Waste organic dye removal using MOF-based electrospun nanofibers of high amine density

) in wastewater treatment is tempered by their polycrystalline, powdery state, and challenges associated with their deployment. In the case of UiO-66, one of the most stable and widely-used MOFs, a low tendency for removing some organic contaminants has been observed on top of the mentioned issues. To address these challenges, herein, we take two complementary steps, i.e., hybridization of UiO-66 with organic nanodiamond (ND) followed by the integration of the hybrid nanoparticles in electrospun polymeric nanofibers based on chitosan/polyvinyl alcohol (PVA). We present the electrospinning of polymer/ MOFs as a promising technique to fabricate highly efficient adsorbents for water remediation. We use the electrospun chitosan/PVA nanofibers (ECPN) as a versatile host for MOF nanoparticles that remove cationic methylene blue and anionic Congo red dyes. Four nanofiber composites containing thermally oxidized nano-diamond (TOND), ND, UiO-66, and TOND@UiO-66 are utilized to unravel the effect of nanoparticles type and loading on dye adsorption capacity. It is shown that incorporation of a small loading of nanoparticles in ECPN significantly enhaces the maximum dye adsorption capacity. More importantly, the rationally engineered hybrid TOND@UiO-66 nanoparticles exhibit the best performance in dye adsorption; for instance, an 80 % increase in maximum dye adsorption capacity, from 769 to 1429 mg/g, is recorded for ECPN loaded with TOND@UiO-66 compared to the unfilled ECPN. On top of that, the designed adsorbent showed appreciable regeneration ability after 6 adsorption – desorption cycles. All in all, this study offers a new generation of engineered advanced materials to remove emerging contaminants from water streams.

While considerable efforts have been devoted to scale-up, chemistry, and application of MOFs at the lab scale level [10,11], exploratory initiatives must be taken to transcend MOFs into practical wastewater treatment.Processability issues and control of MOF superstructure have become critical rate-limiting steps, hindering real-world applications of MOFs [12][13][14], especially their prospective use for wastewater treatment.Shaping MOFs into macrostructures by accommodating them into electrospun nanofibers is deemed a game-changer approach in order to bring powdery MOFs closer to the industrial applications level [15].
Nanofabrication via electrospinning has had a tremendous impact over the past three decades due to its versatility, simplicity, and affordability [16].Notably, numerous enterprises have invested in electrospinning technology and thereby have pushed the nanofibers' yearly production capacity to the ton-scale [17].Placing the atomiclevel functional MOF materials into nanofibers with a high surface-tovolume ratio not only can address the processing obstacles but also can retain the MOFs' active sites for interacting with pollutants in aqueous media [15].For instance, mats comprising ZIF-8/polyvinyl alcohol (PVA) [18] and ZIF-67/polyacrylonitrile (PAN) [19] showed maximum adsorption capacities of 160 and 849 mg/g toward Congo red dye (CR), respectively.
To design novel MOF hybrids, where the coupling agent is immobilized inside the MOF cages, the existence of binding sites for successful synthesis and anchoring of MOFs is vital.Hence, in the present work, thermally oxidized nanodiamond (TOND) was selected as the substrate to produce MOF-based nanocomposite mainly because of its relatively high surface area, good chemical stability, high affinity toward specific organic dyes, non-toxicity, and numerous surface bridging sites due to the existence of large oxygen-containing functional groups on its surface [31][32][33][34].To unravel the benefits of the hybridization of MOFs with other nanomaterials and electrospinning technology to fabricate nanoparticlefilled nanofiber adsorbents, ND, TOND, UiO-66, and their nanocomposites (TOND@UiO-66) were prepared and incorporated into nanofibers formed with a mixture of chitosan and polyvinyl alcohol (CS/ PVA) (Scheme 1).The prepared nanofiber composites were then crosslinked using glutaraldehyde (GA) and tested for cationic MB and anionic CR removal from wastewater.In addition, the systematic adsorbent performance, in terms of adsorption kinetics, isotherms, selectivity, ionic strength, the effect of pH, and reusability, was investigated to verify the applicability of the MOF nanofibers in real wastewater treatment applications.

Preparation of TOND
The oxidation of ND particles was performed to add carboxylic acid groups on their surface, to serve as nuclei sites for UiO-66 growth.Accordingly, 150 mg of ND was heated at 425 • C for 270 min under an air atmosphere [35,36].The oxidized nanoparticles were stored in a desiccator to prevent moisture adsorption from the air.

Synthesis of TOND@UiO-66 nanocomposite
A 'bottle around ship' (BAS) concept was followed to synthesize TOND@UiO-66 nanocomposite [20,26,37].Briefly, 60 mg of TOND nanoparticles were dispersed in 20 mL DMF by sonication for 1 h.Thereafter, 530 mg of ZrCl 4 was added to the TOND suspension, sonicated for 1 h, and stirred at room temperature for 12 h.Meantime, 380 mg of BDC was dissolved in 20 mL DMF and stirred to prepare a transparent solution.Then, the BDC solution was added to the TOND Scheme 1. Schematic illustration of the fabrication of the electrospun TOND@UiO-66 nanofibers.
F. Ahmadijokani et al. suspension, sonicated for 1 h, and the mixture was transferred into a Teflon autoclave and heated at 120 • C for 24 h.The prepared gray TOND@UiO-66 nanocomposite was separated by filtration, washed with DMF, and further with ethanol, and then solvent-exchanged by Soxhlet extraction in ethanol for 5 days.Finally, the synthesized nanocomposite was dried at 120 • C under the vacuum for 12 h.The same procedure was used to synthesize pure UiO-66 nanoparticles without adding TOND to the DMF solution [38].

Preparation of electrospun nanofiber adsorbents
Due to the nature of UiO-66 MOFs, a MOF-first strategy [16] was utilized to prepare nanofiber composites.Accordingly, 7.5 mg of TOND@UiO-66 was mixed with 7.5 mL of distilled water using a bath sonicator for 30 min, followed by stirring at room temperature for 12 h.In the next step, 375 mg of PVA powder was added to the suspension, sonicated for 30 min, and finally stirred at room temperature for 12 h.In the meantime, 125 mg of CS powder was dissolved in 2.5 mL of 2 % v/v HAc solution, and stirring was continued at room temperature for 12 h.In order to prepare a homogeneous suspension for electrospinning, the CS solution was added to the nanoparticle suspensions plus 30 min of sonication and 12 h of stirring at room temperature.
High voltage electrospinning (22-25 kV-syringe tip to collector distance: 15 cm), followed by drying nanofibers (under vacuum for a day) and cross-linking with GA was performed to fabricate the final nanofiber composite for dye removal.To evaluate the performance of the TOND@UiO-66 nanofiber compared to bare UiO-66, ND, and TOND, the same procedure was employed to fabricate their nanofiber composites.The following terminology (Table 1) is used throughout the manuscript for samples with the parent composition of 125 mg CS, 375 mg PVA, 9.8 mL water, and 0.2 mL HAc.

Dye removal experiments
To conduct the adsorption study, measured amounts of each adsorbent were added to 50 mL of dye solutions (20-500 mg/L) and shaken in a dark place for 180 min at room temperature.Afterward, the concentration of dyes was measured by ultraviolet-visible (UV-Vis) spectroscopy.Finally, the removal percentage, adsorption capacity (q t ) at specific intervals (time t), as well as adsorption capacity (q e ) at equilibrium, were calculated using Eqs.S1-S3, respectively [35].In order to fully exploit the adsorbents, they were dried at 80 • C for 12 h under a vacuum before the experiments.
Two sharp peaks observed at 1502 and 1578 cm −1 are attributed to the stretching vibration of C--C bonds of BDC molecules [38].The observation of these characteristic peaks suggests that the signal associated with UiO-66 nanoparticles originated from the presence of BDC ligands in their structures [42,43].The peaks at 661 and 765 cm −1 come from the asymmetric and symmetric stretching vibration of O-Zr-O bonds, confirming the formation of coordination bonds between Zr 4+ ions as the coordination center and the BDC molecules as the bridging ligands [44,45].The emergence of characteristic peaks of ND and UiO-66 in the spectrum of TOND@UiO-66 confirms the successful synthesis of the nanocomposite.It should be noted that the absorption peak position belonging to the stretching vibration of C--O bonds slightly shifted to lower wavenumbers in TOND@UiO-66, mainly due to the coordination of the Zr center of UiO-66 with the carboxylic acid functional groups of TOND.[26].A similar observation was reported by Han et al. [46] in the case of the MIL-68 MOF and CNT (CNT@MIL-68) nanocomposites.

XRD analysis
XRD patterns of the synthesized adsorbent nanoparticles were obtained and shown in Fig. 1b to study the crystallinity of the parent materials and their nanocomposites.The XRD patterns of ND unfold the presence of all the characteristic diffraction peaks of ND (2θ = 16.2 • , 21.4 • , and 43.5 • ), consistent with previous reports [31,47].TOND demonstrated similar XRD patterns to ND with minor peak position and intensity changes.Accordingly, the peak at 2θ = 21.4 • , related to a graphite-type lattice, slightly shifted to lower degrees and its intensity decreased after thermal oxidation, confirming the reduction of graphite layer number on the surface of ND particles.Three sharp diffraction peaks at 2θ = 7.3 • , 8.5 • , and 25.8 • in the XRD patterns of UiO-66 and TOND@UiO-66 belong to the (1 1 1), (2 0 0), and (6 0 0) crystal planes, respectively [43].The good agreement of the XRD patterns with simulated UiO-66 and the literature confirms the successful synthesis of UiO-66 specimens in our study [38,42,43].The XRD observation revealed that the presence of TOND in TOND@UiO-66 did not interrupt the formation of the UiO-66 crystalline structure.

N 2 adsorption-desorption analysis
N 2 adsorption-desorption isotherms assessed the porosity and texture of nanoparticles at 77 K (Fig. 1c).The isotherms of the TOND are similar to Type IV, demonstrating mesoporous (pore diameter > 2 nm) structures [33].Pure UiO-66 nanoparticles showed Type I isotherms with a predominantly microporous structure [48].Fig. 1c demonstrates that the TOND@UiO-66 isotherms are relatively a combination of Type I and Type IV isotherms, showing a sharp increase in N 2 adsorption at sufficiently low relative pressures, continuous growth to high pressures, and hysteresis at high pressures.Moreover, the N 2 gas uptake of TOND@UiO-66 increased continuously with the relative pressure from 0.1 to 0.3, confirming the presence of larger micropores, mainly due to the existence of the TOND in the structure.Table 2 summarizes the pore textural properties of nanoparticles.TOND and UiO-66 nanoparticles have Brunauer-Emmett-Teller (BET) surface areas of 331 and 1050 m 2 / g, respectively, which is in line with previous reports for these materials [38,47].The BET surface area of TOND@UiO-66 experienced a 20 % decrease compared to pure UiO-66, mainly due to the higher density of TOND and its lower porosity.Table 2 shows that the TOND@UiO-66 nanocomposite has a slightly larger pore diameter and pore volume than pure UiO-66, which may be caused by the formation of new pores between UiO-66 crystals and TOND particles.The larger pores in the structure of TOND@UiO-66 nanocomposite facilitate mass diffusion from bulk to intra-particles, which are beneficial to the adsorption of dye molecules [48].

Thermogravimetric analysis
The thermal stability of nanoparticles and their nanocomposite (TOND@UiO-66) was investigated with thermogravimetric analysis (TGA) and displayed in Fig. 1d.Only a slight reduction in the TOND weight was observed at temperatures higher than 500 • C, which might be attributed to the degradation of various oxygen-containing functional groups on the surface of TOND particles [34,49].The somewhat higher thermal stability of TOND in comparison with the as-received ND particles may be due to a reduction in the number of graphite layers on the surface of TOND particles, further confirming the conversion of different functional groups on the surface of the ND particles to carboxylic acid groups during thermal oxidation [47,50].UiO-66 and TOND@UiO-66 showed three weight loss steps by increasing the temperature up to 800 • C. Removing guest water or physically adsorbed water could account for weight loss in the 60-120 • C range [51,52].The second weight loss in the range of 120-250 • C could be attributed to the removal of solvent molecules.The following weight reduction step for UiO-66 and TOND@UiO-66 was observed at temperatures higher than 490 • C, possibly due to the collapse of their framework [25].According to our results, pristine UiO-66 and TOND@UiO-66 exhibited similar thermal stabilities, demonstrating that the thermal stability of UiO-66 was not significantly affected by TOND.However, the final ash content of the nanocomposite is more than that of pristine UiO-66, primarily due to the high thermal stability of TOND in the nanocomposite structure.

SEM analysis
The particle size and morphology of the nanoparticles were investigated by SEM images and are displayed in Fig. 2. Accordingly, both pristine and thermally oxidized ND are nearly spherical, with particle sizes smaller than 40 nm.UiO-66 possesses a cubic shape with a particle size smaller than 200 nm, in line with literature reports [47,53,54].The TOND@UiO-66 morphology is reasonably similar to UiO-66, with minor changes in particle size irregularity.Accordingly, the particle sizes decreased in TOND@UiO-66, and the crystal became more irregular.Coordinating the Zr 4+ center of the UiO-66 framework with the oxygencontaining functional groups (especially carboxylic acid groups) of TOND inhibits MOF nanocrystal aggregation, thereby enhancing their dispersion [25,53].

FTIR analysis
FTIR analysis assessed the interfacial interactions between nanoparticles and matrix in the nanofiber composites.Accordingly, as shown in Fig. 3a, cross-linked CS/PVA (ECPN) has the characteristic peaks of CS and PVA.The broad absorption bands in the range of 3150-3600 cm −1 could originate from stretching vibration of -OH and -NH groups and the formation of intramolecular or intermolecular hydrogen bonds of -OH groups of PVA [39,55].CS and PVA showed two absorption bands between 2870 and 2980 cm −1 , probably due to their aliphatic  -C -H stretching vibrations [56].CS amide groups may be responsible for the strong absorption bands observed in the range of 1500-1650 cm −1 [35,39].The strong absorption band at 1380 cm −1 could correspond to the symmetrical deformation mode of -C-H bonds of both CS and PVA.The sharp absorption bands observed in 1000-1100 cm −1 could be assigned to the stretching vibration of -C-O bonds in both polymers [57].Nanofibers composites exhibited all characteristic peaks of their constituents, including polymers and nanoparticles.However, the broad absorption band in the range of 3150-3600 cm −1 , related to the stretching vibration of hydrogen-bonded -N -H and -O -H groups, shifts to lower wavenumbers.This red shift confirms the formation of hydrogen bonds between the -N -H and -O -H groups of the polymer matrix and the polar functional groups of nanoparticles [58].Delavar and Shojaei observed comparable results in their study on thermally oxidized and untreated ND particles [39].

XRD analysis
The possible effects of the incorporated nanoparticles on the crystallinity of nanofibers composed of semicrystalline CS and PVA were investigated using XRD.Fig. 3b displays the XRD patterns of pure CS/ PVA nanofibers and nanofiber composites.The cross-linked CS/PVA matrix exhibits a broad diffraction peak in the range of 15-30 • due to its amorphous structure, in agreement with the literature [58,59].Nanocomposite nanofibers containing different nanoparticles show XRD patterns nearly similar to pure nanofibers, indicating that the nanoparticles did not significantly affect the crystalline structure of the CS/ PVA matrix.It is worth mentioning that the diffraction patterns of each nanoparticle appeared in the same position as their pure crystals but with decreased peak intensities compared to those of pure nanoparticles.This observation also endorses that the crystalline structure of nanoparticles remains intact under the nanofiber fabrication procedure, i.e., electrospinning.

SEM analysis
The surface morphology, location of nanoparticles, and their effect on the nanofiber diameter and configuration were investigated by SEM analysis and are depicted in Fig. 4. It can be inferred that ECPNs have smooth and uniform surface morphology, with average diameter sizes smaller than 100 nm and without forming bead defects.Including nanoparticles into the nanofiber composites did not disturb the nanofiber integrity and surface morphology.The SEM images showed that the nanoparticles were fully embedded inside the nanofibers, likely due to their interactions with the CS/PVA matrix [16].Furthermore, energy dispersive X-ray spectroscopy (EDS) images (Fig. S1) of ECPN/T@U nanofibers show that these samples are composed of carbon (70 %), oxygen (29 %), and zirconium (1 %) elements.

Adsorption kinetics
Fig. 5 displays the removal efficiency of specimens toward cationic MB and anionic CR dyes.Nanoparticles exhibited fast adsorption kinetics (Fig. 5a, c) in the first ≈ 20 min, followed by a gradual improvement in removal performance and ultimately reaching equilibrium after 1 hour.Compared to nanoparticles, nanofiber composites feature a slower removal rate (Fig. 5 c, d).
The nanofiber composites demonstrated a fast kinetic stage at the beginning (60 min) followed by a slow phase, reaching equilibrium within 180 min.It can be inferred from Fig. 5a that ND showed the lowest adsorption removal toward cationic MB dye, while TOND showed the highest removal efficiency toward this dye.Possibly, this observation can be explained by the presence of more carboxylic acid  functional groups and, consequently, more negative surface charges of TOND particles.That is, electrostatic attraction between cationic MB molecules and the negatively-charged TOND will increase, resulting in higher adsorption of MB [31,60].
As a result of the hybridization of UiO-66 with TOND, MB's removal efficiency increased, primarily due to enhancing the electrostatic attraction between adsorbate-adsorbents as a result of increasing the negative surface charge of UiO-66 nanoparticles (Table 2).Accordingly,   the removal efficiencies of the nanoparticles toward MB dye are in the following order: ND (18.0 %) < UiO-66 (83.8 %) < TOND@UiO-66 (90.9 %) < TOND (99.6 %).However, a completely different trend in the removal performance of nanoparticles toward CR was observed, as shown in Fig. 5b.Accordingly, ND showed the best performance by removing 99.1 % of anionic CR dye, followed by UiO-66 (68.3 %), TOND@UiO-66 (59.9 %), and TOND (22.9 %).This observation can be justified by the different charges of the dye and their chemical structure.
The BET surface area for ND is about 291 m 2 /g (Table 2), which was not too large to account for the high removal performance toward anionic CR, so looking for other contributing factors is imperative.In addition, the negative surface charge of ND could trigger the charge repulsion between ND and CR molecules.It is, therefore, possible that other adsorption mechanisms, such as hydrogen bonding or hydrophobic-hydrophobic interactions, are involved.A comparison between the FTIR spectrum of adsorbents before and after adsorption could provide valuable insight into the adsorption mechanism.As depicted in Fig. S2, the intensities of absorption bands at 1541 and 1683 cm −1 increased after the adsorption of CR onto all adsorbent nanoparticles due to the presence of azo (-N--N-) groups in the structure of CR molecules.Two -SO 3 − groups in the chemical structure of the CR molecule and their oxygen atoms could act as a hydrogen-bonding acceptor and create intramolecular hydrogen bonds with the polar groups of the adsorbent nanoparticles.As a result of this interaction, the positions of characteristic bands of this group (at 1401 and 11120 cm −1 ) slightly shifted to the lower wavenumbers [40].The formation of hydrogen bonds between CR molecules and the synthesized adsorbent nanoparticles was further confirmed by increasing the intensity of the broad absorption bands at 3150-3600 cm −1 due to the formation of hydrogen bonds between the -NH 2 and -SO 3 groups of CR molecules and the polar groups of the adsorbent nanoparticles.Other researchers also observed similar results when removing MO and acid orange 7 from water by ND [40] and modified CNTs [61].
For pure MB, characteristic absorption bands at about 1084, 1316, and 1586 can be ascribed to -CH 3 , -C -N, and -C--S bonds, respectively.The FTIR spectrum of adsorbents showed a slight increase in the intensity of bands in the range of 2850-2950 cm −1 after the adsorption of MB, perhaps caused by stretching vibrations in -C-H bonds [37].The aromatic rings in the structure of both dyes promise their π-π interaction with the π electron-rich regions of the nanoparticles, which can be further seen in the reduction of -C--C stretching vibration intensity of nanoparticles [61].Hence, the nanoparticles are capable of adsorbing anionic CR molecules via hydrogen bonds and electrostatic interactions while adsorbing cationic MB molecules via π-π and electrostatic interactions.
Fig. 5c shows that all nanocomposite nanofibers had relatively low removal efficiencies for cationic MB dye, owing to the protonation of the CS/PVA matrix [59].Accordingly, the removal efficiency of nanofibers toward cationic MB dye is in the order of ECPN/T@U (10.5 %) < ECPN (11.6 %) < ECPN/U (11.7 %) < ECPN/N (14.1 %) < ECPN/T (20.8 %), significantly lower than those observed for their parent nanoparticles (Fig. 5a).There is a possibility that the relatively higher removal efficiency of ECPN/T toward cationic MB dye might be due to the high negative charge of the TOND particles, which reduces the positive charge of the CS/PVA matrix.
As shown in Fig. 5d, adding different nanoparticles significantly improved the removal efficiency of pure nanofibers toward anionic CR dye.Therefore, the removal efficiencies of nanofibers towards anionic CR dye are as follows: ECPN (39.4 %) < ECPN/T (85.2 %) < ECPN/N (86.2 %) < ECPN/U (90.3 %) < ECPN/T@U (95.4 %).The enhancement in effective surface area and porosity of nanocomposite could be responsible for the latter observation.The electrostatic attraction allows CR to interact strongly with positively charged nanofibers.Therefore, electrostatic attraction and hydrogen bonding may be involved in the adsorption of anionic CR molecules by the nanofibers, as schematically illustrated in Scheme 2.
Adsorption of organic dyes on adsorbent materials is a physicochemical process that involves the transfer of the dissolved molecules from the aqueous phase to the surface of adsorbent materials.Consequently, three well-known kinetic models, including pseudo-first-order, pseudo-second-order, and intra-particle diffusion models, were employed to clarify the adsorption behavior of the synthesized adsorbent materials toward dyes [62,63].Table 3 summarizes the determination coefficients (R 2 ) and kinetic parameters determined by fitting the experimental data to the above-stated kinetic models (Figs.S3 and S4).
The R 2 values from the pseudo-second-order kinetic model are greater than 0.99 for the adsorption of both organic dyes on the synthesized adsorbent materials.Moreover, all the calculated adsorption capacities (q e, ca ) are nearly close to those measured experimentally (q e, exp ).This suggests that MB and CR dye adsorption behavior over the synthesized adsorbent materials is better predicted with a pseudosecond-order kinetic model, implying that the rate-controlling step in this adsorption process might be chemisorption [62,64].

Adsorption isotherms
The data related to the adsorption performance of nanofibers for removing MB and CR at various initial dye concentrations (20-500 mg/ L) are depicted in Fig. 6a and b, respectively.All nanofibers showed a sharp increase in adsorption capacity for removing MB before reaching equilibrium (Fig. 6a).Accordingly, the maximum equilibrium adsorption capacity of cationic MB dye over the nanofibers was as follows: ECPN (62 mg/g) < ECPN/U (212 mg/g) < ECPN/N (229 mg/g) < ECPN/T@U (244 mg/g) < ECPN/T (307 mg/g), indicating an approximately four-time increase in the adsorption capacity of ECPN after adding only 1.5 wt% of TOND.
Comparing Fig. 6a and b reveals that the maximum equilibrium adsorption capacities of anionic CR dye over nanofibers are much higher than those of cationic MB dye, possibly due to the prominent role of the dye removal mechanism by electrostatic interactions.Accordingly, the maximum equilibrium adsorption capacity of anionic CR dye over the fabricated nanofibers increased in the order of ECPN (644 mg/g) < ECPN/T (719 mg/g) < ECPN/N (722 mg/g) < ECPN/U (733 mg/g) < ECPN/T@U (1085 mg/g), indicating approximately 68 % improvement in the adsorption performance of ECPN/T@U in comparison with ECPN.
Four well-known isotherms, including Langmuir, Freundlich, Dubinin-Radushkevich, and Tempkin isotherms, were employed to get further insight into the interaction of dye molecules with nanofiber adsorption active sites.The associated data compiled in Fig. S5 (Supplementary Information file) and Table 4 demonstrates that R 2 values belonging to the Langmuir isotherm model for removing CR are among the highest values (ranging from 0.990 to 0.998), suggesting monolayer adsorption of CR molecules onto the homogeneous surface of the nanocomposite nanofibers.
The higher R 2 values of the Tempkin isotherm model for eliminating MB and the corresponding n F values higher than one suggest the favorable nature of the adsorption processes [58].Moreover, the mean energy values for both organic dyes calculated by the Dubinin-Radushkevich isotherm model are below 5000 J/mol, indicating that physical adsorption dominates the adsorption mechanism of both organic dyes onto the fabricated nanofibers [61].The maximum adsorption capacity of ECPN/T@U obtained from the Langmuir model towards CR was compared with those in the literature (Table S1).It is inferred that the maximum adsorption capacity of ECPN/T@U (1428.68 mg/g) surpasses most reported adsorbents in the literature.In addition, as illustrated in Fig. S6, ECPN/T@U showed appreciable selectivity toward anionic CR dye in the CR/MB mixture.Consequently, the designed nanofiber composite appears to be an excellent candidate for effectively and selectively removing the anionic organic dyes.

Thermodynamics of adsorption
To better understand the adsorption process of CR and MB dyes, the effect of temperature on the adsorption performance of ECPN/T@U F. Ahmadijokani et al. toward both CR and MB dyes was investigated by changing the adsorption temperature from 25 to 45 • C (details are summarized in the Supplementary Information).It can be observed from Table S2 that by increasing the temperature, the adsorption capacities of both anionic and cationic dyes onto the ECPN/T@U decreased, indicating the exothermic nature of this adsorption process.Moreover, the negative values of ΔH 0 further confirmed that the adsorption process is exothermic.The negative values of ΔG 0 indicated that the adsorption of these dyes onto the ECPN/T@U is a spontaneous process [65].However, the negative values of ΔS 0 are unfavorable for spontaneous adsorption, demonstrating that the system's randomness decreases during the adsorption of these dyes.

Effect of ionic strength and pH
In industrial wastewater, several interfering agents, such as surfactants, salts, heavy metal ions, etc., can affect the adsorption process [66]; for instance, in its simplest form, by altering the ionic strength of the medium [67].According to two opposite mechanisms, ionic strength can either accelerate or decelerate dye removal, including i) facilitating dye dissociation, which results in an improvement in dye removal efficiency, and ii) reducing the electrostatic attraction between dye molecules and adsorbent particles [35,60].Therefore, the effect of ionic strength on the removal efficiency of ECPN/T@U toward MB and CR dyes was investigated by adding different amounts of NaCl into the dye solutions.Fig. 6c displays the removal efficiency at three different concentrations of NaCl.Accordingly, the MB removal efficiency increased by increasing the NaCl concentration to the maximum removal efficiency of 26.3 % at a concentration of 0.5 mol/L.This observation might be due to MB molecule dissociation, which may enhance the dye removal efficiency of ECPN/T@U.However, it was found that adding NaCl into the CR solution reduced CR removal efficiency, mainly because of the screening of electrostatic interactions between anionic CR molecules and positively-charged surfaces of ECPN/ T@U.
The effect of pH was also investigated on the removal efficiency of ECPN/T@U.The pH of the dye solution can affect the adsorption performance by protonation and deprotonation of the CS/PVA matrix [35].It was reported that acidic solutions are more favorable for removing anionic dyes by pure CS adsorbent compared to basic solutions.However, the cross-linked CS/PVA showed good stability against highly acidic and basic solutions.Accordingly, this study explored the adsorption performance in a wide range of pH, from 3 to 13. Fig. 6d displays the removal efficiency of ECPN/T@U toward MB and CR dyes as a function of solution pH.As shown in Fig. 6d, the removal efficiency of cationic MB decreased by increasing pH from 3.2 to 6.2 and then increased continuously with a further increment of pH from 6.2 to  12.4.At low pH values (lower than the pKa value of MB ≈ 3.8 [64]), MB molecules are in neutral form and unable to interact via electrostatic interactions with the positively-charged surface of adsorbent materials.
As an ideal planar molecule, MB can interact with an adsorbent material's electron-rich regions [61].Therefore, π-π stacking interactions between MB molecules and ECPN/T@U at low pH could be responsible for the removal of the MB, schematically illustrated in Scheme 2. In light of CS's pKa of approximately 6.5, it exists in a protonated state in acidic media and deprotonated in the basic environment.Thus, in the pH range of 3.8-6.5,both MB molecules and ECPN/T@U carry positive charges.Hence, increasing the pH enhances the repulsion between adsorbate and adsorbent and will reduce the ECPN/T@U efficiency for cationic MB dye removal.
Further enhancement of pH will increase the negative charges on both TOND@UiO-66 and CS/PVA, which is desirable for removing MB via strong electrostatic attractions.ECPN/T@U exhibited a maximum removal efficiency of 99.5 % for anionic CR dye at pH 6.2, which is similar to the results obtained with CS and its derivatives for adsorptive dye removal [35,59].However, as mentioned above, increasing pH will induce a more negative charge on the surface of ECPN/T@U and impede the adsorption of anionic CR molecules due to the creation of electrostatic repulsion between adsorbate-adsorbent.As a result, the removal efficiency of ECPN/T@U toward anionic CR dye constantly decreased with increasing the solution pH from 6.2 to 12.4.

Reusability of ECPN/T@U for dye removal
The reusability of ECPN/T@U was evaluated by investigating the removal efficiency of the regenerated nanofibers over repeated adsorption-desorption cycles.Desorption and regeneration of CRloaded ECPN/T@U were achieved by washing the spent adsorbent with an alkaline ethanol solution (containing 0.5 % NaOH), whereas the MB-loaded ECPN/T@U adsorbent was washed with an acidic ethanol solution (containing 0.5 % HCl) [59].Accordingly, after each adsorption cycle, the dye-loaded ECPN/T@U was washed with water and alkaline or acidic ethanol solution to desorb the dye molecules entirely.Then, the regenerated ECPN/T@U was dried at 60 • C under vacuum for 12 h.It can be seen from Fig. 7 that the removal efficiency of ECPN/T@U toward both anionic and cationic dyes slightly decreased with increasing the number of adsorption-desorption cycles.The acceptable reusability of the spent ECPN/T@U and its relatively high removal efficiency of 90.1 % toward CR dye after 6 adsorption-desorption cycles demonstrate the potential of these adsorbent materials for removing anionic CR dye from wastewater.

Conclusion
We demonstrated a facile electrospinning method to prepare rationally desinged (UiO-66 MOFs)-loaded chitosan/polyvinyl alcohol (PVA) nanofibers for dye-contaminated water treatment.According to the prominent role of polymeric substrate in the adsorption performance of MOF nanofibers, chitosan, a functional biobased polymer with abundant amino and hydroxyl functional groups, was added to polyvinyl alcohol to incorporate active functional groups for better dye removal.The unique features of MOF-based polymeric nanofibers, such as ease of handling, straightforward separation from treated water, and regeneration ability, make them appealing compared to pure MOF powder for real-world applications.Moreover, a hybridization approach was developed by binding thermally oxidized nanodiamond (TOND) with UiO-66, named TOND@UiO-66, while keeping an eye on the final morphology and textural properties of nanofiber composites.The posttreatment, i.e., chemical crosslinking of the nanofibers with glutaraldehyde, made the substrate structurally stable for water treatment.Our results show that incorporating 1.5 wt% of TOND@UiO-66 into the pure polymeric nanofiber doubled the maximum Langmuir adsorption capacity of anionic Congo red, from 770 to 1430 mg/g.We also clarified different mechanisms responsible for removing cationic and anionic dyes from aqueous solutions at various pH and in the presence of given counter-ions.Multiple interactions between dyes and the designed adsorbents, including π-π stacking and electrostatic interactions, as well as hydrogen bonding, can develop depending on the media's pH.As a critical factor in applications, the regeneration of adsorbents showed that the designed MOF-based nanofiber composites could survive repeated adsorption-desorption cycles (just a 10 % drop in performance after 6 cycles).Given the unique feature of the MOFs and the electrospinning approach, we believe that this combination will definitely spur the growth of the nascent area of MOF polymer nanocomposites for water remediation.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

F
.Ahmadijokani et al.

Fig. 5 .
Fig. 5. Removal efficiency of the nanoparticles and their electrospun nanofibers composite toward a, c) MB and b, d) CR dyes.

F
.Ahmadijokani et al.

Scheme 2 .
Scheme 2. Plausible adsorption mechanisms as a function of pH.

Fig. 6 .
Fig. 6.Adsorption isotherms of electrospun nanofiber composites for removing a) MB and b) CR dyes; effects of c) NaCl concentration and d) pH on the removal efficiency of ECPN/T@U toward MB and CR dyes.

F
.Ahmadijokani et al.

Table 2
Textural properties of pristine materials and their composite.

Table 3
Kinetic parameters for removing MB and CR dyes by the synthesized adsorbent materials.

Table 4
Isotherm parameters for the adsorption of MB and CR onto the fabricated nanofiber composites.