Adsorption of Carbon Dioxide with Ni-MOF-74 and MWCNT Incorporated Poly Acrylonitrile Nanofibers

Among the new adsorbent forms, nanofiber structures have attracted extra attention because of features such as high surface area, controllable properties, and fast kinetics. The objective of this study is to produce the polyacrylonitrile (PAN) electrospun nanofibers loaded with Ni-MOF-74/MWCNT to obtain maximum CO2 adsorption. The prepared PAN/MWCNT/MOF nanofiber based on the Box–Behnken design (BBD) model suggests the CO2 adsorption of about 1.68 mmol/g (at 25 °C and 7 bar) includes 14.61 w/v%, 1.43 w/w%, and 11.9 w/w% for PAN, MWCNT, and MOF, respectively. The results showed the effective CO2 adsorption of about 1.65 ± 0.03 mmol/g (BET = 65 m2/g, pore volume = 0.08 cm3/g), which proves the logical outcomes of the chosen model. The prepared PAN/MWCNT/MOF nanofiber was characterized using different analyzes such as SEM, TEM, TG, XRD, FTIR, and N2 adsorption–desorption isotherms. More MOF mass loading on the nanofiber surface via secondary growth method resulted in 2.83 mmol/g (BET = 353 m2/g, pore volume = 0.22 cm3/g, 43% MOF mass loading) and 4.35 mmol/g (BET = 493 m2/g, pore volume = 0.27 cm3/g, 65% MOF mass loading) CO2 adsorption at 7 bar for the first and second growth cycles, respectively. This indicates that secondary growth is more effective in the MOF loading amount and, consequently, adsorption capacity compared to the MOF loading during electrospinning.


Introduction
Adsorption is an efficient and low energy-consuming process to capture carbon dioxide. [1,2]. Adsorbent structural properties, such as volumetric working capacity, pressure drop, mass transfer, and thermal properties, regulate the performance of adsorptive gas separation processes [3]. These properties are determined by the parameters, such as specific heat, density, the heat of adsorption, porosity, total pore volume, the specific surface area of the adsorbent composite, and also the adsorbent arrangement in adsorption equipment [3,4].
Adsorbents are conventionally used in the form of beads and granules (pellets) in packed bed columns. Consequently, adsorbent particle size can impose disadvantages, such as mass transfer and pressure drop limitations. Therefore, the reduction in particle size is the most common way to decrease mass transfer resistance. However, it can cause more pressure loss [3]. The balance between pressure drop and mass transfer rate usually results in a bed geometry with a large diameter and small height at a particle size around 0.7 mm, which can induce considerable gas maldistribution and channeling. The novel adsorbent structures, such as monoliths, laminates, foams, and fabric structures, can be Reagent grade of Nickel acetate and 2,5-dihydroxyterephthalic acid were purchased from Sigma-Aldrich. Dimethylformamide (DMF), Ethanol, and tetrahydrofuran (THF) were purchased from Merck. Polyacrylonitrile (PAN) with a molecular weight of 100,000 (g/mol), was purchased from Polyacryl Company, Isfahan, Iran. MWCNT (Diameter <8 nm, Length~30 nm, -COOH content~3.86 wt.%) was provided by Neutrino Company, Tehran, Iran. Deionized (DI) water was used through this work.

MOF Synthesis
The Ni/DOBDC samples were synthesized according to the following hydrothermal method [27,31]. Nickel acetate (0.746 g or 3 mmol) and 2,5-dihydroxyterephthalic acid (0.298 g, 1.5 mmol) were added to 40 mL of 50 vol.% THF in DI water solution. Then the suspension sonicated for half an hour. The mixture was put into the autoclave and heated at 110 • C for 3 days to achieve the yellow crystalline MOF. For activation, the product was rinsed with a mixed solution of methanol, ethanol, and deionized water several times. The resulting MOFs were separated by centrifugation at 6000 rpm. Then MOF particles were dried in an evacuated oven at 100 • C for 2 h and known as Ni-MOF-74.

Electrospinning
To prepare the polymeric solution for electrospinning, the measured amounts of MOF and MWCNT, according to BBD, were mixed with the PAN-DMF solution.
The Polymeric solution injects through a capillary tip with a diameter of 0.7 mm using a 5 mL syringe. Figure 1 demonstrates the schematic diagram of the electrospinning of PAN/MWCNT/MOF nanofibers. The anode of the high voltage power supply has clamped to the metal collector, and the cathode has been connected to the needle tip. The applied voltage was 20 KV, the distance between the tip and the collector was 18 cm, and the flow rate of the spinning solution was 0.5 mL/h. The resulting nanofibers accumulated on an aluminum foil-wrapped drum rotating at approximately 250 rpm. Finally, nanofibers were vacuum dried at 100 °C for 12 h.

Design of Experiment/Response Surface Methodology (RSM)
Response surface methodology (RSM) is a suitable method for finding the effects of several independent variables on the response and determining the optimum conditions. Amongst different design of experiment (DOE) methods, many studies have adopted the Box-Behnken design, BBD, because of its need for fewer experiments [45,46].
To determine the effect of three variables, PAN concentration (X1), MWCNT concentration (X2), and MOF concentration (X3) on the CO2 adsorption capacity, BBD was used. Each parameter has three levels, which result in 17 experiments, while the CO2 adsorption is the response. Table 1 lists variables and levels, which have been chosen through the screening experiments. The standard analysis of variance was employed to analyze the model outputs, and the values of p ≤ 0.05 were considered statistically significant [45,46].

Secondary Growth of Ni-MOF-74 on the Nanofiber
The secondary loading of Ni-MOF-74 was performed via hanging up a PAN/MWCNT/MOF nanofiber (Sections 2-4) in a Teflon-autoclave during the MOF synthesizing process (as mentioned in Section 2.2). This process was repeated for two consecutive cycles until the nanofiber became saturated. The composite nanofiber was rinsed several times with a mixed solution of methanol, ethanol, and deionized water after each cycle to remove the remaining solvent and activate the adsorbent. Finally, it was placed in an evacuated oven at 100 °C for 2 h. The applied voltage was 20 KV, the distance between the tip and the collector was 18 cm, and the flow rate of the spinning solution was 0.5 mL/h. The resulting nanofibers accumulated on an aluminum foil-wrapped drum rotating at approximately 250 rpm. Finally, nanofibers were vacuum dried at 100 • C for 12 h.

Design of Experiment/Response Surface Methodology (RSM)
Response surface methodology (RSM) is a suitable method for finding the effects of several independent variables on the response and determining the optimum conditions. Amongst different design of experiment (DOE) methods, many studies have adopted the Box-Behnken design, BBD, because of its need for fewer experiments [45,46].
To determine the effect of three variables, PAN concentration (X1), MWCNT concentration (X2), and MOF concentration (X3) on the CO 2 adsorption capacity, BBD was used. Each parameter has three levels, which result in 17 experiments, while the CO 2 adsorption is the response. Table 1 lists variables and levels, which have been chosen through the screening experiments. The standard analysis of variance was employed to analyze the model outputs, and the values of p ≤ 0.05 were considered statistically significant [45,46].

Secondary Growth of Ni-MOF-74 on the Nanofiber
The secondary loading of Ni-MOF-74 was performed via hanging up a PAN/MWCNT/MOF nanofiber (Sections 2-4) in a Teflon-autoclave during the MOF synthesizing process (as mentioned in Section 2.2). This process was repeated for two consecutive cycles until the nanofiber became saturated. The composite nanofiber was rinsed several times with a mixed solution of methanol, ethanol, and deionized water after each cycle to remove the remaining solvent and activate the adsorbent. Finally, it was placed in an evacuated oven at 100 • C for 2 h.

CO 2 Adsorption Measurements
An experimental setup based on a volumetric method, Figure 2, was utilized to study the CO 2 adsorption capacities of Ni-MOF-74 in powder and nanofibers forms. Adsorption capacities are determined by measuring the pressure and temperature of the adsorption cell. The difference between the initial and final pressures of the adsorption cell reveals the equilibrium amount of the adsorbed CO 2 . In a typical test run, 100 mg of a previously degassed sample at 100 • C for 12 h was used. Three similar adsorption/regeneration cycles run for each sample. The regeneration took place in an oven at 100 • C for 4 h. The accuracy in measuring pressure and temperature was 1000 Pa and 0.1 K, respectively.

CO2 Adsorption Measurements
An experimental setup based on a volumetric method, Figure 2, was utilized to study the CO2 adsorption capacities of Ni-MOF-74 in powder and nanofibers forms. Adsorption capacities are determined by measuring the pressure and temperature of the adsorption cell. The difference between the initial and final pressures of the adsorption cell reveals the equilibrium amount of the adsorbed CO2. In a typical test run, 100 mg of a previously degassed sample at 100 °C for 12 h was used. Three similar adsorption/regeneration cycles run for each sample. The regeneration took place in an oven at 100 °C for 4 h. The accuracy in measuring pressure and temperature was 1000 Pa and 0.1 K, respectively.

Characterizations
The crystalline structure of the synthesized adsorbent was evaluated by the XRD method using a D8 ADVANCE XRD instrument (Bruker Corporation) in the 2θ range of 1-30° using Co Kα radiation (λ = 1.79 Å). The crystal size of synthesized MOF was calculated by using the Debye Scherrer equation (D = 0.9λ/β cos θ) to the reflections (2θ ~ 7°) [46]. In the mentioned equation, D, λ, β, and θ are the crystal size X-ray wavelength, full width at half-maximum (FWHM) and Bragg angle, respectively. N2 adsorption-desorption isotherms of the samples were calculated at 77 K using a Nova instrument (Quantachrome NovaWin2, Florida, USA) and also a BELSORP MINI instrument (Osaka, Japan). The specific surface area was estimated by the BET method. Field emission-scanning electron microscopy (FESEM, JEOL JSM 6490, JEOL Ltd., Tokyo, Japan) analysis was used to demonstrate the surface morphology of the fabricated samples. Besides, transmission electron microscopy (TEM, Philips CM120, Eindhoven, Netherlands) analysis was used to prove the MWCNT's existence along with the nanofibers. TGA test was performed using a Bahr STA-503 instrument at a heating rate of 10 K min −1 from 298.2 K to 853.2 K under the argon atmosphere. FTIR measurements were performed using a Jasco-Japan (model 6300, Tokyo, Japan) Fourier transform infrared spectrometer, which was configured for transmittance. Spectra were collected from the mid-infrared region (4000−400 cm −1 ) at a resolution of 4 cm −1 with the detector positioned perpendicular to the fiber direction.

Characterizations
The crystalline structure of the synthesized adsorbent was evaluated by the XRD method using a D 8 ADVANCE XRD instrument (Bruker Corporation) in the 2θ range of 1-30 • using Co Kα radiation (λ = 1.79 Å). The crystal size of synthesized MOF was calculated by using the Debye Scherrer equation (D = 0.9λ/β cos θ) to the reflections (2θ~7 • ) [46]. In the mentioned equation, D, λ, β, and θ are the crystal size X-ray wavelength, full width at half-maximum (FWHM) and Bragg angle, respectively. N 2 adsorption-desorption isotherms of the samples were calculated at 77 K using a Nova instrument (Quantachrome NovaWin2, Boynton Beach, FL, USA) and also a BELSORP MINI instrument (Osaka, Japan). The specific surface area was estimated by the BET method. Field emission-scanning electron microscopy (FESEM, JEOL JSM 6490, JEOL Ltd., Tokyo, Japan) analysis was used to demonstrate the surface morphology of the fabricated samples. Besides, transmission electron microscopy (TEM, Philips CM120, Eindhoven, The Netherlands) analysis was used to prove the MWCNT's existence along with the nanofibers. TGA test was performed using a Bahr STA-503 instrument at a heating rate of 10 K min −1 from 298.2 K to 853.2 K under the argon atmosphere. FTIR measurements were performed using a Jasco-Japan (model 6300, Tokyo, Japan) Fourier transform infrared spectrometer, which was configured for transmittance. Spectra were collected from the mid-infrared region (4000−400 cm −1 ) at a resolution of 4 cm −1 with the detector positioned perpendicular to the fiber direction.   [47][48][49]. Two main peaks around 7 and 12 degrees confirm the presence of Ni-MOF-74. The crystal size of MOF by the Scherrer method using the mentioned XRD data is calculated to be about 9 µm [50].

MOF Synthesis
The BET data analysis for synthesized Ni-MOF-74 evaluates the surface specific area of about 788 m 2 g −1 , the pore volume of 0.38 cm 3 g −1 , and the average pore diameter of 1.92 nm, which are in agreement with previous studies [51,52].     [47][48][49]. Two main peaks around 7 and 12 degrees confirm the presence of Ni-MOF-74. The crystal size of MOF by the Scherrer method using the mentioned XRD data is calculated to be about 9 µm [50]. Figure 3 shows the SEM of the synthesized MOF crystals at different magnifications. The size of polyhedral crystalline particles is in the range of 6-9 µm. Figure 4 demonstrates the XRD pattern of Ni-MOF-74 powder [47][48][49]. Two main peaks around 7 and 12 degrees confirm the presence of Ni-MOF-74. The crystal size of MOF by the Scherrer method using the mentioned XRD data is calculated to be about 9 µm [50].
The BET data analysis for synthesized Ni-MOF-74 evaluates the surface specific area of about 788 m 2 g −1 , the pore volume of 0.38 cm 3 g −1 , and the average pore diameter of 1.92 nm, which are in agreement with previous studies [51,52].   The BET data analysis for synthesized Ni-MOF-74 evaluates the surface specific area of about 788 m 2 g −1 , the pore volume of 0.38 cm 3 g −1 , and the average pore diameter of 1.92 nm, which are in agreement with previous studies [51,52].

Nanofiber Characterizations
Considering the appearance of the neat PAN, PAN/MWCNT, PAN/MOF, and PAN/ MWCNT/MOF nanofiber, it can conclude that with adding of MWCNT and MOF to the PAN solution, the PAN white color changes to dark grey, and yellow, respectively as shown in Figure 5. In comparison, the PAN/MWCNT/MOF structure shows a bright gray color, which is evidence of the following materials' existence.

Nanofiber Characterizations
Considering the appearance of the neat PAN, PAN/MWCNT, PAN/MOF, and PAN/MWCNT/MOF nanofiber, it can conclude that with adding of MWCNT and MOF to the PAN solution, the PAN white color changes to dark grey, and yellow, respectively as shown in Figure 5. In comparison, the PAN/MWCNT/MOF structure shows a bright gray color, which is evidence of the following materials' existence. As shown in Figure 6a, the electrospun neat PAN nanofibers have a uniform, smooth, and bead-free morphology. The average diameter of the produced nanofibers is 136 ± 3 nm. These structures with a nanosized scale have a higher surface to volume ratio, resulting in more MOF loading sites.
In Figure 6b, the MWCNT loaded on the PAN nanofibers are shown as darker stains along the nanofiber. Figure 6c shows that MOF particles appear as polyhedral crystals, randomly scatter, and trap in nanofibers. Figure 6d exhibits that MOF particles in MOF/MWCNT/PAN nanofibers, besides linking to fibers surface, in some parts, are impregnated along fibers and made a swollen shape along with them. However, it seems that MWCNTs are loaded inside the fibers because of their small sizes, so they could not be seen on nanofibers surfaces except minor parts, which are colligated.
The TG analysis of Ni-MOF-74, PAN nanofiber, and PAN/MWCNT/MOF nanofiber are demonstrated in Figure 7. According to TG of MOF powder, the maximum loss occurs at 345-520 °C which is about 30%. While PAN and PAN/MWCNT/MOF nanofibers have similar two-step loss curves, they occur in the ranges of 125 to 225 °C with 27% weight and 328 to 383 °C about 32%, respectively. Therefore, it seems that MOF is stable at high temperatures up to 345 °C, but loading these particles in nanofibers structure does not help thermal stability, probably because of the lower percentage of MOF relative to PAN. According to the regeneration temperature of Ni-MOF-74 adsorbent at about 100 °C, no decomposition would happen in PAN/MWCNT/MOF structure during the regeneration process. As shown in Figure 6a, the electrospun neat PAN nanofibers have a uniform, smooth, and bead-free morphology. The average diameter of the produced nanofibers is 136 ± 3 nm. These structures with a nanosized scale have a higher surface to volume ratio, resulting in more MOF loading sites.  In Figure 6b, the MWCNT loaded on the PAN nanofibers are shown as darker stains along the nanofiber. Figure 6c shows that MOF particles appear as polyhedral crystals, randomly scatter, and trap in nanofibers. Figure 6d exhibits that MOF particles in MOF/MWCNT/PAN nanofibers, besides linking to fibers surface, in some parts, are impregnated along fibers and made a swollen shape along with them. However, it seems that MWCNTs are loaded inside the fibers because of their small sizes, so they could not be seen on nanofibers surfaces except minor parts, which are colligated.
The TG analysis of Ni-MOF-74, PAN nanofiber, and PAN/MWCNT/MOF nanofiber are demonstrated in Figure 7. According to TG of MOF powder, the maximum loss occurs at 345-520 • C which is about 30%. While PAN and PAN/MWCNT/MOF nanofibers have similar two-step loss curves, they occur in the ranges of 125 to 225 • C with 27% weight and 328 to 383 • C about 32%, respectively. Therefore, it seems that MOF is stable at high temperatures up to 345 • C, but loading these particles in nanofibers structure does not help thermal stability, probably because of the lower percentage of MOF relative to PAN. According to the regeneration temperature of Ni-MOF-74 adsorbent at about 100 • C, no decomposition would happen in PAN/MWCNT/MOF structure during the regeneration process.    Table 2 shows parameter levels for each test run suggested by BBD and experimental responses which are CO 2 adsorption capacities.

Box-Behnken Design
Additionally, BBD suggests Equation (1) as a quadratic model to relate CO 2 adsorption capacity with model variables.
where q t is adsorption capacity (mmol/g). This equation indicates that MOF concentration plays the most crucial role in increasing the CO 2 adsorption capacity as shown in Table 3. In contrary, the effects of PAN concentration and MWCNT content are negligible.  Table 4 shows the analysis of the variance of this model. As it shows, p-value ≤ 0.0001 refers to suitable fitness and significance of the chosen model. Additionally, the chosen model suggests the highest possible adsorption capacity can be obtained from a solution of PAN (14.61 w/v%), MWCNT (1.43 w/w%), and MOF (11.90 w/w%), which has a CO 2 adsorption capacity of 1.68 mmol/g. Confirmatory experiments have been conducted at these conditions three times which showed an adsorption capacity of 1.65 ± 0.03 mmol/g. This is an indication of the accuracy of the proposed model.  Table 4 illustrates the statistical parameters obtained from ANOVA. The least p-value belongs to MOF concentration, which leads to the importance of this variable. Figure 8 presents the effect of each variable on the response. By comparing three bilateral plots, one can conclude that both PAN and MWCNT concentrations individually have a limited impact on CO 2 adsorption capacity. However, increasing the MOF concentration led to a significant enhancement in the response. The enhancement, as mentioned above, is a consequence of the enormous adsorption capacity of MOF particles. A decrease in CO 2 adsorption, especially in higher MOF concentrations (more than 6%), is attributed to the electrospinning solution limitation for dispersing all of the MOF completely.  Table 4 illustrates the statistical parameters obtained from ANOVA. The least p-value belongs to MOF concentration, which leads to the importance of this variable. Figure 8 presents the effect of each variable on the response. By comparing three bilateral plots, one can conclude that both PAN and MWCNT concentrations individually have a limited impact on CO2 adsorption capacity. However, increasing the MOF concentration led to a significant enhancement in the response. The enhancement, as mentioned above, is a consequence of the enormous adsorption capacity of MOF particles. A decrease in CO2 adsorption, especially in higher MOF concentrations (more than 6%), is attributed to the electrospinning solution limitation for dispersing all of the MOF completely.  Figure 9 explores the response surface plots and counter plots for investigating the relation of any two variables on the response, while the third variable is in its center. According to Figure 9a, increasing the PAN concentration (more than about 11%) in interaction with MWCNT concentration harms CO2 adsorption. Moreover, it shows that in the higher concentration of MWCNT, the lower concentration of PAN is necessary for achieving maximum adsorption. This phenomenon could be related to increasing the solution viscosity induced by MWCNT enhancing. Hence, to adjust the solution viscosity and to  Figure 9 explores the response surface plots and counter plots for investigating the relation of any two variables on the response, while the third variable is in its center. According to Figure 9a, increasing the PAN concentration (more than about 11%) in interaction with MWCNT concentration harms CO 2 adsorption. Moreover, it shows that in the higher concentration of MWCNT, the lower concentration of PAN is necessary for achieving maximum adsorption. This phenomenon could be related to increasing the solution viscosity induced by MWCNT enhancing. Hence, to adjust the solution viscosity and to obtain the highest adsorption, the PAN concentration should be reduced to get the suitable electrospinning without any MWCNT agglomeration and bead on the nanofiber morphology.
nanofiber substrate, which results in higher CO2 adsorption capacities.
The MWCNT/MOF plots in Figure 9c show that in low concentrations of MOF, increasing the MWCNT induces a slight increase in CO2 adsorption. However, at high concentrations of MOF (more than 6%), increasing the MWCNT concentration have a strong positive impact on the CO2 adsorption. It seems that in these concentrations, increasing MWCNT besides its structural improvement has a determining role in compensating solution viscosity, which misses during MOF addition.
(a) (b) (c)  In investigating PAN/MOF plots (Figure 9b), it can be seen that at low concentrations of MOF (less than 6%), increasing the PAN concentration has no significant effect on the CO 2 adsorption capacity. However, at the higher concentrations of MOF, adsorption capacity intensely depends on PAN concentration and enhances with increasing the PAN concentration. This phenomenon is presumably because of the considerable reduction in the solution viscosity due to the higher MOF concentrations. In a low concentration of MOF, viscosity reduction is less, and MOF particles load on nanofibers, completely. While at higher MOF concentrations solution viscosity intensely drops so that a lot of the MOF particles cannot properly load in nanofibers. Thus, an increase in PAN concentration helps the solution to encompass more MOF crystals and, thus, load more MOF crystals on the nanofiber substrate, which results in higher CO 2 adsorption capacities.

Characterization of MOF Secondary Growth on the PAN/MWCNT Nanofiber
The MWCNT/MOF plots in Figure 9c show that in low concentrations of MOF, increasing the MWCNT induces a slight increase in CO 2 adsorption. However, at high concentrations of MOF (more than 6%), increasing the MWCNT concentration have a strong positive impact on the CO 2 adsorption. It seems that in these concentrations, increasing MWCNT besides its structural improvement has a determining role in compensating solution viscosity, which misses during MOF addition.

Characterization of MOF Secondary Growth on the PAN/MWCNT Nanofiber
According to Figure 10, the first cycle of the secondary growth yields a uniform distribution of MOF crystals on the nanofiber mat. In the second cycle, the MOF crystals fill the remaining spaces on the nanofiber mat. They saturate the nanofiber surface, so the extra MOF crystals deposit on each other and induce the agglomerated MOF particles. pressure (0.5 < P/Po< 0.6), respectively, correspond to the relative volume of the micropores and the existence of mesopores [53]. The insignificant adsorption volume of the pure PAN-based nanofiber indicates the meager existence of pores in the nanofibers. With the MOF addition, there is a noticeable adsorption increase in the low relative pressure zone. A steeper increase happens in samples with higher loading of MOF, which is evidence of the presence of more micropores [54]. The isotherms of the MOF powder and the first and second cycles of MOFs growth resembled type I, according to IUPAC classification, indicating micropore domination in this structure. In contrast, the isotherms of the neat PAN nanofiber and PAN/MWCNT/MOF nanofiber are close to type II, which is characteristic of materials with less porosity or large cavities (macropores) [55].  Figure 11 shows the N 2 adsorption-desorption isotherm of PAN-based electrospun nanofiber loaded with different percentages of MOF/MWCNT and Ni-MOF-74 powder. The uptake of N 2 molecules at a low relative pressure range (0 < P/Po < 0.1) and moderate pressure (0.5 < P/Po < 0.6), respectively, correspond to the relative volume of the micropores and the existence of mesopores [53]. The insignificant adsorption volume of the pure PANbased nanofiber indicates the meager existence of pores in the nanofibers. With the MOF addition, there is a noticeable adsorption increase in the low relative pressure zone. A steeper increase happens in samples with higher loading of MOF, which is evidence of the presence of more micropores [54].  Table 5 summarizes the corresponding BET surface area and pore structure information of PAN-based electrospun nanofibers loaded with different percentages of MOF/MWCNT and MOF powder. The improvement of the BET surface area and pore volume is evident in the excess MOF amount addition. However, the BET of the second cycle of MOFs growth compared to the first cycle of MOFs growth shows a poor performance despite more MOF loading, which can be a consequence of a blockage of pores due to MOF crystals agglomeration (as shown in Figure 10).  Figure 12 demonstrates the XRD patterns of Ni-MOF-74 powder, PAN/MWCNT, PAN/MOF, and PAN/MWCNT/MOF nanofiber. In the pattern of Ni-MOF-74 powder, two characteristic peaks, narrow and strong, are detectable at 2θ = 7° (110) and 12° (300) [49]. PAN/MWCNT nanofiber represents a broad peak at the range of 2θ = 15-22° (200) and a weak peak at 2θ = 31° (020), which are attributed to the crystal structure of PAN [56]. Furthermore, there are two characteristic peaks of MWCNT, one weak peak is at 2θ = 42 º indicating the graphite-like structure of MWCNT and another peak is in the range of 2θ = 28-33°, which seems very difficult, due to the low loading of MWCNT and overlapping with the peak of PAN [57]. As shown in the both XRD pattern of PAN/MOF and The isotherms of the MOF powder and the first and second cycles of MOFs growth resembled type I, according to IUPAC classification, indicating micropore domination in this structure. In contrast, the isotherms of the neat PAN nanofiber and PAN/MWCNT/MOF nanofiber are close to type II, which is characteristic of materials with less porosity or large cavities (macropores) [55]. Table 5 summarizes the corresponding BET surface area and pore structure information of PAN-based electrospun nanofibers loaded with different percentages of MOF/MWCNT and MOF powder. The improvement of the BET surface area and pore volume is evident in the excess MOF amount addition. However, the BET of the second cycle of MOFs growth compared to the first cycle of MOFs growth shows a poor performance despite more MOF loading, which can be a consequence of a blockage of pores due to MOF crystals agglomeration (as shown in Figure 10).  Figure 12 demonstrates the XRD patterns of Ni-MOF-74 powder, PAN/MWCNT, PAN/MOF, and PAN/MWCNT/MOF nanofiber. In the pattern of Ni-MOF-74 powder, two characteristic peaks, narrow and strong, are detectable at 2θ = 7 • (110) and 12 • (300) [49]. PAN/MWCNT nanofiber represents a broad peak at the range of 2θ = 15-22 • (200) and a weak peak at 2θ = 31 • (020), which are attributed to the crystal structure of PAN [56]. Furthermore, there are two characteristic peaks of MWCNT, one weak peak is at 2θ = 42 • indicating the graphite-like structure of MWCNT and another peak is in the range of 2θ = 28-33 • , which seems very difficult, due to the low loading of MWCNT and overlapping with the peak of PAN [57]. As shown in the both XRD pattern of PAN/MOF and PAN/MWCNT/MOF, the mentioned characteristic peaks of Ni-MOF-74 and MWCNT remained constant and in agreement with the structure of Ni-MOF-74. It is important to note that, all samples were successfully prepared without any damage to the crystal structure of the MOF. Such observations have been reported by another researcher [58].   Figure 13. In the Ni-MOF-74 spectrum, the characteristic peaks at around 1558, 1419 cm −1 are ascribed to stretching vibrations of symmetric and asymmetric carboxylate groups (-COO-) [59]. Moreover, the peaks are located at 920, 850, and 3415 cm −1 , which are attributed to stretching vibration of Ni-O, C-H, and -OH bonds, respectively [58]. The MWCNT spectrum represents the stretching vibration of C-O, C=O carboxylic acid group, CH2, and -OH bonds at 1070, 1729, 2914, and 3450 cm −1 , respectively [60]. Additionally, Figure 13 depicts the PAN spectrums in presence of Ni-MOF-74 and MWCNT. As shown in the mentioned samples, there are characteristic peaks of PAN at around 3460, 2231, 1459, 1738, 1221, and 1080 cm −1 , which can be attributed to the stretching vibration of the -OH, C≡N, CH2, C=O, C-O bonds from ester groups and the C-O bond from the carboxyl groups, respectively [57]. It is worth mentioning that, when comparing pure PAN nanofiber with MWCNT and Ni-MOF-74 incorporated PAN nanofibers, newly several absorption peaks and changes in the location of characteristic peaks were observed, which is due to the interaction between MWCNT and Ni-MOF-74 with the PAN nanofiber [57,58].  Figure 13. In the Ni-MOF-74 spectrum, the characteristic peaks at around 1558, 1419 cm −1 are ascribed to stretching vibrations of symmetric and asymmetric carboxylate groups (-COO-) [59]. Moreover, the peaks are located at 920, 850, and 3415 cm −1 , which are attributed to stretching vibration of Ni-O, C-H, and -OH bonds, respectively [58]. The MWCNT spectrum represents the stretching vibration of C-O, C=O carboxylic acid group, CH 2 , and -OH bonds at 1070, 1729, 2914, and 3450 cm −1 , respectively [60]. Additionally, Figure 13 depicts the PAN spectrums in presence of Ni-MOF-74 and MWCNT. As shown in the mentioned samples, there are characteristic peaks of PAN at around 3460, 2231, 1459, 1738, 1221, and 1080 cm −1 , which can be attributed to the stretching vibration of the -OH, C≡N, CH 2 , C=O, C-O bonds from ester groups and the C-O bond from the carboxyl groups, respectively [57]. It is worth mentioning that, when comparing pure PAN nanofiber with MWCNT and Ni-MOF-74 incorporated PAN nanofibers, newly several absorption peaks and changes in the location of characteristic peaks were observed, which is due to the interaction between MWCNT and Ni-MOF-74 with the PAN nanofiber [57,58].  Figure 14 presents the CO2 adsorption capacity at 25 °C under 7 bar pressure for PAN nanofiber, PAN/MWCNT/MOF nanofiber, the first cycle of growth MOF, the second cycle of growth MOF, and MOF powder. The PAN nanofiber displays neglectable CO2 uptake, whereas the MOF loaded on PAN nanofibers exhibit the CO2 adsorption capacity comparable to that of MOF powder at the same pressure. The CO2 adsorption capacity of the first and second cycles of the MOFs growth at 25 °C and 7 bar, respectively are 2.83 and 4.35 mmol/g, which are equivalent to 35% and 53% of the adsorption capacity of the MOF powder (8.1 mmol/g). However, their MOF mass loading in the nanofiber structure is about 43 wt.% and 65 wt.%, respectively (Table 5). So, the capacity adsorption was lower than the excepted one, which can be related to MOFs agglomeration and ultimately pores blockage, as confirmed by the BET data (Table 5). For industrial applications, because of disadvantages such as higher pressure drop and high operation costs, the powder form of synthesized MOF must be shaped into common pellets or loaded in new structures, such as monoliths, foams, and nanofibers. In other similar studies, CO2 adsorption capacity of MOF pellet was estimated at about 70% CO2 adsorption capacity of powder [61], while as shown in this investigation, the CO2 adsorption capacity of the second cycle of the MOFs growth was about 20% less than pellet form.

CO2 Adsorption Measurements
On the other hand, to achieve a high-performance cyclic adsorption process, the indexes such as recovery, purity, productivity, and total energy of the cyclic adsorption process must be comprehensive considered [62], which are determined by the structural parameters, such as the mass and heat transfer coefficients, mass transfer zone, pressure drop, and adsorption capacity [3]. Therefore, the effect of structure on the following parameters, other than adsorption capacity, should be further studied. Thus, it seems that  Figure 14 presents the CO 2 adsorption capacity at 25 • C under 7 bar pressure for PAN nanofiber, PAN/MWCNT/MOF nanofiber, the first cycle of growth MOF, the second cycle of growth MOF, and MOF powder. The PAN nanofiber displays neglectable CO 2 uptake, whereas the MOF loaded on PAN nanofibers exhibit the CO 2 adsorption capacity comparable to that of MOF powder at the same pressure. The CO 2 adsorption capacity of the first and second cycles of the MOFs growth at 25 • C and 7 bar, respectively are 2.83 and 4.35 mmol/g, which are equivalent to 35% and 53% of the adsorption capacity of the MOF powder (8.1 mmol/g). However, their MOF mass loading in the nanofiber structure is about 43 wt.% and 65 wt.%, respectively (Table 5). So, the capacity adsorption was lower than the excepted one, which can be related to MOFs agglomeration and ultimately pores blockage, as confirmed by the BET data (Table 5). For industrial applications, because of disadvantages such as higher pressure drop and high operation costs, the powder form of synthesized MOF must be shaped into common pellets or loaded in new structures, such as monoliths, foams, and nanofibers. In other similar studies, CO 2 adsorption capacity of MOF pellet was estimated at about 70% CO 2 adsorption capacity of powder [61], while as shown in this investigation, the CO 2 adsorption capacity of the second cycle of the MOFs growth was about 20% less than pellet form. the nanofiber structure, other than lower adsorption capacity, because of special structural properties compared to the common structure (pellet), can potentially be a promising structure that should be considered further.

CO 2 Adsorption Measurements
The cyclic stability of CO2 adsorption capacities for the first and second cycles of MOFs growth and MOF powder is presented in Figure 15. As can be observed, the samples held their adsorption capacity, and no considerable adsorption loss has occurred after twenty cycles.

Conclusion
PAN electrospun nanofibers loaded with MOF/MWCNT were fabricated and applied as the CO2 adsorbent. According to the BBD model, the maximum CO2 adsorption was about 1.68 mmol/g, at 25 °C and 7 bar, including 14.61 w/v%, 1.43 w/w%, and 11.90 w/w% for PAN, MWCNT, and MOF, respectively. The CO2 adsorption in the experimental On the other hand, to achieve a high-performance cyclic adsorption process, the indexes such as recovery, purity, productivity, and total energy of the cyclic adsorption process must be comprehensive considered [62], which are determined by the structural parameters, such as the mass and heat transfer coefficients, mass transfer zone, pressure drop, and adsorption capacity [3]. Therefore, the effect of structure on the following parameters, other than adsorption capacity, should be further studied. Thus, it seems that the nanofiber structure, other than lower adsorption capacity, because of special structural properties compared to the common structure (pellet), can potentially be a promising structure that should be considered further.
The cyclic stability of CO 2 adsorption capacities for the first and second cycles of MOFs growth and MOF powder is presented in Figure 15. As can be observed, the samples held their adsorption capacity, and no considerable adsorption loss has occurred after twenty cycles. the nanofiber structure, other than lower adsorption capacity, because of special structural properties compared to the common structure (pellet), can potentially be a promising structure that should be considered further. The cyclic stability of CO2 adsorption capacities for the first and second cycles of MOFs growth and MOF powder is presented in Figure 15. As can be observed, the samples held their adsorption capacity, and no considerable adsorption loss has occurred after twenty cycles.

Conclusion
PAN electrospun nanofibers loaded with MOF/MWCNT were fabricated and applied as the CO2 adsorbent. According to the BBD model, the maximum CO2 adsorption was about 1.68 mmol/g, at 25 °C and 7 bar, including 14.61 w/v%, 1.43 w/w%, and 11.90

Conclusions
PAN electrospun nanofibers loaded with MOF/MWCNT were fabricated and applied as the CO 2 adsorbent. According to the BBD model, the maximum CO 2 adsorption was about 1.68 mmol/g, at 25 • C and 7 bar, including 14.61 w/v%, 1.43 w/w%, and 11.90 w/w% for PAN, MWCNT, and MOF, respectively. The CO 2 adsorption in the experimental setup was about 1.65 mmol/g, which is about 20% of the adsorption capacity of the MOF powder. The secondary growth strategy is applied mainly to increase amount of MOF loading, and consequently increase the adsorption capacity. The results of the first and second cycles of the MOF secondary growth displayed 35% and 53% of the adsorption capacity of the MOF powder, respectively, which in comparison to the common form of adsorbent (pellet) is still 20% less. However, other than the adsorption capacity, there are important structural parameters (the mass and heat transfer coefficients, mass transfer zone, pressure drop, etc.) that affect the cyclic adsorption process and, consequently, final performance. Therefore, besides adsorption capacity, the structural properties play a determining role, and to obtain a suitable adsorbent, a trade-off between both of them should be considered. Hence, because of advantages such as high porosity, high surface area, and excellent mechanical properties, the nanofiber structure can potentially be a promising structure that should be given more consideration.

Data Availability Statement:
The data used to support the findings of this study are included within the article.