You are on page 1of 29

Seperation of Olefins/Paraffins using

Polymeric Membranes







Olefins are the most important building blocks of the petrochemical industry. Most olefins
are produced by steam cracking of ethane, propane, naphtha, or gas oil followed by repeated
compression and distillation to separate the complex vapour mixtures. The second step is
considered to be a highly energy-intensive process and have been extensively studied for
possible replacements. Production is highly energy intensive. Current distillation processes
for the separation of ethylene/ethane and propylene/propane mixtureswhich account for
most of the energy consumption and much of the capital cost of the steam crackerconsume
an estimated 120 trillion BTU energy in the US.

Firstly, the new flexible and high performance gas separation membranes were
fabricated by grafting various sizes of cyclodextrin to the cross-linkable co-polyimide
(6FDA-Durene/DABA (9/1)) matrix and then decomposing them at elevated
temperatures. The gas permeability of thermally treated pristine polyimide (referred as
the original PI) and CD grafted co-polyimide (referred as PI-g-CDs for 200 and 300
C and partially pyrolyzed membranes (PPM) - CDs for 350, 400, and 425 C) has
been determined. It was observed that permeability of all tested gases increased with
an increase in thermal treatment temperature from 200 to 425 C. However,
permeability increased more for those grafted with bigger size CD. The permeability
of the original PI thermally treated at 425 C was about 4-6 times higher than that
treated at 200 C. The permeability increase jumped to 8-10 times for PPM--CD and
15-17 times for PPM--CD due to CD decompose at high temperatures and bigger CD
creates bigger micro-pores. Interestingly, the permeability ratios of PPM--CD to
PPM--CD and PPM--CD to PPM--CD at 400 and 425 C were around 0.6 and 0.8,
respectively. These numbers were almost the same as the cavity diameter ratios of -
CD to -CD and -CD to -CD. Permselectivity decreased first with an increase in
thermal treatment temperature up to 350 C
and then increased. Permselectivity of thermally treated CD grafted co-polyimide
membranes were also slightly higher than that of the original PI due to higher
degrees of cross-linking in CD grafted co-polyimide membranes. In addition, for
co-polyimide membranes grafted by CDs, the higher thermal treatment
temperature resulted in membranes with the better plasticization resistance to CO2
and the better separation performance in 50:50 CO2/CH4 mixed gases.

1. Gas separations have always been one of the key processes in the field of chemical engineering. With
industrys demand on lowering operating costs and increasing separation efficiency, more research is
being conducted on process improvements. Gas separation is usually achieved by physical or
physicochemical phenomena. Over the past two decades, gas separation using polymeric membranes
has drawn a great deal of interest from researchers due to many advantages such as low energy costs
and high selectivities. This is especially true for hydrocarbon separations performed by the
petrochemical industry. In particular, olefin/paraffin separations incur a heavy cost to petrochemical
companies. With a growing awareness of the importance of conserving natural resources, companies
are enthusiastic about finding ways to reduce energy consumption and to recycle purge or waste
streams. Traditionally, cryogenic distillation at elevated pressures in trayed fractionators is used to
separate olefins and paraffins. This distillation system is expensive to build and operate, and is currently
only economically attractive for streams containing high quality of olefins. Other available separation
technologies include extractive distillation, physical or chemical adsorption, physical or chemical
absorption, and more recently, membrane separation. A more thorough description of 1 CHAPTER 1.
INTRODUCTION 2 different separation technologies can be found in the review written by Eldridge
The difficulty of the olefin/paraffin separation step stems from the low relative volatilities of the
components. As a result, very large distillation towers with 120-180 trays and high reflux ratios are
required. Even small improvements in these separations could result in significant energy and cost
savings to the petrochemical industry. The use of membranes has been the focal point of research as
a new means of olefin/paraffin separation due to its lower operating and capital costs. In this
article, a comprehensive review is presented for the application of polymeric membranes in light
olefin/paraffin separation. It covers all types of membrane materials starting from the
conventional polymers to more advanced ones such as polyimide and copolyimide membranes.
Permeation and separation characteristics of all polymeric membranes are summarized and
discussed in great details.

Membrane separation, compared to amine absorption, pressure swing adsorption, rectification and
cryogenic distillation, has advantages, including low energy consumption, easy operation and
maintenance, environmental benign and small footprint [19-21]. Many studies have been done on 3
membrane separation for the purpose of replacing traditional separation technologies [8]. Membranes,
especially polymeric membranes have been explored in various kinds of gas separation applications such
as natural gas sweetening [7-9,20-21], and olefin/paraffin separation [22-29]. In order to have a good
gas separation performance, membranes must have high permeability and permselectivity, excellent
chemical resistance (for resistance against corrosive materials such as H2S), great thermal stability (for
high temperature applications), good mechanical properties (for high pressure applications), and
superior plasticization resistance
Two of the most important petrochemicals are the olefins ethylene and propylene. In 2004, 146 and 82
billion pounds of ethylene and propylene, respectively, were produced worldwide [35]. Both are feed
stocks for many other important chemical products; the most important being polyethylene and
polypropylene. Worldwide, 72 billion pounds of polyethylene and 42 billion pounds of polypropylene
were produced in 2004. Propylene and propane are found in the low and middle fractions of the
distillation process, albeit in small amounts. Large amounts are formed when gasoline is made by
cracking or reforming. Other olefins and paraffins can be found throughout the refining process. The
separation of these two has been relatively more difficult than that of other gas pairs, mainly ascribed to
their close thermodynamic and physical properties. The simulated molecular dimension of propylene
and propane is presented in Figure 1.3 while the thermodynamic properties of both are tabulated in
Table 1.1. Table 1.1: Thermodynamic properties of propane and propylene [23] Molecular Formula
Molecular Weight (g/mol) Boiling Point (K) Critical Temp. (K) Propylene C3H6 42 226 365.2 Propane
C3H8 44 230.4 369.9 7 Figure 1.3: Simulated molecular dimension of propylene and propane [23] As can
be seen, both gases have similar molecular dimensions, boiling points and critical temperatures. The
only difference is that propane is saturated hydrocarbon and propylene is unsaturated hydrocarbon.
Compounds of carbon and hydrogen whose adjacent carbon atoms contain only one carbon-carbon
covalent bond are known as saturated hydrocarbons. They are called saturated compounds because all
the four bonds of carbon are fully utilized and no more hydrogen or other atoms can attach to it. These
saturated hydrocarbons are called alkanes and the general formula for an alkane is CnH2n+2. Due to the
presence of all single covalent bonds, these compounds are less reactive and the number of hydrogen
atoms is more when compared to its corresponding unsaturated hydrocarbon. Compounds contain
single carbon-Compounds of carbon and hydrogen that contain one double bond between carbon atoms
(carbon=carbon) or a triple bond between carbon atoms (carboncarbon) are called unsaturated
hydrocarbons. Unsaturated hydrocarbons can be divided into alkenes and alkynes depending on the
presence of double or triple bonds respectively. The general formulae are CnH2n for alkenes and CnH2n-
2 for alkynes. These compounds are more reactive and their high reactivity is due to the presence of pi-
bond in their structure. Then sorption and adsorption of these unsaturated hydrocarbon (alkene) in
polymer matrix or inorganic particles could be higher than saturated hydrocaron (alkan). Although
membranes can lead to extensive energy and cost savings, many researchers envision membrane
separation units fitted in line with current separation systems rather than completely replacing them.
Propylene/Propane Separations

Although there is an increasing demand for on-purpose propylene production,

steam cracking and FCC processes are still the main sources of propylene till date.

Unfortunately none of these processes produces pure enough propylene that can be used

directly. To achieve the desired grade of propylene, additional separation/purification steps

are required. Currently, the separation of olefin and paraffin components is performed by

cryogenic distillation, which is expensive and energy intensive due to the low relative

volatilities of the components (b.p. of propylene is -470C and propane is -42.10C). The

necessary columns are normally 300 feet tall and contain over 200 trays

Due to the height of the column, two splitters are usually required instead of one.

Around 1.2 x 1014 BTU/year energy is used for olefin/paraffin separations. The bottom

product of C3 splitter tower is propane rich, while the overhead product is high purity

propylene. A typical propylene recovery unit from the co-products of FCC unit is

shown in Figure-1.3.

Propylene recovery unit

Due to the cost and complexity of distillation, membrane separation is gradually

gaining popularity for propylene/propane separations. Baker predicts that the membrane

market will reach $30 million by 2010 and $125 by 2020 for vapor/vapor gas separations

In the case of propylene/propane separations, it is unlikely that a membrane unit will

completely replace a C3 splitter because of the purity limitations, but it can obviously be used to

debottleneck a process and increase the purity of the feed-stream of a C3 splitter.

Another attractive application of membrane for propylene/propane separation is

the use of a membrane unit in a polypropylene production plant. In a polypropylene

reactor, propane enters as an impurity (99% propylene and 1% propane) and eventually

builds up in the reactor. Propane buildup is controlled by continuously removing a

recycle stream which is then flared. With every mol of propane purged, 2-3 mol of

propylene monomer is lost incurring a $1 million loss per plant every year [6]. A

membrane unit installed before sending propane to flare is an attractive separation

option. Figure-1and Figure-2 show these two possible membrane applications for

propylene/propane separations.

Membrane unit installed as a hybrid process for propylene/propane separations

Figure-2: Use of membrane unit to recover propylene in a polypropylene production

Studies have shown that several polymeric membranes offer promising potential, but separation
selectivity needs to be improved to be used as a cost effective separating membrane.
Considerable research has been done in glassy polyimides, and polyimide-co-polypyrrolones
with polyimides based on 6FDA (4,4-(hexafluorisopropylidene)dipthalic anhydride) exhibiting
the best performance . On the other hand, carbon molecular sieve and zeolite membranes
potentially offer superior selectivities; however, they are brittle and too costly (> $150/sq foot)
to be commercially useful for large scale applicat Figure-2: Use of membrane unit to recover
propylene in a polypropylene production plant
In this chapter, the materials that were studied, the membrane fabrication
techniques, characterization techniques, and the equipment used to
characterize these materials will be detailed. The specific materials selected
for this work are described in Section 3.1. Membrane formation techniques are
described in section 3.2 and material characterization techniques and
equipment are outlined in Section 3.3.



6FDA-Durene polyimide and 6FDA-Durene/DABA co-polyimide

The polymers selected for this work was a 6FDA-Durene polyimide and
6FDA-Durene/DABA cross-linkable co-polyimide. Figure 3.1 shows the
chemical structure of the monomers that was used in this study. 4,4'-
(hexafluoroisopropylidene) diphthalic anhydride (6FDA), supplied by
Clariant, and 3,5-diaminobenzoic acid (DABA) supplied by Aldrich
(Singapore) were purified by vacuum sublimation before usage. 2,3,5,6-
Tetramethyl-1,4-phenylenediamine (Durene diamine) supplied by Aldrich
(Singapore), was recrystallized two times from methanol. 1-methyl-2-
pyrrolidone (NMP) purchased from Merck Chemicals (Germany), was
distilled at 42 C under 1 mbar and was dried with molecular sieve before use.
Acetic anhydride was received from Aldrich (Singapore) and dried with
molecular sieve before usage. Other chemicals and solvents including
triethylamine, methanol, and p-toluenesulfonic acid were all reagent grades or
better from Aldrich (Singapore) and were used without further purification.

The 6FDA-Durene polyimide, 6FDA-Durene/DABA (9/1) co-polyimide, and

6FDA-Durene/DABA (7/3) co-polyimide were synthesized via a two-step
chemical imidization in an NMP solution. For the preparation of polyamic
acid, a designated molar ratio of Durene and DABA diamine was dissolved in
NMP under a nitrogen atmosphere and then an equimolar of solid 6FDA was
gradually added to the solution in a moisture free flask with a magnetic
stirring under a nitrogen atmosphere at room temperature. The reason of
adding solid 6FDA gradually is to suppress the reaction between water and
6FDA. Aromatic dianhydrides can react with water and other impurities in the
amide solvents, but have a slower reaction rate relative to their reacting with
diamines. Therefore, gradually adding the solid 6FDA reduces its availability
for competing reactions with water or other impurities, thus leads to a higher
conversion and molecular weight After reacted for 24 h, a high molecular
weight polyamic acid was formed. Then for the chemical imidization step, a
mixture of acetic anhydride and triethylamine at 4:1 molar ratio was slowly
added to the polyamic acid solution to perform imidization for 24 h under
nitrogen atmosphere. Finally, the co-polyimide was precipitated in methanol,
filtered, washed and dried at 120 C in vacuum for 24 h. The schemes of co-
polyimide synthesis and chemical structure of this co-polyimide are shown in
6FDA-Durene/DABA (9/1) co-polyimide grafted with Cyclodextrin

In this work, three Cyclodextrins (CDs) were grafted on the 6FDA-

Durene/DABA (9/1) co-polyimide by esterification .The three CDs, namely,
-, -, and -CD, formed by six to eight glucopyranoside units, were all
reagent grade or better from Aldrich (Singapore) and were used without
further purification. Their chemical structure and basic properties have been
presented in Figure 3 and Table 3.1 [3]. For this purpose, co-polyimide and ,
, and -CD were first dried at 120 C in vacuum for 12 h. After that co-
polyimide was dissolved in NMP and then a large excess amount of CD (3-5
times more than stoichiometric balance to the carboxylic acid group) was
added to the NMP solution. The esterification was carried out by adding a
catalytic amount (5 mg per gram of polymer) of p-toluenesulfonic acid and
heating to 120 C under nitrogen atmosphere for 18 h. Finally, the resultant
co-polyimide grafted with , , or -CD (referred as PI-g--CD, PI-g--CD, or
PI-g--CD) was precipitated in methanol, washed to remove unreacted CD and
dried under 120 C in vacuum for 24 h. The chemical structure of co-
polyimide after grafting CD is shown in Figure 3.1.

Figure 3.1: The synthetic scheme and chemical structure of co-polyimide

Figure 3.2: Chemical structure of three kinds of Cyclodextrin [3]

CD Type No. of Mw Cavity Cavity Decomposition

Glucose (g/mol) [3] Diameter Height Temp. (C) a

unit () [3] () [3]

-CD 6 972 4.7-5.7 7.8 3211

-CD 7 1135 6.0-7.8 7.8 3281
-CD 8 1297 7.5-9.5 7.8 3221
Td (5%), at which the mass of the sample is 5.0% less than its mass measured at 50 C.
3. Zeolitic imidazolate frameworks (ZIFs) nanoparticle synthesis

The nanoparticle used in this work to make mixed matrix membranes was zeolitic
imidazolate framework-8 (ZIF-8). The dispersed ZIF-8 nanoparticles were
synthesized at room temperature. In summary, a solution of Zn(NO3)26H2O
(7.332 g, 24.67 mmol) in 500 ml methanol was rapidly poured into a solution of
Hmim (16.222 g, 197.6 mmol) in 500 ml methanol under stirring with a magnetic
bar. The resultant solution slowly turned turbid and after 1 h the nano-crystals
were separated from the milky dispersion by centrifugation and washing with
fresh methanol. After washing and the 2nd centrifugation, the particles were re-
dispersed in fresh NMP before use. The yield of ZIF-8 was ~ 50% based on Zinc

3 Gases

All pure gases and gas mixtures used throughout this research (i.e., CO2, CH4,
C3H6, and C3H8) were research grade (99.999%) and supplied by SOXAL

. Selection Of 6FDA-6FpDA Polyimide For C3H6/C3H8 Separation

6FDA-6FpDA is another polyimide for this particular gas pair with very high

permeability and selectivity. As shown in Table-2.1, the permeability of propylene with

6FDA-6FpDA this polymer is 0.89 at 3.8 atm. The selectivity is 16 in this case. It should

be noted here that propylene pressure reported is 3.8 atm, instead of 2 atm like other

cases mentioned above. Due to dual mode pressure dependence, permeability at 2 atm

should be higher than 0.89 Barrer. Moreover, the permeation test for 6FDA-6FpDA was

done at 350C, instead of 500C like other cases. As mentioned in Chapter 2, high test
temperature also increases the permeability while the selectivity reduces. So effectively,

6FDA-6FpDA is one of the highest performing polymers for C3H6/C3/H8 separations.

Following the difficulties with the DDBT monomer, 6FDA-6FpDA polymer was

selected for the current research work. Another advantage of 6FDA-6FpDA polyimide is the

excellent permeability matching with AlPO-14. As shown later in Chapter-6, polymer and

molecular sieve matching is very important for mixed matrix material success. A very highly

permeable molecular sieve can reduce the overall permeability of the mixed matrix

membrane. This has been demonstrated and discussed in details in Chapter-6.

Although 6FDA-DAM is one of the highest performing polymers for this

particular separation, it is too permeable with respect to AlPO-14 propylene permeability.

Thus using 6FDA-DAM as polymer matrix would have essentially reduced the

permeability of mixed matrix materials. This will be discussed in more details in Chapter-

6 with mixed matrix material results.

.Initial synthesis of 6FDA-6FpDA polyimide

Figure-4.2 shows the structure of 6FDA-6FpDA. Two batches of this polymer

were synthesized by the thermal imidization process described in Chapter-3. Both batches

had low Mw (highest molecular weight observed was ~50k with a PI >5.0) determined by

GPC. A detailed study identified three major areas of difficulty during synthesis process.

First, our 6FDA monomer was 99% pure, whereas 6FpDA was 98% pure creating a

stoichiometric discrepancy in the molar weight ratio. High Mw weight can be achieved

with desired end-group controlled by Carothers Equation [6] which depends strictly on

the stoichiometric ratio.

Structure of 6FDA-6FpDA polyimide

In the case of impure monomers, even if the monomers are added in a 1:1

stoichiometric ratio, impurities disturb the stoichiometric ratio, which is one of the most

important parameters for high Mw and low PI. When the ratio is not actually 1:1, two

possible reactions can take place during the polyamic acid formation step. First, the

excess monomer can form di-mers or tri-mers and stop growing in length. These
relatively small molecules then contribute to the formation of a much broader

polydispersity, thereby resulting in a much higher PI [7]. Second, if there are slow

reacting impurities in the solution, there will be a competitive reaction between the

impurities and monomers where impurities can end the chain growth in a

polycondensation reaction. In this case, low molecular weight oligomers [7] will form

instead of high Mw polymer. In all of these cases, a low molecular weight polyamic

acid solution results in a low Mw polyimide.

Another important problem in the synthesis process is the presence of moisture.

Since 6FDA is an anhydride, which attracts atmospheric moisture even at room

temperature forming carboxylic acids. For the same reason, the number of available sites

for diamine addition for a given amount of anhydride decreases the probability of

creating the desirable repeat unit. This again results in a low molecular weight polymer.

Lastly, one important parameter during synthesis is the local hot zone formation during

6FDA addition. 6FDA is added in a diamine solution (in NMP). As soon as the

dianhydride is added, the polycondensation reaction starts. This reaction is exothermic

and increases the temperature locally. Consecutively, the reaction stops in that particular

region. It is desirable to have an ice bath during polyamic formation step of synthesis thus

reducing the reaction temperature.

Synthesis Of High Molecular Weight 6FDA-6FpDA

The following changes in the polyamic formation process were made to ensure

high molecular weight and low PI polyimide.

Before monomer sublimation, both monomers were dried in a vacuum oven for

overnight, around 200C below its sublimation point. (To increase the

sublimation yield)

Each monomer was sublimed twice to achieve required purity. (Yield increased

from 40% to 90%).

Before the synthesis experiment, sublimed monomers were again dried in the

vacuum oven overnight to eliminate residual water. (To remove any ambient


During synthesis, measures were taken to add monomers rigorously in their

stoichiometric ratios. (High Mw)

NMP was needle-transferred to the reaction flask after being dehydrated by

molecular sieves.

An exact amount of diamine was added, while stirring the reaction flask


An ice bath/cooling system is kept to maintain the solution temperature below

room temperature (RT) after diamine addition.

Once diamine dissolved in the solution completely, a pre-calculated amount

of dianhydride was added portion wise (3-4 equal portion).

The reaction was carried out first at low temperature and then at RT, with constant

temperature monitoring.
In this work three 6FDA-based cross-linkable co-polyimide with high gas separation
performance for C3H6/C3H8 (propane-propylene) were fabricated. Synthesis Of High
Molecular Weight 6FDA-6FpDA is discussed which is the best way for the membrane preparation

As a conventional way is very energy intensive for olefin /paraffin separation we can choose the
middle way and make an hybrid process of utilizing advantages of both system distillation and
membrane system. In these paper we have discussed the material which is best suitable for propane
/propylene separation and its synthesis.

[1] C.E. Sroog, Polyimides, Prog. Polym. Sci. 16 (1991) 561-694.

[2] Y.C. Xiao, T.S. Chung, Grafting thermally labile molecules on cross-

linkable polyimide to design membrane materials for natural gas

purification and CO2 capture, Energy Environ. Sci. 4 (2011) 201-208.

[3] Y. Wang, T.S. Chung, H. Wang, S.H. Goh, Butanol isomer separation
using polyamideimide/CD mixed matrix membranes via pervaporation,
Chem. Eng. Sci. 64 (2009) 5198-5209.

[4] T.X. Yang, T.S. Chung, High performance ZIF-8/PBI nano-composite

membranes for high temperature hydrogen separation consisting of
carbon monoxide and water vapor, Int. J. Hydrogen Energy 38 (2013)

[5] D.F. Li, T.S. Chung, R. Wang, Morphological aspects and structure
control of dual-layer asymmetric hollow fiber membranes formed by a
simultaneous co-extrusion approach, J. Membr. Sci. 243 (2004) 155-175.

[6] L.Y. Jiang, T.S. Chung, D.F. Li, C. Cao, S. Kulprathipanja, Fabrication
of Matrimid/polyethersulfone dual-layer hollow fiber membranes for gas
separation, J. Membr. Sci. 240 (2004) 91-103.

[7] L.Y. Jiang, T.S. Chung, S. Kulprathipanja, An investigation to revitalize

the separation performance of hollow fibers with a thin mixed matrix
composite skin for gas separation, J. Membr. Sci. 276 (2006) 113-125.

[8] B.T. Low, N. Widjojo, T.S. Chung, Polyimide/polyethersulfone dual

layer hollow fiber membranes for hydrogen enrichment, Ind. Eng. Chem.
Res. 49 (2010) 8778-8786.

[9] Y.C. Xiao, T.S. Chung, M.L. Chng, S. Tamai, A. Yamaguchi, Structure
and properties relationships for aromatic polyimides and their derived
carbon membranes: experimental and simulation approaches, J. Phys.
Chem. B 109 (2005) 18741.
[10] A. Bondi, Van der waals volumes and radii, J. Physical Chemistry 68
(1964) 441-451.

[11] T.X. Yang, Y.C. Xiao, T.S. Chung, Poly-/metal-benzimidazole nano-

composite membranes for hydrogen purification, Energy Environ. Sci. 4
(2011) 4171-4180.

[12] H.Z. Chen, Polymeric membranes based on CO2-philic materials for

hydrogen purification and flue gas treatment, Ph.D. thesis, 2012.

[13] T.S. Chung, S.K. Teoh, X.D. Hu, Formation of ultrathin high-
performance polyethersulfone hollow-fiber membranes, J. Membr. Sci.
133 (1997) 161-175.