Probe Into the Influence of Crosslinking on CO2 Permeation of Membranes

Crosslinking is an effective way to fabricate high-selective CO2 separation membranes because of its unique crosslinking framework. Thus, it is essentially significant to study the influence of crosslinking degree on the permeation selectivities of CO2. Herein, we report a successful and facile synthesis of a series of polyethylene oxide (PEO)-based diblock copolymers (BCP) incorporated with an unique UV-crosslinkable chalcone unit using Reversible Addition-Fragmentation Chain Transfer Polymerization (RAFT) process. The membranes of as-prepared BCPs show superior carbon dioxide (CO2) separation properties as compared to nitrogen (N2) after UV-crosslinking. Importantly, the influence of different proportions of crosslinked chalcone on CO2 selectivities was systematically investigated, which revealed that CO2 selectivities increased obviously with the enhancement of chalcone fractions within a certain limit. Further, the CO2 selectivities of block copolymer with the best block proportion was studied by varying the crosslinking time which confirmed that the high crosslinking degree exhibited a better CO2/N2 (αCO2/N2) selectivities. A possible mechanism model revealing that the crosslinking degree played a key role in the gas separation process was also proposed.

work reported that a kind of CO 2 -selective membrane consisting of EO segments and commercially available UV-crosslinker coumarin achieved high CO 2 selectivities compared with N 2 and He after UV-crosslinking, which suggested that the impact of crosslinking on gas permeation properties existed 48 . Based on this, it is our motivation in this work to have an insight on the role of crosslinking degree during the gas separation process of PEO-based membranes.
Chalcone was chosen by Iyoda and his coworkers as an UV-crosslinker to better control the crosslinking degree in their previous work 49,50 . It was because that the chalcone could be crosslinked like "head to head" and "head to tail" and hence forming a higher crosslinking degree of membrane structure as shown in Fig. 1. Therefore, in this work, in order to better understand the correlation of crosslinking degree and CO 2 permeation properties, chalcone was selected as a unique crosslinkable segment instead of coumarin. In detail, chalcone was designed as a UV-crosslinker to study the regulation of gas permeation and crosslinking degree. For this, firstly, the chalcone segments and EO units were synthesized. Further, a series of block copolymers consisting of PEO (poly(ethylene oxide)) and PMA (poly(methacrylate) with chalcone mesogens) with different block ratios (PEO 11 -b-PMA(rChal) 7 , PEO 11 -b-PMA(rChal) 9 , PEO 11 -b-PMA(rChal) 12 , PEO 11 -b-PMA(rChal) 16 ) were fabricated using Reversible Addition-Fragmentation Chain Transfer Polymerization (RAFT) method. The prepared copolymers were further investigated to examine the changes of gas permeabilities and selectivities of the membranes before and after UV crosslinking, respectively. Moreover, the variations of gas permeabilities and selectivities by changing the fraction of the crosslinkable block and crosslinking time were discussed. Based on the obtained results, a plausible mechanism between crosslinking degree and CO 2 permeation properties was also demonstrated.

Results and Discussion
Characterizations of crosslinking degree. The effect of UV-crosslinking time on the prepared membranes was examined by UV-vis absorption spectroscopy. Figure 2 exhibited the typical UV-Vis spectra of the prepared membrane UV-crosslinking for various time intervals. The UV-Vis spectral trend was monitored at  special chalcone absorption band appearing in the range of 300 to 450 nm. The observed UV-vis absorption spectroscopy revealed that the π -π * transition of the chalcone unit had a decrease from the 340 nm absorption band. The observed photoreaction could be explained according to the dimerization of the chalcone moieties through the [2 + 2] cyclization of the double bond 51 . Interestingly, it was observed that with increasing the UV-crosslinking time, the change of absorption spectra tended to balance, which illustrated that the greatest degree of crosslinking reached at 40 min.
Comparison of CO 2 permeation properties of the membranes before and after crosslinking. Table 1 showed the gases permeation and CO 2 /N 2 selectivities of the un-crosslinked film compared with crosslinked film with PEO 11 -b-PMA(rChal) 7 at different temperature. The α CO2/N2 data for other films are demonstrated in Tables S1 and S2. As shown in Table 1, the α CO2/N2 of these two membranes were increasing with rising the temperature. Interestingly, α CO2/N2 of un-crosslinked membrane was 2.76 at 60 °C which was much smaller than α CO2/N2 of crosslinked membrane equal to 12.56. ( Figure S1) The selectivities was mainly related to solubility selectivities and diffusivity selectivities which mainly depended on the interactions of EO unit with CO 2 and the free volume of EO, respectively. To explain this phenomenon, a probable mechanism model was proposed to interpret the increase of α CO2/N2 (Fig. 3). For un-crosslinked membrane, the EO units and the chalcone units had a large excess free volume and increased the mobility of the chain with the temperature rising, which led to higher gas permeabilities. Notably, differed from N 2 , the CO 2 had a strong interaction with EO units, causing the PEO chains more flexible, which also contributed to the final α CO2/N2 . Referred to crosslinked membrane, for the diffusivity selectivities, crosslinking structure limited the free volume and hence the EO units were confined in the hard regions of crosslinked chalcone walls. With increasing the temperature, the crosslinked hard regions were hardly moved but EO units were more flexible in the limited domain. However, for N 2 , the flexible EO units would lead to denser barriers in a limited area 48 , and then the less N 2 molecules went through the free volume which produced a low gas permeabilities for N 2 . For CO 2 , the interactions between EO units and CO 2 increased with rising Permeances at 1 × 10 6 (cm·s −1 ·cmHg −1 ), were calculated by dividing the observed flow rate by the area of the membrane (2.84 cm 2 ) and the pressure gradient (10 psi) employed, using porous Al 2 O 3 membrane supports. The values were obtained from 10 independent measurements and the mean value and standard deviations were determined. The error in each case was < 5%. The membrane PEO-b-PMA (rChal) showed no difference of gas permeation in humid environment. the temperature and plasticized EO regions which lead to more flexible EO fraction, thus further increasing the CO 2 permeabilities 52 . Therefore, with such aforementioned discussion, it can be concluded that the crosslinked membranes showed a favored gas selectivities towards CO 2 with increasing the temperature.
Comparison of gas permeation properties of the membranes composed of various block ratio. In this study, four kinds of block copolymers with different block ratios, i.e. PEO 11 -b-PMA(rChal) 7 , PEO 11 -b-PMA(rChal) 9 , PEO 11 -b-PMA(rChal) 12 , PEO 11 -b-PMA(rChal) 16 were prepared and consequently, four membranes were exposed to UV light (40 min) for complete crosslinking. Table 2 showed the data of crosslinked membranes with the block ratios of PEO 11 -b-PMA(rChal) 9 , PEO 11 -b-PMA(rChal) 12 and PEO 11 -b-PMA(rChal) 16 at different temperatures. It was clearly indicated that all of these block copolymers possessed same tendency with the block ratio of PEO 11 -b-PMA(rChal) 7 . At 60 °C, α CO2/N2 reached to 12.56 when the mole percentage of PMA (PMA%) was about 39%. Further, the α CO2/N2 reached to 14.79 when PMA% was approximately equal to 50%, however, the α CO2/N2 dropped to 10.13 when PMA% was greater than 60% (Fig. 4). For solubility selectivities, the interactions between EO and CO 2 increased with rising the temperature. For diffusivity selectivities, the complicated crosslinking network led to the limited free volume. Moreover, crosslinking segments barely moved with the vary of temperature, so the N 2 denser barriers increased with the more crosslinking units 48 . When the ratio of PEO:PMA varied from 11:7 to 11:12, the N 2 permeabilities decreased as    the data described. In contrast, CO 2 plasticized EO chain to be more flexible which resulted in higher CO 2 /N 2 selectivities. However, the continuous increase of mole percentage of chalcone segments did not represent sustainable rising trend in α CO2/N2 but form a more rigid crosslinking framework instead. The rigid framework structure limited the mobility of EO segments in a large degree, which became obstacles for CO 2 transfer, representing a sharp reduction of α CO2/N2 . As shown in Fig. 5, the highest CO 2 selectivities of crosslinked membranes in this work was much closer to upper bound 53 than un-crosslinked ones. Taking account of these factors, the content of crosslinking segments after fully crosslinking played a key role in CO 2 gas permeable membrane.

Comparison of CO 2 permeation properties of a fixed block ratio BCP under different irradiation time.
For this study, the PEO 11 -b-PMA(rChal) 12 was treated as an example. (The data for other membranes are shown in Tables S3 and S4). Chalcone was an unique UV-crosslinker because of its easy crosslinking degree control by altering the UV irradiation time, which provided us a feasible way to verify the mechanism that crosslinking degree affected the ultimate CO 2 permeation properties. Table 3 presented the system data obtained under different UV irradiation time 0 min, 5 min, 15 min and 40 min (fully crosslinked) at 30 °C and 60 °C, respectively. The observed results ( Figure S2) revealed that the α CO2/N2 was rising with the variation of UV irradiation time. Specifically, the data of α CO2/N2 was 2.43 without UV irradiation at 60 °C and the α CO2/N2 reached to 6.46 after 5 min UV irradiation. Further, the α CO2/N2 value was reached to the maximum of 14.79 after 40 min of UV irradiation, which revealed the full crosslinking. With increasing the irradiation time, the crosslinking degree in chalcone units increased which were minimizing free volume of the framework, resulting in the decreasing of N 2 permeabilities. Meanwhile, harder crosslinked PMA segments may also be contributed to the improvement of CO 2 /N 2 selectivities. For diffusivity selectivities, CO 2 would make EO chains more flexible in the limited area surrounded by the crosslinked wall, which caused high CO 2 permeabilities. On the other hand, the confinement of free volume would result in denser barriers as explained above, leading to the low N 2 permeabilities. Thus, based on the observed results, one can conclude that the demonstrated mechanism is fully consistent with the obtained results.

Conclusion
In summary, we have successfully synthesized a series of diblock copolymers incorporating a novel UV-crosslinkable chalcone based on PEO chains using a facile RAFT process. Interestingly, it was observed that crosslinked membrane exhibited high CO 2 permeabilities over N 2 and hence showing high selectivities of CO 2 , in contrast with un-crosslinked membrane. Further, the detailed studies revealed that the diblock copolymers with different proportion displayed various selectivities. It was researched that the rising amounts of chalcone within certain limits enhanced the crosslinking degree by which the EO fractions become more flexible and thus exhibiting a higher CO 2 permeabilities and selectivities with temperature enhancement. However, excess crosslinking chalcone fragment formed an ultra-rigid framework and confined the transfer of CO 2 through the membrane, which resulted in low CO 2 permeabilities. Thus, tunable CO 2 selectivities could be achieved by monitoring the crosslinking degree of membranes. The presented work provided further applications of UV-crosslinking network for CO 2 separation.

Materials and Methods
Materials. All the chemicals were analytical grade and used as received without any further purifications. Measurements. The prepared materials were characterized in detail using several techniques. The 1 H-NMR measurements were performed on Bruker AV-300 spectrometers in chloroform-d using tetramethylsilane (TMS; δ = 0) as internal reference. All copolymers were examined by gel permeation chromatography (Malvern, GPC 270) as reported in our previous work 48 . The standard sample of GPC is Polystyrene(PS) and Mn = 99385, the measured solvent was THF. The DSC curve was measured in DSC(NETZSCH, STA449F3). Gas permeation measurements were carried out in the similar manner as reported in the literature by the authors; i.e. a home-made stainless steel permeation apparatus as described previously 21,48 . The UV-crosslinking of the films was monitored by UV-Vis absorption spectroscopy. The average thickness of the tested film was examined by ellipsometry. Six sections on each membrane were measured respectively to calculate the average thickness of 1.8 ± 0.1 μ m.
Synthesis of the monomer chalcone. The synthesis of monomer chalcone was done according to the Fig. 6 using organic synthetic procedure. The prepared material was purified with typical process and the purified product was characterized using 1 H-NMR spectroscopy.

Preparation of PM1.
To prepare the PM1, in a typical reaction process, 4-hydroxybenzaldehyde (15.27 g, 125 mmol), 11-bromoundecan-1-ol (32.66 g, 125 mmol) and DMF (100 mL) were added in a 3-necked flask under continuous stirring. The mixture was stirred until the materials were completely dissolved. Consequently, 1.88 g NaI and 34.56 g K 2 CO 3 were added in the resultant solution and stirred again for 30 min. After stirring, the resultant mixture was reflux for 24 h. After desired reaction time, the reaction was terminated and the mixture was cooled at room-temperature and the solvent was removed using rotary evaporation process. Subsequently, the residue was added to water and thus twice extracted with dichloromethane (DCM). Finally, the organic layer was separated and dried over MgSO 4 which was filtered. The obtained precipitate was then purified by column chromatography which finally provided a white solid (43.13 g 29 (12H, m). The typical 1 H-NMR spectrum is shown in Figure S3. 13  Preparation of PM2. To prepare PM2, in a typical process, PM1 (34.9 g, 100 mmol) and 3 namely, 4-butylphenylethylketone (21 g, 100 mmol) were mixed in alcoholic sodium hydroxide (50 mL, 10%) solution under continuous stirring. After 12 h stirring at room-temperature, the reaction mixture was poured into ice-water. The solid precipitate was formed which was collected by filtration and dried. The dried product was the further purified by column chromatography and finally yellow solid was obtained (50.  Figure S4). 13

Preparation of chalcone.
To prepare the chalcone, in a typical reaction process, PM2 (42.8 g, 115 mmol), TEA (1.82 ml, 120 mmol) and 100 ml DCM were added in a dry round-bottom flask. Consequently, methacryloyl chloride (14.57 ml, 115 mmol) was added dropwise to the mixture at 0 °C and the reaction was continued overnight. The obtained product was then added to the water and subsequently extracted with DCM twice. The organic layer was then separated and dried over MgSO 4 . Finally, the obtained product was filtered and purified by column chromatography which gives a light yellow solid ( Figure S5). 13     freeze pump-thaw procedure and the bottle was sealed under vacuum. The sealed bottle was then placed in a preheated oil bath (90 °C) for 12 h. Finally, the solution was precipitated in hexane (Fig. 8).

Preparation of PEO-b-PMA(rChal) diblock copolymers. A series of PEO-b-PMA(rChal) containing
a chalcone mesogen with different content of polymerization were synthesized by RAFT method. The targeted material was prepared as presented in Fig. 9. As an example, a procedure to prepare PEO 11 -b-PMA(rChal) 7 is described here. In a typical reaction process, 0.097 g (1 eqv) PEO macro-initiators, 0.24 g (30 eqv) chalcone and 0.0003 g (0.12 eqv) AIBN were mixed in a 10 mL Shreck bottle with 1.8 ml anhydrous anisole. Then, the mixture was degassed four times using the freeze pump-thaw procedure and sealed under vacuum. The sealed bottle was then placed in a preheated oil bath (90 °C) for 17 h. The solution was precipitated in diethyl ether and finally pure diblock copolymer was obtained. Figure 10 showed the typical 1 H-NMR (a) and GPC (b) results. The observed Mn and PDI of PEO 11 -b-PMA(rChal) 7 , PEO 11 -b-PMA(rChal) 9 , PEO 11 -b-PMA(rChal) 12 12 was measured at a rate of 2 °C/min from 40 °C to 140 °C and shown in Figure S8.
Preparation of thin membranes. The diblock copolymer membrane was made by spin-coating (1000 rpm) of 6 wt% chloroform solutions on Anodic Alumina Oxide (AAO) substrate. The prepared diblock copolymer membrane was placed in vacuum for 4 h at room-temperature. Consequently, the crosslinking of the block copolymer membrane was exposed to UV light (365 nm) for a desirable time and finally the UV-crosslinked membrane was obtained.

Gas permeation measurements.
A home-built gas permeation measurement system was used to estimate the gas permeation as described in our previous work 22,48 . The pure gases with different kinetic diameters such as N 2 and CO 2 were studied in this work. For this, a spin-coated membranes on AAO substrates were placed in the permeation cell with a support screen. The surface area of the tested membrane, available for gas transport, was estimated and found to be 2.84 cm 2 . The membranes were placed in the test system cell to investigate the gas permeation test. The gases, after passing through the membrane in the cell, was directed into a glass U-tube flow meter (Acol = 0.03 cm 2 ) to give the volumetric flow rate of the gas. It was measured by recording the time (t) that was required for a liquid column to travel a distance (X column = 10 cm). All the measurements were taken at ambient temperature and the values were obtained at steady-state (usually last for at least 2 h). The values were obtained from 10 independent measurements and the mean value and standard deviations were determined. The error in each case was < 5%. The same experimental procedure was repeated for other targeted gas. In general, the permeation properties were sequentially measured for He, N 2 and CO 2 , respectively. The permeance (P; 10 6 cm·s −1 ·cmHg −1 ) was calculated based on the equation (1)  And the selectivities (α) of gas A, over gas B, was defined based on the below equation (2) 48 :