A phase conversion method to anchor ZIF-8 onto a PAN nanofiber surface for CO2 capture

Polyacrylonitrile (PAN) nanofibers were prepared by electrospinning and coated with zeolitic imidazolate framework-8 (ZIF-8) by a phase conversion growth method and investigated for CO2 capture. The PAN nanofibers were pre-treated with NaOH, and further coated with zinc hydroxide, which was subsequently converted into ZIF-8 by the addition of 2-methyl imidazolate. In the resulting flexible ZIF-8/PAN composite nanofibers, ZIF-8 loadings of up to 57 wt% were achieved. Scanning electron microscopy and energy-dispersive X-ray spectroscopy (EDS) showed the formation of evenly distributed submicron-sized ZIF-8 crystals on the surface of the PAN nanofibers with sizes between 20 and 75 nm. X-ray photoelectron spectroscopy (XPS) and carbon-13 nuclear magnetic resonance (13C NMR) investigations indicated electrostatic interactions and hydrogen bonds between the ZIF-8 structure and the PAN nanofiber. The ZIF-8/composite nanofibers showed a high BET surface area of 887 m2 g−1. CO2 adsorption isotherms of the ZIF-8/PAN composites revealed gravimetric CO2 uptake capacities of 130 mg g−1 (at 298 K and 40 bar) of the ZIF-8/PAN nanofiber and stable cyclic adsorption performance.


Introduction
Recently, the NET-ZERO 2050 report was published by the European Climate Foundation, 1 which pointed out the urgency to achieve zero carbon emission by 2050 by making the energy system and transportation more sustainable and energyefficient with the goal to eliminate carbon emissions. A part of the activities to achieve these goals is more efficient gas separation, capture and storage technologies for carbon capture and storage (CCS). Adsorption-based methods, such as pressure swing adsorption (PSA) 2 or temperature swing adsorption (TSA) 3 technologies offer opportunities for tailor-made gas separation of CO 2 from gas mixtures 4 and subsequent storage, for example in biogas upgrading 5 and storage as biomethane (CH 4 ). 6 CH 4 produced from sustainable sources such as biogas, can help to reduce anthropogenic CO 2 emissions. One key point to improve adsorption-based technologies is exploring and improving advanced microporous materials (sorbents), 7 such as zeolites, 8 carbon molecular sieves 9 and active carbon. 10 It has been shown that by heteroatom-doping the CO 2 adsorption can be signicantly affected (e.g. in activated carbons) by creating highly polar adsorption sites for the CO 2 molecule and Lewis acid-base interactions. This usually results in improved selectivities in CO 2 separation. [11][12][13] Another promising new class of materials are metal-organic frameworks (MOFs), which are assembled from metal ions/clusters and organic linkers. 14 These materials can be prepared from a variety of chemical compositions, resulting in structures with ultrahigh surface area of several thousand m 2 g À1 , large pore volumes with tailorable pore size and specic chemical sites for capturing guest molecules by adsorption. In these materials, the extremely large internal surface area and pore channels can be modied to introduce specic surface charges, hydrophilicity/ hydrophobicity or uniformity of pore channels. Zeolitic Imidazolate Frameworks (ZIFs), a sub-class of MOFs, have been proposed for gas separation and storage applications. 15 These materials have rst been synthesized by Yaghi's group. 16 The structure consists of zinc(II) cations and 2-methylimidazole anions (2-mim). Within the class of MOF materials, ZIFs have an interesting framework structure with high thermal and chemical stability. According to a report from Park et al., 15 ZIF-8 has a large sodalite (SOD) zeolite-like cavity (11.6Å) and smaller apertures (3.4Å). The size of these pores, the high achievable specic surface area of more than 1000 m 2 g À1 , and large micropore volumes indicate that this material has promising potential for CO 2 capture and separation.
Beneath shaping of microporous materials into useful structures, such as pellets, granules or extrudates, 30 recently, new approaches have been proposed to structure adsorbents (carbons, zeolites, MOFs etc.) into nanobers to enhance gas transport, 17 adsorption capacity 18 and adsorption kinetics. 19 For example, zeolite nanobers of ZSM-5 have been fabricated by electrospinning, post shaping and carbonization of polyvinylpyrrolidone (PVP) 20 the performance has been evaluated for CO 2 separation from CO 2 /CH 4 mixtures in synthetic biogas mixtures.
Due to the temperature instability of MOF materials, this approach to fabricate MOF-nanober composite materials is not feasible for the majority of MOF-based nanobers. However, different approaches for the structuring of MOF-nanober materials for energy and environmental applications have been recently reviewed by Dou et al., 21 including in situ growth of MOF nanocrystals on the surface of polymer nanobers. The preparation of MOF-polymer nanobers through electrospinning is an elegant way to shape MOF materials into hierarchical porous nanobrous sorbent materials at mild conditions with large surface-to-volume ratio, tailored pore size, and high permeability for gas separation processes. In this study we propose a phase inversion method to grow ZIF-8 on the surface of electrospun polyacrylonitrile (PAN) nanobers for CO 2 capture. We investigated the inuence of the growth of the ZIF-8 nanoparticles on PAN nanobers as a function of reaction time to maximize the CO 2 gas uptake, which was expected to increase with the amount of the microporous ZIF-8 sorbent loaded on the inactive PAN nanober as backbone.
The resulting ZIF-8/PAN composite nanobers have been characterized by scanning electron microscope-energy dispersive X-ray spectroscopy (SEM-EDS), X-ray diffraction pattern (XRD), surface area and pore size distribution analysis and by more advanced characterization techniques, such as X-ray photoelectron spectroscopy (XPS) and nuclear magnetic resonance spectroscopy (NMR).
Finally, the CO 2 uptake of the ZIF-8/PAN composite nano-bers have been tested at pressures up to 40 bars and the cyclic adsorption-desorption properties of the composites have been investigated.

Materials
For the in situ growth of ZIF-8, the raw chemicals zinc nitrate hexahydrate (Zn(NO 3 ) 2 $6H 2 O), sodium hydroxide (NaOH) and 2-methylimidazole (2-mim) with purity over 99.0% were purchased from Sigma-Aldrich. For the electrospinning (ES), polyacrylonitrile (PAN) with an average molecular weight of 150 000 was chosen as polymer with a purity of 99.0%. The PAN was purchased from Parchem ne & specialty Co., Ltd N,N-Dimethylformamide (DMF) with purity of 99.0% was purchased from Sigma-Aldrich. Besides, the solvent of methanol and ethanol with purity of 99.9% were purchased from Sigma-Aldrich. The deionized water (D.I.W) was produced in the laboratory by an ultrapure system with a resistivity of 18.0 mU cm À1 . All chemicals used in this work have not been further treated or puried. High purity gases of CO 2 (99.99 mol%), CH 4 (99.99 mol%), He (99.99 mol%) and N 2 (99.99 mol%) (Foshan Huate Gas Co., Ltd, China) were used for gas adsorption experiments.

Preparation of ZIF-8/PAN composite nanobers
The preparation of the ZIF-8/PAN composite nanobers can be divided into four steps: in the 1 st step, the PAN solution was electrospun into PAN nanobers. In the 2 nd step, the PAN nanobers were immerged into a heated sodium hydroxide solution (75 C) to activate the surface. In the 3 rd step, the metalcontained (Zn 2+ ) precursors were attached on the active PAN surface. In the last step, the in situ growth process of ZIF-8 was initiated by the addition of the complementary precursor (2mim solution).
For the preparation of the PAN nanobers, 1 g PAN and 10 g DMF were mixed and sealed into a 50 mL plastic bottle with cylindrical zirconia balls to ball-mill by regular ball milling at 50 rpm for 24 hours. Then the dispersion solution was electrospun by needle-based electrospinning equipment (Linari s.r.l) with the voltage of 30 kV and a distance of 140 mm between the electrospinning needle and receiving substrate to prepare PAN nanobers.
For the activation of the PAN nanober surface with hydroxide groups, 0.5 g of PAN nanobers were immerged into the 50 mL NaOH-D.I.W solution (60 g NaOH dissolved into 500 mL D.I.W) and heated at 75 C for 20 min to activate the ber surface. The treated PAN nanobers were 3 times washed with ethanol and dried aerwards in air at room temperature.
For the application of the metal precursor, the surfacetreated PAN nanobers were immersed into 100 mL of a 120 millimole per liter (mM) Zn(NO 3 ) 2 solution for 20 min to form Zn(OH) 2 on the surface of PAN nanobers. The applied Zn(NO 3 ) 2 solution consisted of 6.29 g Zn(NO 3 ) 2 $6H 2 O in 180 mL of a 1 : 1 volumetric solvent mixture of methanol and ethanol.
For the nal in situ growth process, 100 mL 480 mM 2-mim solution was added to the Zn(NO 3 ) 2 solution directly, where the Zn(OH) 2 -coated PAN ber had been immerged already to complete the in situ growth process. Every 30 minutes, the samples were taken and dried aer ethanol washing to identify the optimum loading and structure of ZIF-8 in the composite. The applied 2-mim solution consisted of 7.49 g 2-methylimidazole in 190 mL of a 1 : 1 volumetric solvent mixture of methanol and ethanol. According to reaction time the samples were labelled ZIF-8/PAN-X (with X resembling 30, 60 and 90 minutes of reaction time, see Table 1 further below).

Preparation of ZIF-8 nanopowder
For the synthesis of ZIF-8 nanocrystals in powder form, 50 mL 120 mM Zn(NO 3 ) 2 solution and 50 mL 480 mL 2-mim solution were prepared (as described in Section 2.2) and mixed by vigorously magnetic stirring for 1 hour at room temperature, then the suspension solution was centrifuged to obtain the ZIF-8 colloids. Followed, the ZIF-8 colloids were washed with ethanol for 3 times. Finally, the ZIF-8 nanocrystals were dried at 100 C to remove the residual ethanol, methanol and moisture.

Characterization of the materials
Scanning electron microscope-energy dispersive X-ray spectroscopy (SEM-EDS) were performed by a Zeisis Merlin FEG-SEM, the samples were sputter-coated with the thin layer of the gold (around 12-16 nm).
X-ray diffraction (XRD) measurements were performed by a Rigaku Smartlab X-ray diffractometer, which was operated at 40 kV and 30 mA using Cu Ka radiation with step speed of 5 min À1 . All of the samples were attached on a silicon-substrate sample holder. The miller indexes were identied by Match! 3 soware.
X-ray photoelectron spectroscopy (XPS) was performed in the range of 200 to 1200 eV with ZIF-8 nanocrystals, pure PAN ber and ZIF-8/PAN composite nanobers in a angle resolved XPS chamber for depth proling of surfaces (equipment from Thermo Scientic Nexsa). 13 C nuclear magnetic resonance spectroscopy (NMR) was performed on a Bruker Avence III 300 Wide Bore solid state NMR spectrometer, using a 4 mm probe. The frequency of 13 C was 75 MHz with the pulse program of CP-Toss, using 4 mm rotors at a MAS spinning frequency of 6500 Hz. The 90 13 C pulse-length was 4 ms, and pulse length in decoupling sequence is 7. Recycle delay is 5 s, while the sweep wide is 30 241.94 Hz. The carbon resonance line of adamantane was used as an external chemical shi standard and was assigned a value of 38.48 ppm. The scans for CP-Toss experiments was 1k. Besides, the samples were frozen and milled into powder with help of liquid nitrogen before it was lled into NMR tube.
The surface area as well as pore-related information (size distribution, total pore volume, micro-pore volume) determination of ZIF-8/PAN composite materials and ZIF-8 powders were measured by nitrogen adsorption isotherms at À196 C (Autosorb-iQ-2, specic analysis for microstructure). The CO 2 uptake measurements were also carried out in this equipment. All samples were degassed to below 10 À6 bar and 383 K for 8 hours prior to the measurement.
The high-pressure CO 2 adsorption performance was measured by an IsoSORP adsorption analyzer (TA instruments, New Castle, DE, USA), on samples with a weight of about 0.5 g. All samples were degassed at 110 C under high vacuum for 4 h, following a buoyancy test with helium at 22 C to determine the mass and volume of each sample. Then, the CO 2 adsorptiondesorption isotherms of the obtained ZIF-8/PAN composite were measured with CO 2 dosing from high vacuum to 40 bar, then to high vacuum with an equilibrium point every 10 bar. The equilibrium for each step was set until the deviation of mass was less than 0.1 mg per 10 min.

Result and discussion
The in situ growth of the ZIF-8 on the surface of electrospun PAN nanobers resulted in exible, paper-like ZIF-8/PAN composite nanober mats with light yellow colour, whereas the solvothermally synthesized ZIF-8 powder showed white nanocrystals.  (Fig. 1a), 60 min (Fig. 1b) and 90 min (Fig. 1c). Small ZIF-8 nanoparticles (with rhombic dodecahedral form) have evenly grown on the surface of the PAN nanobers, forming a core-shell structure. Visual inspection indicate a good mechanical adhesion of the ZIF-8 crystal on the PAN polymer surface and this has also been observed in previous studies on ZIF-8/nanober composites and nanocrystals. 23-25 X-ray diffractograms of the ZIF-8/PAN composites aer 30, 60, 90 minutes reaction time are given  a The micro-pore volumes have been calculated from a T-plot, following a method reported by Galarneau et al. 22 in Fig. 2 together with the solvothermally synthesized ZIF-8 nanopowder.

Microstructure of ZIF-8/PAN nanober structures
The pattern-ray diffractogram of the synthesized ZIF-8/PAN composite nanobers matched closely with the simulated Xray diffractogram of ZIF-8 powder, which can be assigned to cubic I 43m space group, 15 conrming the growth of ZIF-8 nanoparticles on the surface of PAN nanobers. Lower relative diffraction intensity of X-rays and broadening of diffraction peaks at FWHM (full width at half maximum) can be observed for the ZIF-8/PAN composites samples with decreasing reaction time, Fig. 2. This suggests that the crystallinity and crystallite size of the ZIF-8 particles, attached to the PAN nanobers, increases with reaction time. Crystallite sizes increase from 20 to 70 nm for increasing reaction time between 30 to 90 minutes. Effects of crystallite growth and size in ZIF-8 nanopowder has previously been investigated by Tanaka et al. 26 Table 1 summarizes the average crystal size of the ZIF-8 crystals along with the surface textural properties of the ZIF-8/ PAN composite nanober and the ZIF-8 nanopowder, which are discussed in Section 3.3. Fig. 3a and c show the SEM micrographs and Fig. 3b and d the Zinc elemental mappings of the Zn(OH) 2 /PAN composite nanobers (zinc hydroxide attached to the PAN nanober before the in situ growth of the MOF) and of the ZIF-8/PAN composite-90 nanober structure. The larger size and broader distribution of the Zinc in the initial Zn(OH) 2 / PAN composite (Fig. 3b) compared to the nal ZIF-8/PAN-90 composite nanober (Fig. 3d) indicate that there occurs a redistribution of zinc atoms from the initial Zn(OH) 2 phase during the phase conversion process into the ZIF-8 crystallites. The explanation is that during the growth of the ZIF-8 on the surface of nanober, zinc ions from the hydroxide phase are nally incorporated on a molecular level into the crystal structure of the MOF particles, which results in a more even and ner coverage of the PAN nanober surface with Zn.

Chemical interaction between PAN nanober and ZIF-8
In order to reveal chemical interactions between the ZIF-8 phase and the PAN nanobers, X-ray photoelectron spectroscopy (XPS) was carried out. The scan of a XPS spectrum of ZIF-8/PAN-90 and ZIF-8 powder exhibit peaks of Zn 2p 3/2 and Zn 2p 1/2 at 1022.08 and 1044.88 eV, respectively (Fig. 4).
To verify interaction between the PAN nanober and ZIF-8 framework, narrow scans of the N 1s and C 1s were analyzed 27 for the ZIF-8/PAN-90 composite nanober, the ZIF-8 powder and pure PAN nanobers, respectively (Fig. 4). The energy peak of the N 1s exhibits a 0.80 eV red shi from 400.8 to 399.3 eV aer in situ growth of ZIF-8 on the PAN nanober. Similarly, the C 1s signal shows a red shi of 1.9 eV aer the PAN nanober has   been coated with ZIF-8 and resulting in ZIF-8/PAN-90 composite. These shis are related to the change of the atomic chemical environment of carbon and nitrogen, indicating interaction between the PAN nanober and the ZIF-8 structure. 13 C NMR spectroscopy was conducted to further analyse the atomic chemical environment. As shown in Fig. 5, the location of the resonance peaks at 152.0 and 14.6 ppm, respectively, can be assigned to protonated aromatic carbon atoms (C1, carbon) and the methyl group (C3, -CH 3 ). These peaks have no major chemical shi. However, there is a visible chemical shi by 1.7 ppm from 123.7 ppm to 125.4 ppm for the double-bond carbon (C2, -CH] CH-) when comparing the spectra for the ZIF-8 nanopowder with the ZIF-8 crystallites attached to the ZIF-8/PAN-90 composite nanober. Similar chemical shi of the C2 peak in ZIF-8 has been observed when organic molecules, such as caffeine, were encapsulated in the pores of ZIF-8 to form van der Waals bonds. 28 However, more detailed analysis supported by DFT calculations 29 would be required to further proof electrostatic interactions and hydrogen bonding between the ZIF-8 and C-N groups of the PAN nanober (as visualized by the chemical formulas in Fig. 5). A large peak centered around a pore radius of about 11Å can be assigned to micropores, whereas two smaller and broader peaks between 20 to 70 A are related to micro and mesopores. Table 1 summarizes the micropore and total pore volume and BET surface area of all ZIF-8/PAN nanobers and the ZIF-8 nanopowder. The comparably low total pore volume of 0.38 cm 3 g À1 and specic surface area (495 m 2 g À1 ) of the sample obtained aer short reaction time (30 minutes), indicate that the ZIF-8 has not been fully crystallized or has major surface defects. The broad and slightly displaced XRD peaks for this sample in Fig. 2 support this thesis. Aer longer reaction time (90 minutes), the total pore volume increases to 0.51 cm 3 g À1 and the surface area to 888 m 2 g À1 for the sample ZIF-8/PAN-90. Considering a maximum ZIF-8 loading of 57.19 wt% in ZIF-8/PAN-90 achieved in this work, the surface area of the pure ZIF-8 crystallites in the composite can be calculated to 1552 m 2 g À1 , assuming negligible surface area of the PAN nanobers. Using similar estimates (43 wt% PAN in the composite), the actual micropore and total pore volumes of the ZIF-8 crystallites in the ZIF-8/PAN-90 composite should be about 0.54 and 0.88 cm 3 g À1 , respectively. This indicates that by the in situ growth method proposed in this work, ZIF-8/PAN nanober composites with high surface area, pore volume and ZIF-8 particle loadings have been achieved compared to ZIF-8-nanobers in other studies. Table 2, further below, gives a comparision of the CO 2 uptake performance of the ZIF-8 nanober composite materials in this study compared to other studies. Fig. 7 shows the cyclic CO 2 uptake and re-generation of the ZIF-8/PAN-90 at 25 C for 1 bar and for 40 bars for four cycles in Fig. 7a and b, respectively. The CO 2 uptake of ZIF-8/PAN-90 nanobers at 1 bar and 40 bars is 7 cm 3 g À1 and 130 mg g À1 .

Surface textural properties of ZIF-8/PAN composite nanobers
These initial results demonstrate good initial cyclic stability for CO 2 separation with fast adsorption-desorption kinetics. Reasonable high CO 2 uptake of ZIF-8/PAN-90 nanober composites have been achieved compared to similar studies on ZIF-8, 36 suggesting that these materials are interesting for CO 2 separation from gas streams.

Conclusion
In summary, we successfully demonstrated a phase inversion method to grow ZIF-8 crystals on the surface of a PAN polymer nanober matrix without sacricing the surface properties of the material. XPS and 13 C NMR analysis indicate electrostatic interaction and hydrogen bonds between the PAN matrix and the ZIF-8 nanocrystals conned to the PAN nanober surface, resulting in good attachment of the ZIF-8 nanoparticles and relatively high CO 2 uptake. We believe that this method to grow MOF structures into polymer nanober structures offers further opportunities to produce nanobrous adsorbent materials for other applications, such as air or liquid ltration or in medical applications.

Conflicts of interest
There are no conicts to declare.