polyurethane composite membrane for gas separation

Author’s Accepted Manuscript Short fiber/polyurethane composite membrane for gas separation Mohammad Bagher Karimi, Shadi Hassanajili

www.elsevier.com/locate/memsci

PII: DOI: Reference:

S0376-7388(17)31845-8 http://dx.doi.org/10.1016/j.memsci.2017.08.043 MEMSCI15509

To appear in: Journal of Membrane Science Received date: 28 June 2017 Revised date: 14 August 2017 Accepted date: 18 August 2017 Cite this article as: Mohammad Bagher Karimi and Shadi Hassanajili, Short fiber/polyurethane composite membrane for gas separation, Journal of Membrane Science, http://dx.doi.org/10.1016/j.memsci.2017.08.043 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. 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.

Short fiber/polyurethane composite membrane for gas separation

Mohammad Bagher Karimia,b, Shadi Hassanajilib,* a

Department of Polymer Engineering and Color Technology, Amirkabir University of Technology, P. O. Box: 15875-4413,Tehran, Iran

b

Department of Chemical Engineering, School of Chemical and Petroleum Engineering, Shiraz University, Shiraz 71348-51154, Iran

Abstract In this research, short fiber/polymer composite membranes is introduced for gas separation. Our attempt is to increase the membranes performance by using wide interface between fiber and polymer. Short glass wool fiber (SGWF) was used as a polar micro size reinforcement and was mechanically dispersed in polyurethane matrix. Scanning electron microscope (SEM) images showed that fibers have a good dispersion and adhesion to polymer matrix. In order to survey the phase separation and crystallization behavior of polyurethane segments, differential scanning calorimetery (DSC) and dynamic mechanical thermal analysis (DMTA) were used. The results indicated that polar surface of glass wool is a suitable site for attraction of hard segments. Thermal transition of soft segments were assisted using DMTA. Obtained results were showed that the presence of glass wool fiber remarkably reduced the soft segments glass transition temperature (Tg). Gas permeation properties of membranes were assisted using pure CO2, CH4 and N2 gases. Presence of SGWF caused a simultaneous increase in permeability and ideal selectivity (permselectivity) in a way that composite membranes showed high performance. *

[email protected]

1

Thus, using of fiber instead of nanoparticles in polymeric membranes can be more beneficial in view of economic and performance for industrial applications. Keywords: Short glass wool fiber; Polyurethane; Composite; Membrane; Gas selectivity

1. Introduction Recently polymer science has attracted much attention in the field of separation technology. Variety of the chemical structure and properties of polymers makes the possibility of creating different membranes for special applications such as water purification and gas separation [1,2]. There are many different gases with different molecular sizes and polarity degrees which have specific properties, so attempts are to separate them from each other. The main problem of polymeric membrane is the trade-off relationship between permeability and selectivity [3–5]. There is a few research that reported the straight relation between permeability and selectivity [6,7]. One of the main procedures to increase the polymeric membranes performance for gas separation is use of nanoparticles which have important effects on polymer properties[8–10]. There are plenty of nanoparticles with different properties which can be used to improve the polymeric membrane performance [9–12]. Effects of nanoparticles on gas permeation properties of polymeric membranes depend on their weight fraction, size and shape as well as their interaction with the polymer. Nanoparticles have larger ratio of surface area to volume than larger particles [8]. So they can be effectively a bridge between bulk materials and atomic or molecular structures leading to strong interactions with the matrix in nanocomposite materials. But, nanoparticles generally have high price with difficult production method. In addition, their dispersion in the matrix is very difficult particularly at high concentrations. One of the main 2

advantages of nanoparticles is creation of suitable free volumes at their interface with polymers which may be suitable for selective gas transport [8]. Hassanajili et al. [13] studied the effect of surface modification of silica nanoparticles on gas permeation properties of polyurethane membranes. Their results indicated that presence of nanosilica particles in polyurethane simultaneously increased the CO2 permeability and CO2/N2 selectivity. They reported that polymer-inorganic interface is a good site for adsorption of polar gases. Tirouni et al. [14] studied the effect of zeolite particles on hydrocarbon separation in polyurethane membranes. They reported that voids at the polymer-zeolite interface are suitable places for diffusion and dissolution of high molecular weight gases. Tantekin-Ersolmaz et al. [15] studied the effect of zeolite particle size on the performance of silicalite–PDMS mixed matrix membranes. Their results showed that by increasing the particle size, an increase in gas permeability was observed which has no obvious effect on the membrane selectivity. They reported that these behaviors are related to the increase in area and number of the zeolitepolymer interface. All above researches clearly showed that polymer-filler interface can be a good site for dissolution of polar gases and could enhance the permeability of low molecular weight gases in compare to high molecular weight gases such as hydrocarbons. Fiber reinforced polymers are among the most rapidly growing classes of materials [16] and are widely used as high-performance material [17]. In these materials, polymer matrix has attracted much attention due to its wide application in petroleum chemical industry, aeronautics, automotive parts and architecture field [16]. In fiber reinforced composites, polymer is usually the matrix (continues phase) while the fiber is the discontinuous phase [18]. It is known that the

3

mechanical properties of short fiber reinforced polymer composites are governed with fiber content, orientation and fiber-matrix adhesion [19–21]. The reinforcement and the matrix near the interface take different and complex structure which is different from both of the matrix and fiber phase [22,23]. Due to the appearance of interface area, the roles of polymer and fiber are independent but are not isolated [22]. The interface can have two different effects on composites 1- chemical effect, to form a new interfacial layer structure; 2- the mechanical effect, which causes the stress distribution on the interface [22]. The specific surface area of fiber is so big that allows better transfer of load between polymer and fiber which is caused to maintain the strength [24]. As mentioned above interface is different from the bulk matrix; these sites are very important to design composites. Routinely in fiber reinforced polymeric composites moisture diffuse into the polymer and then diffuse along the interface by capillary flow [25]. So adhesion between polymer and fiber has an important role on fiber reinforced polymer composites. Since fiber-polymer interface can reduce the composite durability in water condition [20,26–29], in this work we tried to use this specific trait (interface) in gas separation science. As an initial step, we need to choose a polymer and fiber to create a suitable interface. Besides, the polymer itself should have a good gas permeability behavior. Polyurethanes are a class of blocky copolymers that have specific phase separation behavior[30– 32]. Phase separation in polyurethanes is because of thermodynamic incompatibility between soft and hard segments [31,33,34]. Previous studies have clearly shown that presence of polar component in polyurethane segments or use of nanoparticles can significantly affect the phase separation behavior of hard segments [7,35–38]. 4

In this research, we studied the effects of SGWF on gas permeation properties of polyurethane membrane. Results showed that polar surface of glass wool is a suitable place for adsorption of hard segments. Polyurethane and fibers create a polar interface which is a suitable place for dissolution of polar gases. Glass wool fibers remarkably influenced the gas permeability properties of polyurethanes at low content and simultaneously increased the permeability of CO2 and CO2 over CH4 permselectivity.

2. Experimental 2.1.Material In this research, polyether based polyurethane was provided by Coim S.p.A Co (trade name: LPR9060EF, density: 1.20 g/cm3, Shore hardness A: 90). Its segments composed of poly (oxytetramethylene) (PTMG, 1000 g/mol) as soft segment. The hard segment content was 35 wt% (this hard segments content in polyurethane give it suitable strength for application of its membranes in high pressures, also, due to barrier nature of hard segments for gas transport more increase in its content can restrict polyurethane soft segments to create suitable free volume for gas transport) and composed of 4-4 dimethylphenoldiisocyanate (MDI) and 1,4- butandiol (BDO). Glass wool fiber purchased from Shiraz glass wool Co. Fibers have long length, so mechanically cut to shorten glass fibers (1-2

in length and average of 5

in diameter).

Dimethylformamide (DMF) was used as solvent to prepare polyurethane solution and purchased from Merck Co. CO2, CH4 and N2 gases (purity 99.999) were provided by Aboghadare Co. (Shiraz, Iran). 2.2.Membrane preparation

5

Membranes were prepared using solutions casting and solvent evaporation technic. Pure polyurethane solution was prepared by adding 15 wt. % polyurethane in DMF[39]. Polyurethanes mixed in DMF for 2h at 70

to obtain a homogeneous transparent solution.

Obtained solution was filtered and degassed, after that casted on a clean glass plate by casting knife. Casted solution was dried at ambient temperature for 3 days to remove most of DMF. Formed dense polyurethane film dried at

for 24h and after that vacuumed for 6h at

to

remove residual solvent. For the preparation of composite membranes, the SGWF was added to polyurethane solution and stirred for 1h. The average thickness of prepared membranes based on several measurements was between 80 and 120

. Composition of prepared membranes is

listed in Table 1. Table 1. Composition of SGWF/polyurethane composite membranes sample PU0 PU1 PU3 PU5

Polyurethane (wt. %) 100 99 97 95

SGWF (wt. %) 0 1 3 5

2.3.Membrane characterization In order to identify the effects of SGWF on thermal properties of polyurethane composites, DSC and DMTA analysis were used. Effects of the SGWF on Tm and melting enthalpy of the crystalline structure of polyurethane composites were verified using Mettler-Toledo. Samples were heated at the ratio of 10

/min at a temperature range from -20 to190 . Thermal transition

of polyurethane soft segments was evaluated via DMTA analysis using DMA-PL (model Trinitron 2000, UK). Dynamic mechanical thermal properties were measured in tension mode over the temperature range from -90 to 30 . 6

2.4.Gas permeation Gas permeation test was carried out using a constant pressure method at three different feed pressure of 6, 8 and 10 bar and temperature of 25 . The gas permeability of membrane was calculated based on the following equation:

(1) Where P is permeability expressed in barrer (1 barrer=10‒ 10 cm3 (STP) cm cm‒ 2 s‒ 1 cmHg‒ 1), q is flow rate of the permeated gas passing through the membrane (cm3/s), l is membrane thickness (cm), p1 and p2 are the absolute pressures of feed-side and permeate-side, respectively (cmHg) and A is the effective membrane area (cm2). The permselectivity (αA/B) of membranes was calculated from the ratio of pair gases permeability: (2)

where D and S are diffusion and solution coefficient, respectively and DA/DB and SA/SB are defined as diffusivity and solubility selectivity, respectively. Schematic of gas permeation system is presented in Scheme 1. In this system permeated flow rate was measured by a bubble flow meter. The diffusion coefficient (D) was determined by time-lag method based on the following equation [12]: (3)

In this equation

is the time lag (s). The solubility coefficient (S) was then calculated from the

following equation: 7

(4)

Scheme 1. Schematic of gas permeation system

3. Results and discussion 3.1.Morphology of membranes The morphological structure of the prepared membranes in presence of SGWF in PU matrix were evaluated using SEM. Fig. 1 shows cross-section of prepared membranes. As can be seen in this figure, all prepared membranes have symmetric and dense structure. From the SEM images, aspects of short fiber bounding to polymer matrix and their distribution can be evaluated. According to the SEM images, SGWF have good dispersion in polyurethane matrix at low content. But at a high fraction (5wt%), some agglomerations can be observed. SEM images clearly show that SGWF has good adhesion to the matrix, although some micro-voids at 8

polymer/fiber interface are formed (Fig. 1(5-b)). Due to blocky nature of polyurethanes and thermodynamic incompatibility between soft and hard segments, and the polar, large and available surface of glass wool to adsorb polar hard segments, a good adhesion between fibers and matrix can be observed. This may be related to nucleating sites of glass wool fibers to promote hard segments crystallization which will be discussed in next sections.

Fig. 1. SEM micrographs of SGWF/polyurethane composite membrane (cross section view)

3.2.DSC analysis

9

The thermal properties of polyurethane membranes which are reinforced with SGWF were evaluated by DSC analysis. Due to polar surface of glass wool fibers, their presence in polyurethane can remarkably influence the crystallization behavior of polyurethane blocky segments. Thus, DSC was used to characterize the crystallization and melting points of polyurethane blocks in the presence of SGWF. Some previous researches indicated that organic or inorganic material can act as a nucleation agent for hard segments which can change the melting point of hard segments crystals. Karimi et al. [7] studied the effects of esterified vegetable oil diol on gas permeation properties of polyurethane membrane. Their results indicated that esterified vegetable oil based diol acts as a nucleation agent for crystallization of hard segments, in a way that hard segments create a new crystalline area with higher melting enthalpy. In another work, Hassanajili et al. [40] studied the effect of silica surface modification on polyurethane morphology. Their results showed that unmodified silica nanoparticles behave as a nucleation agent for hard segments crystallization. Fig. 2 shows the DSC thermograms of fiber reinforced polyurethane composites. As can be seen in Fig. 2, there are two endothermic peaks at temperatures of 58 and 160

for pure

polyurethane. Appearance of the broad endothermic peak at high temperature is related to crystallization of hard segments [37]. Emerging of endothermic peak of hard segments indicates the strong hydrogen bonding between hard segments which leads to higher phase separation in polyurethane [41]. Due to the high mobility of PTMG soft segments, they cannot create any obvious crystalline order. So it can be stated that the observed endothermic peak at low temperature is the reflection of weaker interaction of hard segments to soft segments. By addition of SGWF to polyurethane matrix, the observed endothermic peak at low temperature remarkably shifts to higher temperatures. This clearly indicates the creation of a new crystalline 10

structure for hard segments with higher thermal stability [42]. With increasing the fiber content, the endothermic peak at lower temperature is shifted to higher one and the endothermic peak at higher temperature is nearly expanded and shifted to the lower one. In the composite membrane with the highest fiber content (PU5), it seems that these two peaks are unified in a way that we can see just a broad peak with lower Tm and melting enthalpy in comparison with the pure PU. These results confirmed high tendency of polar hard segments towards SGWF to create new crystalline domains. Thermal properties of composite membranes are listed in Table 2. Based on DSC results (Fig. 2), hydrogen bonding between polar hard segments and polar groups on the surface of glass wool fibers can be the reason for observation of good adhesion between the polyurethane matrix and the SGWF as depicted in SEM images.

Fig. 2. DSC thermograms of SGWF/polyurethane composite membranes 11

Table 2. Thermal properties of SGWF/polyurethane composite membranes Melting enthalpy of

Tg1(Tmax1)( ),

Tg2 (Tmax2)( )

hard segments (at

soft segments

soft segments

(DMTA)

(DMTA)

7.04

-31.00

not showen

160.00

6.95

-32.50

-39.20

88.93

160.63

6.82

-32.50

-40.00

Disappeared

155.89

6.72

Disappeared

-42.10

Sample

Tm1 ( ), hard

Tm2 ( ), hard

code

segments

segments

PU0

67.62

159.93

PU1

87.95

PU3 PU5

Tm2)(

)

3.3.Dynamic mechanical thermal analysis (DMTA) DMTA analysis was used to specify the effects of SGWF on dissipation factor (

) of

prepared polyurethane composites. Dissipation factor is defined as the ratio of loss modulus to storage modulus [11], which the value of soft segments Tg is associated with the peak magnitude of

[43,44]. The shape of

peaks such as its height and broadness can also provide

information about the order and mobility of the polyurethane segments [11,43]. Fig. 3 shows the effects of SGWF on

peak of polyurethane composites at low temperatures

which are associated with polyurethane soft segments. As it can be seen in Fig. 3, pure polyurethane shows a peak with a maximum at -33

(Tmax1), indicating that this sample has

homogenous nature [45]. With introducing small amount of short fiber to polyurethane (PU1), observed peak is shifted to lower temperatures and

peak shows two maximums (Tmax1 and

Tmax2). Remarkably by more increase in fiber content in PU3, the height of

peak decreases

and Tmax2 becomes more obvious. It is notable that Tmax2 can be related to soft segments which have lower interaction with hard segments, and Tmax1 is ascribed to the more restricted soft

12

segments due to interaction with hard segments. In PU5 sample, we can just see one maximum in peak (Tmax2). It seems that in PU5, Tmax1 is disappeared. It can be mentioned that in PU0 sample, most of the soft segments are restricted due to interaction with hard segments, and thus Tmax1 is just seen. In PU1 sample some hard segments are adsorbed with glass wool fiber, which reduces the soft segments restriction and increases the phase separation in PU phase. By increase in SGWF content, more hard segments are adsorbed and soft segments become freer. Consequently, in PU5 sample one maximum (Tmax2) in

is appeared.

So, based on the obtained results, the broadening of the

peak by an increase in fiber content

is probably related to the diversity of structure and distribution of free volume [46,47] which is created due to the attraction of hard segments on the SGWF surface.

Fig. 3.

curves of SGWF/polyurethane composite membranes

3.4.Gas permeability

13

Gas permeability in polymeric membranes follows the solution-diffusion mechanism [48]. At first step, gas molecules are dbsorbed with polymer segments and then diffuse through available free volume or pathways [43]. Gas diffusivity in polymeric membranes is controlled by the gas size and the ability of polymer segments to create suitable free volume, while solubility coefficient depends on gas condensability and polarity of polymer segments [49]. So attempts are to increase gas diffusivity and solubility in polymeric membranes. One of the most commonly used methods to increase the membrane performance is use of nanoparticles. The effects of nanoparticle on gas permeability of the polymeric membrane depend on polymer-filer interactions. Non-wetting interaction between polymer and filler due to formation of liquid like layer near the interface [50,51] leads to a reduction of the polymer Tg, while strong interactions can restrict the chain mobility.

14

Fig. 4. Effects of SGWF on gas permeability properties of polyurethane for pure CO2, CH4 and N2 gases at constant 10 bar pressure and ambient temperature

On the other hand, polar or non-polar nature of filler can influence the sorption of gasses at the interface. Fig. 4 shows the effect of micro size glass wool fibers on gas permeability properties of prepared polyurethane membranes at constant pressure (10 bar) and ambient temperature for pure CO2, CH4 and N2 gases. As it can be observed, the permeability of gases through the membranes varied with the following order: CO2 > CH4 > N2 These are three gases with polar, non-polar and inert nature. The critical temperature and kinetic diameter of studied gases are listed in Table 3. According to this table, more permeability of CH4 than N2 is indicative of the substantial effect of gas condensability which facilitates the gas 15

solubility in polymers. CO2 gas shows highest permeability as compared to other gases which is related to its higher critical temperature and lower kinetic diameter. As can be seen in Fig. 4, presence of SGWF greatly affects the permeability of gases. By increase in fiber content, CO2 permeability increased linearly but after a critical fiber content, its permeability was reduced. DSC results depict that polyurethane hard segments show high tendency to interact with polar glass wool fibers, in a way that some new crystalline domains were created for hard segments. Also, SEM images show good adhesion between fiber and polymer matrix and DMTA results confirmed a reduction in soft segments Tg due to the phase separation of hard and soft segments. So based on the results, there is strong interaction between fibers and polymer. The strong interaction between polymer and fiber, can distribute the free volume within interface region which can significantly reduce the gas diffusivity. So it seems that presence of polar glass wool fiber significantly affects the gas solubility. In order to understand the portion of diffusion and solubility of gases on their permeation properties, these coefficients were calculated based on the time lag method. Fig. 5 and Fig. 6 are shown the effects of glass wool content on solubility and diffusivity coefficients of studied gases. As can be seen in these figures polar glass wool fibers have interesting effects on solubility and diffusivity coefficient of CO2 and CH4 gases. Fig. 5 shows that by an increase in fiber content, CO2 and CH4 solubility are continually increased. The results indicated that presence of small content of glass wool remarkably increased the solubility coefficient for CO2 gas, but by more increase in fiber content (PU5), solubility coefficient did not show obvious changes. This behavior could be related to the increase of available polar groups of fiber and hard segments in the polymer/fiber interface which provides a suitable place for interaction with polar CO2 gas. By considering the Fig. 6 it can be seen that by an increase in fiber content in PU1 and PU3, 16

diffusion coefficient does not show intense change for CO2 gas, but by more increase in fiber content (PU5) an obvious reduction in diffusion coefficient can be observed. The diffusion coefficient would be expected to decrease with the content of the fiber owing to the tortuous path of such impermeable reinforcement. On the other hand, the presence of fibers increases the free volume of soft segments as a result of its effect on the raising of PU phase separation. According to obtained results, it can be concluded that the restriction of molecular passing direction due to fiber loading and its effect on reduction of diffusion coefficent is believed to be the dominant factor at higher content of the SGWF. So, considerable reduction in CO2 permeability for PU5 sample can be related to the reduction in diffusion coefficient (See Figs. 4 and 6). Additionally, in polymer/fiber composites due to the presence of wide interface some additional mechanism may influence the gas transport. In the polymer/fiber composites moisture can adsorb into the polymer and then diffuse along the interface by capillary flow [25]. So, in polymer/fiber composites capillary condensation mechanism may occur for condensable gases such as CO2 and CH4 to form the liquid layer of such gases and block the interface [52]. This mechanism occurs when the interaction of gas with interface walls or available voids, leads to its condensation in these places, which can influence the gas diffusivity across the membrane (Scheme 2).

Scheme 2. Mechanism of gas transport in polymer/fiber interface via capillary condensation

17

So, reducing in CO2 and CH4 diffusivities can be related to condensation mechanism and creation of tortuous pathway due to presence of long SGWF. While, inert gases like N2 don’t show condensability and its permeability is increased gradually by pressure. Due to the inert nature of N2 gas changes in its permeability can be related to available pathways for gas transport. As presented in Fig. 4, the low content of SGWF does not have noticeable effects on N2 permeability due to the opposite effect of adding fibers on the diffusion coefficient which was discussed before. But by more raising the fiber content, a slight increase in N2 permeability is shown. It seems that the increase in N2 permeability in PU5 can be related to presence of nonselective micro-voids at the SGWF/polymer interface [53]. The non-selective micro-voids may be created due to interaction of non-polar soft segments with polar SGWF which have poor adhesion at the interface (see Fig.1 (5-b)). Scheme 3 shows the weak interaction of non-polar soft segments with polar SGWF. Table 3. Kinetic diameter and critical temperature of studied gases[54]. Gas CO2 CH4 N2

Kinetic diameter ( ) 3.3 3.8 3.64

Critical temperature (K) 304.12 190.56 126.20

As shown in Fig. 5 the increase in CH4 solubility coefficient by fiber content can be explained by its condensability and phase separation of unfavorable polar groups of hard segments and more availability of soft segments which have lower polarity than hard segments to adsorb this gas.

18

Fig. 5. Effects of SGWF on solubility coefficient ( 10-3 (cm3 (STP) cm3 of polym cmHg)) of CO2 and CH4 in the prepared membranes

19

Fig. 6. Effects of SGWF on diffusivity coefficient ( 10-7 cm2/s) of CO2 and CH4 in the prepared membranes

Scheme 3. Interaction of hard segments and non-polar soft segments with SGWF

3.5.Effect of pressure on gas permeability

20

Figs. 7-9 show the effects of upstream pressure on gas permeation properties of short fiber reinforced polyurethane composite membranes. As can be seen in Fig. 7, the permeability of CO2 increases with increasing upstream gas pressure, while those of CH4 and N2 show more complex behavior. Based on Figs. 8 and 9, a reduction in the CH4 and N2 permeabilities implies that the increase in the upstream pressure causes a decrease in segments ability to create suitable free volume [55]. While the increase in CO2 permeability can be related to increase in concentration of polar and condensable CO2 gas which acts as a plasticizer and causes to increase the free volume. As can be seen in Fig. 8, by increasing the upstream pressure in PU1 and PU3 membranes, CH4 permeability shows a gradually decreasing trend, but by increase in upstream pressure to 10 bar, its permeability linearlly increased for PU0 and more obvious for PU5 samples. The pressure which changes the gas permeability behavior and causes to increase the gas permeability after a reducing trend is called plasticization pressure[56]. In this pressure, gas acts as a plasticizer for polymer segments to increase their segments mobility and create more free volume for gas transport, therefore after plasticization pressure gas permeability shows an increasing trend [57,58]. This behavior has been reported for glassy polymers. In rubbery polymers, because of high mobility of segments, there is high free volume available for gas transport and the solubility of gases follows Henry’s low [59]. While glassy polymers have nonequilibrium nature (due to the formation of excess free volume during cooling) and show additional mode of sorption in comparison with rubbery polymers [55,59,60]. So the plasticization effect of CH4 in PU0 and PU5 can be explained by dual-mobility model based on the following equation [61]:

(5)

21

In this equation CD is the gas concentration based on Henry’s law sorption, CH denotes the gas concentration based on Langmuir sorption, KD is Henry's law constant, b and CH are Langmuir hole affinity parameter and the capacity parameter, respectively. So at the first step, gases are filled excess free volumes [57] which results in a reduction of CH4 and N2 permeabilities, but by more increase in upstream pressure, plasticization effects of more condensable CH4 gas increase the permeability. Based on the results, plasticization pressure was not observed for N2 gas due to the lowest condensability between the mentioned gases. The gas permeability results revealed that the plasticization effect of CH4 was not observed for PU1 and PU3 samples but this effect is more obvious for PU5. Obtained results are indicating that in PU5 sample shows the lowest Tg among the composite membranes due to highly phase separation of polar hard and soft segments of PU in the presence of glass wool fibers. Non polar soft segments can adsorb CH4 gas and increase its solubility. Available excess free volume in non-polar soft segments may remarkably increase the plasticization ability of CH4 gas. So in PU5 sample, the permeability extensively increased due to the more available favorable sites for methane adsorption. Observed behavior for CH4 is in good correlation with solubility coefficients which are reported in Fig. 5.

22

Fig. 7. Effect of pressure on the CO2 permeability of the prepared membranes

Fig. 8. Effect of pressure on CH4 permeability of the prepared membranes 23

Fig. 9. Effects of pressure on N2 permeability of the prepared membranes

3.6. The permselectivity properties of composite membranes Figs. 10 and 11 show the effect of upstream pressure on CO2/CH4 and CO2/N2 permselectivities, respectively. As it can be seen, PU5 sample shows lowest CO2/CH4 and CO2/N2 permselectivities in compare to other samples. Based on the Fig 8, after 8 bar pressure, plasticization effect of CH4 gas caused to increase in its permeability which reduced the CO2/CH4 permselectivity. Based on Figs. 7 and 9 an increase in N2 permeability and reduction in CO2 permeability is the reason for low CO2/N2 permselectivity for PU5 sample. Fig. 12 depicted the permselectivity properties of composite membranes at constant pressure and temperature. Due to the reduction in the CO2 permeability and increase in CH4 permeability for PU5 sample which can be seen in Fig. 7 and 8, this sample shows lowest permselectivity in comparison to 24

other samples. As it can be inferred from the results, the presence of low content of SGWF considerably increased the polyurethane permselectivity properties. Fig. 12 shows that CO2/CH4 and CO2/N2 permselectivities are increased from 7.20 and 25.61 for pure polyurethane (PU0) to 13.19 and 32.99 for polyurethane composite membrane (PU1), respectively. So incorporation of 1wt% SGWF in polyurethane increased the CO2/CH4 and CO2/N2 permselectivities 83.19% and 28.81%, respectively. By more increase in SGWF content, permselectivities reduced to 47.77% and 26.90% for PU3 and PU5, respectively. This obvious reduction in CO2/CH4 permselectivity is related to increase in CH4 solubility. Indeed, by an increase in fiber content based on the DSC and DMTA results, polyurethane phase separation increased and non-polar soft segments become freer to interact with CH4 gas. So by an increase in phase separation, plasticization effect of CH4 gas for soft segments caused to an increase in its permeability and consequently a reduction in CO2/CH4 permselectivity. Also, reduction in CO2/N2 permselectivity is related to increase in soft segments mobility (confirmed by DMTA) to create suitable free volume for gas transport. Obtained permselectivity results for composite membrane indicates that SGWF has comparable effects on gas permeability in comparison to nanoparticles. In the other work which studied the effect of nano-silica on gas permeation properties of polyether based polyurethane membranes [41], 40.5% increase in CO2/CH4 permselectivity for the sample with 20 wt% nanosilica is reported, while in this work only 1 wt% SGWF caused to 83.19% increase in permselectivity. Another advantage of SGWF/polyurethane composite membrane when compared to nanocomposites is the simultaneously increase in permeability and permselectivity. PU1 and PU3 sample shows 24.82% and 59.14% increase in permeability in comparison with PU0. While use of nanoparticles normally cause reduction in permeability [62]. So, applying

25

suitable micro-size reinforcements like short fiber can remarkably increase the membrane performance for gas separation.

Fig. 10. Effect of pressure on CO2/CH4 permselectivity of the prepared membranes

26

Fig. 11. Effect of pressure on CO2/N2 permselectivity of the prepared membranes

27

Fig. 12. Effects of SGWF on CO2/CH4 and CO2/N2 permselectivities of the prepared membranes at 10 bar pressure The obtained results of gas permeability and permselectivity in the composite membranes were compared with Robeson’s upper bound line, 2008 [63] and relevant polyurethane silica nanocomposites [41,64] in Fig. 13. As can be seen, there is trade-off relationship between permeability and permselectivity in nanocomposites, in a way that by an increase in nanoparticle content permeability shows a reducing trend. The results showed that there is a straight relationship between permeability and permselectivity in polyurethane composite membranes. Based on the results increasing in solubility of more condensable and polar CO2 gas in composite membrane results in simultaneous increase in permeability and permselectivity.

28

Fig. 13. The CO2/CH4 separation performance of SGWF/polyurethane composite membranes in comparison with Robeson’s upper bound [63] and other polyurethane/nano-silica composites [41,64]

4. Conclusions Gas permeation properties of short fiber reinforced polyurethane membrane have been studied. Short fibers mechanically dispersed in polyurethane solution. SEM images confirmed good dispersion and adhesion of fibers to polyurethane matrix. Effects of SGWF on crystallization and thermal transition properties of polyurethane segments have been studied by DSC and DMTA analysis. Results showed that polar surface of SGWF acts as nucleation agent for the hard segments crystallization and changed their melting enthalpy and crystallization behavior. DMTA results clearly confirmed that presence of polar fibers considerably caused reduction in soft segment Tg, in a way that a 12 degree reduction in Tg was observed. Effects of polar fibers on gas permeation properties of polyurethane composites have been studied. Gas permeation results 29

showed that presence of polar fibers effectively influenced the solubility coefficients of condensable gases and reduced their diffusion coefficient. Addition of small content of SGWF (1wt %) remarkably increased the CO2 permeability and CO2/CH4 permselectivity from 27.88 to 34.8 barrer and 7.20 to 13.19 respectively. By increase in fiber content, phase separation of nonpolar soft segments caused a raise in permeability of non-polar gases. Comparison of short fiber/polyurethane composites membranes with Robeson’s upper bound indicated that PU1 sample shows better performance for gas separation, also PU3 membrane shows good permeability and permselectivity, but increase in fiber content after a critical point can remarkably reduce the ideal selectivity (PU5). Results showed that suitable micro size reinforcements like polar glass wool fiber for polyurethane can simultaneously increase the polymer permeability and permselectivity. While the dispersion of nanoparticles in polymer matrix is difficult and nanocomposite membranes normally show trade-off relationship between permeability and permselectivity.

References [1]

D.R. Paul, Y.P. ski. Yampol’, Polymeric gas separation membranes .pdf, (1994) 1–640.

[2]

K.C. Khulbe, C.Y. Feng, T. Matsuura, Synthetic Polymeric Membranes, Synth. Polym. Membr. Charact. by At. Force Microsc. (2008). doi:10.1007/978-3-540-73994-4.

[3]

A. Khosravi, M. Sadeghi, H.Z. Banadkohi, M.M. Talakesh, Polyurethane-silica nanocomposite membranes for separation of propane/methane and ethane/methane, Ind. Eng. Chem. Res. 53 (2014) 2011–2021. doi:10.1021/ie403322w.

[4]

S. Kim, Y.M. Lee, High performance polymer membranes for CO2 separation, Curr. Opin. Chem. Eng. 2 (2013) 238–244. doi:10.1016/j.coche.2013.03.006.

[5]

S. Hassanajili, E. Masoudi, G. Karimi, M. Khademi, Mixed matrix membranes based on polyetherurethane and polyesterurethane containing silica nanoparticles for separation of CO2/CH4 gases, Sep. Purif. Technol. 116 (2013) 1–12. doi:10.1016/j.seppur.2013.05.017. 30

[6]

B.P. Ho, K.K. Choon, M.L. Young, Gas separation properties of polysiloxane/polyether mixed soft segment urethane urea membranes, J. Memb. Sci. 204 (2002) 257–269. doi:10.1016/S0376-7388(02)00048-0.

[7]

M.B. Karimi, G. Khanbabaei, G.M.M. Sadeghi, Vegetable oil-based polyurethane membrane for gas separation, J. Memb. Sci. (2016). doi:10.1016/j.memsci.2016.12.008.

[8]

Membrane Modification, n.d. doi:10.1201/b12160.

[9]

G. Ciobanu, O. Ciobanu, Mixed-matrix membranes based on polyurethane containing nanohydroxyapatite and its potential applications, J. Appl. Polym. Sci. 41813 (2015) n/an/a.

[10] H. Cong, M. Radosz, B.F. Towler, Y. Shen, Polymer-inorganic nanocomposite membranes for gas separation, Sep. Purif. Technol. 55 (2007) 281–291. doi:10.1016/j.seppur.2006.12.017. [11] J.M. Herrera-Alonso, E. Marand, J.C. Little, S.S. Cox, Transport properties in polyurethane/clay nanocomposites as barrier materials: Effect of processing conditions, J. Memb. Sci. 337 (2009) 208–214. doi:10.1016/j.memsci.2009.03.045. [12] M. Sadeghi, M.A. Semsarzadeh, H. Moadel, Enhancement of the gas separation properties of polybenzimidazole (PBI) membrane by incorporation of silica nano particles, J. Memb. Sci. 331 (2009) 21–30. [13] S. Hassanajili, M. Khademi, P. Keshavarz, Influence of various types of silica nanoparticles on permeation properties of polyurethane/silica mixed matrix membranes, J. Memb. Sci. 453 (2014) 369–383. doi:10.1016/j.memsci.2013.10.057. [14] I. Tirouni, M. Sadeghi, M. Pakizeh, Separation of C3H8 and C2H6 from CH4 in polyurethane–zeolite 4Å and ZSM-5 mixed matrix membranes, Sep. Purif. Technol. 141 (2015) 394–402. doi:10.1016/j.seppur.2014.12.012. [15] B. Schoeman, J. Sterte, Effect of zeolite particle size on the performance of polymer – zeolite mixed matrix membranes, 175 (2000) 285–288. [16] R. Wang, H. Wang, L. Sun, E. Wang, Y. Zhu, Y. Zhu, The fabrication and tribological behavior of epoxy composites modified by the three-dimensional polyurethane sponge reinforced with dopamine functionalized carbon nanotubes, Appl. Surf. Sci. (2015). 31

doi:10.1016/j.apsusc.2015.10.222. [17] S. Du, F. Li, H. Xiao, Y. Li, N. Hu, S. Fu, Tensile and fl exural properties of graphene oxide coated-short glass fi ber reinforced polyethersulfone composites, Compos. Part B. 99 (2016) 407–415. doi:10.1016/j.compositesb.2016.06.023. [18] Y. Zhang, Y. Zhang, Y. Liu, X. Wang, B. Yang, A novel surface modification of carbon fiber for high-performance thermoplastic polyurethane composites, Appl. Surf. Sci. 382 (2016) 144–154. doi:10.1016/j.apsusc.2016.04.118. [19] S. Fu, Y. Mai, B. Lauke, C. Yue, Synergistic effect on the fracture toughness of hybrid short glass fiber and short carbon fiber reinforced polypropylene composites, 323 (2006) 326–335. [20] A. Kootsookos, A.P. Mouritz, Seawater durability of glass- and carbon-polymer composites, Compos. Sci. Technol. 64 (2004) 1503–1511. doi:10.1016/j.compscitech.2003.10.019. [21] Z. Zhang, H. Song, X. Men, Z. Luo, Effect of carbon fibers surface treatment on tribological performance of polyurethane ( PU ) composite coating, Wear. 264 (2008) 599–605. doi:10.1016/j.wear.2007.05.003. [22] R.-M. Wang, S.-R. Zheng, Y.-P. Zheng, Polymer matrix composites and technology, n.d. [23] V. Cech, E. Palesch, J. Lukes, The glass fiber – polymer matrix interface / interphase characterized by nanoscale imaging techniques, Compos. Sci. Technol. 83 (2013) 22–26. doi:10.1016/j.compscitech.2013.04.014. [24] C.B.Y. Ftir-imaging, A.D.S.C. Techniques, To study the effect of hygrothermal ageing at the interface of glass/epoxy micro-composites by FTIR-imaging and DSC techniques, (2007). [25] C.J. Tsenoglou, S. Pavlidou, C.D. Papaspyrides, Evaluation of interfacial relaxation due to water absorption in fiber-polymer composites, Compos. Sci. Technol. 66 (2006) 2855– 2864. doi:10.1016/j.compscitech.2006.02.022. [26] A. Shalwan, B.F. Yousif, In State of Art  : Mechanical and tribological behaviour of polymeric composites based on natural fibres, J. Mater. (2012). doi:10.1016/j.matdes.2012.07.014. 32

[27] J. Park, P. Shin, Z. Wang, D. Kwon, J. Choi, S. Lee, K.L. Devries, The Change in Mechanical and Interfacial Properties of GF and CF Reinforced Epoxy Composites after Aging in NaCl Solution, Compos. Sci. Technol. (2015). doi:10.1016/j.compscitech.2015.11.011. [28] S.C. Amico, Carbon fiber-reinforced epoxy filament-wound composite laminates exposed to hygrothermal conditioning, (2016). doi:10.1007/s10853-016-9787-9. [29] N. Tual, N. Carrere, P. Davies, T. Bonnemains, E. Lolive, Characterization of sea water ageing effects on mechanical properties of carbon / epoxy composites for tidal turbine blades, Compos. PART A. 78 (2015) 380–389. doi:10.1016/j.compositesa.2015.08.035. [30] M. Ionescu, Chemistry and Technology of Polyols for Polyurethane, 2008. doi:10.1002/pi.2159. [31] M. Szycher, Szycher’s Handbook of Polyurethanes, Second Edition, 2012. doi:10.1002/0471743984.vse5703. [32] D.K. Chattopadhyay, B. Sreedhar, K.V.S.N. Raju, Influence of varying hard segments on the properties of chemically crosslinked moisture-cured polyurethane-urea, J. Polym. Sci. Part B Polym. Phys. 44 (2006) 102–118. [33] S. Velankar, S.L. Cooper, Microphase separation and rheological properties of polyurethane melts. 1. Eff. block length, Macromolecules. 33 (2000) 9181–9192. [34] J.T. Koberstein, R.S. Stein, Small-angle X-ray scattering studies of microdomain structure in segmented polyurethane elastomers, J. Polym. Sci. Polym. Phys. Ed. 21 (1983) 1439– 1472. doi:10.1002/pol.1983.180210814. [35] A.P. Isfahani, B. Ghalei, K. Wakimoto, R. Bagheri, E. Sivaniah, M. Sadeghi, Plasticization resistant crosslinked polyurethane gas separation membranes, J. Mater. Chem. A. 4 (2016) 17431–17439. [36] M. Amrollahi, G.M.M. Sadeghi, Assessment of adhesion and surface properties of polyurethanecoatings based on non-polar and hydrophobic soft segment, Prog. Org. Coatings. 93 (2016) 23–33. [37] 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, J. Memb. 33

Sci. 354 (2010) 40–47. doi:10.1016/j.memsci.2010.02.070. [38] M. Sadeghi, M.A. Semsarzadeh, M. Barikani, B. Ghalei, Study on the morphology and gas permeation property of polyurethane membranes, J. Memb. Sci. 385–386 (2011) 76– 85. doi:10.1016/j.memsci.2011.09.024. [39] A. Woli´nska-Grabczyk, A. Jankowski, Gas transport properties of segmented polyurethanes varying in the kind of soft segments, Sep. Purif. Technol. 57 (2007) 413– 417. [40] S. Hassanajili, M.T. Sajedi, Fumed silica/polyurethane nanocomposites: effect of silica concentration and its surface, Iran Polym J. (2016). [41] M. Sadeghia, M.A. Semsarzadeh, M. Barikani, M. Pourafshari Chenard, Gas separation properties of polyether-based polyurethane–silica nanocomposite membranes, J. Memb. Sci. 376 (2011) 188–195. [42] A. Eceiza, M.D. Martin, K. De La Caba, G. Kortaberria, N. Gabilondo, M.A. Corcuera, I. Mondragon, Thermoplastic Polyurethane Elastomers Based on Polycarbonate Diols With Different Soft Segment Molecular Weight and Chemical Structure: Mechanical and Thermal Properties, Polym. Eng. Sci. 48 (2008). [43] S. Mondal, J.L. Hu, Structural characterization and mass transfer properties of nonporoussegmented polyurethane membrane: Influence of the hydrophilic segment content and soft segment melting temperature, J. Memb. Sci. 276 (2006) 16–22. doi:10.1016/j.memsci.2005.09.029. [44] Q. Wang, Y. Bai, Y. Chen, J. Ju, High performance shape memory polyimides based on p–p interactions, J. Mater. Chem. A. 3 (2014) 352–359. doi:10.1039/C4TA05058D. [45] C. Zhang, Y. Li, R. Chen, M.R. Kessler, Polyurethanes from Solvent-Free Vegetable OilBased Polyols, (2014). [46] X. Kong, G. Liu, J.M. Curtis, International Journal of Adhesion & Adhesives Characterization of canola oil based polyurethane wood adhesives, Int. J. Adhes. Adhes. 31 (2011) 559–564. doi:10.1016/j.ijadhadh.2011.05.004. [47] S. Zhou, L. Wu, J. Sun, W. Shen, The change of the properties of acrylic-based polyurethane via addition of nano-silica, Prog. Org. Coatings. 45 (2002) 33–42. 34

[48] J.G. Wijmans, R.W. Baker, The solution-diffusion model: A review, J. Memb. Sci. 107 (1995) 1–21. [49] A. Khosravi, M. Sadeghi, H.Z. Banadkohi, M.M. Talakesh, Polyurethane-silica nanocomposite membranes for separation of propane/methane and ethane/methane, Ind. Eng. Chem. Res. 53 (2014) 2011–2021. doi:10.1021/ie403322w. [50] H. Koolivand, A. Sharif, M.R. Kashani, M. Karimi, M.K. Salooki, M.A. Semsarzadeh, Functionalized graphene oxide/polyimide nanocomposites as highly CO2-selective membranes, J Polym Res. 599 (2014) 1–12. [51] B.J. Ash, R.W. Siegel, L.S. Schadler, Glass-Transition Temperature Behavior of Alumina / PMMA Nanocomposites, J. Polym. Sci. Part B Polym. Phys. (2004) 4371–4383. doi:10.1002/polb.20297. [52] A. Javaid, Membranes for solubility-based gas separation applications, Chem. Eng. J. 112 (2005) 219–226. doi:10.1016/j.cej.2005.07.010. [53] M. a. Aroon, a. F. Ismail, T. Matsuura, M.M. Montazer-Rahmati, Performance studies of mixed matrix membranes for gas separation: A review, Sep. Purif. Technol. 75 (2010) 229–242. doi:10.1016/j.seppur.2010.08.023. [54] B. Freeman, Y. Yampolskii, I. Pinnau, Mater. Sci. Membr. Gas. Vapor Sep, 2006. doi:10.1002/047002903X. [55] L.S. Teo, C.Y. Chen, J.F. Kuo, The gas transport properties of amine-containing polyurethane and poly(urethane-urea) membranes, J. Memb. Sci. 141 (1998) 91–99. doi:10.1016/S0376-7388(97)00293-7. [56] S. Kanehashi, G.Q. Chen, D. Danaci, P.A. Webley, E. Sandra, Can the addition of carbon nanoparticles to a polyimide membrane reduce plasticization  ?, Sep. Purif. Technol. 183 (2017) 333–340. doi:10.1016/j.seppur.2017.04.013. [57] R.R. Tiwari, J. Jin, B.D. Freeman, D.R. Paul, Physical aging, CO2 sorption and plasticization in thin films of polymer with intrinsic microporosity (PIM-1), J. Memb. Sci. (2017). doi:10.1016/j.memsci.2017.04.069. [58] C.A. Scholes, G. Dong, J.S. Kim, H.J. Jo, J. Lee, Y.M. Lee, Permeation and separation of SO2, H2S and CO2 through thermally rearranged (TR) polymeric membranes, Sep. Purif. 35

Technol. 179 (2016) 449–454. doi:10.1016/j.seppur.2016.12.039. [59] D.R. Robeson, L. M., Liu, Q., Freeman, B. D., & Paul, Comparison of transport properties of rubbery and glassy polymers and the relevance to the upper bound relationship, J. Memb. Sci. 476 (2015) 421–431. doi:10.1016/j.memsci.2014.11.058. [60] E. Ricci, M. Minelli, M.G. De, A multiscale approach to predict the mixed gas separation performance of glassy polymeric membranes for CO2 capture: the case of CO2/CH4 mixture in Matrimid®, J. Memb. Sci. (2017). doi:10.1016/j.memsci.2017.05.068. [61] L.S. Teo, C.Y. Chen, J.F. Kuo, The gas transport properties of amine-containing polyurethane and poly(urethane-urea) membranes, J. Memb. Sci. 141 (1998) 91–99. [62] S.R. Reijerkerk, M.H. Knoef, K. Nijmeijer, M. Wessling, Poly(ethylene glycol) and poly(dimethyl siloxane): Combining their advantages into efficient CO2 gas separation membranes, J. Memb. Sci. 352 (2010) 126–135. doi:10.1016/j.memsci.2010.02.008. [63] L.M. Robeson, The upper bound revisited, 320 (2008) 390–400. doi:10.1016/j.memsci.2008.04.030. [64] M. Sadeghi, M.M. Talakesh, B. Ghalei, M.R. Shafiei, Preparation, characterization and gas permeation properties of a polycaprolactone based polyurethane-silica nanocomposite membrane, J. Memb. Sci. 427 (2013) 21–29.

36

Graphical Abstract

Highlights    

Short glass wool fiber (SGWF) was used as a polar micro size reinforcement for polyurethane membrane. Polar surface of glass wool fiber is a suitable site for attraction of hard segments and increased the polyurethane micro-phase separation. Presence of glass wool fiber remarkably reduced the soft segments glass transition temperature Presence of SGWF caused a simultaneous increase in permeability and ideal selectivity (permselectivity) of polyurethane membrane.

37