Performance of PEI/BMI semi-IPN membranes for separations of various binary gaseous mixtures☆

TL;DR: In this article, a semi-interpenetrating polyetherimide-bismaleimide (PEI-BMI) semi-IPN with solvent phase inversion was used to prepare asymmetric flat membranes, which were coated with silicon rubber.

Abstract: Synthesis of polyetherimide–bismaleimide (PEI–BMI) semi-interpenetrating polymer networks (semi-IPNs) combined with solvent phase inversion was used to prepare asymmetric flat membranes, which were coated with silicon rubber. These membranes were evaluated for production of O 2 -enriched air and separation of CO 2 from its mixture with CH 4 as well as with N 2 . Using different preparation schemes membranes with varying skin and supported layer characteristics were prepared. These different morphologies of membranes were responsible for a trade-off performance between gas permeance and permselectivity. These new composite, PEI–BMI semi-IPN membranes showed suitable performance for production of O 2 -enriched air and separation of CO 2 from natural gas or flue gas relevant to greenhouse gas emission control. Membrane performance was explained in terms of the intrinsic gas transport properties of the coated silicon layer and membrane glassy material, which determine the limitations of permeance–permselectivity trade-off. It was also found that the permselectivity for CO 2 over CH 4 or N 2 increases with increasing CO 2 feed concentration. This might be exploited to arrange for more cost-efficient multistage gas separation systems.

Summary

1. Introduction

• There has been significant growth in the production of natural gas as an efficient and environmentally clean fuel supply [1].
• There are challenges in preparing membranes with the desirable combination of high selectivity and high permeability that could achieve competitive gas separation processes in all applications of natural gas industry [3].
• Tailoring useful nanocomposite polymeric membrane still faced with many challenges such as the formation of micro-scale defects, inadequate particle dispersion and poor polymer–particle interfacial adhesion [5,8].
• The authors found that in situ polymerization of BMI monomer inside PEI/NMP solution forms hard phase of thermoset BMI resin that interpenetrates soft phase of thermoplastic PEI networks, i.e. formation of physical interlock between the two phases.

2.1. Materials

• Aromatic polyetherimide (PEI) Ultem® 1000 was supplied by General Electric Plastics, USA in pellet form and was dried in an oven at 150 ◦C for 8 h before use to remove any possible absorbed water vapors.
• Anhydrous ethyl alcohol was received from Commercial Alcohols Inc., Ont., Canada.
• All solvents were used as supplied under a dry nitrogen atmosphere.
• Ultra high purity helium, medical air, CO2 and CH4 were supplied by BOC Gases Canada Ltd. and were used as received without further purification.

2.2. Membrane preparation

• Casting solutions with various compositions listed in Table 2 were prepared.
• PEI was completely dissolved in NMP solvent by rolling the bottle of each sample slowly then, anhydrous EtOH was added with slow mixing until a homogenous solution was obtained.
• Membranes were cast at room temperature on clean glass plates placed in a glove box equipped with a gas filter.
• After casting each sample with a doctor knife having a gap of 250 m, the plate was quickly immersed in distilled water at ambient temperature.
• Three circular coupons of 7.4 × 10−2 m diameter were cut from each sample to be used in the permeation test while other pieces were cut from the same sample for SEM characterization.

2.3. Membrane morphology

• Membrane samples without silicon rubber coating were examined by scanning electron microscope (SEM) using JEOL 840A equipment at an accelerating voltage of 10 kV.
• Samples were prepared by cutting a strip from membrane, freezing in liquid nitrogen and fracturing to obtain a representative sample.
• They were mounted on carbon tape on 45◦ SEM stubs and sputter coated with gold.

2.4. Permeation test

• A cross-flow test cell having a permeation surface area of 9.6 cm2 was used.
• Pure O2, N2 and medical air were used to study O2/N2 separation.
• The permeate flow rate was measured by a soap bubble flow meter and O2, CH4 and CO2 concentrations of feed and permeate gas mixtures were determined by gas chromatography.

3. Results and discussion

• Synthesis procedure for PEI–BMI semi-IPNs combined with membrane formation was finalized after several preliminary experiments.
• The progress of polymerization was followed by measuring the viscosity, color change and spectroscopic analysis as discussed in details in their earlier work [11].
• Absorption of water vapor during preparation of the solution might also have an influence on the rate of BMI polymerization.
• For their samples the authors observed that at room temperature, the solution need from 11–24 days for onset of color change while at 60 ◦C, color changed in less than 24 h.
• All polymer solutions were suitable for producing gas separation membranes as long as there is no large phase separation or precipitation.

3.1. Morphology analysis

• Membrane morphology plays an important role in determining membrane performance for gas separations.
• This means that the shrinkage of membrane (a) due to coagulation is higher, which led to a thinner structure and larger number of fingers (see Fig. 1) than those in membrane (b).
• The density of two polymer solutions used to produce the above mentioned membranes (a and b) were determined using the same balance and 100 ml volumetric flasks with class A glass stopper.
• The thickness of the solution was calculated from its volume and area (i.e. 16 cm2).
• According to Ismail and Hassan [16] shear rate is a function of the velocity of casting knife and membrane thickness.

3.2.1. Air separation

• Semi-IPN PEI–BMI membranes were evaluated for permeation of pure oxygen and nitrogen as well as for air.
• Therefore, both procedures for membrane preparation in addition to membrane material characteristics are of great importance for improving performance of gas separation membranes.
• It was worth noting that the semi-IPN PEI–BMI membranes had an improved performance for O2/N2 separation compared to PEI membranes without BMI.
• It was further observed that membrane labeled (b) in this work had O2-enriched air permeance of 9 GPU and O2/N2 permselectivity of 7.
• Conversely, the authors observed an increase in the gas permeance due to a decrease in the gas transport resistance of the membranes without formation of defects, which are not small enough to be caulked by the silicon rubber.

3.2.2. CO2/CH4 separation

• Performance of semi-IPN PEI–BMI membranes were studied through the permeation of CO2, CH4 and their mixtures.
• The trade-off curve of CO2 permeance and CO2/CH4 permselectivity based on permeation of pure gases is shown in Fig.
• On the other hand, a membrane with CO2 permeance of 458 GPU and ideal CO2/CH4 permselectivity of 3.5 was also obtained.
• There is a slight improvement in the performance of semi-IPN PEI–BMI membranes over PEI membranes without BMI (see Fig. 5).
• The permeation test for pure CH4 was carried out before using pure CO2 permeation test.

3.2.3. Case study for CO2/CH4 separation using four semi-IPN PEI–BMI membranes

• S2, s3, s4 and s10 were selected for further studies, also known as Four semi-IPN PEI–BMI membranes.
• The subscript indicates the sample numbers shown in Table 2.
• When a mixture of CO2 and CH4 is used, the gas permeance decreases slightly with the increase in the CO2 concentrations in the feed gas mixture as shown in Fig.
• The phenomenon of increasing the productivity of the more permeable gas (i.e. CO2) upon increasing its driving force through the membrane was reported by Ismail and Yaacob [30], however, they did not refer to the simultaneous change in the driving force of other gases in the mixture.
• These results confirm that the change in CO2 feed concentration has greater influence on the CO2/CH4 permselectivity for s2 and s3 membranes than those for s4 and s10 membranes while it has a greater influence on the gas permeance for s4 and s10 membranes than those for s2 and s3 membranes.

3.2.4. CO2/N2 separation

• The trade-off curve of CO2 permeance versus CO2/N2 permselectivity based on permeation of pure gases is shown in Fig. 13.
• By selecting suitable membrane preparation procedures and conditions, it is possible to have a semiIPN PEI–BMI membrane with CO2 permeance of 52 GPU and CO2/N2 ideal permselectivity of 26 or a membrane with CO2 permeance of 392.7 GPU and CO2/N2 ideal permselectivity of 13.4.
• The silicon rubber shows an ideal CO2/N2 permselectivity of 13.0 as found elsewhere [20].
• The behavior of these two membranes for CO2/N2 separation is similar for CO2/CH4 separation as shown in Fig. 11.
• The observations were similar to earlier discussion on CO2 and CH4 mixtures.

4. Conclusion

• It was possible to prepare semi-IPN PEI–BMI membranes that have a higher performance for gas separations than PEI membranes.
• Regardless of membrane materials, changing membrane morphology especially for the skin layer or the supported layer, it was possible to produce membranes with a high gas permeance but a low permselectivity or membranes with a high permselectivity but a low gas permeance.
• Using cast solution containing 19.5% (w/w) of PEI and BMI polymers, it was possible to decrease significantly the gas transport resistance of the supported membrane layer but it was difficult to obtain improved skin integrity, which is responsible for high permselectivity.
• It was concluded that there would be a slight decrease in the gas permeance with increasing CO2 feed concentration.
• Using their semi-IPN PEI–BMI membranes, the increase of CO2/N2 or CO2/CH4 permselectivity with increasing CO2 feed concentration might be economically exploited to arrange more cost-efficient separation systems to capture CO2 from natural gas or flue gas.

Figures

Fig. 3. Trade-off curve plot of O2 permeance and O2/N2 permselectivity based on the permeation of pure gases. Filled square symbols are for PEI membranes without BMI. Labels correspond to the sample number listed in Table 2.

Fig. 2. Cross-section SEM of membranes produced from casting solutions containing 19.5% polymer concentration (a) and 29% polymer concentration (b). Top two photographs represent the top layer of each membrane and the two bottom photographs represent the bottom layer of each membrane.

Fig. 4. Trade-off curve plot of O2-enriched air permeance and O2/N2 permselectivity based on the permeation of air. Filled square symbols are for PEI membranes without BMI. Labels correspond to the sample number listed in Table 2.

Fig. 5. Trade-off curve plot of CO2 permeance and CO2/CH4 permselectivity based on the permeation of pure gases. Filled square symbols are for PEI membranes without BMI. Labels correspond to the sample number listed in Table 2.

Fig. 14. Permselectivity of CO2 over N2 at different CO2 feed concentrations.

Fig. 13. Trade-off curve plot of CO2 permeance and CO2/N2 permselectivity based on the permeation of pure gases. Filled square symbol is for PEI membranes without BMI. Labels correspond to the sample number listed in Table 2.

Fig. 15. Permselectivity of CO2 over N2 versus gas permeance at different CO2 feed concentrations.

Table 1 Chemical structure of PEI, BMI and NMP

Table 2 Compositions of polymeric solutions used in this study

Fig. 8. Total gas permeance of CO2 and CH4 mixture at different feed concentrations.

Fig. 6. Trade-off curve plot of permeance of the mixture of CO2 and CH4 vs. CO2/CH4 permselectivity based on the permeation of gas mixture containing 26.5% (v/v) CO2 and 73.5% (v/v) CH4. Filled square symbols are for PEI membranes without BMI. Labels correspond to the sample number listed in Table 2.

Fig. 7. Permeance of pure CO2 at different feed pressures for different semi-IPN membranes.

Fig. 9. Partial pressure differences (driving force) for CO2 and CH4 at different feed concentrations.

Fig. 11. Permselectivity of CO2 over CH4 at different CO2 feed concentrations.

Fig. 12. Permselectivity of CO2 over CH4 vs. gas permeance at different CO2 feed concentrations.

Fig. 10. CO2 concentration of permeate gas mixture of CO2 and CH4 at different feed concentrations.

Fig. 1. Morphologies of membranes produced from casting solutions containing 19.5% polymer concentration (a) and 29% polymer concentration (b). SEM was obtained at the same magnification 500× and the same scale distance was used for both membranes.

Table 3 Apparent volumea and thickness of two membranes (a and b) as well as for their corresponding solutions with 19.5% and 29% (w/w) polymers, respectively

Frequently asked questions

Q1. What are the contributions mentioned in the paper "Performance of pei/bmi semi-ipn membranes for separations of various binary gaseous mixtures" ?

In this paper, the authors used semi-interpenetrating polyetherimide-bismaleimide ( PEI-BMI ) semi-IPN networks to produce asymmetric flat membranes, which were coated with silicon rubber.