Temperature and Pressure Dependence of Gas Permeation in a Microporous Tröger’s Base Polymer

Gas transport properties of PIM-EA(H2)-TB, a microporous Tröger’s base polymer, were systematically studied over a range of pressure and temperature. Permeability coefficients of pure CO2, N2, CH4 and H2 were determined for upstream pressures up to 20 bar and temperatures up to 200 °C. PIM-EA(H2)-TB exhibited high permeability coefficients in absence of plasticization phenomena. The permeability coefficient of N2, CH4 and H2 increased with increasing temperature while CO2 permeability decreased with increasing temperature as expected for a glassy polymer. The diffusion and solubility coefficients were also analysed individually and compared with other polymers of intrinsic microporosity. From these results, the activation energies of permeation, diffusion and sorption enthalpies were calculated using an Arrhenius equation.


Introduction
Membranes are one of the most promising technologies to compete with conventional separation processes for gas separations including post-and pre-combustion carbon capture. Studies on the use of polymeric membranes in an Integrated Gasification Combined Cycle (IGCC) power plant [1][2][3] show their viability and their competitiveness with the currently more developed solvent-based technology. Process simulations [1] have shown an advantage for hydrogen selective materials for this application and new membrane materials are currently under development [3]. The performance of the materials in the relatively harsh conditions of the separation (50 bars and 200 • C) needs to be investigated before production can be scaled up [4].
The increase of gas pressure can have a negative impact on membrane performance, due to plasticization effects. For glassy polymers, many gases, such as O 2 , N 2 and H 2 , can permeate through the polymer without modifying the polymer's properties due to their relatively low solubility in the polymer. Therefore, with the pressure's increase, the gas permeability slightly decreases, as expected from the dual sorption-dual mobility model [5]. On the contrary, highly sorbing gases such as CO 2 can induce a swelling of the polymer matrix, that is, plasticization, leading to a large increase of the gas permeability with increasing pressure. In addition, the influence of temperature on gas separation performance has been investigated for a large number of polymers. Depending on the polymer, the membrane performance can be improved by an increase of temperature as shown by the Robeson plot in Figure 1. Most polymers, including ultrahigh-free volume polymers such as PTMSP and Teflon Figure 1. Influence of temperature on membrane performance (calculated from [7]) [6,[8][9][10][11].
For the first Polymer of Intrinsic Microporosity, PIM-1, Budd et al. [8] showed that the CO2 permeability coefficient decreased gradually as the temperature increased, whereas the H2 permeability coefficient increased. Thus, PIM-1 also becomes slightly more H2 selective at higher temperature. Recently, Fuoco et al. [12] studied the temperature dependence of gas permeation in triptycene-based ultrapermeable PIMs, such as PIM-TMN-Trip. With increasing temperature, the permeability coefficient increased for the bulkier penetrants (N2 and CH4), while for the faster penetrants (CO2 and O2) it decreased and for the very small penetrants (H2 and He) it was constant. Therefore, PIM-TMN-Trip became more selective to H2 at high temperature; these ultrapermeable polymers behave as microporous solids, in which the pore dimensions are rather large in comparison with the diffusing gas molecules. Such studies of the temperature and pressure dependence of transport properties are essential for understanding the behaviour of membranes over a wide range of conditions, in order to assist any consideration of industrial use.
For the first Polymer of Intrinsic Microporosity, PIM-1, Budd et al. [8] showed that the CO 2 permeability coefficient decreased gradually as the temperature increased, whereas the H 2 permeability coefficient increased. Thus, PIM-1 also becomes slightly more H 2 selective at higher temperature. Recently, Fuoco et al. [12] studied the temperature dependence of gas permeation in triptycene-based ultrapermeable PIMs, such as PIM-TMN-Trip. With increasing temperature, the permeability coefficient increased for the bulkier penetrants (N 2 and CH 4 ), while for the faster penetrants (CO 2 and O 2 ) it decreased and for the very small penetrants (H 2 and He) it was constant. Therefore, PIM-TMN-Trip became more selective to H 2 at high temperature; these ultrapermeable polymers behave as microporous solids, in which the pore dimensions are rather large in comparison with the diffusing gas molecules. Such studies of the temperature and pressure dependence of transport properties are essential for understanding the behaviour of membranes over a wide range of conditions, in order to assist any consideration of industrial use.
PIM-EA(H 2 )-TB differs from PIM-EA(Me 2 )-TB only by the absence of methyl groups at the bridgehead (9,10) position of the ethanoanthracene (EA) unit, which modifies its chain packing in the solid state. PIM-EA(H 2 )-TB presents an inter-chain distance, d-space, of 7.7 Å and 32% free volume, whereas PIM-EA(Me)-TB has values of 11 Å and 30%, respectively [18]. With these differences, a higher separation performance for PIM-EA(H 2 )-TB is expected. However, few papers have been published on this polymer. Bernardo et al. [15] developed thin film composite based on PIM-EA(H 2 )-TB and they studied the impact of the residual casting solvent on the separation performance at 25 • C and 1 bar. In addition, Benito et al. [19] studied composite membranes based on a ultrathin layer of PIM-EA(H 2 )-TB for CO 2 /N 2 separation at 35 • C and 3 bar.
Here we report a novel study on the permeation properties of PIM-EA(H 2 )-TB over a large temperature and pressure range for a series of gases (CO 2 , H 2 , N 2 and CH 4 ).
polymers behave as microporous solids, in which the pore dimensions are rather large in comparison with the diffusing gas molecules. Such studies of the temperature and pressure dependence of transport properties are essential for understanding the behaviour of membranes over a wide range of conditions, in order to assist any consideration of industrial use.

Experimental Section
The detailed synthetic procedure for making PIM-EA(H 2 )-TB and its structural characterization are reported elsewhere [15]. Robust flat films of thickness between 130 and 200 µm were cast from chloroform with their thickness determined using a digital micrometre (Mitutoyo, Kawasaki, Japan). The permeation properties were measured in a constant volume-variable pressure apparatus ( Figure 3) using pure CO 2 , N 2 , CH 4  PIM-EA(H2)-TB differs from PIM-EA(Me2)-TB only by the absence of methyl groups at the bridgehead (9,10) position of the ethanoanthracene (EA) unit, which modifies its chain packing in the solid state. PIM-EA(H2)-TB presents an inter-chain distance, d-space, of 7.7 Å and 32% free volume, whereas PIM-EA(Me)-TB has values of 11 Å and 30%, respectively [18]. With these differences, a higher separation performance for PIM-EA(H2)-TB is expected. However, few papers have been published on this polymer. Bernardo et al. [15] developed thin film composite based on PIM-EA(H2)-TB and they studied the impact of the residual casting solvent on the separation performance at 25 °C and 1 bar. In addition, Benito et al. [19] studied composite membranes based on a ultrathin layer of PIM-EA(H2)-TB for CO2/N2 separation at 35 °C and 3 bar.
Here we report a novel study on the permeation properties of PIM-EA(H2)-TB over a large temperature and pressure range for a series of gases (CO2, H2, N2 and CH4).

Experimental Section
The detailed synthetic procedure for making PIM-EA(H2)-TB and its structural characterization are reported elsewhere [15]. Robust flat films of thickness between 130 and 200 μm were cast from chloroform with their thickness determined using a digital micrometre (Mitutoyo, Kawasaki, Japan). The permeation properties were measured in a constant volume-variable pressure apparatus ( Figure  3) using pure CO2, N2, CH4 and H2 (Table 1) with pressures up to 20 bar (10 bar for H2) and temperatures up to 200 °C. For each measurement campaign (i.e., one gas and either variable T or variable P), the sample was carefully treated with methanol prior to the measurement in order to start from the same ageing history. The methanol treatment consists of soaking the sample in methanol for 2 h, drying it under ambient conditions for 20 min and under vacuum at 30 °C overnight. At the end of the campaign, the gas permeability at 30 °C and 1 bar was re-measured in order to check the absence of physical/chemical ageing. Moreover, each campaign's duration was short, carried out over a maximum of 3 days in order to limit physical ageing. By using this procedure, the physical ageing was minimised and had no apparent impact on the results for permeability and selectivity.   For each measurement campaign (i.e., one gas and either variable T or variable P), the sample was carefully treated with methanol prior to the measurement in order to start from the same ageing history. The methanol treatment consists of soaking the sample in methanol for 2 h, drying it under ambient conditions for 20 min and under vacuum at 30 • C overnight. At the end of the campaign, the gas permeability at 30 • C and 1 bar was re-measured in order to check the absence of physical/chemical ageing. Moreover, each campaign's duration was short, carried out over a maximum of 3 days in order to limit physical ageing. By using this procedure, the physical ageing was minimised and had no apparent impact on the results for permeability and selectivity.
The permeability was obtained from the evolution of pressure of the downstream side (MKS Baratron 615A (Andover, MA, USA)). The permeability coefficient, P, was determined from the slope of the pressure vs. time curve under steady state condition. Before each experiment, the apparatus was vacuum-degassed and a leak rate determined from the pressure increase in the downstream part. Three different downstream volumes could be selected accordingly to the permeation rate of the gas.
The time lag, θ, was used to determine the diffusivity coefficient D (Equation (1)).
The solubility coefficient, S, for the gas in the polymer was evaluated indirectly, assuming the validity of the diffusion-solution mechanism (Equation (2)): The ideal selectivity between two gas species i and j is the ratio of the two single gas permeabilities (Equation (3)).

Permeability
Permeation measurements on methanol treated films of PIM-EA(H 2 )-TB were carried out using pure N 2 , H 2 , CO 2 and CH 4 at several pressures (1 to 20 bar) and temperatures (30 • C to 200 • C). Table 2 reports the results from the time lag experiment at 30 • C and 1 bar. PIM-EA(H 2 )-TB presents high CO 2 and H 2 permeability coefficients and good ideal selectivity over N 2 and CH 4 . The order of gas permeabilities for PIM-EA(H 2 )-TB is CO 2 > H 2 > CH 4 > N 2 , the same as that for PIM-1. CO 2 , which is the most condensable gas, is the most permeable due to the predominant role of solubility in PIMs [8]. In comparison with PIM-EA(Me 2 )-TB, the permeability coefficients obtained for PIM-EA(H 2 )-TB are lower. This can be explained by the methyl groups increasing the distance between polymer chains of PIM-EA(Me 2 )-TB, relative to PIM-EA(H 2 )-TB, which ensures higher free volume and, hence, higher permeability [16] but reduces selectivity.  As shown on Figure 4, the data for PIM-EA(H2)-TB are located above the 2008 upper bound for all five gas pairs. For H2/CH4 and H2/N2, they are clearly higher than for PIM-1 and PIM-EA(Me)-TB. This demonstrates the potential of PIM-EA(H2)-TB for industrial applications, such as carbon capture (CO2/N2 mixture), natural gas sweetening and biogas treatment (CO2/CH4 mixture) or hydrogen recovery (H2/CH4 mixture).

Diffusivity and Solubility Coefficients
The gas transport in PIM-EA(H2)-TB was analysed using the solution-diffusion model (Equation (2)), to provide the diffusivity and sorption coefficients (Table 3).
The diffusivity and solubility values of PIM-EA(H2)-TB are similar to those of polymers from the same family (PIM-EA(Me)-TB) [16] with a very high value of CO2 solubility coefficient. This affinity towards CO2 may be enhanced by the presence of the amine groups in the TB moiety.
Diffusivity and solubility data are plotted in Figure 5 as correlations of log D versus d 2 and log S versus Tc, respectively, where d is the kinetic diameter and Tc is the critical temperature of the gases.  As shown on Figure 4, the data for PIM-EA(H2)-TB are located above the 2008 upper bound for all five gas pairs. For H2/CH4 and H2/N2, they are clearly higher than for PIM-1 and PIM-EA(Me)-TB. This demonstrates the potential of PIM-EA(H2)-TB for industrial applications, such as carbon capture (CO2/N2 mixture), natural gas sweetening and biogas treatment (CO2/CH4 mixture) or hydrogen recovery (H2/CH4 mixture).

Diffusivity and Solubility Coefficients
The gas transport in PIM-EA(H2)-TB was analysed using the solution-diffusion model (Equation (2)), to provide the diffusivity and sorption coefficients (Table 3). The diffusivity and solubility values of PIM-EA(H2)-TB are similar to those of polymers from the same family (PIM-EA(Me)-TB) [16] with a very high value of CO2 solubility coefficient. This affinity towards CO2 may be enhanced by the presence of the amine groups in the TB moiety.
Diffusivity and solubility data are plotted in Figure 5 as correlations of log D versus d 2 and log S versus Tc, respectively, where d is the kinetic diameter and Tc is the critical temperature of the gases.  As shown on Figure 4, the data for PIM-EA(H2)-TB are located above the 2008 upper bound for all five gas pairs. For H2/CH4 and H2/N2, they are clearly higher than for PIM-1 and PIM-EA(Me)-TB. This demonstrates the potential of PIM-EA(H2)-TB for industrial applications, such as carbon capture (CO2/N2 mixture), natural gas sweetening and biogas treatment (CO2/CH4 mixture) or hydrogen recovery (H2/CH4 mixture).

Diffusivity and Solubility Coefficients
The gas transport in PIM-EA(H2)-TB was analysed using the solution-diffusion model (Equation (2)), to provide the diffusivity and sorption coefficients (Table 3). The diffusivity and solubility values of PIM-EA(H2)-TB are similar to those of polymers from the same family (PIM-EA(Me)-TB) [16] with a very high value of CO2 solubility coefficient. This affinity towards CO2 may be enhanced by the presence of the amine groups in the TB moiety.
Diffusivity and solubility data are plotted in Figure 5 as correlations of log D versus d 2 and log S versus Tc, respectively, where d is the kinetic diameter and Tc is the critical temperature of the gases.
) at 30 • C and 1 bar. The lines represents the 2008 upper bound for each gas pair [21].
As shown on Figure 4, the data for PIM-EA(H 2 )-TB are located above the 2008 upper bound for all five gas pairs. For H 2 /CH 4 and H 2 /N 2 , they are clearly higher than for PIM-1 and PIM-EA(Me)-TB. This demonstrates the potential of PIM-EA(H 2 )-TB for industrial applications, such as carbon capture (CO 2 /N 2 mixture), natural gas sweetening and biogas treatment (CO 2 /CH 4 mixture) or hydrogen recovery (H 2 /CH 4 mixture).

Diffusivity and Solubility Coefficients
The gas transport in PIM-EA(H 2 )-TB was analysed using the solution-diffusion model (Equation (2)), to provide the diffusivity and sorption coefficients (Table 3).
The diffusivity and solubility values of PIM-EA(H 2 )-TB are similar to those of polymers from the same family (PIM-EA(Me)-TB) [16] with a very high value of CO 2 solubility coefficient. This affinity towards CO 2 may be enhanced by the presence of the amine groups in the TB moiety.
Diffusivity and solubility data are plotted in Figure 5 as correlations of log D versus d 2 and log S versus T c , respectively, where d is the kinetic diameter and T c is the critical temperature of the gases.   Figure 5a shows that the diffusivity coefficient of PIM-EA(H2)-TB decreases with increasing molecular size of the permeate. Larger molecules interact with more segments of the polymer chains than do smaller molecules and thus the mobility selectivity always favours the passage of smaller molecules over larger ones [20]. Moreover, this decrease is large due to the glassy state of the polymer where the highly rigid polymer chains of PIM-EA(H2)-TB are essentially fixed and do not move readily to accommodate the transport of larger molecules. It is noteworthy that the value of diffusivity for CO2 is slightly lower than for N2. Generally, in polymers, the smaller molecule, that is, CO2, is expected to diffuse faster than N2, which is a larger molecule. This unusual inversion is found for polymer with high CO2 affinity [13,17,22] and is caused by the specific interaction between CO2 and amine groups slowing diffusion [23].
The sorption coefficient of the gas within PIM-EA(H2)-TB increases with its critical temperature (i.e., its condensability) as is usually observed for polymers (Figure 5b).

Effect of Pressure
The permeability coefficients of each gas were measured as a function of upstream feed pressure. The measurements were carried out with H2, CO2, CH4 and N2 at 30 °C and pressures up to 20 bar (10 bar for H2) ( Figure 6).  Figure 5a shows that the diffusivity coefficient of PIM-EA(H 2 )-TB decreases with increasing molecular size of the permeate. Larger molecules interact with more segments of the polymer chains than do smaller molecules and thus the mobility selectivity always favours the passage of smaller molecules over larger ones [20]. Moreover, this decrease is large due to the glassy state of the polymer where the highly rigid polymer chains of PIM-EA(H 2 )-TB are essentially fixed and do not move readily to accommodate the transport of larger molecules. It is noteworthy that the value of diffusivity for CO 2 is slightly lower than for N 2 . Generally, in polymers, the smaller molecule, that is, CO 2 , is expected to diffuse faster than N 2 , which is a larger molecule. This unusual inversion is found for polymer with high CO 2 affinity [13,17,22] and is caused by the specific interaction between CO 2 and amine groups slowing diffusion [23].
The sorption coefficient of the gas within PIM-EA(H 2 )-TB increases with its critical temperature (i.e., its condensability) as is usually observed for polymers (Figure 5b).

Effect of Pressure
The permeability coefficients of each gas were measured as a function of upstream feed pressure. The measurements were carried out with H 2 , CO 2 , CH 4 and N 2 at 30 • C and pressures up to 20 bar (10 bar for H 2 ) ( Figure 6). The permeability of nitrogen is constant with increasing pressure while CO2 and CH4 permeabilities decrease with increasing pressure, which is classical behaviour for glassy polymers [24] and is due to the filling of Langmuir sorption sites. At higher pressures, the contribution of the Langmuir region to the overall permeability is weaker and gas permeability approaches a constant value associated with simple dissolution (Henry's law) transport. In contrast to the majority of glassy polymers, PIM-EA(H2)-TB does not exhibit the typical increase in CO2 permeability associated with "plasticization" in the high pressure range for CO2. A similar behaviour has been also noted for other polymers of intrinsic microporosity, such as PIM-1 or PIM-EA(Me)-TB [17,24,25]. However, the decrease in H2 permeability is higher than expected [25].
Despite the decrease of permeability coefficients, the ideal selectivities of PIM-EA(H2)-TB stay constant with the increase of the feed pressure (Table 4). However, it should be noted that ideal selectivity is usually not representative of behaviour at high pressure in mixed gas systems due to the interactions between different gases.

Effect of Temperature
The temperature effect on gas permeability through PIM-EA(H2)-TB was studied over a temperature range of 30-200 °C for pure gas at different pressures. The values of the permeability coefficients are summarised in the Table S1. Figure 7 shows the permeability coefficient of N2, CO2, H2 and CH4 as a function of the inverse absolute temperature at 1 bar. The permeability of nitrogen is constant with increasing pressure while CO 2 and CH 4 permeabilities decrease with increasing pressure, which is classical behaviour for glassy polymers [24] and is due to the filling of Langmuir sorption sites. At higher pressures, the contribution of the Langmuir region to the overall permeability is weaker and gas permeability approaches a constant value associated with simple dissolution (Henry's law) transport. In contrast to the majority of glassy polymers, PIM-EA(H 2 )-TB does not exhibit the typical increase in CO 2 permeability associated with "plasticization" in the high pressure range for CO 2 . A similar behaviour has been also noted for other polymers of intrinsic microporosity, such as PIM-1 or PIM-EA(Me)-TB [17,24,25]. However, the decrease in H 2 permeability is higher than expected [25].
Despite the decrease of permeability coefficients, the ideal selectivities of PIM-EA(H 2 )-TB stay constant with the increase of the feed pressure (Table 4). However, it should be noted that ideal selectivity is usually not representative of behaviour at high pressure in mixed gas systems due to the interactions between different gases.  1  22  14  25  16  2  5 bar  1  20  -24  --10 bar  1  20  14  25  16  2  20 bar  ---23 14 2

Effect of Temperature
The temperature effect on gas permeability through PIM-EA(H 2 )-TB was studied over a temperature range of 30-200 • C for pure gas at different pressures. The values of the permeability coefficients are summarised in the Table S1. Figure 7 shows the permeability coefficient of N 2 , CO 2 , H 2 and CH 4 as a function of the inverse absolute temperature at 1 bar. The permeability coefficient of N2, CH4 and H2 increases with increasing temperature while for CO2 it decreases with increasing temperature. In order to explore the temperature dependence of the gas permeability, the data were correlated with the Arrhenius equation.
where P0 is the pre-exponential factor ((cm 3 (STP)·cm)/(cm 2 ·s·cmHg)), Ep is the activation energy of permeation (J/mol), T is the temperature (K) and R is the ideal gas constant (8.314 kJ/(mol·K)). Ep for the transport of each gas through PIM-EA(H2)-TB were determined from the slopes (−EP/R) of the best curve-fits through the permeation data in Figure 7. The Ep values at 1 bar are summarized in Table 5. PIM-EA(H2)-TB presents high values for the activation energy of permeation for N2 and CH4, which means that the permeability coefficients depend strongly on the temperature. On the contrary, for the smaller gases, such as H2, EP is close to zero as the dependence on temperature is much weaker. For CO2, the activation energy of permeation is negative. This behaviour is routinely observed for microporous solids, such as PIM-1, PIM-TMN-Trip and PTMSP, in which the pore dimensions are relatively large in comparison with the diffusing gas molecules [11].
Since the gas transport in a microporous membrane is based on a solution-diffusion mechanism, the impact of temperature on the permeation can be better understood when looking at diffusion and solubility individually. The activation energy of permeation can be represented as the sum of the activation energies of diffusion, ED and sorption ΔHs. Table 6 lists the activation energies of gas permeation and diffusion as well as the enthalpy of sorption of all the gases in PIM-EA(H2)-TB. For all the gases at 1 bar, the activation energy of diffusion, ED, is positive, which means that the diffusivity increases with the temperature, which is expected as the main effect of increasing the The permeability coefficient of N 2 , CH 4 and H 2 increases with increasing temperature while for CO 2 it decreases with increasing temperature. In order to explore the temperature dependence of the gas permeability, the data were correlated with the Arrhenius equation.
where P 0 is the pre-exponential factor ((cm 3 (STP)·cm)/(cm 2 ·s·cmHg)), E p is the activation energy of permeation (J/mol), T is the temperature (K) and R is the ideal gas constant (8.314 kJ/(mol·K)). E p for the transport of each gas through PIM-EA(H 2 )-TB were determined from the slopes (−E P /R) of the best curve-fits through the permeation data in Figure 7. The E p values at 1 bar are summarized in Table 5. PIM-EA(H 2 )-TB presents high values for the activation energy of permeation for N 2 and CH 4 , which means that the permeability coefficients depend strongly on the temperature. On the contrary, for the smaller gases, such as H 2 , E P is close to zero as the dependence on temperature is much weaker. For CO 2 , the activation energy of permeation is negative. This behaviour is routinely observed for microporous solids, such as PIM-1, PIM-TMN-Trip and PTMSP, in which the pore dimensions are relatively large in comparison with the diffusing gas molecules [11].
Since the gas transport in a microporous membrane is based on a solution-diffusion mechanism, the impact of temperature on the permeation can be better understood when looking at diffusion and solubility individually. The activation energy of permeation can be represented as the sum of the activation energies of diffusion, E D and sorption ∆H s . Table 6 lists the activation energies of gas permeation and diffusion as well as the enthalpy of sorption of all the gases in PIM-EA(H 2 )-TB. For all the gases at 1 bar, the activation energy of diffusion, E D , is positive, which means that the diffusivity increases with the temperature, which is expected as the main effect of increasing the temperature is an increase of molecular vibrations which facilitates diffusion. In contrary, the sorption enthalpy, ∆H s , is negative as expected since the sorption is an exothermic process. For CH 4 , N 2 and H 2 , the absolute value of E D is greater than ∆H s and so the energy of activation E p is positive, which means that diffusion rather than sorption dominates the response of permeation to temperature. For CO 2 , the absolute value of E D is smaller than ∆H s , which induces a negative activation energy E P . The CO 2 transport is mainly influenced by the gas solubility, which is characteristic of microporous polymer, with similar results being found for PIM-1 and PTMSP [5,8,11].
Based upon these effects, the increase of temperature improves H 2 /CO 2 selectivity modestly moving the data for PIM-EA(H 2 )-TB close to the 200 • C upper bound (Figure 8, however, even its enhanced high temperature selectivity (~2) is insufficient for viable pre-combustion application. In contrast, the selectivity for CO 2 or H 2 over N 2 or CH 4 decreases dramatically at higher temperatures suggesting that optimal performance is obtained at lower temperatures ( Figure 9). temperature is an increase of molecular vibrations which facilitates diffusion. In contrary, the sorption enthalpy, ΔHs, is negative as expected since the sorption is an exothermic process. For CH4, N2 and H2, the absolute value of ED is greater than ΔHs and so the energy of activation Ep is positive, which means that diffusion rather than sorption dominates the response of permeation to temperature. For CO2, the absolute value of ED is smaller than ΔHs, which induces a negative activation energy EP. The CO2 transport is mainly influenced by the gas solubility, which is characteristic of microporous polymer, with similar results being found for PIM-1 and PTMSP [5,8,11].
Based upon these effects, the increase of temperature improves H2/CO2 selectivity modestly moving the data for PIM-EA(H2)-TB close to the 200 °C upper bound (Figure 8, however, even its enhanced high temperature selectivity (~2) is insufficient for viable pre-combustion application. In contrast, the selectivity for CO2 or H2 over N2 or CH4 decreases dramatically at higher temperatures suggesting that optimal performance is obtained at lower temperatures ( Figure 9).   temperature is an increase of molecular vibrations which facilitates diffusion. In contrary, the sorption enthalpy, ΔHs, is negative as expected since the sorption is an exothermic process. For CH4, N2 and H2, the absolute value of ED is greater than ΔHs and so the energy of activation Ep is positive, which means that diffusion rather than sorption dominates the response of permeation to temperature. For CO2, the absolute value of ED is smaller than ΔHs, which induces a negative activation energy EP. The CO2 transport is mainly influenced by the gas solubility, which is characteristic of microporous polymer, with similar results being found for PIM-1 and PTMSP [5,8,11].
Based upon these effects, the increase of temperature improves H2/CO2 selectivity modestly moving the data for PIM-EA(H2)-TB close to the 200 °C upper bound (Figure 8, however, even its enhanced high temperature selectivity (~2) is insufficient for viable pre-combustion application. In contrast, the selectivity for CO2 or H2 over N2 or CH4 decreases dramatically at higher temperatures suggesting that optimal performance is obtained at lower temperatures ( Figure 9).

Conclusions
Transport properties of permeability, diffusivity and solubility of PIM-EA(H 2 )-TB have been determined for H 2 , N 2 , CH 4 and CO 2 over a range of pressures and temperatures. This PIM presents high CO 2 and H 2 permeability coefficients, which allows it to have good ideal selectivity over N 2 . PIM-EA(H 2 )-TB exhibits the classical behaviour of a glassy polymer, with the decrease of diffusivity coefficient with increasing penetrant molecular size and the increase of sorption coefficient gas with increasing condensability of the permeant. However, no increase in CO 2 permeability due to plasticization is noted over the range of pressure tested. The permeability coefficient of N 2 , CH 4 and H 2 increase with increasing temperature while for CO 2 the permeability decreases with increasing temperature, which is classically observed for microporous materials. Therefore, the separation performance of PIM-EA(H 2 )-TB for H 2 /CO 2 is reversed at high temperature and maintained also at high pressure. This suggests that, after further development to enhance absolute selectivity of H 2 over CO 2 , PIMs could become good candidates for membrane materials for use in pre-combustion CO 2 capture. For other gas separations, better performance is obtained at lower temperatures.