Retarded O₂ transport in Co²⁺-coordinated supramolecular polymer networks for membrane CO₂/O₂ separations

Abstract

Dissociated Co²⁺ ions in liquids and polymers have been demonstrated to reversibly react with O₂ and increase O₂ permeability. However, we find a series of Co²⁺-coordinated supramolecular polymer networks (SPNs) with enormous O₂ sorption but retarded diffusion, leading to superior CO₂/O₂ separation properties. Specifically, Co(BF₄)₂ is dissolved by cross-linked poly(ethylene oxide) (XLPEO), as validated by Fourier transform infrared (FTIR) spectroscopy, X-ray diffraction (XRD), and differential scanning calorimetry (DSC). The dissociated Co²⁺ ions increase O₂/CO₂ solubility selectivity but decrease its diffusivity selectivity. For example, adding 6.4 mass% Co(BF₄)₂ in XLPEO increases O₂ solubility by 35 times and O₂/CO₂ solubility selectivity from 0.12 to 5.0, but it decreases O₂/CO₂ diffusivity selectivity from 0.40 to 0.0058, leading to a CO₂/O₂ permeability selectivity of 35, above Robeson's upper bound and superior to that of state-of-the-art polymers. This study unravels an exciting platform of metal ion-coordinated supramolecular networks for various molecular separations by harnessing strong affinity but retarded diffusion despite their stability challenge.

---

Introduction

CO₂/O₂ separations are of great interest for healthcare and industrial applications. For instance, artificial lungs for CO₂ and O₂ exchange are critical for blood oxygenation;^1,2 atmospheric CO₂ and O₂ compositions can be manipulated for storing and transporting vegetables and fruit to improve their shelf lifetime;³ CO₂ capture from fossil fuel-derived flue gas has been proposed as a critical way to mitigate CO₂ emissions to the atmosphere, and the flue gas contains 4–16% O₂ that needs to be separated from CO₂.^4,5 Membrane technology offers a cost- and energy-efficient approach for gas separations. However, CO₂ has a kinetic diameter of 3.3 Å, very close to that of O₂ (3.46 Å), and therefore, most polymers do not have sufficiently strong size-sieving ability to obtain high diffusivity selectivity and thus permeability selectivity. For example, Matrimid is widely investigated for membrane separations due to its strong size-sieving ability, and it showed a CO₂/CH₄ selectivity of 35 but a CO₂/O₂ selectivity of only 3.8 at 35 °C.⁶

Gas separation properties can be often manipulated through interactions between polymers and targeted gases. For example, polymers containing poly(ethylene oxide) (PEO) demonstrate good CO₂/O₂ separation properties because of their affinity towards CO₂, leading to excellent CO₂/O₂ solubility selectivity.^7–9 PEO exhibited a CO₂/O₂ selectivity of 18, though its CO₂/CH₄ selectivity was only 20,⁸ much lower than that of Matrimid.

Co²⁺ forms complexes with O₂, leading to high O₂ sorption, and thus, Co²⁺-based O₂-binding carriers have been reported, such as cobalt porphyrins,¹⁰ cobalt(II) complexes,¹¹ cobalt(II)-based metal-containing ionic liquid (MCIL),^12,13 and Co-based metal–organic frameworks (MOFs).¹⁴ These carriers were incorporated into liquid membranes to improve O₂/gas separation performance,^12,15,16 as shown in Table 1. For example, introducing Co (3-MeOsaltmen) in dimethylacetamide (DMAc) increased O₂/N₂ selectivity from 1.9 to 25 and achieved O₂ permeability as high as 270 barrer (1 barrer = 10⁻¹⁰ cm³(STP) cm cm⁻² s⁻¹ cm_Hg⁻¹).¹¹ However, liquid membranes face instability challenges because of solvent evaporation, pressure-induced liquid loss, and carrier activity loss.

Table 1 Enhanced O₂ transport properties in conventional liquid membranes and polymers containing Co²⁺-based carriers

Facilitated transport membranes			Temp. (°C)	Pressure (bar)	O2 permeability (barrer)	O2/N2 selectivity
Polymers/liquids	Cobalt carriers	Carrier content (wt%)
a GPU, 1 GPU = 10^−6 cm^3(STP) cm^−2 s^−1 cmHg^−1.
DMAc^11	Co (3-MeOsaltmen)	15	−10	0.19	270	25
Pebax 1657 (ref. 16)	CoPc	1	25	2	224	8.5
Polycarbonate^17	CoSalen	3	5	3	0.33	15
PDMS^18	CoSalen	20	30	0.05	46^a	7.7
PIM-1 (ref. 19)	mim2Co (NCS)4	2.0	30	2	113	5.3

---

To overcome the instability of liquid membranes, Co²⁺-based carriers were dissolved in solid polymers by ion coordination to form supramolecular polymer networks (SPNs) with good mechanical properties.^20–22 The dissociated Co²⁺ can reversibly bind O₂, improving O₂/N₂ separation properties (Table 1). For instance, adding 1 wt% cobalt(II) phthalocyanine (CoPc) in Pebax 1657 (a microphase-separated block copolymer containing PEO and nylon) increased O₂/N₂ selectivity from 2.9 to 8.5.¹⁶ Similar approaches have been adopted to develop Ag⁺-coordinated SPNs for olefin/paraffin separations. As Ag⁺ forms complexes with olefins, AgBF₄ was dissolved in Pebax to dramatically increase ethylene/ethane selectivity from 1 to 30.^23,24

In striking contrast, we report retarded O₂ transport in the SPNs prepared from Co(BF₄)₂ dissociated in crosslinked PEO (XLPEO) synthesized from poly(ethylene glycol) methyl ether acrylate (PEGMEA) and poly(ethylene glycol) diacrylate (PEGDA) (Fig. 1). The Co ions complex with ether oxygens, enabling complete dissociation of the Co salts, as evidenced by transparent and uniform films obtained (Fig. 1c), leading to extremely high O₂ sorption. However, instead of improving O₂ permeability, increasing the Co(BF₄)₂ content in XLPEO dramatically decreases O₂ permeability and increases CO₂/O₂ selectivity. The effect of Co ions on the structures and CO₂/O₂ transport properties of the SPNs is thoroughly investigated to derive structure and property relationships. This study unveils a new mechanism for harnessing strong binding capabilities with a penetrant to lower its diffusivity and permeability and enable its separation from other penetrants.

Fig. 1 Co²⁺-coordinated SPNs with superior CO₂/O₂ separation properties. (a) Schematic of O₂-retarded transport facilitating CO₂/O₂ separation. (b) Schematic of XLPEO prepared from PEGDA and PEGMEA and its interactions with Co(BF₄)₂. (c) Photos of XLPEO/salt films.

Experimental

Materials

PEGDA (n = 13; M_n = 700 g mol⁻¹), PEGMEA (n = 9; M_n = 480 g mol⁻¹), 1-hydroxycyclohexyl phenyl ketone (HCPK), and Co(ClO₄)₂·6H₂O were received from Sigma Aldrich Corporation (St. Louis, MO). Co(BF₄)₂·6H₂O was purchased from Thermo Fisher Scientific (Waltham, MA). Gas cylinders of O₂, N₂, and CO₂ (99.99%) were obtained from Airgas Inc. (Buffalo, NY).

Preparation of Co²⁺-coordinated SPNs

To prepare SPNs, PEGDA, PEGMEA, HCPK (0.1 mass% relative to PEGDA and PEGMEA), and Co salt were dissolved in water. The mass ratio of PEGDA to PEGMEA was set at 1 : 4 to obtain XLPEO with high gas permeability and good mechanical properties.²⁵ The prepolymer solution was sandwiched between two quartz plates and photopolymerized using UV light with a wavelength of 254 nm at 3.0 mW cm⁻² for 5 min.²⁶ The obtained solid films were vacuumed at ≈23 °C for three days or more to remove water and then kept under vacuum before use. The samples are named XLPEO/salt-x, where x represents the mass% of the salt in the dried films.

Characterization of SPNs

Film thickness was measured using a digital micrometer (Mitutoyo Corporation, Kanagawa, JP). An attenuated total reflection-Fourier transform infrared (ATR-FTIR) spectrometer (Vertex 70, Billerica, MA) was utilized at a resolution of 4 cm⁻¹ to investigate the conversion of PEGDA and PEGMEA and the salt state in XLPEO. The SPN structure was examined using an Ultima IV X-ray diffractometer (XRD, Rigaku Corporation, Tokyo, JP) with CuKa radiation (a wavelength of 1.54 Å) at ≈23 °C. Thermal transitions were obtained by Differential Scanning Calorimetry (DSC, Q2000, TA Instruments, New Castle, DE). Two heating cycles were carried out from −90 to 100 °C at 10 °C min⁻¹ in a N₂ atmosphere. Glass transition temperature (T_g) was determined as the inflection point of a step change in the second heating cycle. SPN density, ρ_SP (g cm⁻³), was determined using an analytical balance equipped with a density kit.^26,27

Pure-gas permeability of the SPN films was determined using a constant-volume and variable-pressure apparatus at different feed pressures (4.4, 7.9, and 11.3 bar) and 35 °C.²⁸ The measurement followed an order of N₂, CO₂, and O₂. N₂ was also re-tested for some samples after O₂ to ensure that the samples remained intact after O₂ exposure. Gas permeability has an uncertainty of less than 10%, estimated using an error propagation analysis.²⁹ Pure-gas sorption in the SPNs (∼2 g) was determined using a dual-volume and dual-transducer apparatus based on the pressure-decay method at 35 °C and three pressures (∼4.5, ∼7.9, and ∼11.4 bar).³⁰

Results and discussion

Physical and chemical properties of Co²⁺-coordinated SPNs

Fig. 2a exhibits FTIR spectra of XLPEO/Co(BF₄)₂ SPNs. The peak at 1096 cm⁻¹ representing C–O vibration in XLPEO redshifts with increasing Co(BF₄)₂ content, indicating the weakened C–O bonds caused by the interactions with Co²⁺.^26,31 For example, adding 17 mass% Co(BF₄)₂ lowers the C–O stretching frequency from 1097 to 1057 cm⁻¹. By contrast, increasing the Co(BF₄)₂ content has minimal impact on the C=O stretching at 1732 cm⁻¹, indicating the absence of strong interactions with Co²⁺. Similar behaviors have been observed for XLPEO/Co(Cl₄)₂ (Fig. S1a†). Additionally, the SPNs exhibit a new peak at 623 cm⁻¹ characteristic of free ClO₄⁻ anions, confirming the dissociation of Co(ClO₄)₂ by XLPEO.³¹

Fig. 2 Physical properties of SPNs. (a) FTIR spectra and (b) XRD patterns of XLPEO/Co(BF₄)₂. (c) Correlation between ρ_SP and the salt content using eqn (1). (d) Correlation between T_g and the salt content using eqn (2).

Fig. 2b and S1b† show that both series of SPNs are amorphous at ≈23 °C, validating complete dissociation of the Co salts by XLPEO. Adding salts has a minimal effect on the d-spacing (4.2 Å). Fig. 2c shows that increasing salt content increases ρ_SP (Tables 2 and S2†), which can be described using an additive model:³¹

where w is the mass fraction, and the subscripts of P and S represent properties of XLPEO and the amorphous salt, respectively. The fittings yield a ρ_S value of 3.84 g cm⁻³ for amorphous Co(ClO₄)₂ and 2.83 g cm⁻³ for amorphous Co(BF₄)₂. However, there are no reported values for anhydrous crystalline or dissociated amorphous Co(ClO₄)₂ or Co(BF₄)₂ in the literature for comparison.

Table 2 Physical properties of XLPEO/Co(BF₄)₂-x including r (molar ratio of ether oxygens to Co²⁺ ions), ρ_SP, T_g,SP, and pure-gas permeability and selectivity at 35 °C

x (mass%)	r	ρ SP (g cm^−3)	T g,SP (°C)	P A (barrer)			Selectivity
				N2	O2	CO2	CO2/N2	CO2/O2	O2/N2
0	—	1.149	−63	10	26	510	51	20	2.6
2	259	1.179	−63	2.9	2.2	109	38	50	0.76
3.4	150	1.182	−62	1.2	0.82	33	28	40	0.68
6.4	77	1.224	−57	0.49	0.40	15	31	38	0.82
17	26	1.282	−27	0.24	0.14	8.4	35	60	0.58

---

Fig. S1c and d† present the DSC thermograms of the SPNs. All samples do not show degradation at the temperatures ranging from −90 to 100 °C, indicating their potential applications for typical O₂ purification (∼23 °C). XLPEO shows a crystallization peak at −36 °C and a melting peak at −7 °C, which disappear after adding Co(BF₄)₂ because of the complexation between ethylene oxides and Co²⁺. The absence of melting peaks at around 0 °C also suggests the absence of water in these films.³² Additionally, increasing salt loading increases T_g,SP (Tables 2 and S2†), which can be described using an empirical equation:^26,33

1/T_g,SP = 1/T_g,P − ac_S (2)

where T_g,P is the glass transition temperature of XLPEO (K), a is an adjustable constant, and c_S is Co²⁺ content (kmol m⁻³). Fig. 2d shows that the best fit yields a value of 8.1 × 10⁻⁴ m³ kmol⁻¹ K⁻¹, close to that for XLPEO/Ni(BF₄)₂ (9.3 × 10⁻⁴ m³ kmol⁻¹ K⁻¹, Table S2†).²⁶

Gas sorption properties of SPNs

To elucidate the effect of the salt content on gas transport properties, CO₂ and O₂ sorption isotherms for XLPEO/Co(BF₄)₂ were obtained at 35 °C and are displayed in Fig. 3a and b. CO₂ and O₂ solubility can be calculated (Fig. 3c and Table S3†). Increasing Co(BF₄)₂ loading decreases CO₂ solubility and significantly increases O₂ solubility. For instance, adding 6.4 mass% Co(BF₄)₂ in XLPEO increases O₂ solubility from 0.19 to 6.7 cm³(STP) cm⁻³ atm⁻¹ because of the Co²⁺–O₂ complexation and decreases CO₂ solubility from 1.5 to 1.3 cm³(STP) cm⁻³ atm⁻¹ due to the competitive interactions with XLPEO from the Co²⁺ ions. More importantly, increasing the Co(BF₄)₂ content decreases CO₂/O₂ solubility selectivity and significantly increases O₂/CO₂ solubility selectivity (Fig. 3d and e). In contrast, adding Co(ClO₄)₂ has minimal impact on O₂ solubility (Fig. S2a†). Although the underlying mechanism for missing O₂ affinity in XLPEO/Co(ClO₄)₂ is unclear, it is widely reported that anions and ligands can significantly influence the complexation between Co²⁺ and O₂. For example, Co²⁺-based carriers with different anions exhibited a wide range of O₂ sorption values between 0.06 and 0.90 mol O₂ per mol carrier at 16 cm_Hg.¹¹ Similarly, when various silver salts were dissolved in PEO for C₃H₆ sorption, the SPNs showed a molar ratio of C₃H₆ to Ag⁺ of 0.49 for AgBF₄, 0.13 for AgCF₃SO₃, but only 0.03 for AgNO₃;³⁴ anions of bis(trifluoromethylsulfonyl)imide (Tf₂N) stabilized Ag⁺ against H₂, but not NO₃²⁻ anions.³⁵

Fig. 3 Enormous O₂ sorption in XLPEO/Co(BF₄)₂ SPNs at 35 °C. Effect of the Co(BF₄)₂ loading on (a) CO₂ sorption isotherms, (b) O₂ sorption isotherms, (c) CO₂ and O₂ solubility, and (d) CO₂/O₂ solubility selectivity.

Gas permeation properties of the SPNs

Fig. 4a and b present the effect of the Co(BF₄)₂ and Co(ClO₄)₂ content in XLPEO on pure-gas permeability at 35 °C (a typical temperature used for membrane applications), respectively. Increasing the salt loading decreases gas permeability because of the increased T_g,SP. For instance, adding 17 mass% Co(BF₄)₂ decreases CO₂ permeability by 98.4% from 520 to 8.4 barrer and O₂ permeability by 99.5% from 26 to 0.14 barrer. Similar behaviors have been observed for XLPEO with the addition of LiClO₄, Ni(BF₄)₂, and Cu(BF₄)₂ were correlated with the increased T_g,SP.²⁶ Increasing the salt content also decreases CO₂/N₂ selectivity because of the interactions between Co²⁺ and XLPEO, which reduce the favorable interactions between CO₂ and XLPEO. Importantly, XLPEO/Co(BF₄)₂ SPNs exhibit unexpected CO₂/O₂ selectivity. For instance, adding only 3.4 mass% Co(BF₄)₂ increases CO₂/O₂ selectivity from 20 to 40, while adding Co(ClO₄)₂ consistently decreases CO₂/O₂ selectivity (Table S1†). Interestingly, XLPEO/Co(BF₄)₂ SPNs even show O₂/N₂ selectivity less than 1 (Fig. S2b†), which is very unusual for polymeric films.

Fig. 4 Pure-gas transport properties of the SPNs at 35 °C. Dependence of gas permeability on (a) Co(BF₄)₂ loading and (b) Co(ClO₄)₂ loading. Dependence of CO₂/gas selectivity on (c) Co(BF₄)₂ loading and (d) Co(ClO₄)₂ loading. Effect of the Co(BF₄)₂ loading on (e) CO₂ and O₂ diffusivity and (f) CO₂/O₂ diffusivity selectivity.

CO₂ and O₂ diffusion coefficients (D_A, cm² s⁻¹) can be calculated by using D_A = P_A/S_A, and the results are shown in Fig. 4e, f, S2c and Table S3.† Increasing the Co salt content decreases CO₂ diffusivity because of the decreased chain flexibility, as indicated by the increased T_g,SP. Importantly, adding 17 mass% Co(BF₄)₂ increases CO₂/O₂ diffusivity selectivity from 2.3 to 650, validating the retarded O₂ diffusion in the XLPEO/Co(BF₄)₂ SPNs. In contrast, adding 17 mass% Co(ClO₄)₂ increases CO₂/O₂ diffusivity selectivity only to 5.7 (Fig. S2c†).

CO₂/O₂ separation properties

Fig. 5 exhibits superior CO₂/O₂ separation properties of the XLPEO/Co(BF₄)₂ SPNs in Robeson's upper bound plot.^36,37 All XLPEO/Co(BF₄)₂ SPNs (x = 2, 3.4, and 17) exhibit CO₂/O₂ separation properties above the upper bound.

Fig. 5 Superior CO₂/O₂ separation properties in XLPEO/Co(BF₄)₂ (x = 2, 3.4, and 17) compared with Robeson's upper bound,^36,37 and representative polymers, including Matrimid,⁶ Pebax 1657 and 2533,⁹ 6FDA-DAM,³⁸ PDMS,¹ and PIM-1.³⁹

Our preliminary data show that the SPNs may lose separation properties over time, as shown in Table S4.† After storage in a vacuum for ∼4 months, a freestanding film (XLPEO/Co(BF₄)₂-5) was tested with CO₂ at 50 psig and O₂ at 100 psig alternatively for 9 days. CO₂ permeability decreased from 75 to 12 barrer, and CO₂/O₂ selectivity decreased from 100 to 48, presumably because of the instability of Co²⁺ ions. Though CO₂/O₂ selectivity is still above the upper bound, the practical applications of these SPNs will need more investigation.

Conclusion

We began this work with a hypothesis that Co²⁺-based carriers can be incorporated into XLPEO to increase O₂ sorption and permeability because of the affinity of O₂ towards Co²⁺. Co(BF₄)₂ at loadings as high as 17 mass% can be dissociated by XLPEO to form SPNs, as indicated by DSC and XRD results. As expected, increasing Co(BF₄)₂ content increases T_g,SP and density and decreases gas permeability. Importantly, it dramatically increases O₂ solubility and O₂/CO₂ solubility selectivity. Surprisingly, the XLPEO/Co(BF₄)₂ SPNs exhibit extremely low O₂ permeability due to the retarded O₂ diffusion and thus unexpectedly high CO₂/O₂ permeability selectivity, surpassing Robeson's upper bound and current leading materials for CO₂/O₂ separation. This study unveils a new series of SPNs with the potential for CO₂/O₂ separations, and the discovered retarded diffusion may be harnessed to design high-performance materials for various gas and vapor separations.

Interestingly, adding Co(ClO₄)₂ into XLPEO does not improve O₂ sorption and decreases CO₂/O₂ permeability selectivity. We expect that the anion type influences the dissociation of the Co salts and the ability of Co²⁺ ions to interact with polymers and O₂. Additionally, Co²⁺ is subjected to oxidation and converted to Co³⁺ without affinity towards O₂. Therefore, experimental investigation and computational simulations will be needed to unravel the underlying mechanisms and determine the carrier stability. Future work should also focus on preparing thin-film composite membranes and evaluating them with real gas streams to determine long-term stability.

Author contributions

T. A.: conceptualization, investigation, data curation, writing – original draft preparation; N. E.: data curation and discussion, writing-reviewing and editing; G. Z.: data curation and discussion, writing-reviewing and editing; H. L.: conceptualization, writing-reviewing and editing, supervision.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work received financial support from the U.S. Department of Energy Small Business Technology Transfer Program (DE-SC0020730), the National Energy Technology Laboratory (DE-FE0031736), and the New York State Foundation for Science, Technology and Innovation (NYSTAR).

References

1. J. Wang, F. Kang, Y. Chen, X. Zhang, X. Jiang, G. He and X. Ruan, Sep. Purif. Technol., 2024, 337, 126461.
2. Z. Zhao, W. Gao, Y. Chang, Y. Yang, H. Shen, T. Li and S. Zhao, Adv. Healthcare Mater., 2023, 690, 122183.
3. B. S. Kirkland, R. Clarke and D. R. Paul, J. Membr. Sci., 2008, 324, 119–127.
4. T. C. Merkel, X. Wei, Z. He, L. S. White, J. G. Wijmans and R. W. Baker, Ind. Eng. Chem. Res., 2013, 52, 1150–1159.
5. J. Wu, F. Hillman, C.-Z. Liang, Y. Jia and S. Zhang, J. Mater. Chem. A, 2023, 11, 17452–17478.
6. T.-S. Chung, S. S. Chan, R. Wang, Z. Lu and C. He, J. Membr. Sci., 2003, 211, 91–99.
7. B. Zhu, X. Jiang, S. He, X. Yang, J. Long, Y. Zhang and L. Shao, J. Mater. Chem. A, 2020, 8, 24233–24252.
8. H. Lin and B. D. Freeman, J. Membr. Sci., 2004, 239, 105–117.
9. V. I. Bondar, B. D. Freeman and I. Pinnau, J. Polym. Sci., Part B: Polym. Phys., 2000, 38, 2051–2062.
10. X. Chen, H. Nishide, K. Oyaizu and E. Tsuchida, J. Phys. Chem. B, 1997, 101, 5725–5729.
11. B. M. Johnson, R. W. Baker, S. L. Matson, K. L. Smith, I. C. Roman, M. E. Tuttle and H. K. Lonsdale, J. Membr. Sci., 1987, 31, 31–67.
12. Y. Kohno, M. G. Cowan, A. Okafuji, H. Ohno, D. L. Gin and R. D. Noble, Ind. Eng. Chem. Res., 2015, 54, 12214–12216.
13. H. Zhao, T. Song, X. Ding, R. Cai, X. Tan and Y. Zhang, J. Membr. Sci., 2023, 679, 121713.
14. D. E. Jaramillo, A. Jaffe, B. E. Snyder, A. Smith, E. Taw, R. C. Rohde, M. N. Dods, W. DeSnoo, K. R. Meihaus and T. D. Harris, Chem. Sci., 2022, 13, 10216–10237.
15. H. Nishide, H. Kawakami, S. Toda, E. Tsuchida and Y. Kamiya, Macromolecules, 1991, 24, 5851–5855.
16. H. Nagar, P. Vadthya, N. S. Prasad and S. Sridhar, RSC Adv., 2015, 5, 76190–76201.
17. S. H. Chen and J. Y. Lai, J. Appl. Polym. Sci., 1996, 59, 1129–1135.
18. W. Choi, P. G. Ingole, H. Li, S. Y. Park, J. H. Kim, H.-K. Lee and I.-H. Baek, Microchem. J., 2017, 132, 36–42.
19. J. Han, L. Bai, S. Luo, B. Yang, Y. Bai, S. Zeng and X. Zhang, Sep. Purif. Technol., 2020, 248, 117041.
20. E. Khare, N. Holten-Andersen and M. J. Buehler, Nat. Rev. Mater., 2021, 6, 421–436.
21. L. Hu, V. T. Bui, S. Fan, W. Guo, S. Pal, Y. Ding and H. Lin, J. Mater. Chem. A, 2022, 10, 10872–10879.
22. Y. Fu, L. Chen, F. Xu, X. Li, Y. Li and J. Sun, J. Mater. Chem. A, 2022, 10, 4695–4702.
23. A. Morisato, Z. J. He, I. Pinnau and T. C. Merkel, Desalination, 2002, 145, 347–351.
24. T. C. Merkel, R. Blanc, I. Ciobanu, B. Firat, A. Suwarlim and J. Zeid, J. Membr. Sci., 2013, 447, 177–189.
25. H. Lin, E. Van Wagner, J. S. Swinnea, B. D. Freeman, S. J. Pas, A. J. Hill, S. Kalakkunnath and D. S. Kalika, J. Membr. Sci., 2006, 276, 145–161.
26. T. Alebrahim, A. Chakraborty, L. Hu, S. Patil, S. Cheng, D. Acharya, C. M. Doherty, A. J. Hill, T. R. Cook and H. Lin, J. Membr. Sci., 2022, 644, 120063.
27. T. Alebrahim, L. Huang, H. K. Welgama, N. Esmaeili, E. Deng, S. Cheng, D. Acharya, C. M. Doherty, A. J. Hill, C. Rumsey, M. Trebbin, T. R. Cook and H. Lin, ACS Appl. Mater. Interfaces, 2024, 16, 11116–111124.
28. K. M. Rodriguez, W. Wu, T. Alebrahim, Y. Cao, B. D. Freeman, D. Harrigan, M. Jhalaria, A. Kratochvil, S. Kumar, W. Lee, Y. Lee, H. Lin, J. M. Richardson, Q. Song, B. Sundell, R. Thur, I. Vankelecom, A. Wang, C. Wiscount, J. Xingming and Z. P. Smith, J. Membr. Sci., 2022, 659, 120746.
29. P. R. Bevington and D. K. Robinson, Data reduction and error analysis for the physical sciences, McGraw-Hill, Inc., New York, 2nd edn, 1992.
30. T. Tran, Y. Fu, D. Jiang and H. Lin, Macromolecules, 2022, 55, 9860–9867.
31. N. Paranjape, P. C. Mandadapu, G. Wu and H. Lin, Polymer, 2017, 111, 1–8.
32. T. Tran, C. Lin, S. Chaurasia and H. Lin, J. Membr. Sci., 2019, 574, 299–308.
33. C. J. Hawker, F. K. Chu, P. J. Pomery and D. J. T. Hill, Macromolecules, 1996, 29, 3831–3838.
34. S. Sunderrajan, B. D. Freeman, C. K. Hall and I. Pinnau, J. Membr. Sci., 2001, 182, 1–12.
35. S. Park, O. Morales-Collazo, B. Freeman and J. F. Brennecke, Angew. Chem., Int. Ed., 2022, 61, e202202895.
36. B. D. Freeman, Macromolecules, 1999, 32, 375–380.
37. H. Lin and M. Yavari, J. Membr. Sci., 2015, 475, 101–109.
38. L. Shao, T.-S. Chung, S. H. Goh and K. P. Pramoda, J. Membr. Sci., 2005, 267, 78–89.
39. C. G. Bezzu, A. Fuoco, E. Esposito, M. Monteleone, M. Longo, J. C. Jansen, G. S. Nichol and N. B. McKeown, Adv. Funct. Mater., 2021, 31, 2104474.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ta02376e

---