. Sep Purif Tech 1999;15:1219.

[99] Centeno TA, Fuertes AB. Carbon molecular sieve gas separation

membranes based on poly(vinylidene chloride-co-vinyl chloride).

Carbon 2000;38:106773.

[100] Sedigh MG, Jahangiri M, Liu PKT, Sahimi M, Tsotsis TT.

Structural characterization of polyetherimide-based carbon

molecular sieve membranes. AIChE J 2000;46:224555.[101] Corbin DR, Foley HC, Shiflett MB. Mixed matrix nanoporous

carbon membranes. WO patent 01/97956, 2001.

[102] Shiflett MB, Foley HC. Reproducible production of naono-

porous carbon membranes. Carbon 2001;39:142146.

[103] Strano MS, Foley HC. Synthesis and characterization of catalytic

nanoporous carbon membranes. AIChE J 2001;47:6678.

[104] Strano MS, Foley HC. Temperature- and pressure-dependent

transient analysis of single component permeation through

nanoporous carbon membranes. Carbon 2002;40:102941.

[105] Wang H, Zhang L, Gavalas GR. Preparation of supported

carbon membranes from furfuryl alcohol by vapor deposition

polymerization. J Membr Sci 2000;177:2531.

[106] Hayashi J, Yamamoto M, Kusakabe K, Morooka S. Effect of

oxidation on gas permeation of carbon molecular sieve mem-


fabrication of carbon membranes for gas separation

Documents