Elucidating the effect of chain extenders substituted by aliphatic side chains on morphology and gas separation of polyurethanes

Journal Pre-proofs Elucidating the effect of chain extenders substituted by aliphatic side chains on morphology and gas separation of polyurethanes Afsaneh Fakhar, Morteza Sadeghi, Mohammad Dinari, Mohammadmahdi Zarabadipoor, Rob Lammertink PII: DOI: Reference:

S0014-3057(19)31664-7 https://doi.org/10.1016/j.eurpolymj.2019.109346 EPJ 109346

To appear in:

European Polymer Journal

Received Date: Revised Date: Accepted Date:

14 August 2019 4 October 2019 31 October 2019

Please cite this article as: Fakhar, A., Sadeghi, M., Dinari, M., Zarabadipoor, M., Lammertink, R., Elucidating the effect of chain extenders substituted by aliphatic side chains on morphology and gas separation of polyurethanes, European Polymer Journal (2019), doi: https://doi.org/10.1016/j.eurpolymj.2019.109346

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Elsevier Ltd. All rights reserved.

Elucidating the effect of chain extenders substituted by aliphatic side chains on morphology and gas separation of polyurethanes Afsaneh Fakhar a, Morteza Sadeghi a,*, Mohammad Dinari b, Mohammadmahdi Zarabadipoor a, Rob Lammertink c,** aDepartment

of Chemical Engineering, Isfahan University of Technology, Isfahan 84156-83111, Isfahan, Iran

bDepartment cSoft

of Chemistry, Isfahan University of Technology, Isfahan 84156-83111, Isfahan, Iran

matter, Fluidics and Interfaces (SFI), Faculty of Science and Technology (TNW), University of Twente, P.O. Box 217217 7500 AE Enschede, The Netherlands *[email protected] **[email protected]

Abstract Novel poly(urethane-urea) (PUU) membranes were developed as the aim of a structure-property relationship study to enhance gas permeation. Designing the PUUs was followed by three synthesized chain extenders with different length-alkyl side chains, polytetramethylene glycol, and isophorone and hexamethylene diisocyanates. The longest substituted PUU indicated higher phase separation and lower glass transition temperature. Pure and mixed gas permeabilities of prepared membranes grew as phase separation of PUU material increased, while fractional free volume decreased by lengthening the side chain of the PUUs. The reasons for this event were the migration of the side chains to the surface of hard domain, thereby shielding it and filling the soft/hard interface. Enhanced permeability of materials with longer side chains is attributed to its plasticizing effect. The highest CO2 permeability (287 barrer) was obtained for the longest substituted PUU. The findings revealed an increase in gas permeation without a significant reduction of selectivity by longer substituted PUUs.

Keywords: Polyurethane, aliphatic side chain, gas separation membranes, phase separation, chain mobility

1. Introduction

Polymeric gas separation membranes possessing economic benefits are considered a competitive technology relative to the conventional methods. However, their application is mostly challenged by permeability and selectivity trade-off relations [1-7]. Based on structureproperty relations, significant approaches have been adopted to design polymeric membranes with high efficiency for gas separation [8-12]. Considering CO2 capture, polyurethane (PU) 1

membranes containing polar urethane and urea groups are appropriate candidates. Segmented PUs contain hard domains (formed by diol or diamine chain extender and diisocyanate reactive component) and soft domains (formed by polyol) and are attractive materials due to the possibility of manipulating their gas separation properties via changing the microstructure. Microphase segregation arises in PUs because of thermodynamic inconsistency between hard and soft domains, which alters PU physical and gas separation properties [13, 14]. Since gas transport through polyurethane membranes is the duty of the soft segments, more phase separation between the domains provides higher gas permeability [15]. Phase separation for poly(urethane urea) (PUU) containing urea groups is higher than PU containing only urethane groups. The microstructure and gas separation relationship of polyurethane membranes was explored in other studies via various alterations to chemical structure [9, 16]. The distribution of cyanuric chloride nanoparticles and their amine derivatives in the different segments of polyurethane membranes is another attempt to make a structure-property relationship [17]. Enhanced contribution of hard domains in gas permeation and separation (simultaneous enhanced selectivity and permeability) using chain extenders with aromatic side chains in the hard segment was our achievement in our previous study [18]. Simultaneous increased phase separation and fractional free volume (FFV) using bulkier aromatic side chain-based chain extender could warrant the improved gas separation behavior of the studied bulky PUUs [18]. The results of previous studies indicate that permeability grows with higher urea content and that HDI-based polymers give more phase segregation and higher permeability than IPDI-based ones [18-20]. One approach is based on increasing the free volume of the polymers. Some studies suggest that incorporation of side or bulky groups decreases the chain packing density and causes higher gas permeability [5, 21-24]. In contrast, using side groups in the structure of some polymers resulted in decreased FFV value [25-27]. The increase in chain mobility by lengthening the side chain was expressed as a main reason for enhanced gas permeability in such polymers 2

[26]. Beside FFV value and phase separation extent, chain mobility of hard domains would impress gas diffusivity, permeability, and gas separation behavior of polyurethanes. One example is the increase in phase separation rate by using linear diisocyanate compared to aromatic diisocyanates, as reported by Hood et al. [19, 28, 29]. Then, it seems that changing the structure of hard segments in order to change these parameters can determine PU gas separation properties. To elucidate the side chain influence on polyurethane phase separation, Tan et al. reported the migration of fluorinated side chain of hard domain possessing low surface energy to the surface. This phenomenon led to reduced and increased phase separation of fluorinated poly(carbonate urethane) and poly(ether urethane), respectively [30-32]. The weak intermolecular interactions between aliphatic hydrocarbon and fluorocarbon pendant groups are reflected in surface tension, resulting in shielding the main chain [33]. Regarding the polymer structure-controlled gas separation behavior, the main intention of this study is to increase the knowledge of the structureproperty relation of polyurethane membranes, thereby achieving higher performance membranes. It is well established that more permeable polymers are preferred for flue gas separation because of the low pressure differential arising from the high volume emission of flue gas from power plants [34]. Since the hard segments are impermeable for gas and provide only the mechanical strength of polyurethane materials, the main purpose of this research is to alter the structure of the hard domains to enhance chain mobility and phase separation favoring gas permeation and also to lighten their inhibitory impact for gas transport [15]. For this purpose, the introduction of aliphatic hydrocarbon side chains to the hard domains with high affinity to the soft segment was used. The extensive studies on gas permeation reveal that many different factors including hole size, free volume, chain flexibility, interchain interactions, diffusion barriers, and chemistry of the free volume walls would impress gas permeation behavior [35-40]. The presence of polar and H-bridging groups in polymer structure, which enhance interchain interactions, 3

induces additional energy barriers for gas diffusion. Therefore, using hydrocarbon aliphatic side chains is seemed to be a good idea for reducing interchain interactions and also for giving rise to higher chain flexibility [36]. Herein, as a continuation of our previous study [18], three novel chain extenders containing different-length aliphatic side chains, instead of the aromatic side chain based-chain extenders in the previous study, were used to synthesize polyurethanes. The influence of these groups on gas separation and mechanical properties was investigated. The novelty of our study is the preparation of novel structures and a proposed mechanism for the structure-property relationship. We believe understanding the structure-property relationship is a fundamental approach to improve gas separation performance of membranes including polyurethanes. This paper attempts to elucidate this relationship of the novel PUU membranes by the purpose of gas permeability increment. 2. Experimental 2.1 Materials

Hexylamine, dodecylamine and octadecylamine, cyanuric chloride (CC), sodium hydrogen carbonate, and hydrazine were provided by Merck to synthesize three bulky chain extenders containing different side chain lengths. Ameretat Shimi Pharmaceutical Co. supplied acetone and dichloromethane. Poly(tetramethylene glycol) with molecular weight of 2000 g/mol (PTMG2000) was supplied by Arak Petrochemical. Isophorone diisocyanate (IPDI), hexamethylene diisocyanate (HDI), dibutyltin dilaurate (DBTDL) as a catalyst, dehydrated dimethyl acetamide (DMAc), and dimethyl formamide (DMF) were supplied by Merck. The linear chain extenders including 1,4-butanediol (BDO) and 1,4-butanediamine (BDA) were provided by Sigma-Aldrich.

4

2.2 Synthesis of chain extenders A two-step reaction was served to synthesize three bulky diamine chain extenders. As the first step, one of the chlorine atoms of triazine ring of CC was substituted by a linear amine group (three different amines with different lengths), and in the second step, other two chlorine atoms of triazine ring were substituted by reacting with hydrazine [41] (schemes 1-3). The detailed procedures to synthesize the chain extenders are as follows: Hexylamine-based chain extender: To the solution of 1.30 g (15.5 mmol) sodium hydrogen carbonate in 40 ml of deionized water upon cooling using an ice-water bath, 2.86 g CC (15.5 mmol) in acetone (50 ml) was added. Afterwards, 1.57 g (15.5 mmol) hexylamine dissolved in 7 ml acetone was added via dropwise addition. The solution was kept for 1.5 hours below 2°C, after which the reaction was continued for 12 hours at ambient temperature. The solid was collected and washed thoroughly with deionized water [42]. Excess hydrazine reacted with the product overnight under reflux as the second step [43, 44]. Then the solid product was filtered and washed (Scheme 1). Dodecylamine-based chain extender: To the stirred and cooled suspension of 2.84 g CC (15.4 mmol) and 1.554 g sodium bicarbonate in dichloromethane (50 ml), 2.86 g dodecylamine (15.4 mmol) dissolved in dichloromethane (10 ml) was added dropwise. After two hours, the ice bath was removed, and the reacting mixture was stirred at ambient temperature overnight. The collected solid was washed thoroughly by deionized water to remove unreacted reactants. The reaction of the product with hydrazine was the next step performed as mentioned above (Scheme 2) [45]. Octadecylamine-based chain extender: The procedure of this product was performed as with the dodecylamine-based chain extender. For this reaction 4.16 g (15.4 mmol) of

5

octadecylamine was consumed for the same values of sodium bicarbonate and CC used to synthesize dodecylamine-based chain extender (Scheme 3).

NH(C5H10)CH3 Cl N Cl

N N

0-2 ◦C, Acetone/water Sodium bicarbonate

+ C6H13NH2 Cl

Cyanuric chloride

NH(C5H10)CH3 N Cl

N

N

Hydrazine, reflux

N

N

N HN NH2

Cl

NH NH2

Hexylamine-based chain extender

Hexylamine

Scheme 1. Synthesis route for hexylamine-based chain extender

NH(C11H22)CH3 Cl N Cl

0-2 C, Dichloromethane/water Sodium bicarbonate ◦

N N

+ C12H25NH2 Cl

Cyanuric chloride

NH(C11H22)CH3 N N

Cl

N

Hydrazine, reflux

N

N

HN N NH2

Cl

NH NH2

Dodecylamine-based chain extender

Dodecylamine

Scheme 2. Synthesis route for dodecylamine-based chain extender NH(C17H34)CH3 Cl N Cl

N N

+ C18H37NH2 Cl

Cyanuric chloride

0-2 ◦C, Dichloromethane/water Sodium bicarbonate

Cl

NH(C17H34)CH3 N

N N

N

Hydrazine, reflux

N HN NH2

Cl

N NH NH2

Octadecylamine-based chain extender

Octadecylamine

Scheme 3. Synthesis route for octadecylamine-based chain extender

2.3 Polyurethane synthesis

A two-step method of polymerization was applied to synthesize all polyurethanes. First, a vacuum oven was used to remove any residual water from PTMG for 24 hours at 80°C. Prior to the synthesis, all equipment was dried. A round-bottom vessel with an impeller (anchor-type) and temperature controlled under nitrogen were used for the reaction. In order to prepare the macro-prepolymer, PTMG was reacted with excess diisocyanate (PTMG: IPDI/HDI 1:3 molar ratio) at 80°C for 2 h in the presence of catalyst (DBTDL). Afterwards, the chain extenders (dissolved in DMAc) were added to the vessel at room temperature. The reaction between the prepolymer and chain extender was continued overnight under nitrogen at 70°C. It is notable that the NCO:OH molar ratio was kept at 1:1 to prepare linear polymer and the PTMG: diisocyanate: 6

chain extender molar ratio was selected as 1:3:2 [16]. Two polyurethanes based on BDO/IPDI and BDA/IPDI was prepared as materials for reference membranes. PUU-α-β notation was used for the synthesized polyurethanes, where α and β are assigned to chain extender and diisocyanate types, respectively (α: H, D and O for hexylamine-, dodecylamine-, and octadecylamine-based chain extenders, β: I and H for IPDI-based and HDI-based polymers, respectively). The notations of PU and PUU were applied for PTMG/BDO/IPDI- and PTMG/BDA/IPDI-based polyurethane and poly(urethane-urea), respectively. Scheme S1 in Supplementary Information presents the polyurethanes’ chemical structures. 2.4 Characterization of chain extenders The products of the two-stage reaction toward chain extenders’ preparation were characterized by FTIR (ALPHA Bruker) (wave number range: 4000–400 cm-1) at room temperature by 128 scans. In addition, 1H NMR (DMSO) and elemental analysis were used to characterize the chain extenders. Table S1 presents the elemental analysis data (Supplementary Information). 2.5 Fabrication and characterization of membranes Polyurethane membranes were fabricated by casting the solutions of 10 wt% synthesized polymers in DMF followed by evaporation at 60°C for one day. To completely remove DMF, the membranes (thickness ≅ 100 μm) were kept in a vacuum oven at 60°C overnight. ATR-FTIR was applied to the membranes using an ALPHA Bruker spectrometer (wave number range: 400−4000 cm−1). Solubility parameters of soft and different hard segments and also three side chains of different lengths were calculated to estimate thermodynamic tendency of the synthesized PU and PUUs to phase segregation, theoretically (Table 2). To explore membrane’s gas separation behavior, FFV was estimated (Table 2) [46]. The equations used for estimation of solubility parameters (δ) and FFV values were defined in detail in our previous study [18]. The

7

density of different polyurethanes was measured by a Micromeritics AccuPyc II 1340 Gas Pycnometer under helium atmosphere. Differential scanning calorimeter (DSC 8000, PerkinElmer) was used to study the polyurethanes’ thermal behavior under nitrogen in temperature range -100°C to 200°C with 20°C/min heating rate (after removing thermal history) . Tensile measurements (Instron 5942) were done at ambient temperature at a 5 mm/min rate to explore the mechanical behavior of the membranes. A specimen’s gauge length of 10 mm was used. Tensile measurements were repeated at least 5 times on the membranes and the average data are reported. Pure and mixed gas permeation tests were done on the membranes at room temperature and 4 bar. Single gas measurements were performed at different pressures to specify optimal pressure for the gas permeability study. Pressure of 4 bar, below the plasticization of the membranes, was applied in the main measurements. Mixed gases of CO2/N2 and CO2/CH4 containing various compositions (50:50, 10:90, and 30:70 vol%) were used to study the gas composition effect on gas transport performance of the membranes at 30°C and 4 bar. For comparative study of the partial pressure and feed pressure effects in the mixed gas study, 50:50 mixed gas permeation tests were also performed at 8 bar (partial pressure of 4 bar). The measurements were taken three times on three samples of each type of polyurethane membrane and the average data are presented. All equations used for these measurements were defined in our previous study [18]. 3. Results and discussion 3.1 Characterization of chain extenders

Fig. S1 indicates the FTIR spectra of products derived from the two-stage reaction toward chain extenders. The presence of characteristic peaks confirms the correctness of the chemical reaction done in each step. Fig. S2 and Table 1 illustrate 1H NMR data for the synthesized 8

chemical structures. For the first step of the reactions, the hydrogens of the NH group on triazine ring attached on linear hydrocarbons appear at 9.0-9.5 ppm for hexyl- and dodecylamine-based chain extenders and at 7.5-8.0 ppm for the octadecylamine-based chain extender. The hydrogens of the aliphatic chains are present at 0.5-3.5 ppm. In the second stage, the hydrogens at 6.5-8.0 and 4.0-4.5 are attributed to NH and NH2 groups, respectively, on the triazine ring. The hydrogens of NH groups on the triazine ring attached on aliphatic hydrocarbons shift to 10-10.5 ppm in the next step. Consequently, the resulting chemical structures of the products are affirmed by the presence of characteristic hydrogens in the NMR spectra. It should be noted that due to the insolubility of the final product with the octadecylamine-based chain extender reaction in DMSO at room temperature, NMR study of this product was not possible. In addition, the elemental analysis (Table S1, Supplementary Information) revealed good accord between experimental and theoretical data, confirming the resultant structures as well.

Table 1. 1H-NMR data of the aliphatic side chain substituted chain extenders (both reaction steps) Hydrogen NH attached on hydrocarbon

Hexylamine-based chain extender (ppm) Stage 1 Stage 2 9-9.5 10-10.5

Dodecylamine-based chain extender (ppm) Stage 1 Stage 2 9-9.5 10-10.5

Octadecylamine-based chain extender (ppm) Stage 1 Stage 2 7.5-8 No data

NH on triazine ring

-

6.5-8

-

6.5-8

-

No data

NH2 on triazine ring

-

4-4.5

-

4-4.5

-

No data

3.2 Polyurethanes characterization 3.2.1 FTIR analysis Fig. 1a presents the FTIR results of the studied polyurethanes. The reaction completion can be verified by the disappearance of the NCO peak at 2250 cm-1. The C=O stretching band at 1600-1800 cm-1, the NH stretching band at about 3300 cm-1, the ether (C-O-C) band at around 1110 cm-1, and the CH2 peaks of polytetramethylene glycol at 2860 and 2940 cm-1 as the basic characteristic bands demonstrate the polyurethane structure. The characteristic peaks which 9

appear at 2850-2920 cm-1 are related to the (-CH2)aliph signals of linear amines with different lengths. There are two types of hydrogen bonding in polyurethanes: hydrogen bonding of C=O and NH groups existing in the hard domains, and hydrogen bonding of the hard domain’s NH groups with the soft segment. The first one enhances phase segregation between soft and hard domains, increasing the storage modulus, while the second results in phase mixing [47, 48]. The carbonyl groups interacting with NH groups by hydrogen bonding (C=O,bonded) and the free carbonyl groups (C=O,free) appear at lower and higher wave numbers, respectively. Phase separation degree can be predicted by the hydrogen bonding index (HBI): 𝐻𝐵𝐼 =

𝐴𝐶 = 𝑂, 𝑏𝑜𝑛𝑑𝑒𝑑

(1)

𝐴𝐶 = 𝑂, 𝑓𝑟𝑒𝑒

In Equation 1, AC=O,bonded and AC=O,free refer to the bonded and free carbonyls’ absorbance, respectively. Higher hydrogen bonding of the hard domain’s carbonyl and NH groups can be obtained by polyurethane with increased HBI, meaning low interaction between the soft domain’s ethereal oxygen and urethane or urea NH groups [49]. Table 2 illustrates the HBI values for all synthesized polyurethanes. In order to accurately estimate the HBI values, the deconvolution of two overlapped free and bonded carbonyl peaks in FTIR spectra was done by Origin software, and the area ratio of the peaks was considered as the HBI value [15]. Fig. S3 displays the deconvolution data. Phase separation in both groups of synthesized polyurethanes increases by the length of the aliphatic side chains in the hard domains (Table 2). It is well established that segmental incompatibility between two domains of polyurethane would result in microphase separation [48]. In both categories of diisocyanates, octadecylamine-based PUUs show higher phase separation, although the calculated solubility parameter difference between soft and hard segments reported in Table 2 shows lower difference (more difference between two domains’ solubility parameters means more incompatibility between them). The aliphatic side chains are not compatible with the main chain in the hard

10

segments (by comparing their solubility parameters in Table 2), so it may be concluded that phase segregation would occur in the poly(urethane-urea) structures even between main and side chain parts in hard domains [50, 51]. Since the HBI only relates to the extent of free and bonded carbonyl groups, it is intriguing how the longer chain side enhances the HBI value. As reported in the literature, the side chains with low surface energy can migrate to the surface of hard segment main chain [30-32]. To explore this, the surface energy of different side chains and main chains for both diisocyanate-based PUUs were estimated based on group contributions using this equation [52]: 2 3

𝛾 ≈ 0.75𝑒𝑐𝑜ℎ = 0.75

𝐸𝑐𝑜ℎ

2 3

( )

(2)

𝑉

where γ, ecoh, Ecoh, and V are surface tension (MJ/m2), cohesive energy density (MJ/m3), cohesive energy (J/mol), and molar volume (cm3/mol), respectively [52] (Table 3). It is important to note that errors in surface tension calculations should not severely affect the comparison of various pendant groups due to the presence of the same chemical groups in their structures. It is suggested that the aliphatic substituents would stick out of and shield the rigid main chain of the hard segments (Scheme 4). Since these aliphatic pendants are intrinsically hydrocarbons, the weak intermolecular interactions between them are reflected in their low surface tension [33]. The shielding effect of aliphatic substituents on the main chain of hard segment would result in less accessible NH groups in the hard domains to make hydrogen bonding with ethereal groups of the soft segment and, as a result, a higher degree of phase segregation. As evident from Tables 2 and 3, HBI value increased as the side chain lengthened due to less surface tension and higher flexibility of longer side chains. Additionally, close solubility parameters of the side chains and soft domain suggests that aliphatic side chains are compatible with the soft segments. This causes the aliphatic side chains to act as emulsifiers, which likely gives rise to filling the hard/soft domain interfaces with side chains. It leads to covering the functional groups of hard domains 11

with aliphatic side chains, which can confirm the proposed shielded hard domain structure; thereby, higher phase separation of longer side chain substituted PUUs is observed. It seems that along with the higher phase separation degree, soft and hard phases with higher purity are consequently obtained.

(b)

(a)

Fig. 1. a) FTIR spectra and b) DSC curves of all synthesized polyurethane materials

Table 2. Some measured or calculated parameters of all obtained polyurethanes Sample

HBIa

Tg (°C) b

Hard segment (%)a

Solubility parameter (δ) (J/cm3)0.5 a Soft segment

Total

Main chain

Density (g/cm3) b

FFVa

Hard segment Total

Main chain

Side chain

PU

0.07

-76.79

29.75

29.75

18.411

20.91

-

-

1.1453

0.13

PUU

1.17

-77.25

29.66

29.66

18.41

23.90

-

-

1.1525

0.12

PUU-H-I

2.96

-79.50

36.46

30.14

18.41

24.05

25.43

17.86

1.1163

0.19

PUU-H-H

3.12

-80.77

33.00

26.34

18.41

26.47

28.65

17.86

1.0974

0.23

PUU-D-I

3.05

-80.03

39.68

28.61

18.41

23.21

25.43

17.70

1.1316

0.15

PUU-D-H

3.86

-81.62

36.58

24.93

18.41

25.23

28.65

17.70

1.1073

0.19

PUU-O-I

3.69

-81.80

42.60

27.23

18.41

22.57

25.43

17.64

1.1404

0.13

PUU-O-H

4.12

-82.44

39.79

23.67

18.41

24.30

28.65

17.64

1.0869

0.18

1

This parameter has been expressed as 17.6 or 19.13 (J/cm3)0.5 [52, 53]. a

These parameters were calculated.

b

These parameters were measured.

12

Table 3. Surface tension of soft and hard domains of the synthesized PUUs Sample

Surface tension (MJ/m2) Soft segment

Hard segment Main chain

Side chain

PUU-H-I

36.45

56.09

35.00

PUU-H-H

36.45

65.74

35.00

PUU-D-I

36.45

56.09

34.61

PUU-D-H

36.45

65.74

34.61

PUU-O-I

36.45

56.09

34.46

PUU-O-H

36.45

65.74

34.46

The HBI values of all synthesized PUUs are higher than those of the linear polyurethanes (PU and PUU), which agrees with the greater difference between the solubility parameter of their hard segment main chains and soft segment. Furthermore, PUU containing more NH and urea groups yielded higher HBI compared to PU. Polyurethanes containing longer side chains in their hard segments for both used diisocyanate show lower FFV values (Table 2), despite the greater phase segregation between soft and hard segments [54, 55]. A decrease in FFV with increased side chain length for different polymers has been reported by others, as well [26, 27, 56]. Alkyl substituent incorporation to polymer backbone can lead to an increase or decrease of the chain packing density and phase separation extent. The incorporation of polyesters with various –CH3 side chain contents in polyurethane resulted in higher phase separation and free volume [57]. Lengthening the side chain in methacrylate copolymers increased FFV, whereas for polynorbornenes, FFV decreased [25, 58]. Reduced FFV values using longer side chains for the studied PUUs could be attributed to some simultaneous effects of the aliphatic side chains. As previously evaluated, the incompatibility of aliphatic side groups with hard domain main chains leads to side chains’ 13

migration to the surface of the hard domains. This phenomenon may cause the following results in polymer morphology: a) The main chains of hard segments are forced to order in hard domains. b) The interface of the hard and soft domains would be filled by aliphatic side chains which are interacted with soft domain chains, due to the compatibility of aliphatic side chains and the soft segment. c) Shielding the hard domains by the aliphatic side chains reduces the access to NH groups for ether links of polyol, which leads to more purity of both phases. All probable effects mentioned could increase the phase separation and reduce FFV, simultaneously. Based on Table 2, for the same chain extender, HDI-based PUUs resulted in higher HBI than IPDI-based PUUs, which is consistent with previous researches. The linearity of HDI compared to IPDI can reduce the hindrance effect of chains and lead to higher chain packing in HDI-based PUs. This results in more hydrogen bonding in hard domains of HDI-based polyurethane, which reduces hydrogen bonding between hard and soft domains and consequently results in higher HBI and higher phase-separated structures. The FFV values of HDI-based PUUs are higher than IPDI-based PUUs with the identical chain extender, which agrees with the typical behavior based on higher FFV for higher phase-separated polyurethanes.

14

Scheme 4. The proposed shielding effect of aliphatic side chains on the main chain of hard segments.

3.2.2. DSC analysis The thermal behavior of all polyurethanes is presented in Fig. 1b and the Tg of soft segments for different polyurethanes are reported in Table 2. PUUs having chain extenders containing longer aliphatic side chains and enhanced phase separation possess slightly lower Tgs than do shorter side chains. The glass transition temperatures of HDI-based PUUs are lower compared to IPDI-based PUUs, consistent with FTIR results. Higher phase separated polyurethanes have less Tgs for their soft domain because of the larger molecular mobility caused by phase separation [59, 60]. The substituted PUUs showed lower glass transition temperatures compared with PUU and PU, which confirms their stronger phase-separated structures. Also, this can be referred to as covering the hard domain functional groups at hard/soft phases, which reduces the connection of mobile soft chains to rigid hard chains. Regarding PU and PUU (linear polymers), the presence of urea linkages in PUU promoting phase segregation leads to lower Tg compared to PU. 3.2.3. The analysis of mechanical properties

15

Tensile analysis was performed to assess the mechanical behavior of the polyurethanes, the obtained data are presented in Table 4. Unlike soft domains which provide chain flexibility, hard domains are responsible for mechanical performance. Different mechanical behaviors of the synthesized polyurethanes can be assigned to the different types of hard domains. As the aliphatic side chain length increases, modulus and strength decrease in both categories of diisocyanate. Longer substituents act as plasticizers [61].

Table 4. Mechanical properties of studied membranes based on tensile analysis Young’s modulus

Ultimate tensile strength (UTS)

Elongation at break

(MPa)

(MPa)

(%)

PU

4.63±0.35

4.13±0.19

180.15±6.62

PUU

5.37±0.65

7.29±0.57

499.27±12.16

PUU-H-I

11.65±0.45

4.88±0.32

414.73±13.20

PUU-H-H

9.10±0.30

2.58±0.15

203.40±7.34

PUU-D-I

10.58±0.33

3.75±0.28

805.20±19.15

PUU-D-H

8.95±0.18

2.45±0.19

510.4±14.41

PUU-O-I

9.96±0.27

3.04±0.25

907.47±23.10

PUU-O-H

8.66±0.17

2.24±0.22

621.78±20.77

Sample

Octadecylamine-based chain-extended PUUs exhibited lower UTS and Young’s modulus but higher elongation for both categories of diisocyanate because of greater plasticizing effect. Despite the direct relationship between phase segregation and mechanical performance, reported by others [62, 63], HDI-based polyurethanes with higher phase segregation indicated weaker mechanical properties than IPDI-based PUUs in tensile test. Stronger phase separation may reduce the number of physical crosslinks between two domains. Higher degree of phase separation results in enhanced segmental motions, lowering the modulus and UTS. Fu et al. reported enhanced mechanical properties for their studied polyurethane until HBI=3, but found a reverse trend for higher HBI values [62]. 16

Another parameter affecting the mechanical performance of polyurethanes is the content of the hard domain. The hard domains provide the mechanical strength of polyurethanes, so the modulus and tensile strength would be improved with hard segment content [59]. IPDI-based PUUs contain more hard segment contents than HDI-based PUUs with the same chain extenders, which explains their improved mechanical properties. The longer side chain–based polyurethanes (e.g. PUU-O-I and PUU-O-H) containing more hard segment content are expected to exhibit enhanced mechanical performance; however, the plasticizing effect of the longer side chains leads to a decrease in mechanical performance of the PUUs. Also, it is evident from Table 2 that longer side chain-based PUUs have lower main chain content of hard segments, which is responsible for providing mechanical properties. Due to the high affinity of side chains to the soft domain, the overall soft segment content would increase by lengthening the side chain, which is reflected in enhanced elongation at break and reduced modulus. Also, the greater purity of soft and hard phases can be concluded as more disjunction (detachment) of hard and soft domains. 3.2.4. Gas permeation analysis The gas permeation through the membranes was defined with pure O2, N2, CH4, and CO2 at 30°C and various pressures (Fig. S4). Also, CO2/CH4 and CO2/N2 mixed gas permeation measurements were performed at 4 bar and 30°C. To analyze the effect of upstream pressure in single gas permeation and partial pressure (4 bar) in mixed gas permeation measurement on permeability and selectivity, mixed gas permeation measurements with the composition of 50:50 vol% were also performed at 8 bar. The results of single gas permeation measurements at 30°C and 4 bar are reported in Table 5. The decreasing order of permeabilities through the membranes is: P(CO2) >> P(CH4) > P(O2) > P(N2). Gas transport in dense membranes can be defined by the solution-diffusion model [64, 65]. The higher CO2 permeation than other studied gases arises from its lower kinetic diameter 17

and higher sorption to the polymer [20, 66]. The higher CH4 permeability than O2 and N2 (despite the larger kinetic diameter of CH4) confirms the dominant solution mechanism for permeation through the polyurethane membranes, with regard to different gas condensabilities expressed in literature such as [67]. The gas permeation through various polyurethanes indicates that for both diisocyanates, the polyurethanes having longer side chains in the chain extender structures provide higher permeabilities. This complies with stronger phase separation between their hard and soft domains, as discussed above. As mentioned earlier, gas permeation is mainly governed by the soft domain-rich volume of polyurethane membranes. Higher overall soft segment content (including side chain content) in longer substituents-based PUUs due to the higher affinity of longer side chains to the soft domain, and their plasticization effect promotes gas permeation. By lengthening the side chain from hexyl to dodecyl and octadecylamine by 6 and 12 methylene, respectively, the growth in CO2 permeability for IPDI-based PUUs toward PUU as a linear poly(urethane-urea) is in following order: PUU-O-I (198.76 barrer: 82%) > PUU-D-I (182.36 barrer: 67%) > PUU-H-I (179.97 barrer: 64%) At first glance, lower FFV in longer aliphatic side chain–based PUUs with higher gas permeability displays a contradiction between polymer structure and permeation properties. However, other research results show that increase in phase separation, chain mobility of soft segment (Tg reduction), and purity of the phases are also impressive factors on the resultant permeation properties of polyurethane membranes [15, 19, 68]. Additionally, Thran et al. [38] believe that FFV as a non-optimum property of polymer cannot anticipate gas permeation through polymers alone, which is confirmable by other contributing factors illustrated by others [35-37]. Enhanced gas permeability through longer side chain-substituted PUUs reveals the 18

dominant plasticizing effect by the aliphatic substituents and microphase separation influence on gas permeation over reduced FFV values. Table 5. Single gas separation performance of all prepared membranes at 30 °C and 4 bar. Sample

Permeability (barrer)

Selectivity

N2

O2

CH4

CO2

CO2/CH4

CO2/N2

O2/N2

PU

3.2±0.2

7.9±0.3

10.2±0.5

85.3±1.0

8.4±0.3

27.0±0.3

2.5±0.02

PUU

5.00±0.1

11.3±0.3

14.0±0.7

109.4±3.7

7.8±0.4

21.9±0.2

2.3±0.05

PUU-H-I

5.5±0.2

15.6±0.4

19.6±0.4

180.0±1.5

9.1±0.2

32.5±0.3

2.8±0.03

PUU-H-H

8.2±0.2

22.0±0.3

25.9±0.5

231.0±2.8

8.9±0.1

28.1±0.3

2.7±0.05

PUU-D-I

5.7±0.2

15.9±0.3

20.0±0.3

182.4±1.9

9.1±0.2

32.1±0.4

2.8±0.02

PUU-D-H

8.5±0.2

21.9±0.3

27.6±0.6

234.9±2.6

8.3±0.3

27.5±0.2

2.6±0.04

PUU-O-I

7.0±0.2

19.0±0.4

24.1±0.6

198.8±2.4

8.2±0.2

28.3±0.4

2.7±0.06

PUU-O-H

10.5±0.3

25.1±0.5

35.0±0.7

286.5±5.6

8.0±0.2

27.2±0.1

2.4±0.03

By comparing IPDI and HDI-based membranes, it is obvious that the HDI-based membranes provide higher permeabilities, likely due to their higher phase segregation and lower Tg values. The higher permeability through PUU compared to PU originates from more NH and urea groups in PUU, which provide stronger hydrogen bonding reactions in hard domains and consequently higher phase segregation (HBI value). The higher chain mobility of PUU based on Tg values also implies the higher gas permeability of PUU. All prepared PUUs containing alkylsubstituted chain extenders are more permeable than linear PU and PUU membranes. This can be attributed to their higher HBI and lower glass transition temperatures, caused by the shielding effect, and the filled soft/hard interface by the long side chains, which promote the phase segregation degree between two domains. It seems that for octadecylamine-based PUUs, possessing lower FFV, the plasticizing and shielding effects of the long side chain are crucial for increasing gas permeability. According to the Shahrooz et al. study on polyurethane gas separation membranes, two important factors which would affect transport properties are the volume fraction and purity of the phases [15]. Longer side chain–based PUUs would induce a

19

higher fraction of the soft segment as the permeable domain and the likely higher purity of the phases favoring gas permeability. The CO2/N2 selectivity in comparison to other pair gases is the highest for all membranes due to the considerable difference of N2 and CO2 kinetic diameters and higher critical temperature of CO2. Since CO2 and CH4 are both condensable, CO2/CH4 selectivity is lower than that of CO2/N2. Due to the approximately equal and low condensabilities of O2 and N2, O2/N2 selectivity is the lowest selectivity. The selectivity decreased slightly as permeability increased. This accords with the tradeoff between permeability and selectivity based on Robeson’s report. It can be concluded that the presence of longer alkyl side chains improves gas permeability but reduces the screening character of the membrane due to the plasticizing effect caused by the long aliphatic substituent. Since the phase separation degree is a crucial factor affecting selectivity through polyurethanes, polyurethanes with less phase segregation (lower HBI) present higher molecular sieving properties and thus higher selectivities [19, 20]. To explore this further, diffusivities and solubilities, and diffusivity and solubility selectivities for all membranes, were estimated using the time lag method [69] (Table 6). Fig. S5 displays the gas diffusivity against kinetic diameter and HBI values, and solubility against critical temperature. Since gas diffusion of smaller molecules is faster in the membrane, gas diffusivity ensures the molecular sieving character of the membranes. Moreover, polyurethanes with a higher extent of phase separation gave higher diffusivity coefficients because of enhanced chain mobility. The calculated solubility coefficients agree with condensability of the gases, i.e. higher solubility coefficients for gases having higher critical temperatures. Because of the trade-off between permeability and selectivity, higher permeable membranes are expected to indicate lower selectivity [19, 70]. According to Table 6, the gas

20

diffusivity increased but diffusivity selectivity decreased by longer substituents, consistent with the higher plasticizing effect of longer side chains. In contrast, a small decrease in solubility coefficient is evident as the NH groups of the main chain of hard segments would be less accessible for CO2 sorption as a result of reduced FFV for longer side chain extender–based PUUs. In the case of other gases, the reduced FFV values would be the reason for the decreased solubility coefficient [22]. Hard domains containing polar functional groups have higher affinity toward the studied gases than soft segments and likely contribute to gas sorption [71, 72]. Since gas permeation is provided by the soft domain-rich area of polyurethane membrane, the presence of hard domains in soft domains gives rise to increased gas sorption n [15]. The reduced gas sorption in longer side chain–based PUUs implies the higher purity of the soft phase. Increased purity of soft segments enhances the mobility of these domains, which promotes gas diffusivity (decreases diffusivity selectivity). In contrast, the small improvement in CO2/N2 and CO2/CH4 solubility selectivities as the side chain is lengthened can be assigned to more CO2 sorption in competition with CH4 and N2 in higher chain-packed PUUs. On the other hand, CH4 and N2 with larger kinetic diameters are more restricted to be absorbed into PUUs having lower FFV. Since O2/N2 separation as non-condensable gases is only provided by diffusivity selectivity, the more reduced selectivities of CO2/CH4 and CO2/N2 in comparison with O2/N2 of longer side chain– based PUUs (Table 5) confirm more shielding and less accessible NH groups for CO2 sorption. No significant change in O2/N2 solubility selectivity by lengthening the substituents confirms the diffusion-dominant mechanism for this separation. The reduced diffusivity selectivity of O2/N2 illustrates the reduced screening character of the PUUs with longer side chain (Table 6).

Table 6. Diffusivities and solubilities, diffusivity and solubility selectivities through the prepared membranes at 30 °C and at 4 bar. Sample

N2

𝟕 Diffusivity ×𝟏𝟎

𝟑 Solubility×𝟏𝟎

(cm2/s)

(cm3 (STP)/cm3poly.cmHg)

O2

CH4

CO2

N2

O2

CH4

Diffusivity selectivity

CO2

21

O2/N2

CO2/CH4

CO2/N2

Solubility selectivity

O2/N2

CO2/CH4

CO2/N2

PU

3.8

7.3

3.1

5.2

0.8

1.1

3.3

16.3

1.9

1.7

1.4

1.3

4.9

19.9

PUU

5.7

11.2

4.5

6.1

0.9

1.0

3.1

18.0

1.9

1.4

1.1

1.1

5.8

20.0

PUU-H-I

6.0

11.5

5.3

9.5

0.9

1.4

3.7

18.9

2.1

1.8

1.6

1.5

5.1

20.5

PUU-H-H

9.1

15.1

6.6

10.4

0.9

1.5

3.9

22.2

1.7

1.6

1.1

1.6

5.77

24.7

PUU-D-I

7.2

13.5

6.7

10.5

0.8

1.2

3.0

17.3

1.9

1.6

1.5

1.6

5.9

21.9

PUU-D-H

10.2

16.6

9.7

11.5

0.8

1.3

3.5

20.6

1.6

1.2

1.6

5.9

24.8

PUU-O-I

9.4

15.8

8.6

11.7

0.7

1.2

2.8

16.9

1.8

1.4

1.6

6.0

22.6

PUU-O-H

12.9

19.1

11.6

14.0

0.8

1.3

3.0

20.4

1.5

1.2

1.6

6.8

24.9

1.1 1.3 1.1

Since polyurethanes are desirable membranes for separation of CO2/N2, the capability of the aliphatic side chain–based membranes for CO2 separation was compared with other polyurethane membranes, the Pebax family, PU/Pebax, and Pebax/PEG copolymers as rubbery membranes from the literature and the Robeson’s upper bound limit in Fig. S6 (the same reference data of the previous study were used [18]). The membranes containing aliphatic side chains appear closer to the Robeson line than do linear ones (PU and PUU). Although slightly lower selectivities are obtained, these membranes provide high permeabilities. Fig. S6 indicates the suitability of the prepared PUUs among other polyurethanes and of Pebax as a commercial rubbery copolymer. 3.2.5. Mixed gas measurements Since single gas measurements do not provide realistic gas separation ability of membranes for practical purposes, mixed gas measurement was performed on the prepared membranes [73]. The mixed gas separation results are illustrated in Figs. 2 and 3. Plasticization and penetrant competition can lead to reduced mixed-gas selectivity compared to single gas measurement [74-76]. Enhanced gas permeability along with significant reduced selectivity defines the plasticization [77, 78]. It is evident from Figs. 2 and 3 that the selectivities of CO2/CH4 and CO2/N2 in the case of mixed gases including various compositions are less than those of single gas data because of plasticization by CO2 and penetrate competition of both pair gases. CO2/CH4 and CO2/N2 selectivities decreased as upstream pressure increased from 4 to 8 bar as a result of stronger plasticization effect by CO2 at 8 bar, leading to decreased screening 22

character of the membrane. By increasing the feed pressure, a decrease or increase in gas permeability would be obtained by competitive sorption or plasticization, respectively [79]. Enhanced N2 permeability by increasing upstream pressure during CO2/N2 mixed gas measurement (50:50 vol%) originates from the plasticization effects by CO2. However, the two mentioned opposing effects caused no significant effect on CO2 permeability as upstream pressure increased. No change in CO2 permeability along with increase in N2 permeability by increasing the upstream pressure results in the reduction of CO2/N2 selectivity (Fig. 2). The selectivity variation trend by changing the gas composition for some membranes is as follows: CO2/N2 (50:50) < CO2/N2 (10:90) < CO2/N2 (30:70). Higher CO2/N2 (30:70) selectivity than CO2/N2 (10:90) is found as the predominantly preferable driving force of CO2 permeation over plasticization effect. The same trend in selectivity reduction was found in single gas measurements. Upon further increase of the CO2 composition, the plasticization becomes more pronounced, resulting in lower selectivity. However, the higher CO2/N2 separation factor of the PUUs with different side chains than PUU and PU as the reference membranes in the mixed gas permeation reveals the improved CO2/N2 separation capability of alkyl side chain–based PUUs compared with linear polyurethanes. For CO2/CH4 mixed gas study, the presence of condensable CH4 besides CO2 makes the explanation of gas separation performance more complex. Generally, CO2/CH4 selectivity and CO2 permeability for mixed gas measurements decrease in comparison with single gas data due to the plasticization applied by CO2 and competitive sorption effects, respectively. CH4 permeability increases at higher feed pressure (for CO2/CH4 (50:50)) and higher concentration of CO2 due to the plasticization effect of CO2 (Fig. 3).

23

(a)

(c)

(b)

Fig. 2. CO2:N2 mixed gas separation properties with different compositions through the membranes and the influence of upstream pressure for CO2:N2 (50:50) mixed gas on gas permeation: a) CO2/N2 selectivities, b) N2 permeabilities and c) CO2 permeabilities at 30 °C. (a)

24

(c)

(b)

Fig. 3. CO2:CH4 mixed gas separation properties with different compositions through the membranes and the influence of upstream pressure for CO2:CH4 (50:50) mixed gas on gas permeation: a) CO2/CH4 selectivities, b) CH4 permeabilities and c) CO2 permeabilities at 30 °C.

Comparing this research and the previous study [18], it is deduced that the gas separation mechanisms of the studied PUUs are different, which confirms the controlling effect of microstructure on the resultant properties. In our previous study, the bulky chain extender caused higher phase separation and FFV, simultaneously. The bulky chain extender gave higher intersegmental spacing in the hard domains of the polyurethanes and consequently more significant contribution in gas permeation by the hard segment. Simultaneous increased gas permeability and selectivity using bulkier chain extenders are the characteristic features of the bulky membranes used in our previous study [18] . In the present study, using alkyl substituents with different lengths attached on the hard domains led to increased high separation and decreased FFV as a result of the shielding mechanism of the hard domains by the side chains. In these series of PUUs, increased phase separation and purity of the domains are responsible for higher permeability. This means that in this study, the soft domain-rich area of polyurethane is more of a determinant for gas permeation and the longer side chain–based PUUs give lower selectivity. 4. Conclusions

25

The intention of this study was to design the hard segment of polyurethanes to increase chain mobility and gas permeation. For this purpose, aliphatic-substituted diamine chain extenders with varying side chain lengths were synthesized. The aliphatic side chains are not compatible with the main chain of the hard segment, but they do possess more affinity to the soft domain. This results in side chains’ migration to the surface of the hard domains. Therefore, incorporation of longer side chains to the hard segment evinced several consequences, including: a) shielding the main chain of the hard segment and reducing accessible NH groups for ethereal links of polyol, b) higher degree of phase separation and lower FFV values, c) filling the hard/soft interfaces and reducing the concentration of functional groups at the interface, d) more likely formation of purer phases, e) more ordering of the main chains of hard segments, f) enhanced chain mobility which promotes gas permeation, but reduces mechanical properties, and g) lower Tgs. Increased gas permeability through longer side chain based PUUs having lower FFV values indicates that the higher chain mobility and plasticizing effect by longer aliphatic side chains are likely more dominant. The results reveal that the complexity of the gas transport through the polymeric membranes arises from many effective factors, including chain mobility, diffusion barriers, phase separation degree, the purity of phases, and free volume. The highest gas permeability was provided by PUU-O-H at 286.5 barrer during single gas measurement. The principal trait of the PUU membranes is increasing permeability through the substituted chain extender–based membranes, without a significant reduction in selectivity. Plasticization and penetrant competition are crucial factors for the reduced selectivity in mixed gas study, remarkable at higher CO2 proportion. As a general conclusion, the molecular designing of these polyurethane membranes allows a further tuning of the gas permeation of the studied PUUs. 5. Conflict of interest The authors have no conflict of interest to declare. Acknowledgment

26

The authors express their sincere gratitude to the Iranian Ministry of Science, Research and Technology (MSRT) for the monetary support of research visit for accomplishment of this work at University of Twente, The Netherlands.

References [1] P.M. Budd, N.B. McKeown, Highly permeable polymers for gas separation membranes, Polymer Chemistry 1(1) (2010) 63-68. [2] Z. Kang, L. Fan, D. Sun, Recent advances and challenges of metal–organic framework membranes for gas separation, J. Mater. Chem. A 5 (2017) 10073–10091 [3] A. Arabi Shamsabadi, A. Kargari, F. Farshadpour, S. Laki, Mathematical modeling of CO2/CH4 separation by hollow fiber membrane module using finite difference method, Journal of Membrane and Separation Technology 1(1) (2012) 19-29. [4] T. Wang, C. Cheng, L.-g. Wu, J.-n. Shen, B. Van der Bruggen, Q. Chen, D. Chen, C.-y. Dong, Fabrication of polyimide membrane incorporated with functional graphene oxide for CO2 separation: The effects of GO surface modification on membrane performance, Environmental Science & Technology 51(11) (2017) 6202-6210. [5] A. Arabi Shamsabadi, F. Seidi, M. Nozari, M. Soroush, A new pentiptycene‐based dianhydride and its high‐Free‐volume polymer for carbon dioxide removal, ChemSusChem 11(2) (2018) 472-482. [6] X. Ma, S. Williams, X. Wei, J. Kniep, Y.S. Lin, Propylene/propane mixture separation characteristics and stability of carbon molecular sieve membranes, Industrial & Engineering Chemistry Research 54(40) (2015) 9824-9831. [7] W. Dujardin, C. Van Goethem, Z. Zhang, R. Verbeke, M. Dickmann, W. Egger, E. Nies, I. Vankelecom, G. Koeckelberghs, Fine-tuning the molecular structure of binaphthalene polyimides for gas separations, European Polymer Journal 114 (2019) 134-143. [8] X. Zhou, M.M. Obadia, S.R. Venna, E.A. Roth, A. Serghei, D.R. Luebke, C. Myers, Z. Chang, R. Enick, E. Drockenmuller, H.B. Nulwala, Highly cross-linked polyether-based 1,2,3-triazolium ion conducting membranes with enhanced gas separation properties, European Polymer Journal 84 (2016) 65-76. [9] A. Pournaghshband Isfahani, M. Sadeghi, K. Wakimoto, B.B. Shrestha, R. Bagheri, E. Sivaniah, B. Ghalei, Pentiptycene-based polyurethane with enhanced mechanical properties and CO2-plasticization resistance for thin film gas separation membranes, ACS Applied Materials & Interfaces 10(20) (2018) 17366-17374. [10] M. Galizia, W.S. Chi, Z.P. Smith, T.C. Merkel, R.W. Baker, B.D. Freeman, 50th anniversary perspective: polymers and mixed matrix membranes for gas and vapor separation: A review and prospective opportunities, Macromolecules 50(20) (2017) 7809-7843. [11] T. Wang, L. Zhao, J.-n. Shen, L.-g. Wu, B. Van der Bruggen, Enhanced performance of polyurethane hybrid membranes for CO2 separation by incorporating graphene oxide: The relationship between membrane performance and morphology of graphene oxide, Environmental Science & Technology 49(13) (2015) 8004-8011. [12] M. Monga Mulunda, C. van Goethem, Z. Zhang, E. Nies, I. Vankelecom, G. Koeckelberghs, Influence of the molecular structure of polybinaphthalene on the membrane separation performance, European Polymer Journal 101 (2018) 248-254. [13] M. Sadeghi, M.M. Talakesh, A. Arabi Shamsabadi, M. Soroush, Novel application of a polyurethane membrane for efficient separation of hydrogen sulfide from binary and ternary gas mixtures, ChemistrySelect 3(11) (2018) 3302-3308. [14] A. Pournaghshband Isfahani, M. Sadeghi, K. Wakimoto, A.H. Gibbons, R. Bagheri, E. Sivaniah, B. Ghalei, Enhancement of CO2 capture by polyethylene glycol-based polyurethane membranes, Journal of Membrane Science 542 (2017) 143-149. [15] M. Shahrooz, M. Sadeghi, R. Bagheri, M. Laghaei, Elucidating the significance of segmental mixing in determining the gas transport properties of polyurethanes, Macromolecules 49(11) (2016) 4220-4228. [16] A. Pournaghshband Isfahani, B. Ghalei, R. Bagheri, Y. Kinoshita, H. Kitagawa, E. Sivaniah, M. Sadeghi, Polyurethane gas separation membranes with ethereal bonds in the hard segments, Journal of Membrane Science 513 (2016) 58-66. 27

[17] M. Sadeghi, A. Arabi Shamsabadi, A. Ronasi, A.P. Isfahani, M. Dinari, M. Soroush, Engineering the dispersion of nanoparticles in polyurethane membranes to control membrane physical and transport properties, Chemical Engineering Science 192 (2018) 688-698. [18] A. Fakhar, M. Sadeghi, M. Dinari, R. Lammertink, Association of hard segments in gas separation through polyurethane membranes with aromatic bulky chain extenders, Journal of Membrane Science 574 (2019) 136-146. [19] M. Sadeghi, M.A. Semsarzadeh, M. Barikani, B. Ghalei, Study on the morphology and gas permeation property of polyurethane membranes, Journal of Membrane Science 385-386 (2011) 76-85. [20] M. Sadeghi, M.A. Semsarzadeh, M. Barikani, B. Ghalei, The effect of urethane and urea content on the gas permeation properties of poly(urethane-urea) membranes, Journal of Membrane Science 354(1) (2010) 40-47. [21] S. Luo, K.A. Stevens, J.S. Park, J.D. Moon, Q. Liu, B.D. Freeman, R. Guo, Highly CO2-selective gas separation membranes based on segmented copolymers of poly(Ethylene oxide) reinforced with pentiptycene-containing polyimide hard segments, ACS Applied Materials & Interfaces 8(3) (2016) 2306-2317. [22] K.-S. Chang, C.-C. Tung, K.-S. Wang, K.-L. Tung, Free volume analysis and gas transport mechanisms of aromatic polyimide membranes: A molecular simulation study, The Journal of Physical Chemistry B 113(29) (2009) 9821-9830. [23] Y. Dai, M.D. Guiver, G.P. Robertson, Y.S. Kang, Effect of hexafluoro-2-propanol substituents in polymers on gas permeability and fractional free volume, Macromolecules 38(23) (2005) 9670-9678. [24] P. Wisian-Neilson, G.-F. Xu, Gas permeation studies of silylated derivatives of poly(methylphenylphosphazene), Macromolecules 29(10) (1996) 3457-3461. [25] S.R. Thrasher, M.E. Rezac, Transport of water and methanol vapors in alkyl substituted poly(norbornene), Polymer 45(8) (2004) 2641-2649. [26] I. Pinnau, A. Morisato, Z. He, Influence of side-chain length on the gas permeation properties of poly(2alkylacetylenes), Macromolecules 37(8) (2004) 2823-2828. [27] K.D. Dorkenoo, P.H. Pfromm, M.E. Rezac, Gas transport properties of a series of high Tg polynorbornenes with aliphatic pendant groups, Journal of Polymer Science Part B: Polymer Physics 36(5) (1998) 797-803. [28] M.A. Hood, B. Wang, J.M. Sands, J.J. La Scala, F.L. Beyer, C.Y. Li, Morphology control of segmented polyurethanes by crystallization of hard and soft segments, Polymer 51(10) (2010) 2191-2198. [29] Y. Li, T. Gao, B. Chu, Synchrotron SAXS studies of the phase-separation kinetics in a segmented polyurethane, Macromolecules 25(6) (1992) 1737-1742. [30] H. Tan, M. Guo, R. Du, X. Xie, J. Li, Y. Zhong, Q. Fu, The effect of fluorinated side chain attached on hard segment on the phase separation and surface topography of polyurethanes, Polymer 45(5) (2004) 1647-1657. [31] H. Tan, X. Xie, J. Li, Y. Zhong, Q. Fu, Synthesis and surface mobility of segmented polyurethanes with fluorinated side chains attached to hard blocks, Polymer 45(5) (2004) 1495-1502. [32] H. Tan, J. Li, M. Guo, R. Du, X. Xie, Y. Zhong, Q. Fu, Phase behavior and hydrogen bonding in biomembrane mimicing polyurethanes with long side chain fluorinated alkyl phosphatidylcholine polar head groups attached to hard block, Polymer 46(18) (2005) 7230-7239. [33] A.C. Savoca, A.D. Surnamer, C.F. Tien, Gas transport in poly(silylpropynes): the chemical structure point of view, Macromolecules 26(23) (1993) 6211-6216. [34] H. Feng, T. Hong, S.M. Mahurin, K.D. Vogiatzis, K.R. Gmernicki, B.K. Long, J.W. Mays, A.P. Sokolov, N.G. Kang, T. Saito, Gas separation mechanism of CO2 selective amidoxime-poly(1-trimethylsilyl-1-propyne) membranes, Polymer Chemistry 8(21) (2017) 3341-3350. [35] M.F. Ferreira Marques, C. Lopes Gil, P.M. Gordo, Z. Kajcsos, A.P. de Lima, D.P. Queiroz, M.N. de Pinho, Free-volume studies in polyurethane membranes by positron annihilation spectroscopy, Radiation Physics and Chemistry 68(3) (2003) 573-576. [36] C. Nagel, K. Günther-Schade, D. Fritsch, T. Strunskus, F. Free volume and transport properties in highly selective polymer membranes, Macromolecules 35(6) (2002) 2071-2077. [37] J. Qiu, J.M. Zheng, K.V. Peinemann, Gas transport properties of poly(trimethylsilylpropyne) and ethylcellulose filled with different molecular weight trimethylsilylsaccharides:  Impact on fractional free volume and chain mobility, Macromolecules 40(9) (2007) 3213-3222. [38] A. Thran, G. Kroll, F. Faupel, Correlation between fractional free volume and diffusivity of gas molecules in glassy polymers, Journal of Polymer Science Part B: Polymer Physics 37(23) (1999) 3344-3358. [39] L.A. Pessan, W.J. Koros, Isomer effects on transport properties of polyesters based on bisphenol-A, Journal of Polymer Science Part B: Polymer Physics 31(9) (1993) 1245-1252. [40] C.M. Zimmerman, W.J. Koros, Comparison of gas transport and sorption in the ladder polymer BBL and some semi-ladder polymers, Polymer 40(20) (1999) 5655-5664. 28

[41] L.R. Edwin M. Smolin, s-Triazines and derivatives, Interscience Publishers, INC., 250 Fifth Avenue, New York , 48-58 (1959). [42] S. A. Brewer, H. T. Burnell, I. Holden, B. G. Jones, C. R. Willis, Synthesis of a series of dichloroamino- and dihalosulfonamido-1,3,5-triazines and investigation of their hindered rotation and stereodynamic behaviour by NMR spectroscopy, Journal of the Chemical Society, Perkin Transactions 2 (6) (1999) 1231-1234. [43] J.A. Chaudhari, R.P. Patel and M.V. Hathi, Studies on novel bisaryl hydrazino-s-triazine derivatives - Part 2, EJournal of Chemistry 4(3) (2007) 385-389. [44] F. Yaoting, L. Gang, L. Zifeng, H. Hongwei, M. Hairong, Synthesis, structure and third-order nonlinear optical properties of 1,3,5-triazine-based Zn(II) three-dimensional supramolecule, Journal of Molecular Structure 693(1) (2004) 217-224. [45] L. Tei, M. Benzi, F. Kielar, M. Botta, C. Cavallotti, G. B. Giovenzana, S. Aime, Synthesis and relaxometric properties of gadolinium(III) complexes of new triazine-based polydentate ligands, Helvetica Chimica Acta 92 (2009) 2414-2426. [46] M.M. Khan, G. Bengtson, S. Shishatskiy, B.N. Gacal, M. Mushfequr Rahman, S. Neumann, V. Filiz, V. Abetz, Cross-linking of polymer of intrinsic microporosity (PIM-1) via nitrene reaction and its effect on gas transport property, European Polymer Journal 49(12) (2013) 4157-4166. [47] D.K. Chattopadhyay, B. Sreedhar, K.V.S.N. Raju, Effect of chain extender on phase mixing and coating properties of polyurethane ureas, Industrial & Engineering Chemistry Research 44(6) (2005) 1772-1779. [48] J.A. Miller, S.B. Lin, K.K.S. Hwang, K.S. Wu, P.E. Gibson, S.L. Cooper, Properties of polyether-polyurethane block copolymers: effects of hard segment length distribution, Macromolecules 18(1) (1985) 32-44. [49] A. Pournaghshband Isfahani, B. Ghalei, K. Wakimoto, R. Bagheri, E. Sivaniah, M. Sadeghi, Plasticization resistant crosslinked polyurethane gas separation membranes, Journal of Materials Chemistry A 4(44) (2016) 1743117439. [50] S. Hiller, O. Pascui, H. Budde, O. Kabisch, D. Reichert, M. Beiner, Nanophase separation in side chain polymers: new evidence from structure and dynamics, New Journal of Physics 6(1) (2004) 10. [51] S. Hugger, R. Thomann, T. Heinzel, T. Thurn-Albrecht, Semicrystalline morphology in thin films of poly(3hexylthiophene), Colloid and Polymer Science 282(8) (2004) 932-938. [52] D.W. Van Krevelen, K.T. Nijenhuis, Properties of polymers: Their correlation with chemical structure; Their numerical estimation and prediction from additive group contributions, Fourth ed., Delft University of Technology, The Netherlands, 2009 (completely revised edition). [53] C.L. Yaws, Chapter 13: solubility parameter and liquid volume - organic compounds, thermophysical properties of chemicals and hydrocarbons, Lamar University, Beaumont, Texas, 2008. [54] Z.F. Wang, B. Wang , Y.R. Yang, C.P. Hu, Correlations between gas permeation and free-volume hole properties of polyurethane membranes, European Polymer Journal 39 (2003) 2345–2349. [55] M. Amani, S. Amjad-Iranagh, K. Golzar, G.M.M. Sadeghi, H. Modarress, Study of nanostructure characterizations and gas separation properties of poly(urethane–urea)s membranes by molecular dynamics simulation, Journal of Membrane Science 462 (2014) 28-41. [56] Z. Mogri, D.R. Paul, Gas sorption and transport in poly(alkyl (meth)acrylate)s. I. Permeation properties, Polymer 42(18) (2001) 7765-7780. [57] S.L. Chang, T.L. Yu, C.C. Huang, W.C. Chen, K. Linliu, T.L. Lin, Effect of polyester side-chains on the phase segregation of polyurethanes using small-angle X-ray scattering, Polymer 39(15) (1998) 3479-3489. [58] P. Tiemblo, A. Fernández-Arizpe, E. Riande, J. Guzmán, Transport of gases and vapors in methacrylic copolymers of varying side chain length, Polymer 44(3) (2003) 635-641. [59] A. Eceiza, M. D. Martin, K. de la Caba, G. Kortaberria, N. Gabilondo, M. A. Corcuera and I. Mondragon, Thermoplastic polyurethane elastomers based on polycarbonate diols with different soft segment molecular weight and chemical structure: Mechanical and thermal properties, Polymer Engineering & Science 48(2) (2008) 297-306. [60] H. Li, B.D. Freeman, O.M. Ekiner, Gas permeation properties of poly(urethane-urea)s containing different polyethers, Journal of Membrane Science 369(1) (2011) 49-58. [61] T. Koichi, M. Hidehiko, M. Toshio, H. Toshinobu, Gas permeability of polyacetylenes carrying substituents, Journal of Applied Polymer Science 30(4) (1985) 1605-1616. [62] H. Fu, K. Liu, c. Yan, W. Chen, Y. Wang, Phase morphology and mechanical properties of aliphatic waterborne polyurethane-ureas: Effect of 1,6-hexamethylene diisocyanate (HDI)/isophorone diisocyanate (IPDI) ratio, Polymers & Polymer Composites 23(3) (2015).

29

[63] A.A. Deschamps, D.W. Grijpma, J. Feijen, Poly(ethylene oxide)/poly(butylene terephthalate) segmented block copolymers: the effect of copolymer composition on physical properties and degradation behavior, Polymer 42(23) (2001) 9335-9345. [64] V. Mozaffari, M. Sadeghi, A. Fakhar, G. Khanbabaei, A.F. Ismail, Gas separation properties of polyurethane/poly(ether-block-amide) (PU/PEBA) blend membranes, Separation and Purification Technology 185 (2017) 202-214. [65] S.-H. Huang, C.-C. Hu, K.-R. Lee, D.-J. Liaw, J.-Y. Lai, Gas separation properties of aromatic poly(amideimide) membranes, European Polymer Journal 42(1) (2006) 140-148. [66] M. Sadeghi, M. Mehdi Talakesh, B. Ghalei, M. Shafiei, Preparation, characterization and gas permeation properties of a polycaprolactone based polyurethane-silica nanocomposite membrane, Journal of Membrane Science 427 (2013) 21-29. [67] M. Mulder, basic principles of membrane technology London (1997). [68] M.M. Talakesh, M. Sadeghi, M. Pourafshari Chenar, A. Khosravi, Gas separation properties of poly(ethylene glycol)/poly(tetramethylene glycol) based polyurethane membranes, Journal of Membrane Science 415-416 (2012) 469–477. [69] V. Giel, J. Kredatusová, M. Trchová, J. Brus, J. Žitka, J. Peter, Polyaniline/polybenzimidazole blends: Characterisation of its physico-chemical properties and gas separation behaviour, European Polymer Journal 77 (2016) 98-113. [70] L.M. Robeson, The upper bound revisited, Journal of Membrane Science 320(1) (2008) 390-400. [71] R.-C. Ruaan, W.-C. Ma, S.-H. Chen, J.-Y. Lai, Microstructure of HTPB-based polyurethane membranes and explanation of their low O2/N2 selectivity, Journal of Applied Polymer Science 82 (2001) 1307–1314 [72] S.-H. Chena, K.-C. Yua, S.-L. Houng and J.-Y. Lai, Gas transport properties of HTPB based polyurethane/cosalen membrane, Journal of Membrane Science 173 (2000) 99–106. [73] E.S. Sanders, Penetrant-induced plasticization and gas permeation in glassy polymers, Journal of Membrane Science, 37 (1988) 63-80. [74] J.D. Wind, C. Staudt-Bickel, D.R. Paul, W.J. Koros, The effects of crosslinking chemistry on CO2 plasticization of polyimide gas separation membranes, Industrial & Engineering Chemistry Research 41(24) (2002) 6139-6148. [75] B. Ghalei, Y. Kinoshita, K. Wakimoto, K. Sakurai, S. Mathew, Y. Youfeng, H. Kusuda, H. Imahori, E. Sivaniah, Surface functionalization of high free-volume polymers as a route to efficient hydrogen separation membranes, J. Mater. Chem. A, 5 (2017) 4686-4694. [76] S. Yi, X. Ma, I. Pinnau, W.J. Koros, A high-performance hydroxyl-functionalized polymer of intrinsic microporosity for an environmentally attractive membrane-based approach to decontamination of sour natural gas J. Mater. Chem. A 3 (2015) 22794-22806. [77] L. Escorial, M. de la Viuda, S. Rodríguez, A. Tena, A. Marcos, L. Palacio, P. Prádanos, A.E. Lozano, A. Hernández, Partially pyrolized gas-separation membranes made from blends of copolyetherimides and polyimides, European Polymer Journal 103 (2018) 390-399. [78] S. Saedi, S.S. Madaeni, A. Arabi Shamsabadi, PDMS coated asymmetric PES membrane for natural gas sweetening: Effect of preparation and operating parameters on performance, The Canadian Journal of Chemical Engineering 92(5) (2014) 892-904. [79] T. Visser, N. Masetto, M. Wessling, Materials dependence of mixed gas plasticization behavior in asymmetric membranes, Journal of Membrane Science 306(1) (2007) 16-28.

30

Highlights 

A series of PUU membranes was designed by three synthesized chain extenders containing different length-side chains.



The migration of side chains to the surface of the main chain of hard segment, thereby shielding it increased phase separation and reduced FFV, simultaneously.



Increased gas permeability through longer side chain based-PUUs, indicated dominant plasticizing effect by the aliphatic substituents over FFV.



The molecular designing of the PUUs allowed a further tuning of the gas permeation of the studied membranes.

31