Redefining the Robeson upper bounds for CO2/CH4 and CO2/N2 separations using a series of ultrapermeable benzotriptycene-based polymers of intrinsic microporosity

This jou nal is ©The Royal Socie y o Chemis y 2019 Ene gy En i on. Sci., 2019, 12, 2733--2740 | 2733 Ci e his: Ene gy En i on. Sci., 2019, 12,2733 Rede ining he Robeson uppe bounds o CO 2 / CH 4 and CO 2 /N 2 sepa a ions using a se ies o ul ape meable benzo ip ycene-based polyme s o in insic mic opo osi y† Bibiana Comesan ˜a-Ga ´nda a, ‡ a Jie Chen,‡ a C. G azia Bezzu, a Ma iolino Ca a, b Ian Rose, a Ma ia-Chia a Fe a i, c Elisa Esposi o, d Alessio Fuoco, d Johannes C. Jansen * d and Neil B. McKeown * a Memb anes composed o Polyme s o In insic Mic opo osi y (PIMs) ha e he po en ial o ene gy efficien indus ial gas sepa a ions. He e we epo he syn hesis and gas pe meabili y da a o a se ies o ul ape meable PIMs, o wo-dimensional chain con o ma ion and based on benzo ip ycene s uc u al uni s, ha demons a e ema kable ideal selec i i y o mos gas pai s o impo ance. In pa icula , he CO 2 ul ape meabili y and high selec i i y o CO 2 o e CH 4 , o key impo ance o he upg ading o na u al gas and biogas, and o CO 2 o e N 2 , o impo ance o cos -effec i e ca bon cap u e om powe plan s, exceed he pe o mance o he cu en s a e-o - he-a polyme s. All o he gas pe meabili y da a om his se ies o benzo ip ycene-based PIMs a e placed well abo e he cu en 2008 Robeson uppe bounds o CO 2 /CH 4 and CO 2 /N 2 . Indeed, he da a o some o hese polyme s all in o a linea co ela ion on he benchma k Robeson plo s [i.e. log(P CO2 /P CH4 ) e sus log P CO2 and log(P CO2 /P N2 ) e sus log P CO2 ], which a e pa allel o, bu signi ican ly abo e, ha o he 2008 CO 2 /CH 4 and CO 2 /N 2 uppe bounds, allowing hei e ision. The ede ini ion o hese uppe bounds se s new aspi a ional a ge s o polyme chemis s o aim o and will esul in mo e a ac i e pa ame ic es ima es o ene gy and cos efficiencies o ca bon cap u e and na u al/bio gas upg ading using s a e-o - he-a CO 2 sepa a ion memb anes. B oade con ex The low-cos and ene gy-effec i e emo al o ca bon dioxide (CO 2 ) om na u al gas and biogas would help he supply o me hane as he cleanes bu ning and lowes ca bon-emi ing hyd oca bon uel. In addi ion, ca bon cap u e and s o age (CCS) om powe plan emissions will be equi ed o achie e he goals o he 2015 Pa is Ag eemen , which aspi es o main ain global wa ming o less han 1.5 1C abo e ha o he p e-indus ial age by he end o he 21s Cen u y. Indeed, he combined use o bio uels, such as biogas, and CCS echnology is ega ded as he key nega i e emissions echnology equi ed in o de o each he Ag eemen ’s ambi ious a ge s o educed emissions. Despi e he u gen need o CCS, he bes echnology pla o m o i s deli e y is s ill unclea due o he difficul ies in he es ima ion o cos s and he complex e alua ion o he ad an ages and disad an ages associa ed wi h each echnology. Highly pe meable memb anes ha a e selec i e o CO 2 o e me hane (CO 2 /CH 4 )andCO 2 o e ni ogen (CO 2 /N 2 ) a e o inc easing in e es o na u al gas/biogas upg ading and ca bon cap u e, espec i ely, due o he inhe en efficiency o memb ane sepa a ions. He e we epo he syn hesis o a se ies o ul ape meable polyme s ha de ine he s a e-o - he-a in he ade-off be ween pe meabili y and selec i i y o all impo an gas sepa a ions and, in pa icula , o CO 2 /CH 4 and CO 2 /N 2 . The da a om hese polyme s we e used o ede ine he benchma k Robeson uppe bounds o hese wo gas sepa a ions a much highe alues o selec i i y. This enhancemen will imp o e he c edibili y o polyme memb anes o CO 2 sepa a ions when e alua ed agains compe ing p ocesses. Hope ully, his will help o s imula e he undamen al polyme science and applied enginee ing equi ed o de elop memb ane sys ems o hese CO 2 sepa a ions o key impo ance o ene gy and he en i onmen . In oduc ion Memb anes based on polyme s as he selec i e laye a e used o he ene gy efficien sepa a ion o gas mix u es including hose o key ele ance o ene gy and he en i onmen . 1–4 The de elop- men o new polyme s wi h g ea e gas pe meabili y and selec- i i y would u he enhance he e iciency o memb ane gas sepa a ions o cu en indus ial in e es , 5 including hyd ogen eco e y du ing ammonia p epa a ion (H 2 om N 2 ), oxygen o a EaS CHEM, School o Chemis y, Uni e si y o Edinbu gh, Da id B ews e Road, Edinbu gh, EH9 3FJ, UK. E-mail: [email p o ec ed] b Depa men o Chemis y, Swansea Uni e si y, College o Science, G o e Building, Single on Pa k, Swansea, SA2 8PP, UK c Ins i u e o Ma e ials and P ocesses, School o Enginee ing, The Uni e si y o Edinbu gh, May ield Road, Edinbu gh EH9 3JL, UK d Ins i u e on Memb ane Technology, ITM-CNR, Via P. Bucci 17/C, 87036 Rende (CS), I aly †Elec onic supplemen a y in o ma ion (ESI) a ailable. See DOI: 10.1039/c9ee01384a ‡The i s wo au ho s con ibu ed equally. Recei ed 29 h Ap il 2019, Accep ed 11 h July 2019 DOI: 10.1039/c9ee01384a sc.li/ees Ene gy & En i onmen al Science PAPER Open Access A icle. Published on 23 July 2019. Downloaded on 1/26/2024 10:26:39 AM. This a icle is licensed unde a C ea i e Commons A ibu ion 3.0 Unpo ed Licence. View A icle Online View Jou nal | View Issue 2734 |Ene gy En i on. Sci., 2019, 12, 2733--2740 This jou nal is ©The Royal Socie y o Chemis y 2019 ni ogen en ichmen o ai (O 2 om N 2 ) 6 and na u al gas o biogas upg ading (p edominan ly CO 2 om CH 4 ). 7–10 Inc easingly, polyme memb anes a e also being conside ed as a p ac ical al e na i e o sol en abso p ion o la ge-scale cap u e o CO 2 om powe plan lue gas (p edominan ly CO 2 om N 2 ). 7,9,11–14 Fo gas sepa a ions on such a massi e scale, memb anes wi h e y high pe meance (i.e. lux) a e desi able o minimise ene gy cos s o gas comp ession and o educe he ac i e su ace a ea o he memb ane, he eby, op imising he o e all size and manu ac u e cos o he memb ane sys em. 5,15 Howe e , polyme memb ane ma e ials su e om he well-es ablished ade-o be ween gas pe meabili y (P x ) and selec i i y o one gas o e ano he (P x /P y ), 16,17 so ha es ablished ul ape meable polyme s, such as he polyace ylene poly( ime hylsilylp opyne) (PTMSP), 18,19 and ecen ly epo ed examples 20 a e insu icien ly selec i e o use in gas sepa a ions. The gene al ade-off be ween polyme pe meabili y and selec i i y was i s quan i ied by Robeson in 1991 when he iden i ied uppe bounds in plo s o log(P x /P y ), e sus log P x o O 2 /N 2 ,H 2 /N 2 , He/N 2 ,H 2 /CH 4 , He/CH 4 ,CO 2 /CH 4 , and He/H 2 gas pai s based on he gas pe meabili y o he bes pe o ming polyme s a ha ime. 21 Subsequen ly, o a newly p epa ed polyme (o a mixed ma ix memb ane) 22,23 he posi ion o i s gas pe meabili y da a ela i e o he uppe bounds on Robeson plo s allows o i s po en ial o gas sepa a ions o be es ima ed. Robeson upda ed all o he uppe bounds in 2008 using ini ial da a o wo spi obisindane-based Polyme s o In insic Mic o- po osi y (PIM-1 and PIM-7; Table S1, ESI†), 24 whose igid and con o ed mac omolecula s uc u es p o ided excep ionally high pe meabili y wi h mode a e selec i i y. 25 In addi ion, da a o hese woPIMswe ealsoused ode ineanuppe bound o he CO 2 /N 2 gas pai , which is o key impo ance o pos -combus ion ca bon cap u e bu had been conside ed o no p ac ical in e es in 1991. 24 Since 2008, many PIMs wi h enhanced igidi y ha e demon- s a ed gas pe meabili y da a ha lie well abo e some o he 2008 uppe bounds. 26 These highly shape-pe sis en PIMs we e ob ained by eplacing he ela i ely lexible spi obisindane s uc u al uni wi h spi obi luo ene 27,28 uni s o highly igid b idged bicyclic componen s such as e hanoan h acene, 29–32 ip ycene, 33–36 me hanopen acene 37 and T o ¨ge s base. 29,35 Indeed, in 2015 Pinnau e al. 38 p oposed ha he O 2 /N 2 ,H 2 /N 2 and H 2 /CH 4 uppe bounds should be upda ed using pe meabili y da a om aged ilms o highly selec i e ip ycene-based PIMs (e.g. PIM-T ip-TB 35 and TPIM-1 33 ). Howe e , e isions o he uppe bound o CO 2 /N 2 and CO 2 /CH 4 we e no p oposed a ha ime due o he da a o hese polyme s and o he high-pe o ming PIMs being close o he exis ing 2008 CO 2 /N 2 and CO 2 /CH 4 uppe bounds (Table S1, ESI†). Recen ly, we in oduced a new PIM de i ed om a benzo- ip ycene monome , PIM-TMN-T ip, which p o ed o be as ul ape meable o gases as PTMSP due o enhanced in insic mic opo osi y a ising om i s 2D chain s uc u e. 39 PIM-TMN- T ip demons a es highe selec i i y han PTMSP due o i s g ea e chain igidi y p o iding enhanced molecula sie ing (i.e. diffusi i y selec i i y). Fu he mo e, i was ound ha he unsubs i u ed benzo ip ycene-based PIM (PIM-BT ip) demon- s a es e en g ea e selec i i y placing i s da a abo e he p oposed 2015 O 2 /N 2 ,H 2 /N 2 and H 2 /CH 4 uppe bounds and e en abo e Robeson’s 2008 uppe bounds o CO 2 /N 2 and CO 2 / CH 4 . 40,41 He e we epo on he syn hesis and p ope ies o some new membe s o he benzo ip ycene-based PIM se ies (Fig. 1), all o which demons a e high pe meabili y and selec i i y. In pa icula , his polyme se ies demons a es pe me- abili y da a o CO 2 /N 2 and CO 2 /CH 4 ha sugges new posi ions o he Robeson uppe bound o hese impo an gas pai s ha a e o key in e es o sepa a ions o ele ance o ene gy and he en i onmen . Resul s and discussion Polyme design and syn hesis A u he ou benzo ip ycene PIMs we e syn hesised along wi h new ba ches o PIM-TMN-T ip and PIM-BT ip o allow o di ec compa ison o hei gas pe meabili ies. The no el poly- me s include PIM-HMI-T ip, o which he s e ically c owded hexame hylindane (HMI)-solubilising g oup 42 would be expec ed o be mo e igid han he e ame hylnaph halene (TMN) g oup o PIM-TMN-T ip. P e iously o spi obi luo ene-based PIMs, 43 he in oduc ion o adjacen me hyl subs i uen s had been Fig. 1 S uc u e and syn hesis o he benzo ip ycene PIMs. Reagen s and condi ions: i. B 2 ,Fe,DCM, ,3h;ii.n-BuLi, u an, THF, 78 1C, 1.5 h; iii. 9,10-Dime hyl- 2,3,6,7- e ame hoxyan h acene, DMF, 250 1C, 7 ba , 2 h, mic owa e i adia ion, i . TFA o MeSO 4 H, , 24 h.; . BB 3 , DCM. (See ESI† o de ails). Pape Ene gy & En i onmen al Science Open Access A icle. Published on 23 July 2019. Downloaded on 1/26/2024 10:26:39 AM. This a icle is licensed unde a C ea i e Commons A ibu ion 3.0 Unpo ed Licence. View A icle Online This jou nal is ©The Royal Socie y o Chemis y 2019 Ene gy En i on. Sci., 2019, 12, 2733--2740 | 2735 shown o be bene icial o pe o mance, he e o e, a PIM based on dime hylbenzo ip ycene was p epa ed (PIM-DM-BT ip). In addi ion, he po en ial bene i o in oducing one o wo i luo ome hyl (TFM) solubilising g oups on o he benzo- ip ycene uni was e alua ed by he syn hesis o PIM-TFM- BT ip and PIM-DTFM-BT ip, espec i ely. Each polyme was p epa ed om i s e ahyd oxy benzo- ip ycene monome (1a– ) using he well-es ablished benzodioxin- o ming polyme isa ion eac ion de ised o PIM syn hesis (Fig. 1). 44 Monome s we e p epa ed by adap a ion o he classic benzo ip ycene syn hesis, in ol ing he Diels–Alde eac ion be ween 2,3,6,7- e ame hoxy-9,10-dime hylan h acene and he app op ia e 1,4-dihyd o-1,4-epoxynaph halene 39 – wi h he la e p epa ed om he Diels–Alde eac ion be ween he app op ia e benzyne in e media e and u an. 45–47 PIM-TMN-T ip and PIM-HMI-T ip a e bo h soluble in chlo o- o m, acili a ing analysis using Gel Pe mea ion Ch oma og aphy (GPC) ha con i med ha high molecula mass polyme was achie ed o bo h polyme s (Table 1). In con as , PIM-DM-B ip, PIM-TFM-BT ip and PIM-DTFM-BT ip p o ed soluble only in quinoline. The success o his high-boiling a oma ic sol en o dissol ing hese o he wise in ac able polyme s p omp ed a e-in es iga ion o he solubili y o unsubs i u ed PIM-BT ip, which we had p e iously desc ibed as insoluble. 39 Pleasingly, his polyme also p o ed soluble in quinoline. Al hough quino- line is no an app op ia e sol en o GPC analysis, solu ions o PIM-DM-BT ip, PIM-TFM-BT ip, PIM-DTFM-BT ip and PIM- BT ip could be used o cas mechanically lexible and obus ilms, implying ha a easonably high molecula mass had been achie ed du ing he syn hesis. Syn he ic and s uc u al cha ac e isa ion de ails, including solid s a e NMR (Fig. S1) a e gi en in he ESI.† Gas adso p ion and gas anspo p ope ies. In hei powde o m, all benzo ip ycene-based PIMs adso b a la ge amoun o ni ogen (N 2 , 77 K) a low ela i e p essu e. Analysis o he N 2 adso p ion iso he ms (Fig. S1, ESI†) gi es appa en B unaue –Emme –Telle (BET) su ace a eas (SA BET ) wi hin he ange o 848–1034 m 2 g 1 (Table 1), which a e amongs he highes ob ained om solu ion p ocessable polyme s. 29,39 The shapes o he N 2 iso he ms a e simila o all polyme s excep o PIM-TMN-T ip and PIM-DTFM-BT ip, o which he e is la ge up ake a highe p essu es associa ed wi h a la ge hys e esis be ween he adso p ion and deso p ion iso he ms. This migh be ela ed o he TMN and CF 3 subs i uen s p o uding ou o he 2D plane o he polyme chain and hus in e e ing wi h he elec os a ic ni ile–ni ile in e ac ions which a e likely o domina e polyme cohesion. Adso p ion o CO 2 a 273 K (Fig. S2, ESI†) shows simila up akes o he benzo ip ycene-PIMs (2.5–3.3 mmol g 1 ). The up ake o PIM-BT ip is sligh ly highe a lowe p essu es, which may be asc ibed o a g ea e concen a ion o ul amic opo es (diame e o0.7 nm in i s po e size dis ibu ion (Fig. S3, ESI†)). Sol en cas ilms (Fig. S4, ESI†) o he benzo ip ycene- based PIMs all demons a e excep ionally high gas pe meabili y (Table 2). Howe e , he e alua ion o gas pe meabili y da a o a new polyme equi es ca e ul conside a ion o i s ilm his o y and hickness as hese ac o s in luence g ea ly he obse ed alues. 32 Gene ally, he highes epo ed alues o gas pe meabili y o high ee olume polyme s such as he PTMSP and PIMs we e ob ained om ilms eshly ea ed wi h me hanol (o e hanol), which emo es any esidual cas ing sol en bu also induces addi ional ee olume. 31,48 The alues o gas pe meabili y om eshly me hanol ea ed hick ilms (135–176 mm) o he benzo ip ycene PIMs a e some o highes epo ed o a pu e polyme ilm (e.g.,P CO 2 =21–5310 3 Ba e ) and a e compa able o hose om e hanol ea ed ul ape me- able polyace ylenes (e.g.,P CO 2 = 28–47 10 3 Ba e ). 19,42 Fo each o he me hanol ea ed ilms he o de o dec easing gas pe meabili y is CO 2 4H 2 4O 2 4He 4CH 4 4N 2 wi h he excep ion o hose om he less pe meable and mo e size- selec i e PIM-BT ip o which He pe mea es as e han O 2 . The ideal selec i i ies o all o he me hanol ea ed ilms a e signi ican ly highe han hose ob ained o he ul ape meable polyace ylenes and all in he ange o hose epo ed o me hanol ea ed ilms o less pe meable PIMs such as PIM-1 (e.g.,P O 2 /P N 2 = 2.6–3.6). 48 As no ed o all PIMs and highly pe meable polyme s, 31,32,49–51 he ex emely high alues o gas pe meabili y measu ed ini ially om he eshly me hanol ea ed ilms a e no main ained on ageing. 52 Howe e , he educ ion in pe meabili y is accompa- nied by an inc ease in ideal selec i i y o all gas pai s. In addi ion, on ageing, He pe meabili y su passes he alue o O 2 o all he polyme s, indica ing enhanced size selec i i y. Compa ing da a om app oxima ely like- o -like samples (i.e. B120 day aged and 110–180 mm hick ilms) he o de Table 1 Yield, molecula mass and gas adso p ion p ope ies o he benzo ip ycene-based PIMs Polyme Yield (%) Solubili y M n (g mol 1 )M w /M n Z a (cm 3 g 1 ) SA BETb (m 2 g 1 ) V To alc (ml g 1 ) V Md (ml g 1 ) CO 2 up ake e (mmol g 1 ) PIM-TMN-T ip 67 CHCl 3 52 300 3.8 74 1034 0.87 0.38 3.3 PIM-HMI-T ip 58 CHCl 3 61 300 2.4 58 1033 0.71 0.38 3.0 PIM-BT ip 78 Quinoline — g — g 66 911 0.63 0.33 3.2 PIM-DM-BT ip 82 Quinoline — g — g 72 920 0.72 0.33 3.0 PIM-TFM-BT ip 79 Quinoline — g — g 37 848 0.66 0.31 2.5 PIM-DTFM-BT ip 84 Quinoline — g — g 65 964 1.02 0.33 2.5 a Inhe en iscosi y in quinoline a 25 1C. b BET su ace a ea calcula ed om N 2 adso p ion iso he m ob ained a 77 K. c To al po e olume es ima ed om N 2 up ake a P/P o = 0.98. d Mic opo e olume es ima ed om N 2 up ake a P/P o = 0.05. e CO 2 adso p ion a 1 ba and 273 K. Rela i e o polys y ene s anda ds. g No measu ed due o insolubili y in sol en s compa ible wi h GPC analysis. Ene gy & En i onmen al Science Pape Open Access A icle. Published on 23 July 2019. Downloaded on 1/26/2024 10:26:39 AM. This a icle is licensed unde a C ea i e Commons A ibu ion 3.0 Unpo ed Licence. View A icle Online 2736 |Ene gy En i on. Sci., 2019, 12, 2733--2740 This jou nal is ©The Royal Socie y o Chemis y 2019 o dec easing pe meabili y and inc easing selec i i y o he benzo ip ycene PIMs is PIM-TMN-T ip 4PIM-DTFM-BT ip 4 PIM-HMI-T ip 4PIM-TFM-T ip 4PIM-BT ip EPIM-DM- BT ip. I can be deduced ha he bulky TMN and HMI subs i uen s bo h enhance pe meabili y g ea ly, wi h he mo e igid HMI subs i uen p o iding sligh ly highe selec i i y o e TMN. The ela i ely small –CF 3 subs i uen s o PIM-TFM-BT ip and PIM-DTFM-BT ip also enhance pe meabili y ela i e o unsubs i u ed PIM-BT ip. In e es ingly, he –CF 3 subs i uen s appea o slow ageing, wi h 54% o he alue o P O 2 o he me hanol ea ed ilm o PIM-DTFM-BT ip e ained a e one yea , and 56% o PIM-TFM-BT ip, as compa ed o only 30–36% o ilms wi hou –CF 3 subs i uen s. Depending on he gas, he s anda d de ia ion o he pe me- abili y is in he ange 4–18% o he eshly MeOH ea ed PIM-HMI-T ip and PIM-DM-BT ip ilms, and 3–6% o he aged PIM-BT ip ilm. These a e small compa ed o he e ec o he ageing in his wo k, and almos negligible when ep esen ed on he double-loga i hmic Robeson diag ams (Fig. S6, ESI†). A hinne ilm o PIM-BT ip (64 mm) demons a es lowe ini ial pe meabili y a e me hanol ea men , consis en wi h he well-es ablished end ha hinne ilms age mo e apidly han hicke ilms. 32,52,53 I is also mo e size selec i e han he hicke ilm o he same polyme wi h H 2 4CO 2 4He 4O 2 4 CH 4 4N 2 he o de o dec easing gas pe meabili y. Due o he commonly encoun e ed a iabili y o gas pe meabili y om diffe ing ilm hicknesses and his o y, da a o a new polyme a e bes compa ed o hose o exis ing polyme s by using Robeson plo s (Fig. 2). As no ed, he posi ion o he da a om a new polyme ela i e o he Robeson uppe bounds p o ides a use ul indica o o i s po en ial pe o mance as gas sepa a ion memb anes. All da a poin s o he benzo ip ycene polyme s lie a abo e he 2008 uppe bounds o O 2 /N 2 (Fig. 2a), H 2 /N 2 (Fig. 2b), H 2 /CH 4 ,CO 2 /N 2 (Fig. 2c) and CO 2 /CH 4 (Fig. 2d). Da a o he B1 yea aged ilms o all o he polyme s lie close o he p oposed 2015 uppe bound o O 2 /N 2 . In pa icula , aged PIM-BT ip demons a es excep ional selec i i y o a highly pe meable polyme so ha i s da a lie well abo e he p oposed Table 2 Thickness (l, mm), ideal gas pe meabili ies (Px, Ba e ) and selec i i ies o eshly me hanol ea ed and aged ilms measu ed a 25 1C and 1 ba o eed p essu e PIM- a l b P N2 P O2 P CO2 P CH4 P H2 P He P O2 /P N2 P H2 /P N2 P CO2 /P N2 P CO2 /P CH4 BT ip 160 1190 4330 21 500 1690 12 100 4540 3.64 10.2 18.1 12.7 (130) c,d 160 522 2570 13 200 570 8440 3110 4.92 16.2 25.3 23.2 (253) c,d 160 401 2170 10 700 411 8930 3400 5.41 22.3 26.7 26.0 (365) c,d 160 280 1580 8020 282 7160 2810 5.65 25.6 28.6 28.4 (490) c,d 160 195 1240 6060 203 6380 2650 6.34 32.6 31.0 29.9 (633) c,d 160 127 935 4350 130 5100 2180 7.36 40.1 34.2 33.5 (718) e,g 160 112 (4) 838 (48) 3770 (166) 113 (4) 4820 (186) 2150 (64) 7.51 (0.19) 43.2 (0.53) 33.8 (0.53) 33.5 (0.33) BT ip d 64 339 1800 9200 412 9430 3960 5.31 27.8 27.1 22.3 (120) 64 200 1160 6040 237 7180 3020 5.79 35.8 30.2 25.5 (253) d 64 190 1143 5990 225 8080 3490 6.01 42.5 31.5 26.6 (371) c,d 64 154 997 5150 163 7730 3620 6.47 50.2 33.4 31.6 TMN-T ip 166 3540 10 400 52 800 7250 18 800 6490 2.94 5.31 14.9 7.28 (120) 166 1970 6620 33 300 3130 15 300 5600 3.36 7.77 16.9 10.6 (253) 166 1470 5440 25 900 2030 14 100 5190 3.71 9.59 17.6 12.8 (358) 166 1289 5082 23 648 1751 14 118 5290 3.94 11.0 18.4 13.5 (426) 166 1100 4620 20 400 1440 14 100 5420 4.20 12.8 18.5 14.2 HMI-T ip d 135 2560 8540 44 200 4870 16 600 5700 3.34 6.48 17.3 9.08 (1) ,g 135 2120 (330) 7380 (989) 39 000 (3680) 3990 (708) 18 400 (1765) 6500 (762) 3.49 (0.14) 8.95 (2.16) 18.6 (1.7) 9.94 (1.44) (120) 135 1440 5180 26 900 2150 11 800 4240 3.60 8.19 18.7 12.5 (253) 135 972 3930 18 900 1220 10 700 3960 4.04 11.0 19.5 15.6 (358) 135 907 3760 17 404 1083 11 141 4245 4.15 12.3 19.2 16.1 (426) 135 804 3580 16 400 967 11 000 4150 4.45 13.7 20.4 16.9 TFM-BT ip c,d 176 1830 6210 33 700 2280 13 600 5150 3.39 7.43 18.4 14.8 (123) c 176 1090 4230 22 100 1250 10 700 4120 3.88 9.82 20.3 17.7 (255) c 176 875 3640 18 400 953 9870 4050 4.15 11.3 21.0 19.3 (367) c 176 791 3450 17 000 873 10 100 4170 4.36 12.7 21.5 19.5 (496) 176 722 3260 15 600 792 9760 3920 4.51 13.5 21.6 19.7 DTFM-BT ip 112 3000 7770 42 600 4340 14 700 5860 2.59 4.90 14.2 9.82 (119) 112 1800 5410 29 000 2150 11 300 4690 3.01 6.28 16.1 13.5 (366) 112 1300 4460 22 900 1390 10 700 4590 3.41 8.23 17.5 16.4 (490) 112 864 3490 16 900 890 10 400 4770 4.04 12.1 19.6 19.0 (636) 112 741 3170 14 800 728 10 200 4730 4.27 13.8 20.0 20.3 DM-BT ip d, 114 1020 (133) 3950 (374) 22 000 (1071) 1570 (85) 11 400 (482) 4000 (354) 3.90 (0.16) 11.3 (1.07) 21.8 (2.5) 14.0 (1.5) (128) d 114 521 2640 12 200 599 9870 3650 5.07 18.9 23.4 20.4 a Numbe in pa en heses is he ageing ime in days a e me hanol ea men . b Thickness did no exhibi signi ican changes upon ageing. c Da a de ining he p oposed CO 2 /CH 4 uppe bound. d Da a de ining he p oposed CO 2 /N 2 uppe bound. e A e age and s anda d de ia ion (in pa en heses) o ou independen measu emen s o he same aged sample. A e age and s anda d de ia ion (in pa en heses) o ou independen samples. g Da a no included on Robeson plo s (Fig. 2). Pape Ene gy & En i onmen al Science Open Access A icle. Published on 23 July 2019. Downloaded on 1/26/2024 10:26:39 AM. This a icle is licensed unde a C ea i e Commons A ibu ion 3.0 Unpo ed Licence. View A icle Online This jou nal is ©The Royal Socie y o Chemis y 2019 Ene gy En i on. Sci., 2019, 12, 2733--2740 | 2737 2015 uppe bounds o O 2 /N 2 (Fig. 2a), H 2 /N 2 (Fig. 2b), and H 2 /CH 4 . A no able ea u e o he pe meabili y da a om aged samples o he benzo ip ycene-PIMs on he O 2 /N 2 and H 2 /N 2 Robeson plo s is he nea linea co ela ion a a s eepe slope han ha o he uppe bounds (Fig. S5, ESI†). This e lec s he a la ge educ ion o pe meabili ies on ageing o gases composed o la ge molecules such as N 2 and CH 4 as compa ed o hose composed o he smalle O 2 and H 2 molecules. Gas anspo h ough a polyme is desc ibed by he solu ion- diffusion model 54 wi h P x =D x S x ,whe eD x is he diffusi i y coefficien (Table S2, ESI†) and S x is he solubili y coefficien o gas x (Table S3, ESI†). The e o e, he ideal selec i i y (P x /P y ) o a polyme comes om a combina ion o diffusi i y selec i i y (D x /D y ) and solubili y selec i i y (S x /S y ). The ema kable posi ions o he da a o he benzo ip ycene-PIMs on he H 2 /N 2 ,andO 2 /N 2 Robeson plo s a e due o e y high diffusi i y selec i i y o igina ing om he size-sie ing beha iou o he polyme s, which diffe en ia es be ween gas molecules o diffe ing effec- i e diame e s (d x ). 40 This is bes illus a ed by he co ela ion be ween d x2 and he diffusi i y coefficien (D x ), 55 which is s eepes o PIM-BT ip and less s eep o benzo ip ycene PIMs ha possess a subs i uen , al hough he absolu e alue o he diffusion coefficien is la ge (Fig. 3). Ageing dec eases he diffusion coefficien o all polyme s bu s eepens he co ela ion be ween d x2 and D x , especially o PIM-BT ip, which is e idence o i s u he enhanced size selec i i y (Fig. S7, ESI†). 40 The ex ao dina y pe o mance o PIM-BT ip can be a ibu ed o i s ul amic opo osi y, which acili a es he diffusi i y o small Fig. 2 Robeson plo s o he (a) O 2 /N 2 ,(b)H 2 /N 2 , (c) CO 2 /N 2 and (d) CO 2 /CH 4 gas pai s showing he posi ion o he gas pe meabili y da a o ilms o PIM-BT ip ( ), PIM-TMN-T ip ( ), PIM-HMI-T ip ( ), PIM-DM-BT ip ( ), PIM-TFM-BT ip (K) and PIM-DTFM-BT ip ( ). P e iously epo ed da a a e also shown o non-PIM polyme s (&) and PIMs ( ). Uppe bounds a e ep esen ed by black lines (1991), blue lines (2008), and ed lines o he p e iously p oposed (2015) uppe bounds o O 2 /N 2 and H 2 /N 2 . The p oposed e ised uppe bounds o CO 2 /N 2 and CO 2 /CH 4 a e shown as do ed ed lines. Ene gy & En i onmen al Science Pape Open Access A icle. Published on 23 July 2019. Downloaded on 1/26/2024 10:26:39 AM. This a icle is licensed unde a C ea i e Commons A ibu ion 3.0 Unpo ed Licence. View A icle Online 2738 |Ene gy En i on. Sci., 2019, 12, 2733--2740 This jou nal is ©The Royal Socie y o Chemis y 2019 gas molecules, oge he wi h e y high chain igidi y, 16,54 which hinde s he ac i a ed anspo o la ge gas molecules by educing he mal mo ions ha allow gaps o o m be ween oids. The ex eme igidi y o PIM-BT ip accoun s o he e y high ac i a ion ene gy o he diffusion o la ge gases such as N 2 and CH 4 . 40 The gas anspo p ope ies o PIM-BT ip appea s simila o hose epo ed o he wo ip ycene-de i ed polyme s, PIM-T ip-TB 35 and TPIM-1, 33 which we e used o de ine he p oposed 2015 uppe bounds o O 2 /N 2 ,H 2 /N 2 and H 2 /CH 4 . 38 I should be no ed ha he da a om PIM-T ip-TB used o de ine he 2015 uppe bounds we e aken om a ilm ha was aged o only 100 days a e me hanol ea men . 35 Recen emeasu e- men o he gas pe meabili y o his ilm a e 1900 days gi es da a ha a e also well o e he p oposed 2015 uppe bounds o O 2 /N 2 (i.e. P O 2 = 532 Ba e ; P O 2 /P N 2 =8.2)andH 2 /N 2 (i.e. P H 2 = 4430 Ba e ; P H 2 /P N 2 = 65). The e o e, he design concep s used o ob ain he ex ao dina y size selec i i y demons a ed by PIM-BT ip and PIM-T ip-TB a e likely o p o ide PIMs ha will p o oke u u e signi ican e isions o he O 2 /N 2 ,H 2 /N 2 and H 2 /CH 4 Robeson uppe bounds. Rede ining he CO 2 /N 2 and CO 2 /CH 4 uppe bounds Sepa a ions in ol ing CO 2 a e mechanis ically mo e complex han hose go e ned p edomina ely by diffusi i y selec i i y (e.g. O 2 /N 2 o H 2 /N 2 )becauseS CO 2 domina es anspo , especially o CO 2 /N 2 due o he simila effec i e diame e s o he wo gas molecules. Typically o PIMs, alues o S CO 2 /S N 2 lie in he ange 15–20 whe eas hose o D CO 2 /D N 2 lie be ween 0.9–1.5 and hese alues a e simila o PIMs wi h bo h highe and lowe P CO 2 pe meabili y. In gene al, solubili y selec i i y ends o emain ai ly cons an du ing ageing, in con as o he inc eases obse ed o ideal selec i i y alues o anspo domina ed by di usi i y selec i i y. 52 Thus, plo ing da a o p e iously epo ed PIMs on he Robeson plo o CO 2 /N 2 shows many da a poin s sligh ly abo e he 2008 uppe bound a highe pe meabili y (P CO 2 43000 Ba e ) bu ew a lowe alues o pe meabili y. Indeed, e y ew highly pe meable polyme s possess a CO 2 /N 2 selec i i y 430, 56–59 which is he lowe limi o in e es o a i s -pass polyme memb ane o pos -combus ion ca bon cap u e (Table S1, ESI†). 12 Al hough all o he da a o he benzo ip ycene PIMs a e abo e he 2008 uppe bound o CO 2 /N 2 , he da a om PIM-BT ip a e pa icula ly p omising wi h bo h hick and hinne aged ilms p o iding P CO 2 44000 Ba e and P CO 2 /P N 2 430. The imp essi e pe o mance o PIM-BT ip appea s o be due o an unusually high D CO 2 /D N 2 o 2.0, whe eas ha o he subs i u ed membe s o he se ies elies on g ea e S CO 2 /S N 2 esul ing om he g ea e numbe o CO 2 adso p ion si es p o ided by he la ge amoun o in insic mic opo osi y (Table S3, ESI†).Theele enda apoin son he Robesonplo om ou diffe en polyme s ha allin oalinea co ela ion pa allel o ha o he 2008 uppe bound allows us o p opose a subs an ially imp o ed new uppe bound o CO 2 /N 2 (Fig. 2c and Tables 2 and 3). These da a poin s a e dis ibu ed o e ala geP CO 2 ange o 4400–52 000 Ba e . In addi ion, he da a o all o he benzo ip ycene PIMs lie well abo e he 2008 uppe bound o CO 2 /CH 4 a a highe selec i i y han hose o p e iously epo ed polyme s. Indeed, only da a o he highly igid ‘‘in e molecula ly-locked’’ de i a- i e o PIM-1 (PIM-C1) 60 and PIM-SBF-2 43 come close o hose o he benzo ip ycene PIMs (Table S1, ESI†). This excep ional pe o mance appea s due o a combina ion o bo h high diffusi i y selec i i y, wi h D CO 2 /D CH 4 in he ange 5.7–9.5 o aged ilms, and good solubili y selec i i y (S CO 2 /S CH 4 43). Ten da a poin s om wo di e en polyme s allows us o p opose a new uppe bound o CO 2 /CH 4 pa allel o ha o 2008 (Fig. 2d and Tables 2 and 3). The benzo ip ycene PIMs ha ei he de ine o p o ide da a ha a e e y close o his e ised uppe bound a e ei he unsubs i u ed (PIM-BT ip) o possess only small subs i uen s (i.e. PIM-DM-BT ip; PIM-TFM-BT ip and PIM-DTFM-BT ip). In con as , hose possessing la ge cyclic solubilising g oups (i.e. PIM-TMN-T ip and PIM-HMI-T ip) a e sligh ly less selec i e. When de ining his 2008 CO 2 /CH 4 uppe bound, Robeson no ed ha da a o a se ies o The mally Rea anged (TR) Fig. 3 Plo o diffusi i y coefficien (D x ) e sus d x2 (whe e d x = effec i e diame e o gas molecule x: He = 1.78; H 2 = 2.14; O 2 = 2.89; CO 2 = 3.02; N 2 =3.04;CH 4 = 3.18 Å) 55 o eshly me hanol ea ed ilms o PIM-BT ip (), PIM-TMN-T ip ( ), PIM-HMI-T ip ( ), PIM-DM-BT ip ( ). Da a o PIM-TFM-BT ip and PIM-DTFM-BT ip a e no shown o cla i y bu a e e y simila o hose o PIM-TMN-T ip and PIM-HMI-T ip, espec i ely. Table 3 Fi ing pa ame e s o he 2008 and p oposed CO 2 /N 2 and CO 2 / CH 4 uppe bounds using he o mula P x =ka xy n (whe e P x is pe meabili y (Ba e ) o he mos pe meable x-gas, kis he on ac o (Ba e ), a xy is he selec i i y o x/y gas pai , and nis he slope) k(Ba e ) n Robeson 2008 uppe bounds 24 CO 2 /CH 4 5.369 10 6 2.636 CO 2 /N 2 30.967 10 6 2.888 P oposed uppe bounds CO 2 /CH 4 22.584 10 6 2.401 CO 2 /N 2 755.58 10 6 3.409 Pape Ene gy & En i onmen al Science Open Access A icle. Published on 23 July 2019. Downloaded on 1/26/2024 10:26:39 AM. This a icle is licensed unde a C ea i e Commons A ibu ion 3.0 Unpo ed Licence. View A icle Online This jou nal is ©The Royal Socie y o Chemis y 2019 Ene gy En i on. Sci., 2019, 12, 2733--2740 | 2739 polyme s, epo ed by Pa k e al., 15,61 ‘‘wi h excep ional CO 2 /CH 4 sepa a ion capabili ies’’, 24 appea ed o o m an uppe bound abo e ha p oposed o solu ion p ocessable polyme s. Such insoluble ne wo k polyme s as he TR polyme s o en pe o m abo e he 2008 uppe bounds de ined o solu ion p ocessable polyme s due o hei igidi y app oaching ha o ca bon mole- cula sie es (i.e. polyme s ca bonised a high empe a u es). Rema kably, he CO 2 /CH 4 uppe bound de ined by he solu ion p ocessable benzo ip ycene-based PIMs lies a he same posi- ion as ha o Robeson’s en a i ely p oposed TR polyme uppe bound wi h a selec i i y 2.5 imes highe han ha o he 2008 uppe bound. Conclusions The benzo ip ycene-based PIMs p o ide excep ional gas pe meabili y da a o mos impo an gas pai s and allow o he ede ini ion o he CO 2 /CH 4 and CO 2 /N 2 Robeson uppe bounds. This is impo an in o de o se aspi a ional a ge s o chemis s in he design and syn hesis o no el polyme s. In addi ion, i will help pa ame ic s udies o ene gy and cos efficiency o ca bon cap u e and na u al/bio gas upg ading by p o iding enhanced bu ealis ic s a e-o - he-a alues o memb ane pe meabili y and selec i i y. The esul ing es ima es o ene gy efficiencies and cos s will be mo e a ac i e ela i e o bo h p e ious calcula ions o memb ane sys ems and o compe i i e CO 2 sepa a ion p ocesses. The esul ing imp o ed c edibili y o polyme memb anes o hese c ucial sepa a ions will s imula e esea ch ac i i y in his echnological a ea o p ime impo ance o ene gy and he en i onmen . Con lic s o in e es The e a e no con lic s o in e es o decla e. Acknowledgemen s The esea ch leading o hese esul s has ecei ed unding om he EU FP7 F amewo k P og am unde g an ag eemen no. 608490, p ojec M 4 CO 2 and om he EPSRC (UK) g an numbe s EP/M01486X/1, EP/R000468/1 and EP/K008102/2. Re e ences 1 P. Be na do, E. D ioli and G. Golemme, Ind. Eng. Chem. Res., 2009, 48, 4638–4663. 2 Y. Yampolskii, Mac omolecules, 2012, 45, 3298–3311. 3 R. W. Bake and B. T. Low, Mac omolecules, 2014, 47, 6999–7013. 4 M. Galizia, W. S. Chi, Z. P. Smi h, T. C. Me kel, R. W. Bake and B. D. F eeman, Mac omolecules, 2017, 50, 7809–7843. 5 P. M. Budd and N. B. McKeown, Polym. Chem., 2010, 1, 63–68. 6 R. S. Mu ali, T. Sanka shana and S. S idha , Sep. Pu i . Re ., 2013, 42, 130–186. 7 S. F. Wang, X. Q. Li, H. Wu, Z. Z. Tian, Q. P. Xin, G. W. He, D. D. Peng, S. L. Chen, Y. Yin, Z. Y. Jiang and M. D. Gui e , Ene gy En i on. Sci., 2016, 9, 1863–1890. 8 J. K. Adewole and A. L. Ahmad, J. Polym. Res., 2017, 24, 17. 9 N. Y. Du, H. B. Pa k, M. M. Dal-Cin and M. D. Gui e , Ene gy En i on. Sci., 2012, 5, 7306–7322. 10 E. Esposi o, L. Dellamuzia, U. Mo e i, A. Fuoco, L. Gio no and J. C. Jansen, Ene gy En i on. Sci., 2019, 12, 281–289. 11 M. C. Fe a i, D. Boccia do and S. B andani, G een Ene gy En i on., 2016, 1, 211–221. 12 T. C. Me kel, H. Q. Lin, X. T. Wei and R. Bake , J. Memb . Sci., 2010, 359, 126–139. 13 R. W. Bake , B. F eeman, J. Kniep, X. T. Wei and T. Me kel, In . J. G eenhouse Gas Con ol, 2017, 66, 35–47. 14 L. S. Whi e, K. D. Amo, T. Wu and T. C. Me kel, J. Memb . Sci., 2017, 542, 217–225. 15 S. Kim and Y. M. Lee, P og. Polym. Sci., 2015, 43, 1–32. 16 B. D. F eeman, Mac omolecules, 1999, 32, 375–380. 17 H. B. Pa k, J. Kamce , L. M. Robeson, M. Elimelech and B. D. F eeman, Science, 2017, 356, eaab0530. 18 T. Masuda, E. Isobe, T. Higashimu a and K. Takada, J. Am. Chem. Soc., 1983, 105, 7473–7474. 19 Y. Yampolskii, Polym. Re ., 2017, 57, 200–212. 20 Y. He, F. M. Benede i, S. Lin, C. Liu, Y. Zhao, H.-Z. Ye, T. Van Voo his, M. G. De Angelis, T. M. Swage and Z. P. Smi h, Ad . Ma e ., 2019, 31, 1807871. 21 L. M. Robeson, J. Memb . Sci., 1991, 62, 165–186. 22 S. Budha hoki, O. Ajayi, J. A. S eckel and C. E. Wilme , Ene gy En i on. Sci., 2019, 12, 1255–1264. 23 C. Y. Chuah, K. Goh, Y. Q. Yang, H. Q. Gong, W. Li, H. E. Ka ahan, M. D. Gui e , R. Wang and T. H. Bae, Chem. Re ., 2018, 118, 8655–8769. 24 L. M. Robeson, J. Memb . Sci., 2008, 320, 390–400. 25 P. M. Budd, K. J. Msayib, C. E. Ta e shall, B. S. Ghanem, K. J. Reynolds, N. B. McKeown and D. F i sch, J. Memb . Sci., 2005, 251, 263–269. 26 M. D. Gui e and Y. M. Lee, Science, 2013, 339, 284–285. 27 X. Ma, O. Salinas, E. Li wille and I. Pinnau, Mac omolecules, 2013, 46, 9618–9624. 28 X. Ma, B. Ghanem, O. Salines, E. Li wille and I. Pinnau, ACS Mac o Le ., 2015, 4, 231–235. 29 M. Ca a, R. Malpass-E ans, M. C oad, Y. Rogan, J. C. Jansen, P. Be na do, F. Bazza elli and N. B. McKeown, Science,2013, 339, 303–307. 30 Y. Rogan, L. S a anniko a, V. Ryzhikh, Y. Yampolskii, P. Be na do, F. Bazza elli, J. C. Jansen and N. B. McKeown, Polym. Chem., 2013, 4, 3813–3820. 31 E. Tocci, L. De Lo enzo, P. Be na do, G. Cla izia, F. Bazza elli, N. B. McKeown, M. Ca a, R. Malpass-E ans, K. F iess, K. Pilnacek, M. Lanc, Y. P. Yampolskii, L. S a anniko a, V. Shan a o ich, M. Mau i and J. C. Jansen, Mac omolecules, 2014, 47, 7900–7916. 32 X. H. Ma and I. Pinnau, Mac omolecules,2018,51, 1069–1076. 33 B. S. Ghanem, R. Swaidan, X. H. Ma, E. Li wille and I. Pinnau, Ad . Ma e ., 2014, 26, 6696–6700. Ene gy & En i onmen al Science Pape Open Access A icle. Published on 23 July 2019. Downloaded on 1/26/2024 10:26:39 AM. This a icle is licensed unde a C ea i e Commons A ibu ion 3.0 Unpo ed Licence. View A icle Online 2740 |Ene gy En i on. Sci., 2019, 12, 2733--2740 This jou nal is ©The Royal Socie y o Chemis y 2019 34 B. S. Ghanem, R. Swaidan, E. Li wille and I. Pinnau, Ad . Ma e ., 2014, 26, 3688–3692. 35 M. Ca a, M. C oad, R. Malpass-E ans, J. C. Jansen, P. Be na do, G. Cla izia, K. F iess, M. Lanc and N. B. McKeown, Ad . Ma e ., 2014, 26, 3526–3531. 36 I.Rose,M.Ca a,R.Malpass-E ans,M.-C.Fe a i,P.Be na do, G.Cla izia,J.C.JansenandN.B.McKeown,ACS Mac o Le ., 2015, 4, 912–915. 37 R. Williams, L. A. Bu , E. Esposi o, J. C. Jansen, E. Tocci, C. Rizzu o, M. Lanc, M. Ca a and N. B. McKeown, J. Ma e . Chem. A, 2018, 6, 5661–5667. 38 R. Swaidan, B. Ghanem and I. Pinnau, ACS Mac o Le ., 2015, 4, 947–951. 39 I. Rose, C. G. Bezzu, M. Ca a, B. Comesana-Ganda a, E. Lasseugue e, M. C. Fe a i, P. Be na do, G. Cla izia, A. Fuoco, J. C. Jansen, K. E. Ha , T. P. Liyana-A achchi, C. M. Colina and N. B. McKeown, Na . Ma e ., 2017, 16, 932–937. 40 A. Fuoco, B. Comesana-Ganda a, M. Longo, E. Esposi o, M. Mon eleone, I. Rose, C. G. Bezzu, M. Ca a, N. B. McKeown and J. C. Jansen, ACS Appl. Ma e . In e aces, 2018, 10, 36475–36482. 41 Y. Yin and M. D. Gui e , Na . Ma e ., 2017, 16, 880–881. 42 Y. Hu, M. Shio suki, F. Sanda, B. D. F eeman and T. Masuda, Mac omolecules, 2008, 41, 8525–8532. 43C.G.Bezzu,M.Ca a,M.C.Fe a i,J.C.Jansen, M. Mon eleone, E. Esposi o, A. Fuoco, K. Ha , T. P. Liyana- A achchi, C. M. Colina and N. B. McKeown, J. Ma e . Chem. A, 2018, 6, 10507–10514. 44 P. M. Budd, B. S. Ghanem, S. Makhseed, N. B. McKeown, K. J. Msayib and C. E. Ta e shall, Chem. Commun., 2004, 230–231. 45 H.Ha ,C.Lai,G.C.NwokoguandS.Shamouilian,Te ahed on, 1987, 43, 5203–5224. 46 F. Bailly, F. Co e and M. Schlosse , Syn hesis, 2005, 791–797, DOI: 10.1055/s-2005-861813. 47 R. S. Luo, J. H. Liao, L. Xie, W. J. Tang and A. S. C. Chan, Chem. Commun., 2013, 49, 9959–9961. 48 P. M. Budd, N. B. McKeown, B. S. Ghanem, K. J. Msayib, D. F i sch, L. S a anniko a, N. Belo , O. San i o a, Y. Yampolskii and V. Shan a o ich, J. Memb . Sci., 2008, 325, 851–860. 49 P.Be na do,F.Bazza elli,F.Tasselli,G.Cla izia,C.R.Mason, L. Mayna d-A em, P. M. Budd, M. Lanc, K. Pilnacek, O. Vopicka, K. F iess, D. F i sch, Y. P. Yampolskii, V. Shan a o ich and J. C. Jansen, Polyme , 2017, 113, 283–294. 50 R. Swaidan, B. Ghanem, E. Li wille and I. Pinnau, Mac o- molecules, 2015, 48, 6553–6561. 51 L. S a anniko a, V. Khodzhae a and Y. Yampolskii, J. Memb . Sci., 2004, 244, 183–191. 52 Z. X. Low, P. M. Budd, N. B. McKeown and D. A. Pa e son, Chem. Re ., 2018, 118, 5871–5911. 53 K. D. Do kenoo and P. H. P omm, Mac omolecules, 2000, 33, 3747–3751. 54 L. M. Robeson, B. D. F eeman, D. R. Paul and B. W. Rowe, J. Memb . Sci., 2009, 341, 178–185. 55 V. Teplyako and P. Mea es, Gas Sep. Pu i ., 1990, 4, 66–74. 56 R. Swaidan, B. S. Ghanem, E. Li wille and I. Pinnau, J. Memb . Sci., 2014, 457, 95–102. 57 J. Wu, J. T. Liu and T. S. Chung, Ad . Sus ainable Sys ., 2018, 2, 1800044. 58 N. Du, H. B. Pa k, G. P. Robe son, M. M. Dal-Cin, T. Visse , L. Scoles and M. D. Gui e , Na . Ma e ., 2011, 10, 372–375. 59 C. R. Mason, L. Mayna d-A em, N. M. Al-Ha bi, P. M. Budd, P. Be na do, F. Bazza elli, G. Cla izia and J. C. Jansen, Mac omolecules, 2011, 44, 6471–6479. 60 J. Zhang, H. Kang, J. Ma in, S. Zhang, S. Thomas, T. C. Me kel and J. Jin, Chem. Commun., 2016, 52, 6553–6556. 61 H. B. Pa k, C. H. Jung, Y. M. Lee, A. J. Hill, S. J. Pas, S. T. Mudie, E. Van Wagne , B. D. F eeman and D. J. Cookson, Science, 2007, 318, 254–258. Pape Ene gy & En i onmen al Science Open Access A icle. Published on 23 July 2019. Downloaded on 1/26/2024 10:26:39 AM. This a icle is licensed unde a C ea i e Commons A ibu ion 3.0 Unpo ed Licence. View A icle Online