Torlon blend membranes

Journal of Membrane Science 462 (2014) 119–130

Contents lists available at ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

Molecular interaction, gas transport properties and plasticization behavior of cPIM-1/Torlon blend membranes Wai Fen Yong, Fu Yun Li, Tai Shung Chung n, Yen Wah Tong Department of Chemical & Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117565, Singapore

art ic l e i nf o

a b s t r a c t

Article history: Received 6 January 2014 Received in revised form 9 March 2014 Accepted 17 March 2014 Available online 25 March 2014

Polymers of intrinsic microporosity, specifically PIM-1, have emerged as promising materials for gas separation due their high gas permeability. However, its insolubility in common polar aprotic solvents like N-Methyl-2-pyrrolidone (NMP) limits its full potential and possible industrial applications. In this study, the solubility of PIM-1 in such solvents has been modified by carboxylation via hydrolysis reaction in a short period of 1 h. The success of carboxylation was verified by nuclear magnetic resonance (NMR) spectroscopy, Fourier transform infrared spectroscopy (FTIR) and water contact angles. The carboxylated PIM-1 (cPIM-1) was subsequently blended with Torlon to enhance the intrinsic permeability of Torlon rich membranes and the intrinsic selectivity of cPIM-1 rich membranes. The additions of 5, 10 and 30 wt% cPIM-1 into Torlon increase its CO2 permeability by 26%, 128% and 791%, respectively, from the original value of 0.541–0.682, 1.233 and 4.822 (1 Barrer¼ 1  10  10 cm3(STP) cm/cm2 s cmHg¼3.348  10  19 kmol m/m2 s Pa) with minor sacrifices in CO2/CH4 selectivity. These permeability improvements are attributed to the formation of hydrogen bonding and charge transfer complexes (CTC) between cPIM-1 and Torlon, which promotes better interactions in the blends. In addition, all the cPIM-1/Torlon membranes exhibit a great plasticization resistance up to 30 atm. This is ascribed to the incorporation of the rigid Torlon that may lead to restricting chain mobility in CO2 environments. The overall separation performance has been driven closer to the Robeson upper bound for O2/N2, CO2/CH4, CO2/N2 and H2/N2 separations. Therefore, the newly developed membranes may have great potential for energy development and industrial applications. & 2014 Elsevier B.V. All rights reserved.

Keywords: Polymer blend Polyimide PIM-1 CO2 plasticization Membrane gas separation

1. Introduction Anthropogenic CO2 emissions arise from forest burning, industrial manufacturing and power plants. According to the IPCC Climate Change 2007 report, the CO2 emission from energy sectors is predicted to increase from 40% to 110% between 2000 and 2030 [1]. A recent study by Yamasaki stated that CO2 contributes about 60% of global warming as compared to other greenhouse gases [2]. To mitigate the global warming, CO2 capture has been discussed aggressively. The main purpose of CO2 capture is to concentrate CO2 before it can be transported to CO2 storage sites [3,4]. Among all the available technologies, the polymer-based membrane separation process has played a significant role in CO2 capture because of its environmentally benign nature, easy processability, simple operation, small footprint and cost competitiveness [5–10]. Compared to existing polymer materials, polyimide emerges as a promising polymer because it exhibits high selectivity in major gas

n

Corresponding author. Tel.: þ 65 6516 6645; fax: þ 65 6779 1936. E-mail address: [email protected] (T.S. Chung).

http://dx.doi.org/10.1016/j.memsci.2014.03.046 0376-7388/& 2014 Elsevier B.V. All rights reserved.

pairs (e.g., CO2/CH4 and O2/N2), high thermal stability and reasonable mechanical strength [11–15]. Nevertheless, polyimide membranes generally have the plasticization issue when the separation involves a highly condensable gas (e.g., CO2) or is operated in aggressive conditions (e.g., high pressure and temperature). Plasticization happens with an increase in CO2 pressure that causes certain structure dissolution within the polymer matrix [11,12,16]. As a result, the inherent selectivity of polymer chains deteriorates. To enhance the membrane stability against the CO2-induced plasticization, many researchers have modified their membranes through cross-linking [12,13,17–19]. However, the cross-linked membranes yield a high selectivity but a reduced permeability in most cases. Other than cross-linking, polyimide polymers with an intrinsic high anti-plasticization property such as Torlon have been utilized. Torlon is a type of polyamide-imide that dissolves in polar aprotic solvents such as N-Methyl-2-pyrrolidone (NMP) and N,Ndimethylformamide (DMF). It has an impressive plasticization pressure up to 30–40 atm [20]. However, this polymer has a very low permeability which makes it impractical for industrial CO2 capture. Attempts have been made to modify Torlon by blending it with other polymers such as Matrimid and PBI [21]. Miscible

120

W.F. Yong et al. / Journal of Membrane Science 462 (2014) 119–130

Sigma Aldrich were used as received. Methanol (MEOH, Z99.9%) and N,N-dimethylformamide (DMF, 4 99.5%) from Merck were utilized without further purification. Hydrochloric acid (HCl, 37.5%), ethanol (EtOH, Z 99.9%), dichloromethane (DCM, 99.99%) and tetrahydrofuran (THF, 99.99%) from Fisher Scientific were used as received. The chemical structures of PIM-1 and Torlon are depicted in Fig. 1.

blends were formed owing to the existence of hydrogen bonding between polymers, but no gas permeability of these blends was reported. Compared to Torlon, PIM-1 is a type of polymer of intrinsic microporosity well recognized for its superior gas permeability recently [22–29]. However, blending PIM-1 with Torlon is not feasible because PIM-1 only dissolves in dichloromethane (DCM), tetrahydrofuran (THF), chloroform (CHCl3), while Torlon dissolves in NMP or DMF. In the recent work of Du et al.'s group, PIM-1 was chemically modified to replace nitrile groups by carboxylic groups through hydrolysis [30–32]. With the aid of carboxylic groups, the carboxylated PIM-1 (cPIM-1) became soluble in polar aprotic solvents. The resultant cPIM-1 membrane (treated at 120 1C for 5 h) showed an increase in CO2/N2 selectivity of about 136% (from the initial value of 11 to 26) but a decrease in CO2 permeability of approximately 93% (from the initial value of 8310 Barrer to 620 Barrer) [28]. Since Torlon has superior gas-pair selectivity and plasticization resistance while PIM-1 has outstanding gas permeability, we aim to explore a simple and novel method in this paper to combine their strengths by blending Torlon and the modified PIM-1 as new materials for gas separation membranes with enhanced gas transport properties and plasticization resistance. Different from Du et al.'s work, PIM-1 powders instead of PIM-1 membranes would be used for the chemical modification. Therefore, the objectives of this work are to: (1) molecularly tailor cPIM-1/Torlon membranes with synergistic separation efficiency and antiplasticization property, (2) fundamentally understand the properties and interactions of the polymer blends and (3) investigate the effects of blend ratio on separation performance and plasticization behavior. To our best knowledge, this is the first paper investigating the polymer blends of cPIM-1 and polyamide-imide for gas separation with enhanced separation performance and superior anti-plasticization behavior.

2.2. Modification of PIM-1 to cPIM-1 The cPIM-1 was formed by the hydrolysis of the nitrile group of PIM-1 [30–32]. PIM-1 powders were first immersed in a 20 wt% NaOH solution with the ratio of ethanol/water of 1:1 at 120 1C for 1 h. After that, the polymer was filtered from the solution and then boiled in deionized water with the addition of HCl (pH 4–5) for 1 h. The product was then filtered, rinsed with deionized water and washed again with MEOH to remove the HCl residue. The cPIM-1 powders were dried at 70 1C under vacuum for 24 h prior to use. The modification mechanism of PIM-1 to cPIM-1 is depicted in Fig. 2. The molecular weight of cPIM-1 was analyzed by gel permeation chromatography (GPC). The GPC system comprises a Waters 1515 iscoratic HPLC pump, a Waters 717 plus autosampler and a Waters 2414 refractive index detector. Polystyrene standards were used for calibration. DMF was used as the solvent at a flow rate of 1 mL/min with the testing temperature of 35 1C. The concentration of PIM-1 powders dissolved in DMF was 0.005 wt%. The GPC results showed the weight-average molar mass (Mw) of 193,500. The apparent surface areas of cPIM-1 were characterized from Brunauer–Emmett–Teller (BET) using a BET model NOVA4200e. The apparent surface area of cPIM-1 was 486 m2 g  1. 2.3. Membrane fabrication Dense films with different ratios of cPIM-1/Torlon were prepared via the solution casting method. Torlon was first dissolved in NMP and stirred overnight at 65 1C. Subsequently, cPIM-1 was added into the solution and stirred overnight. The final solution containing 2 wt% cPIM-1/Torlon was then filtered through a 1 mm PTFE filter and cast onto a silicon wafer at ambient temperature. The polymer solution was heated under vacuum at 40 1C for 12 h and then increased to 75 1C for 24 h. The formed dense films were peeled off from the silicon wafer and dried under vacuum with a temperature ramp of 25 1C/30 min to 250 1C and held for 12 h. The resultant dense films were labeled as “cPIM-1/Torlon (weight composition ratio)”, for example, cPIM-1/Torlon (5:95). The

2. Experimental 2.1. Materials A PIM-1 polymer with the weight-average molar mass (Mw) of 108,500 was synthesized in our lab while Torlon 4000TF powders were supplied by Solvay Advanced Polymers. Torlon was dried overnight under vacuum at 120 1C before use. N-methyl-2pyrrolidone (NMP, 499.5%) from Merck was further purified via vacuum distillation before usage. Anhydrous potassium carbonate (K2CO3, 499.5%) and sodium hydroxide (NaOH, Z98%) from

O

O

O

O

N

N

N

0.7n

O

H3C

CH3

O

CN O

O H3C

O CH3

N H

H

O

O

CN

n

Fig. 1. Chemical structures of (a) Torlon and (b) PIM-1.

0.3n

W.F. Yong et al. / Journal of Membrane Science 462 (2014) 119–130

O

H3C

O

CH3

CN

O CH3

COOH O

H3C

CN

CH3

O

EtOH / H2O H3C

H3C

NaOH

O

O

121

n

O CH3

COOH

n

PIM-1

cPIM-1 Fig. 2. Mechanism of the hydrolysis reaction from PIM-1 to cPIM-1.

average thickness of the membranes was measured from 10 different points using a Digimatic indicator (IDC-112B-5) with an accuracy of 1 μm. The thicknesses of the cast films were about 50 75 μm. 2.4. Characterizations To verify the success of the hydrolysis reaction, the solubility of cPIM-1 membranes in DCM (the solvent that can fully dissolve PIM-1) was determined through its gel content. The gel content was evaluated by extracting the insoluble portion of the membranes immersed in DCM for 24 h and drying them under vacuum at 80 1C for 24 h. The gel content was calculated by using the following equation: Gel content ð%Þ ¼ m1 =m0  100%

ð1Þ

where m0 and m1 are the masses of the original membranes and the insoluble portions, respectively. The chemical structures of PIM-1 and cPIM-1 were characterized using liquid state 1H (400 MHz) nuclear magnetic resonance (NMR) spectroscopy. The PIM-1 powders were dissolved in CDCl3 while cPIM-1 powders were dissolved in DMSO-d6. The CDCl3 and DMSO-d6 signals appeared at 1H 7.25 ppm and 1H 2.50 ppm, respectively. The changes in chemical structure of the polymeric membranes before and after the hydrolysis modification were analyzed by an attenuated total reflectance (ATR) mode using a Shimadzu Fourier transform infrared spectroscopy (FTIR) 8400 spectrometer under the range of 500–4000 cm  1. The spectra were obtained with an average of 16 scans at a resolution of 4 cm  1. The characteristic band that appeared at 2350 cm  1 was a contribution from the difference of carbon dioxide between background and samples. The hydrophilicity of the membranes was determined by their water contact angles, which were measured by Milli-Q DI water using a Contact Angle Geniometer (Rame Hart, USA) at room temperature. The UV absorbance spectra of dense membranes were performed using Shimadzu UV-3600 with the range of 200– 800 nm. The thermal stability of the membranes was analyzed by thermogravimetric analyses (TGA) using a Shimadzu Themogravimetry Analyzer DTG-60AH. The membranes were heated between 50 1C and 700 1C at a rate of 10 1C/min. During analyses, nitrogen was used as the purging gas and the flow rate was maintained at 50 mL/min. The surface morphology of cPIM-1/Torlon blend membranes was investigated using an Olympus BX50 polarized light microscope (PLM). The PLM micrographs were further analyzed by Image Pro Plus 3.0 software. The glass transition temperature Tg of dense cPIM-1/Torlon membranes was characterized by using differential scanning calorimetry (DSC, DSC822e, Mettler Toledo). The samples were tested with two consecutive scans at a heating rate of 10 1C min  1

from 40 to 50 1C. The first cycle of ramping and cooling was to eliminate any thermal history of the samples. The Tg of each sample was determined based on the mid-point transition temperature of the second heating curve. The density of dense membranes was determined at room temperature using a Mettler Toledo balance with a density kit applying the Archimedean principle by measuring the membrane weights in air and hexane. The pre-dried membranes were first weighed in air and then weighed again while immersed in hexane at room temperature. The density of the membranes was calculated using the following relationship:

ρpolymer ¼

wair ρ wair  whexane hexane

ð2Þ

where wair and whexane are the film weights in air and hexane, respectively. The relationship between FFV of dense membranes and blend ratio was studied using the following correlation: FFV ¼

ðV  V o Þ V

ð3Þ

The specific volume, V, is calculated from the measured density and the occupied volume, Vo, is determined by the equation: Vo ¼ 1.3Vw, where Vw is the van der Waals volume obtained using the group contribution proposed by Bondi [33]. For the polymer blends, Vw is predicted by using the correlation: V w ¼ w1 V w1 þ w2 V w2 , where w1 and w2 are the molar fractions and V w1 and V w2 are the van der Waals volumes of the pristine components 1 and 2, respectively. The specific free volume, Vf, is calculated as the difference between the specific volume and the occupied volume (e.g., Vf ¼V  Vo). The ideal specific free volume, Vf,mix, is employed to estimate the discrepancy between the real and ideal cases for the polymer blends: V f ;mix ¼ w1 V f 1 þ w2 V f 2 . 2.5. Measurements of gas transport properties and plasticization behavior Both pure gas and binary gas were employed to study the gas transport properties. The pure gas permeability of cPIM-1/Torlon membranes was measured using a variable-pressure constantvolume gas permeation cell reported elsewhere [34]. Pure gas permeability was tested in the sequence of H2, O2, N2, CH4, and CO2 at 35 1C and 3.5 atm. The rate of pressure increase (dp/dt) at steady state was used to calculate gas permeability as follows:   273  1010 Vl dp ð4Þ P¼ ATðp2 ð76=14:7ÞÞ dt 760 where P is the gas permeability of a membrane in Barrer (1 Barrer¼1  10  10 cm3 (STP) cm/cm2 s cmHg ¼3.348  10  19 kmol m/m2 s Pa), V is the volume of the downstream chamber (cm3), A refers to the effective membrane area (cm2), l is the membrane thickness (cm), T is the operating temperature (K), and

122

W.F. Yong et al. / Journal of Membrane Science 462 (2014) 119–130

p2 is defined as the upstream operating pressure (psia). The permeability tests were repeated at least three times with different membranes and the average deviation obtained was less than 5%. The ideal selectivity between two different gases in a polymeric membrane is the ratio of the permeability of single gas as described in Eq. (5): P α¼ A PB

ð5Þ

where PA and PB are the gas permeability of gases A and B, respectively. The CO2 induced plasticization behavior of the membranes was tested from 0.1 to 30 atm. The mixed gas permeation properties of cPIM-1/Torlon membranes were evaluated using a binary 50 mol% CO2 and 50 mol% CH4 mixtures. The membranes were tested at 35 1C and 7 atm. The detailed experimental design and procedures have been reported by Tin et al. [35]. The gas permeabilities of CO2 and CH4 are expressed by the following equations:   yCO2 V l 273  1010 dp ð6Þ P CO2 ¼ 760 ATð76=14:7ÞðX CO2 P 2 Þ dt P CH4 ¼

  ð1  yCO2 ÞVl 273  1010 dp 760 ATð76=14:7Þ½ð1  xCO2 ÞP 2  dt

Table 1 Solubility tests and contact angle measurements of PIM-1 and cPIM-1 at room temperature. Solventa

Polymer

PIM-1 cPIM-1

Contact angle (deg)

DCM

THF

NMP

DMF

Yes No

Yes No

No Yes

No Yes

71.5 72.6 58.7 71.7

a DCM, THF, NMP and DMF refer to dichloromethane, tetrahydrofuran, N-methyl-2-pyrrolidone and N,N-dimethylformamide, respectively.

ð7Þ

where P CO2 and P CH4 are the permeability (Barrer) of CO2 and CH4, respectively. P2 denotes the upstream feed gas pressure (psia), x refers to the mole fraction in the feed gas and y is the mole fraction in the permeate. The other symbols retain the same meanings as described earlier.

3. Results and discussion 3.1. Characterization of cPIM-1 polymer To analyze the structure of cPIM-1, solubility tests, NMR, FTIR, contact angle measurements and TGA analyses were performed. Table 1 shows the solubility tests of PIM-1 and cPIM-1 in different solvents; namely, dichloromethane (DCM), tetrahydrofuran (THF), N-methyl-2-pyrrolidone (NMP) and N,N-dimethylformamide (DMF). The original PIM-1 dissolves only in DCM and THF while cPIM-1 dissolves in those non-solvents for PIM-1 such as NMP and DMF. Their gel content measurements in DCM reveal that the conversion of PIM-1 to cPIM-1 is 98.6%. Compared to Du et al.'s work using PIM-1 films for the modification [30–32], the high degree of conversion in this study in a short period of reaction time (i.e., 1 h reaction time) may be attributed to the sufficient contact area between PIM-1 powders and chemical agents during the reaction. The obvious solubility enhancement of PIM-1 after the hydrolysis modification provides a clear evidence of structural change. To verify this change, NMR and FTIR analyses were also employed and their results will be discussed in the following section. Fig. 3 depicts the 1H NMR spectra of PIM-1 in CDCl3 and cPIM-1 in DMSO-d6. A typical PIM-1 spectrum is achieved, where the proton at 6–7 ppm is ascribed to the aromatic H-a and H-b and the doublet at 1.8–2.8 ppm is assigned to the aliphatic H-h [31,32]. After hydrolysis, a broad signal at 7.2–8.3 ppm appears in the cPIM-1 spectrum, which is attributed to the formation of –COOH groups. Moreover, the relative peak intensity ratio of the aromatic groups (e.g. H-a, H-b and –COOH) to the aliphatic groups (e.g. H-h and –CH3) is about 6H:16H, which indicates the correct structure of cPIM-1 [31,32]. A similar 1H NMR spectrum of cPIM-1 was reported by Du et al. [31,32].

Fig. 3. 1H NMR of (a) PIM-1 and (b) cPIM-1.

PIM-1

cPIM-1

3000-3600 (-COOH)

4000

3500

2250 (C≡N)

3000

2500

1700 (C=O)

2000

1500

1000

500

Wavenumbers-1)

Fig. 4. FTIR spectra of PIM-1 and cPIM-1 membranes.

Fig. 4 displays the structural changes of PIM-1 after hydrolysis by ATR-FTIR. A characteristic stretching band of nitrile (CRN) groups appears near 2250 cm  1 in the spectrum of PIM-1 [36]. Compared to the pristine PIM-1, the nitrile band is not visible in cPIM-1. In contrast, a prominent broad band at the range of 3000–3600 cm  1 and a small band at 1700 cm  1 are observed in

W.F. Yong et al. / Journal of Membrane Science 462 (2014) 119–130

cPIM-1, which are attributed to the hydroxyl (–OH) and the carbonyl (CQO) stretching of the newly formed carboxylic acid groups, respectively [31,32]. These results are consistent with the NMR spectra and re-confirm the approximately complete conversion of nitrile groups to carboxylic groups in cPIM-1. To study the surface hydrophilicity, water contact angle measurements were carried out on the surfaces of PIM-1 and cPIM-1 membranes. Hydrophilicity is an indication of the affinity of a material with water molecules. A greater hydrophilicity comes with a smaller water contact angle value. As revealed in Table 1, the water contact angles of the original PIM-1 and cPIM-1 are 71.572.61 and 58.7 71.71, respectively. An increase in hydrophilicity for cPIM-1 is probably due to the introduction of the highly polar carboxylic group in the polymer chains. Fig. 5 shows the thermal degradation curves of the pristine PIM-1 and cPIM-1 membranes. Generally, the thermal degradation temperature is defined based on a 5 wt% loss of the original weight. From the TGA curves, it can be clearly seen that the cPIM-1 membrane exhibits lower thermal stability (e.g., 350 1C) than the PIM-1 membrane (e.g., 450 1C). The lower degradation

123

temperature in cPIM-1 results from the loss of the strong dipolar interaction of the nitrile groups [37] in PIM-1. On the other hand, the TGA analyses indirectly imply the successful modification of nitrile groups to carboxylic groups in cPIM-1.

3.2. Characterizations of cPIM-1/Torlon membranes The cPIM-1/Torlon membranes were characterized by PLM, FTIR, DSC, TGA analyses and UV absorbance tests. Generally, optical inspection provides immediate information about the homogeneity of the polymer blends. A PLM was employed as the optical tool in this study. A PLM image with a clear and single phase may indicate a homogenous blend. Fig. 6 shows the PLM micrographs of the cPIM-1/Torlon membranes tested at room temperature. A homogenous morphology could be observed for the blend membranes consisting of 5–10 wt% cPIM-1 or 5 wt% Torlon. The membranes with other blend ratios seem to be partially miscible, where two phases coexist in the PLM images. These results suggest that only a small portion (i.e., up to 10 wt%) of cPIM-1 in Torlon and up to 5 wt% Torlon in cPIM-1 could be fully miscible in the blends.

100

Residual Weight (wt%)

95

90

85 PIM-1 cPIM-1

80 0

100

200

300

400

500

600

700

4000

Fig. 5. TGA curves of PIM-1 and cPIM-1 membranes.

3500

3000

2500

2000

1500

1000

500

Wavenumbers (cm-1)

Temperature (°C)

Fig. 7. FTIR spectra of PIM-1, cPIM-1, Torlon and cPIM-1/Torlon blend membranes.

100μm

Fig. 6. PLM images of the dense membranes at room temperature: (a) Torlon; (b) cPIM-1/Torlon (5:95); (c) cPIM-1/Torlon (10:90); (d) cPIM-1/Torlon (30:70); (e) cPIM-1/ Torlon (50:50); (f) cPIM-1/Torlon (70:30); (g) cPIM-1/Torlon (90:10); (h) cPIM-1/Torlon (95:5); (i) cPIM-1.

124

W.F. Yong et al. / Journal of Membrane Science 462 (2014) 119–130

Other than PLM analyses, FTIR were employed to examine the structural changes. Fig. 7 depicts the FTIR spectra of the cPIM-1/ Torlon blend membranes. The absorption band in the range of 3000–3600 cm  1 corresponds to –COOH group. The typical bands of imide rings are at 1700 cm  1, 1360 cm  1 and 720 cm  1, which reflect the symmetric CQO stretching, C–N stretching and CQO bending, respectively. The band appearing at 1360 cm  1 represents the transverse stretching of the C–N–C bond in imide groups. The CQO and C–N stretching in the amide groups have absorption at 1650 cm  1 and 1590 cm  1, respectively. It is clearly seen that there are two bands in the region of 1720–1650 cm  1 which represent the imide CQO and amide CQO groups for Torlon rich membranes (e.g., the blends consist of 50–100% Torlon). The enlarged spectra in the region of 3900–2500 cm  1 of the selected cPIM-1/Torlon blend membranes are shown in Fig. 8. It is worth noting that the CQO band of cPIM-1 at 1700 cm  1 slightly shifts to a lower band with an increase in Torlon loading. This is probably due to the formation of hydrogen bonding interaction between cPIM-1 and Torlon.

3000-3600 (-COOH)

λb ¼ ϕ1 λ1 þ ϕ2 λ2

cPIM-1

cPIM-1/Torlon (90:10)

cPIM-1/Torlon (70:30) cPIM-1/Torlon (50:50)

2500

520

Fig. 8. FTIR spectra in the region of 3900–2500 cm  1 of the cPIM-1/Torlon blend membranes.

Membranes ID

Glass transition temperature, Tg (1C)

Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon

273.9 278.3 284.5 286.3, 415.4 297.1, 403.4 319.4, 399.2 359.8, 398.5

500

λ (UV) (nm)

Table 2 Tg value of cPIM-1/Torlon blend system obtained from DSC.

(5:95) (10:90) (30:70) (50:50) (70:30) (90:10)

ð8Þ

where λb is the wavelength of UV absorbance band of the blend, λ1 and λ2 are the wavelengths of UV absorbance bands of the pristine components 1 and 2, respectively, ϕ1 and ϕ2 are the respective volume fractions of components 1 and 2, respectively. Fig. 9 displays the UV absorbance data of cPIM-1/Torlon blend membranes. Compared to the pristine cPIM-1 or Torlon, the wavelength of UV absorbance band of cPIM-1/Torlon blends exceeds the predicted data calculated from the additive law. This provides clear evidence that the formation of charge transfer complexes (CTC) occurs between cPIM-1 and Torlon in their blends. The CTCs are formed due to two electron transfer mechanisms: intramolecular and intermolecular charge transfers. The intramolecular charge transfer comes from the high electron density in the nitrogen atoms of Torlon and oxygen atoms in the ether groups of PIM-1. They behave as electron donors to their own carbonyl groups as electron acceptors [38,41–43]. On the other hand, the intermolecular charge transfer is formed by the electron transfer from the oxygen atoms of the ether groups in one polymer to the carbonyl groups of another polymer. As a result, there are UV absorption shifts in the cPIM-1/Torlon blend membranes. Thus, the formation of CTCs may enhance the entanglement of the cPIM-1 and Torlon, as occurred in other blend systems [37,40–43].

cPIM-1/Torlon (95:5)

3500 3000 Wavenumbers (cm-1)

The partially miscible behavior of the membranes was further validated by measuring their glass transition temperatures (Tgs) using DSC. Table 2 tabulates the DSC results of cPIM-1/Torlon membranes. The Torlon membrane displays a characteristic Tg of 273.9 1C [21]. Similar to PIM-1, cPIM-1 exhibits a high Tg value which is hardly detected in the testing range of 50–450 1C. This is mainly due to the high stiffness and low rotational freedom of the polymer chains in cPIM-1. There are two distinguishable Tgs for membranes consisting of 30–70 wt% cPIM-1. The two Tgs appear at 286–360 1C and 399–415 1C which may represent the Torlon rich phase and cPIM-1 rich phase, respectively. Interestingly, these two Tgs shift towards each other when the cPIM-1 concentration increases. This suggests that there are two partial miscible phases entangled in the blends; namely, Torlon rich and cPIM-1 rich phases [39]. The entanglement in the blends may be caused by the hydrogen bonding interaction between these polymer molecules. UV absorbance tests were carried out to examine the development of charge transfer complexes (CTCs) in the cPIM-1/Torlon blend membranes. The formation of CTCs is attributed to the electron transfer interaction between the electron acceptor and donor [38,40–43]. The electron donor tends to donate electrons to the acceptor, which shortens the distance between the donor and the acceptor. The wavelength of UV absorbance band of the blend membranes may be predicted using the additive law as follows [44]:

480

460

440 Experimental data Additive law

420 0

0.2

0.4

0.6

0.8

Volume fraction of cPIM-1

Fig. 9. UV absorbance values of cPIM-1/Torlon polymer blends.

1

W.F. Yong et al. / Journal of Membrane Science 462 (2014) 119–130

Hydrogen bonding could be formed from various intermolecular interactions between cPIM-1 and Torlon. These interactions include the carboxyl group of cPIM-1 and nitrogen atoms of Torlon; the hydroxyl group of cPIM-1 and nitrogen atoms of Torlon; and the hydroxyl group of cPIM-1 and oxygen atoms of Torlon. Strong dipole–dipole interactions existing in the hydrogen bonding are much stronger compared to the other physical interactions such as CTCs interactions. Different from our previous study on the PIM-1/ Matrimid blends [38], strong hydrogen bonding interactions can be formed for the cPIM-1/Torlon blend systems. It is believed that the strong hydrogen bonding coupled with the CTCs interactions promote a better compatibility between the two polymers. 3.3. Effect of different blend compositions to the gas transport properties and plasticization

1.36

0.20

1.34

0.18

1.32

0.16 0.14

1.30

0.12

1.28

0.10 1.26

FFV

Density (g/cm3)

FFV refers to the fractional free volume which is not occupied by polymer chains. FFV can be calculated from the density of the polymer and it is inversely proportional to polymer density [11,18,45,46]. Fig. 10 shows the density and calculated FFV of cPIM-1, Torlon and cPIM-1/Torlon membranes. cPIM-1 has a lower density and thus a higher FFV than Torlon. As a consequence, the FFV of the resultant blend membranes increases with an increase in cPIM-1 amount. It is fairly understandable that the hydrogen bonding and CTCs interactions between the two polymers promote the polymer chain packing, hinder the formation of FFV, and thus a less experimental FFV than the predicted is expected. Interestingly, a good linear relationship between the experimental FFV and the predicted FFV was obtained. This might be due to the discrepancy in the predicted FFV compared to the actual FFV of Torlon in all the blend ratios. In the prediction line, the FFV of Torlon is based on the assumption that it remains unchanged. In fact, the incorporation of rigid structure of cPIM-1 has retarded

0.08

1.24

0.06

Density (experimental) Density (prediction) FFV (experimental) FFV (prediction)

1.22 1.20

0.04 0.02 0.00

1.18 0

0.2

0.4 0.6 Volume fraction of cPIM-1

0.8

1

Fig. 10. Density and FFV of cPIM-1/Torlon polymer blends.

125

the chain packing of Torlon during blends. Thus, the actual FFV of Torlon might be higher than the pristine Torlon and it changes with the increasing loading of cPIM-1. Table 3 presents the pure gas transport properties of pristine cPIM-1, pristine Torlon and cPIM-1/Torlon membranes. On the one hand, the membranes incorporated with cPIM-1 show a remarkable improvement in permeability as compared to the pristine Torlon. On the other hand, compared to pristine cPIM-1, the addition of Torlon in the polymer blends increases the membrane selectivity. Interestingly, by incorporating 5, 10 and 30 wt% cPIM-1 in Torlon, the CO2 permeability increases from the original 0.541 to 0.682, 1.233 and 4.822 Barrers, respectively, which is equivalent to about 26%, 128% and 791% increments, respectively. Meanwhile, the CO2/CH4 selectivity of these membranes drops slightly from the original 41.6 to 40.1, 39.8 and 34.4, respectively, which corresponds to a decrease of 3.6%, 4.4% and 17.2%, respectively. A similar trend could be observed for the H2/N2, O2/N2, CO2/N2, H2/CH4, and H2/CO2 separation. The improvement in the permeability of all gases is mainly attributed to the fact that the incorporation of cPIM-1 with a higher FFV reduces chain packing and thus allows more gases to permeate through the membrane. Since the membranes exhibit a partially miscible behavior, the rule of semi-logarithmic addition can be used to estimate the gas permeability and selectivity as follows [45,46]: ln P b ¼ ϕ1 ln P 1 þ ϕ2 ln P 2

ln

ð9Þ

      P1 P1 P1 ¼ ϕ1 ln þ ϕ2 ln P2 P2 1 P2 2

ð10Þ

where Pb is the permeability of the polymer blend, P1 and P2 are the permeability of components 1 and 2, ϕ1 and ϕ2 are the respective volume fractions of components 1 and 2. Fig. 11a shows the experimental and predicted permeability data for the cPIM-1/ Torlon membranes. It can be seen that the experimental data of all polymer blends are in good agreement with the values calculated from the rule of semi-logarithmic addition. This good agreement is a direct result of the partially miscible behavior and the hydrogen bonding and CTCs interactions between cPIM-1 and Torlon. As seen from Fig. 11b and c, the selectivity of all membranes is higher than the predictions possibly due to the effect of hydrogen bonding and CTCs [38]. This is due to the physical interactions of hydrogen bonding and CTCs that provide different interaction sites of the functional groups to the gases. Interestingly, the blend with 90 wt% of cPIM-1 in Torlon has a higher selectivity in both O2/N2 and CO2/CH4 separation as compared to the predicted data. A similar result has been reported in the PIM-1/Matrimid blend system [38]. This is likely attributed to both hydrogen bonding and CTCs effect that provides different interaction sites to the gas molecules.

Table 3 Pure gas permeability and selectivity of the Torlon, cPIM-1 and cPIM-1/Torlon blend membranes tested at 35 1C and 3.5 atm. Permeability (Barrera)

Membranes ID

Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1 PIM-1 [36] a

(5:95) (10:90) (30:70) (50:50) (70:30) (90:10) (95:5)

Ideal selectivity

H2

O2

N2

CH4

CO2

H2/N2

O2/N2

CO2/N2

CO2/CH4

H2/CH4

H2/CO2

3.730 4.324 6.761 20.23 52.06 480.2 909.1 1044 1619 2918

0.130 0.163 0.327 1.152 4.191 79.30 185.0 244.4 462.2 735.0

0.018 0.023 0.047 0.191 0.854 17.61 42.17 67.42 142.9 192.0

0.013 0.017 0.031 0.140 0.701 18.21 42.31 86.42 208.6 268.0

0.541 0.682 1.233 4.822 21.42 440.6 1013 1382 2654 3825

207.2 188.0 143.9 105.9 61.0 27.3 21.6 15.5 11.3 15.2

7.2 7.1 7.0 6.0 4.9 4.5 4.4 3.6 3.2 3.8

30.1 29.7 26.2 25.2 25.1 25.0 24.0 20.5 18.6 19.9

41.6 40.1 39.8 34.4 30.6 24.2 23.9 16.0 12.7 14.3

286.9 254.4 218.1 144.5 74.3 26.4 21.5 12.1 7.8 10.9

6.9 6.3 5.5 4.2 2.4 1.1 0.9 0.8 0.6 0.8

1 Barrer ¼ 1  10  10 cm3 (STP)cm/cm2 s cmHg.

126

W.F. Yong et al. / Journal of Membrane Science 462 (2014) 119–130

Permeability (Barrer)

10000 1000 100 10

CO2 O2 N2

1 0.1

Rule of semi-logarithmic

CH4

0.01 0.0

0.2

0.4

0.6

0.8

1.0

Volume fraction of cPIM-1

1000

8

45

5 10

4 3 2

1

selectivity selectivity (prediction) permeability

1 0 0

0.2 0.4 0.6 0.8 Volume fraction of cPIM-1

CO2/CH4 selectivity

100

1000

35 100

30 25

100

20

10

15 selectivity selectivity (prediction) permeability

10 5

0.1

1 0.1

0

1

1

CO2 permeability

6

O2 permeability

O2/N2 selectivity

10000

40

7

0

0.2 0.4 0.6 0.8 1 Volume fraction of cPIM-1

Fig. 11. Comparison between experimental data and prediction data for: (a) O2, N2, CH4, CO2 permeability; (b) O2/N2 selectivity; (c) CO2/CH4 selectivity (dashed lines estimated using the rule of semi-logarithmic addition).

30

3000

Permeability (Barrer)

Permeability (Barrer)

3500 cPIM cPIM-11

2500 2000 1500 1000

cPIM-1/Torlon (95:5) cPIM-1/Torlon (90:10)

500

cPIM-1/Torlon (70:30)

0

25 20

cPIM-1/Torlon (50:50)

15 10 cPIM-1/Torlon (30:70)

5 0

0

5

10

15

20

30

25

0

5

10 15 20 Pressure (atm)

Permeability (Barrer)

Pressure (atm) 1.8 1.6 1. 4 1.2 1.0 0.8 0.6 0.4 0.2 0.0

25

30

cPIM-1/Torlon (10:90) cPIM-1/Torlon (5:95) Torlon 0

5

10 15 20 Pressure (atm)

25

30

Fig. 12. CO2 plasticization behavior of cPIM-1/Torlon membranes in the pressure range of 0.1–30 atm.

Fig. 12 shows the CO2 plasticization behavior of the cPIM-1/ Torlon membranes. In general, the plasticization pressure refers to the pressure where the lowest permeability exists in a membrane before the dissolution of CO2 causes a sudden increase in permeability. It can be seen that all the cPIM-1/Torlon membranes show impressive anti-plasticization properties. The pristine cPIM-1

membrane exhibits a plasticization pressure at around 20 atm, but the Torlon/cPIM-1 blends display no plasticization pressures up to 30 atm. The hydrogen bonding promotes polymer chain packing and reduces the FFV. It is somewhat improving plasticization resistance. Moreover, the enhancement in plasticization resistance is likely due to the introduction of Torlon that not only

W.F. Yong et al. / Journal of Membrane Science 462 (2014) 119–130

forms partial miscible blends with cPIM-1 but also reduces the inter-segment mobility in the polymer matrix.

3.4. Mixed gas separation performance and the Robeson upper bound comparison Table 4 compares the CO2/CH4 (50%/50%) mixed gas and pure gas transport properties of cPIM-1/Torlon membranes. The binary gas data of cPIM-1/Torlon (90:10) shows a CO2 permeability of 963 Barrer, a CH4 permeability of 43 Barrer and a CO2/CH4 selectivity of 22.2. Generally, the CO2/CH4 selectivity of the mixed gas is similar to the pure gas. However, there is a slight difference in CO2 permeability between mixed and pure gas tests and this is ascribed to the sorption competition between CO2 and CH4 [8,9,11,12,47]. A Robeson upper bound performance of the O2/N2, CO2/CH4, CO2/N2 and H2/N2 separation is plotted in Fig. 13. The gas separation performance goes along with the upper bound line with an increase in cPIM-1 loading. Loading a small amount (i.e., 5–30 wt%) of cPIM-1 into Torlon drives the overall gas separation performance close to the upper bounds for the O2/N2, CO2/CH4, CO2/N2 and H2/N2 separation.

Table 4 Binary gas permeability and selectivity of cPIM-1/Torlon membranes tested with a CO2/CH4 (50:50 mol%) at 35 1C and 7 atm. Membranes ID

Permeability (Barrer) CO2

cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon

a

(10:90) (30:70) (50:50) (70:30) (90:10)

1.061 (1.233) 4.52 (4.82) 19.48 (21.44) 383.2 (440.6) 946 (1013)

Selectivity

CH4

CO2/CH4

0.027 (0.017) 0.13 (0.14) 0.65 (0.72) 16.3 (18.2) 43 (42)

39.7 33.6 30.0 23.4 22.2

(39.8) (34.4) (29.8) (24.2) (24.1)

a Number in parentheses is the permeability and ideal selectivity obtained from pure gas test.

3.5. Comparison among the polymer blends incorporated with PIM-1 or cPIM-1 PIM-1 is a highly permeable polymer attractive for gas separation but with a drawback of low selectivity. Blending it with highly selective polymers may expand and excel its potential. However, there are only few PIM-1 blend studies with other polymers such as Matrimid [28,38,48] and Ultem [48,49]. This is likely due to the difficulty in obtaining a fully miscible blend system. The first polymer blend with PIM-1 [48] was patented by the researchers of UOP but only the gas separation performance of the blend was reported. The detailed studies on the evolution of phase behaviors, prediction of gas transport properties [38,49] and extention to hollow fiber configurations [28] have been recently carried by Chung and his co-workers. Up to date, to our best knowledge, this work is the first paper that reports the polymer blend with cPIM-1. Fig. 14 compares the phase behavior of PIM-1 blending with Matrimid and Ultem, and cPIM-1 blending with Torlon. Among these three blend systems, the blends of cPIM-1 with Torlon show the most compatible system in most blend ratios. This is likely attributed to the hydrogen bonding and CTCs induced interactions between polymers that promote greater miscibility of the blends. Meanwhile, the PIM-1/Ultem blends give the least compatibility with the existence of two phases in all the blend ratios. Table 5 also shows a comparison of gas transport properties of the membranes made from these the blends. The addition of 10 wt% cPIM-1 in Torlon has the highest CO2/CH4 selectivity of 39.8 and a CO2 permeability of 1.233 Barrer among all the blends with low loadings of PIM-1 or cPIM-1 (e.g., less than 10 wt%). On the other hand, the high loading of 90 wt% PIM-1 in Matrimid shows the highest CO2 permeability of 1953 Barrer with a moderate CO2/CH4 selectivity of 16 compared to the other blends. However, as a type of polyimide, the main drawback for the Matrimid blend system is its intrinsic poor plasticization behaviors [8,10,11,50]. Therefore, the cPIM-1/Torlon membrane reveals as a good candidate for natural gas separation or CO2 capture because of its exceptional high

CO2/CH4 selectivity

1000

10 Torlon cPIM-1/Torlon (5:95) cPIM-1/Torlon (10:90) cPIM-1/Torlon (30:70) PIM 1/T l (50:50) cPIM-1/Torlon cPIM-1/Torlon (70:30) cPIM-1/Torlon (90:10) cPIM-1/Torlon (95:5) cPIM-1

1

0.1 0.1

1 10 100 O2 permeability (Barrer)

100

1

0.1

1000

1000

100

100

Torlon cPIM-1/Torlon (5:95) cPIM-1/Torlon (10:90) cPIM-1/Torlon (30:70) cPIM-1/Torlon (50:50) cPIM-1/Torlon (70:30) cPIM-1/Torlon (90:10) cPIM-1/Torlon (95:5) cPIM-1

10 1

Torlon cPIM-1/Torlon (5:95) cPIM-1/Torlon (10:90) cPIM-1/Torlon (30:70) cPIM-1/Torlon cPIM 1/Torlon (50:50) cPIM-1/Torlon (70:30) cPIM-1/Torlon (90:10) cPIM-1/Torlon (95:5) cPIM-1

10

0.1

1000

H2/N2 selectivity

O2/N2 selectivity

100

CO2/N2 selectivity

127

10 1000 CO2 permeability (Barrer)

Torlon cPIM-1/Torlon (5:95) cPIM-1/Torlon cPIM 1/Torlon (10:90) cPIM-1/Torlon (30:70) cPIM-1/Torlon (50:50) cPIM-1/Torlon (70:30) cPIM-1/Torlon (90:10) cPIM-1/Torlon (95:5) cPIM-1

10 1 0.1

0.1 0.1

10 1000 CO2 permeability (Barrer)

1

10

100

1000

10000

H2 permeability (Barrer)

Fig. 13. Comparison with Robeson upper bound for cPIM-1/Torlon blend membranes: (a) O2/N2; (b) CO2/CH4; (c) CO2/N2; (d) H2/N2 separation.

128

W.F. Yong et al. / Journal of Membrane Science 462 (2014) 119–130

PIM-1 or cPIM-1content 10 wt%

30 wt%

50 wt%

70 wt%

90 wt%

(i) PIM-1/Matrimida

(ii) PIM-1/Ultemb

(iii) cPIM-1/Torlon 100μm Fig. 14. Comparison of PLM images for different blend systems: (i) PIM-1/Matrimid [38]; (ii) PIM-1/Ultem [49]; (iii). [38,49] copyright 2011 and 2013. Reproduced with the permission from Elsevier B.V. Table 5 A comparison of PIM-1 and cPIM-1 based polymer blends [38,48,49]. Membranes ID

Testing T/P (1C/atm)

Permeability (Barrera) O2

Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon cPIM-1/Torlon

(5:95) (10:90) (30:70) (50:50) (70:30) (90:10) (95:5)

Matrimid PIM-1/Matrimid PIM-1/Matrimid PIM-1/Matrimid PIM-1/Matrimid PIM-1/Matrimid PIM-1/Matrimid PIM-1/Matrimid Ultem PIM-1/Ultem PIM-1/Ultem PIM-1/Ultem PIM-1/Ultem PIM-1/Ultem PIM-1/Ultem PIM-1/Ultem PIM-1/Ultem PIM-1/Ultem a

(5:95) (10:90) (30:70) (50:50) (70:30) (90:10) (95:5)

(5:95) (10:90) (20:80) (30:70) (50:50) (70:30) (80:20) (90:10) (95:5)

35/3.5 35/3.5 35/3.5 35/3.5 35/3.5 35/3.5 35/3.5 35/3.5

0.13 0.163 0.327 1.152 4.191 79.3 185.0 244.4

35/3.5 (50/6.9) 35/3.5 35/3.5 (50/6.9) 35/3.5 (50/6.9) 35/3.5 35/3.5 35/3.5 35/3.5

2.1 2.6 3.4 11 31 116 400 632

35/3.5 (50/6.9) 35/3.5 35/3.5 (50/6.9) 35/3.5 (50/6.9) 35/3.5 35/3.5 35/3.5 35/3.5 35/3.5 35/3.5

0.38 0.58 1.1 1.6 2.2 10.4 93.9 230.1 512.4 556.5

N2 0.018 0.023 0.047 0.191 0.854 17.61 42.17 67.42

Ideal selectivity CH4

CO2

O2/N2

CO2/N2

CO2/CH4

Refs.

0.013 0.017 0.031 0.14 0.701 18.21 42.31 86.42

0.541 0.682 1.233 4.822 21.42 440.6 1013 1382

7.2 7.1 7.0 6.0 4.9 4.5 4.4 3.6

30.1 29.7 26.2 25.2 25.1 25.0 24.0 20.5

41.6 40.1 39.8 34.4 30.6 24.2 23.9 16

This This This This This This This This

(0.355)

9.6 (10) 12 17 (20.3) 56 (35.9) 155 558 1953 3355

6.4 6.6 6.1 5.8 5.4 5 4 3.8

30 29 30 28 27 24 20 20

36 (28.2) 35 34 (27.1) 31 (24.8) 28 25 16 14

[38,48] [38] [38,48] [38,48] [38] [38] [38] [38]

1.48 (0.064) 2.18 3.95 (0.092) 6.58 (0.191) 9.27 51.7 477.1 1260 2877 3276

7.1 7.0 6.8 6.1 5.8 5.4 4.4 3.5 3.6 3.6

27.4 26.2 25.2 25.7 24.8 26.9 21.8 19.3 20.0 21.1

37 (30.3) 36.5 33.8 (31.6) 34.6 (30.2) 34.7 23.2 16.7 13.0 11.6 11.8

[49,48] [38] [49,48] [49,48] [49] [49] [49] [49] [49] [49]

(0.749) (1.45)

0.054 0.083 0.16 0.26 0.38 1.9 21.9 65.2 143.7 155

0.04 (1.95) 0.06 0.12 (2.89) 0.19 (5.77) 0.28 2.2 28.5 97.1 247.6 277

work work work work work work work work

1 Barrer¼ 1  10  10 cm3 (STP)cm/cm2 s cmHg.

plasticization resistance and reasonably good compatibility in the blend systems. The blend consists of 90 wt% cPIM-1 in Torlon with a CO2 permeability of 1013 Barrer and a CO2/CH4 selectivity of 23.9 may be suitable for gas separations that require a high flux.

4. Conclusions In this study, PIM-1 powders was carboxylated by a hydrolysis reaction and blended with Torlon. The newly developed cPIM-1/ Torlon blends were investigated in terms of miscibility and gas separation properties. The anti-plasticization properties of the

blend membranes with different blend ratios were also examined in-depth. The following conclusions could be drawn from this work: 1. The carboxylation of PIM-1 powders was successfully performed by the hydrolysis reaction, which was verified by solubility tests, gel content analyses, NMR, FTIR and water contact angle measurements. By introducing highly polar carboxylic groups in the polymer chains, the cPIM-1 displays a greater hydrophilicity and becomes soluble in polar aprotic solvents, which makes it ready for blending with other polymers soluble in DMF and NMP.

W.F. Yong et al. / Journal of Membrane Science 462 (2014) 119–130

2. The cPIM-1/Torlon blend membranes with a small amount of cPIM-1 or Torlon (5–10 wt%) hold a homogeneous morphology, which was validated by PLM images and Tg values. This good miscibility is attributed to the strong interchain of hydrogen bonding interactions and the intramolecular and intermolecular charge transfers formed among cPIM-1 and Torlon, which promote better entanglement of cPIM-1 and Torlon in the polymer matrix. FTIR and UV absorbance measurements have confirmed our hypothesis. 3. The addition of cPIM-1 in Torlon results in a significant increment in gas permeability of Torlon-rich membranes with a slight decrease in selectivity. The overall separation performance of cPIM-1/Torlon membranes has been driven close to the Robeson upper bound for O2/N2, CO2/CH4, CO2/N2 and H2/N2 separations. 4. Remarkably, all the cPIM-1/Torlon membranes show impressive high anti-plasticization properties with no plasticization pressures up to 30 atm. The improved anti-plasticization property is attributed to the incorporation of Torlon, which not only has a great rigidity but also form partial miscible blends with cPIM-1, thus restricting the chain movement in the polymer matrix. 5. Compared to PIM-1/Matrimid and PIM-1/Ultem blend systems, the cPIM-1/Torlon blend shows the most compatible system. The interactions between cPIM-1 and Torlon are likely promoted by hydrogen bonding and CTCs.

Acknowledgments This research is supported by the National Research Foundation, Prime Minister’s Office, Singapore under its Competitive Research Program (CRP Award No. NRF-CRP 5-2009-5 (NUS grant number R-279-000-311-281)). The authors also express their appreciation to Miss Ngoc Lieu Le and Dr. You Chang Xiao for their valuable suggestions to this work. References [1] B. Metz, O.R. Davidson, P.R. Bosch, R. Dave, L.A. Meyer, (Eds.), IPCC, 2007: Climate change 2007: mitigation. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2007. [2] A. Yamasaki, An overview of CO2 mitigation options for global warming, J. Chem. Eng. Jpn. 36 (2003) 361–375. [3] B. Metz, O. Davidson, H.C. de Coninck, M. Loos, L.A. Meyer, (Eds.), IPCC special report on Carbon dioxide Capture and Storage. Prepared by Working Group III of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, United kingdom and New York, NY, USA, 2005. [4] M.E. Boot-Handford, J.C. Abanades, E.J. Anthony, M.J. Blunt, S. Brandani, N. Mac Dowell, J.R. Fernández, M.-C. Ferrari, R. Gross, J.P. Hallett, R.S. Haszeldine, P. Heptonstall, A. Lyngfelt, Z. Makuch, E. Mangano, R.T.J. Porter, M. Pourkashanian, G.T. Rochelle, N. Shah, J.G. Yao, P.S. Fennell, Carbon capture and storage update, Energy Environ. Sci. 7 (2014) 130–189. [5] P. Bernardo, E. Drioli, G. Golemme, Membrane gas separation: a review/state of the art, Ind. Eng. Chem. Res. 48 (2009) 4638–4663. [6] R.W. Baker, Future directions of membrane gas separation technology, Ind. Eng. Chem. Res. 41 (2002) 1393–1411. [7] V. Abetz, T. Brinkmann, M. Dijkstra, K. Ebert, D. Fritsch, K. Ohlrogge, D. Paul, K.V. Peinemann, S. Pereira-Nunes, N. Scharnagl, M. Schossig, Developments in membrane research: from material via process design to industrial application, Adv. Energy Mater. 8 (2006) 328–358. [8] Y. Yampolskii, B. Freeman, Membrane Gas Separation, John Wiley & Sons, Chichester, 2010. [9] D.R. Paul, Y.P. Yampol'skii, Polymeric Gas Separation Membranes, CRC Press, Boca Raton, FL, 1994. [10] Y. Yampolskii, Polymeric gas separation membranes, Macromolecules 45 (2012) 3298–3311. [11] Y.C. Xiao, B.T. Low, S.S. Hosseini, T.S. Chung, D.R. Paul, The strategies of molecular architecture and modification of polyimide-based membranes for CO2 removal from natural gas—a review, Prog. Polym. Sci. 34 (2009) 561–580.

129

[12] M.R. Coleman, W.J. Koros, Conditioning of fluorine-containing polyimides. 2. Effect of conditioning protocol at 8% volume dilation on gas-transport properties, Macromolecules 32 (1999) 3106–3113. [13] M. Askari, T.S. Chung, Natural gas purification and olefin/paraffin separation using thermal cross-linkable co-polyimide/ZIF-8 mixed matrix membranes, J. Membr. Sci. 444 (2013) 173–183. [14] W. Qiu, L. Xu, C.-C. Chen, D.R. Paul, W.J. Koros, Gas separation performance of 6FDA-based polyimides with different chemical structures, Polymer 54 (2013) 6226–6235. [15] C.Y. Soo, H.J. Jo, Y.M. Lee, J.R. Quay, M.K. Murphy, Effect of the chemical structure of various diamines on the gas separation of thermally rearranged poly(benzoxazole-co-imide) (TR-PBO-co-I) membranes, J. Membr. Sci. 444 (2013) 365–377. [16] E.S. Sanders, Penetrant-induced plasticization and gas permeation in glassy polymers, J. Membr. Sci. 37 (1988) 63–80. [17] 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. [18] S.L. Liu, L. Shao, M.L. Chua, C.H. Lau, H. Wang, S. Quan, Recent progress in the design of advanced PEO-containing membranes for CO2 removal, Prog. Polym. Sci. 8 (2013) 1089–1120. [19] C.E. Powell, X.J. Duthie, S.E. Kentish, G.G. Qiao, G.W. Stevens, Reversible diamine cross-linking of polyimide membranes, J. Membr. Sci. 291 (2007) 199–209. [20] J. Vaughn, W.J. Koros, Effect of the amide bond diamine structure on the CO2, H2S, and CH4 transport properties of a series of novel 6FDA-based polyamideimides for natural gas purification, Macromolecules 45 (2012) 7036–7049. [21] Y. Wang, S.H. Goh, T.S. Chung, Miscibility study of Torlons polyamide-imide with Matrimids 5218 polyimide and polybenzimidazole, Polymer 48 (2007) 2901–2909. [22] A.F. Bushell, M.P. Attfield, C.R. Mason, P.M. Budd, Y. Yampolskii, L. Starannikova, A. Rebrov, F. Bazzarelli, P. Bernardo, J. Carolus Jansen, M. Lanč, K. Friess, V. Shantarovich, V. Gustov, V. Isaeva, Gas permeation parameters of mixed matrix membranes based on the polymer of intrinsic microporosity PIM-1 and the zeolitic imidazolate framework ZIF-8, J. Membr. Sci. 427 (2013) 48–62. [23] N.B. McKeown, P.M. Budd, Exploitation of intrinsic microporosity in polymerbased materials, Macromolecules 43 (2010) 5163–5176. [24] C.G. Bezzu, M. Carta, A. Tonkins, J.C. Jansen, P. Bernardo, F. Bazzarelli, N.B. McKeown, A spirobifluorene-based polymer of intrinsic microporosity with improved performance for gas separation, Adv. Mater. 24 (2012) 5930–5933. [25] N. Du, M.M. Dal-Cin, I. Pinnau, A. Nicalek, G.P. Robertson, M.D. Guiver, Azidebased cross-linking of polymers of intrinsic microporosity (PIMs) for condensable gas separation, Macromol. Rapid Commun. 32 (2011) 631–636. [26] M.M. Khan, V. Filiz, G. Bengtson, S. Shishatskiy, M.M. Rahman, J. Lillepaerg, V. Abetz, Enhanced gas permeability by fabricating mixed matrix membranes of functionalized multiwalled carbon nanotubes and polymers of intrinsic microporosity (PIM), J. Membr. Sci. 436 (2013) 109–120. [27] F.Y. Li, Y.C. Xiao, T.S. Chung, S. Kawi, High-performance thermally self-crosslinked polymer of intrinsic microporosity (PIM-1) membranes for energy development, Macromolecules 45 (2012) 1427–1437. [28] W.F. Yong, F.Y. Li, Y.C. Xiao, T.S. Chung, Y.W. Tong, High performance PIM-1/ Matrimid hollow fiber membranes for CO2/CH4, O2/N2 and CO2/N2 separation, J. Membr. Sci. 443 (2013) 156–169. [29] P. Li, T.S. Chung, D.R. Paul, Gas sorption and permeation in PIM-1, J. Membr. Sci. 432 (2013) 50–57. [30] N. Du, M.M. Dal-Cin, G.P. Robertson, M.D. Guiver, Decarboxylation-induced cross-linking of polymers of intrinsic microporosity (PIMs) for membrane gas separation, Macromolecules 45 (2012) 5134–5139. [31] N. Du, M.D. Guiver, G.P. Robertson, J. Song, Carboxylated Polymers of Intrinsic Microporosity (PIMs) With Tunable Gas Transport Properties, WO 2010/ 124359 A1, 2010. [32] N. Du, G.P. Robertson, J. Song, I. Pinnau, M.D. Guiver, High-performance carboxylated polymers of intrinsic microporosity (PIMs) with tunable gas transport properties, Macromolecules 42 (2009) 6038–6043. [33] A. Bondi, Van der Waals volumes and radii, J. Phys. Chem. 68 (1964) 441–451. [34] W.H. Lin, R.H. Vora, T.S. Chung, Gas transport properties of 6FDA-durene/1,4phenylenediamine (pPDA) copolyimides, J. Polym. Sci. B: Polym. Phys. 38 (2000) 2703–2713. [35] P.S. Tin, T.S. Chung, Y. Liu, R. Wang, S.L. Liu, K.P. Pramoda, Effects of crosslinking modification on gas separation performance of Matrimid membranes, J. Membr. Sci. 225 (2003) 77–90. [36] W.F. Yong, F.Y. Li, T.S. Chung, Y.W. Tong, Highly permeable chemically modified PIM-1/Matrimid membranes for green hydrogen purification, J. Mater. Chem. A 1 (2013) 13914–13925. [37] N. Du, G.P. Robertson, J. Song, I. Pinnau, S. Thomas, M.D. Guiver, Polymers of intrinsic microporosity containing trifluoromethyl and phenylsulfone groups as materials for membrane gas separation, Macromolecules 41 (2008) 9656–9662. [38] W.F. Yong, F.Y. Li, Y.C. Xiao, P. Li, K.P. Pramoda, Y.W. Tong, T.S. Chung, Molecular engineering of PIM-1/matrimid blend membranes for gas separation, J. Membr. Sci. 407–408 (2012) 47–57. [39] L.M. Robeson, Polymer Blends: A Comprehensive Review, Hanser Gardener Publications, Cincinnati, OH, 2007.

130

W.F. Yong et al. / Journal of Membrane Science 462 (2014) 119–130

[40] M. Hasegawa, M. Kochi, I. Mita, R. Yokota, Molecular aggregation and fluorescence spectra of aromatic polyimides, Eur. Polym. J. 25 (1989) 349–354. [41] Y.C. Xiao, L. Shao, T.S. Chung, D.A. Schiraldi, Effects of thermal treatments and dendrimers chemical structures on the properties of highly surface crosslinked polyimide films, Ind. Eng. Chem. Res. 44 (2005) 3059–3067. [42] M. Hasegawa, K. Horie, Photophysics, photochemistry and optical properties of polyimides, Prog. Polym. Sci. 26 (2001) 259–335. [43] L. Shao, T.S. Chung, S.H. Goh, K.P. Pramoda, Polyimide modification by a linear aliphatic diamine to enhance transport performance and plasticization resistance, J. Membr. Sci. 256 (2005) 46–56. [44] D.R. Paul, Gas transport in homogeneous multicomponent polymers, J. Membr. Sci. 18 (1984) 75–86. [45] M. Hasegawa, I. Mita, M. Kochi, R. Yokota, Miscibility of polyimide/polyimide blends and charge-transfer fluorescence spectra, Polymer 32 (1991) 3225–3232.

[46] A.E. Barnabeo, W.S. Creasy, L.M. Robeson, Gas permeability characteristics of nitrile-containing block and random copolymers, J. Polym. Sci.: Polym. Chem. Ed. 13 (1975) 1979–1986. [47] J.H. Kim, S.Y. Ha, Y.M. Lee, Gas permeation of poly(amide-6-b-ethylene oxide) copolymer, J. Membr. Sci. 190 (2001) 179–193. [48] C. Liu, S.T. Wilson, Mixed Matrix Membranes Incorporating Microporous Polymers as Fillers, US Patent, 7,410,525B1, 2008. [49] L. Hao, P. Li, T.S. Chung, PIM-1 as an organic filler to enhance the gas separation performance of Ultem polyetherimide, J. Membr. Sci. 453, 2014, 614-623. [50] A. Bos, I. Pünt, H. Strathmann, M. Wessling, Suppression of gas separationn membrane plasticization by homogeneous polymer blending, AlChE J. 47 (2001) 1088–1093.