TiO2-Doped Chitosan Microspheres Supported on Cellulose Acetate Fibers for Adsorption and Photocatalytic Degradation of Methyl Orange

Chitosan/cellulose acetate (CS/CA) used as a biopolymer systema, with the addition of TiO2 as photocatalyst (C-T/CA) were fabricated by alternating electrospinning/electrospraying technology. The uniform dispersion of TiO2 and its recovery after the removal of methyl orange (MO) was achieved by incorporating TiO2 in CS electrosprayed hemispheres. The effects of pH values, contact time, and the amount of TiO2 on adsorption and photocatalytic degradation for MO of the C-T/CA were investigated in detail. When TiO2 content was 3 wt %, the highest MO removal amount for fiber membranes (C-T-3/CA) reached 98% at pH value 4 and MO concentration of 40 mg/L. According to the data analysis, the pseudo-second-order kinetic and Freundlich isotherm model were well fitted to kinetic and equilibrium data of MO removal. Especially for C-T-3/CA, the fiber membrane exhibited multiple layers of adsorption. All these results indicated that adsorption caused by electrostatic interaction and photocatalytic degradation were involved in the MO removal process. This work provides a potential method for developing a novel photocatalyst with excellent catalytic activity, adsorbing capability and recycling use.


Introduction
With the quick progress of industrial technology, water pollution generated from organic dye contaminants has become an increasing global environmental concern in recent years [1][2][3][4]. Organic dyes are mainly generated from manufacturing, such as textile finishing, plastics, paper, cosmetics, pharmaceuticals and food processing. It can cause serious problems for both the environment and to human beings [5][6][7][8][9][10]. Up until now, several methods such as chemical oxidation, coagulation, adsorption, biodegradation, photocatalysis, and reverse osmosis, have been developed to treat dye contaminated wastewater [11][12][13][14]. Particularly, photocatalysis is considered a "green" treatment method which uses solar energy effectively for the degradation of organic pollutants [15][16][17][18][19]. Adsorption is also considered a valid method for the removal of dye in aqueous solutions with high efficiency, low price and easy operation [20][21][22].
At present, researchers have found that the combination of adsorption and photocatalysis had excellent potential applications for dye treatment [23]. TiO 2 is a type of common semiconductor photocatalyst on account of its prominent characteristics of chemical resistance, excellent catalytic activity, mechanical robustness, and low cost [24][25][26][27][28]. While the low dispersibility, inconvenient recovery, low adsorption performance of TiO 2 limit its use in photocatalysis applications [29]. Absorbent

Fabrication of C-T/CA Fiber Membranes
The fabrication processes including (I) electrospinning and (II) electrospraying are shown in Figure 1. The solution was loaded into a 5 mL plastic syringe. The electrospinning conditions included the applied voltage of 20 kV, the 15 cm distance between the tip of the needle and the collector, and the flow rate of 1.0 mL/h. The condition for electrospraying was 18 kV voltage with a distance of 15 cm between the tip of the needle and the collector, and the flow rate was 0.3 mL/h. Then the membrane was dried at 25 • C for 24 h, and peeled off aluminum foil for further study. The fabrication processes including (I) electrospinning and (II) electrospraying are shown in Figure 1. The solution was loaded into a 5 mL plastic syringe. The electrospinning conditions included the applied voltage of 20 kV, the 15 cm distance between the tip of the needle and the collector, and the flow rate of 1.0 mL/h. The condition for electrospraying was 18 kV voltage with a distance of 15 cm between the tip of the needle and the collector, and the flow rate was 0.3 mL/h. Then the membrane was dried at 25 °C for 24 h, and peeled off aluminum foil for further study.

Characterizations
The viscosity of precursor solutions was measured by a viscosimeter (NDJ-1, Yutong, Shanghai, China). The electrical conductivity of the solutions was measured using a conductometer (DDS-11A, Shengci, Shanghai, China). All measurements were taken 3 times at room temperature. The morphology and structure of membranes were characterized by SEM (JSM-6700F, JEOL, Tokyo, Japan). All samples were sputtered with platinum under vacuum before assessment. The average microsphere diameters were measured by Image J software. Surface area, pore volume and pore size were characterized by N2 adsorption and desorption using a Brunauer-Emmett-Teller analyzer (BET, Autosorb-iQ2, Quantachrome Instruments, Shanghai, China). The structures of the composite fibers were analyzed by FTIR spectroscopy (FTIR-4100, Jasco, Shanghai, China), within the wavenumber ranging from 400 to 4000 cm −1 and a wide angle X-ray diffractometer (XRD, D/Max, Rigaku, Beijing, China), under mode of 40 kV and 50 mA, λ = 1.5406 Å radiation using copper-K alpha. The samples were scanned in the 2θ range from 5° to 60° with a scanning rate of 30 min −1 . The chemical composition of membrane surfaces were analyzed by XPS (ESCALab220i-XL, VG Scientific, Waltham, MA, USA) with Al Kα X-ray source, Ni-filtered radiation, 40 kV of tube electric pressure and 30 mA of tube electric current.

Investigations on MO Adsorption and Photocatalysis Study
Organic dye MO was used as the adsorbate for adsorption and photocatalysis studies of fiber mats. 20 mg of fiber mats were immersed in 40 mg/L MO aqueous solution (20 mL). The solution with fiber mats was stirred in dark conditions for 1 h to reach the adsorption equilibrium. After that, the mix solution was irradiated for 30 min in a photodegradation reactor under UV light (wavelength of 365 nm) by mercury lamp (25 W). The mixed solution (6 mL) was taken out every 5 min and centrifuged. The centrifugation was used to separate the solid particles in the suspension from the liquid to obtain a supernatant for subsequent absorbance detection. The UV spectrophotometer of the residual MO content was recorded. The process was repeated six times. The other set of photocatalytic experiments were performed under visible light (wavelength of 600 nm) using a 350

Characterizations
The viscosity of precursor solutions was measured by a viscosimeter (NDJ-1, Yutong, Shanghai, China). The electrical conductivity of the solutions was measured using a conductometer (DDS-11A, Shengci, Shanghai, China). All measurements were taken 3 times at room temperature. The morphology and structure of membranes were characterized by SEM (JSM-6700F, JEOL, Tokyo, Japan). All samples were sputtered with platinum under vacuum before assessment. The average microsphere diameters were measured by Image J software. Surface area, pore volume and pore size were characterized by N 2 adsorption and desorption using a Brunauer-Emmett-Teller analyzer (BET, Autosorb-iQ2, Quantachrome Instruments, Shanghai, China). The structures of the composite fibers were analyzed by FTIR spectroscopy (FTIR-4100, Jasco, Shanghai, China), within the wavenumber ranging from 400 to 4000 cm −1 and a wide angle X-ray diffractometer (XRD, D/Max, Rigaku, Beijing, China), under mode of 40 kV and 50 mA, λ = 1.5406 Å radiation using copper-K alpha. The samples were scanned in the 2θ range from 5 • to 60 • with a scanning rate of 30 min −1 . The chemical composition of membrane surfaces were analyzed by XPS (ESCALab220i-XL, VG Scientific, Waltham, MA, USA) with Al K α X-ray source, Ni-filtered radiation, 40 kV of tube electric pressure and 30 mA of tube electric current.

Investigations on MO Adsorption and Photocatalysis Study
Organic dye MO was used as the adsorbate for adsorption and photocatalysis studies of fiber mats. 20 mg of fiber mats were immersed in 40 mg/L MO aqueous solution (20 mL). The solution with fiber mats was stirred in dark conditions for 1 h to reach the adsorption equilibrium. After that, the mix solution was irradiated for 30 min in a photodegradation reactor under UV light (wavelength of 365 nm) by mercury lamp (25 W). The mixed solution (6 mL) was taken out every 5 min and centrifuged. The centrifugation was used to separate the solid particles in the suspension from the liquid to obtain a supernatant for subsequent absorbance detection. The UV spectrophotometer of the residual MO content was recorded. The process was repeated six times. The other set of photocatalytic experiments were performed under visible light (wavelength of 600 nm) using a 350 W xenon lamp. The operation was the same as the process using UV light. The concentration of the residual MO was recorded every 20 min for 120 min. The adsorption and photocatalytic degradation R (%) of membranes to MO was calculated by the Equation (1): where A 0 and A t are the original and equilibrium absorbances of the MO solutions.
The batch experiments of MO onto the fiber membranes were carried out as functions of TiO 2 concentration (0-5 wt %), pH (4-9), contact time (0-85 min), and initial concentration (10-80 mg/L) at room temperature. The adsorption and photocatalysis ability was computed as follows: where q t (mg/g) represented the removal amount at time t, C 0 (mg/L) and C t (mg/L) are the original and equilibrium concentrations of the MO solutions, V (L) is the volume of MO solutions, m (g) is the quantity of fiber membrane.
In the reusability experiment, the used adsorbent was washed by deionized water. Then the supernatant of the solution, after the reaction was replaced by fresh MO solution. The photodegradation reaction was carried out under UV illumination with magnetic stirring for 1 h. This experiment was cycled five times.

Morphology Observation
The morphology and diameter distributions of the CA substrate are displayed in Figure 2. The surfaces of the CA fibers were smooth without beads. The mean diameter of CA fibers was 0.55 ± 0.24 µm (shown in Figure 2b). W xenon lamp. The operation was the same as the process using UV light. The concentration of the residual MO was recorded every 20 min for 120 min. The adsorption and photocatalytic degradation R (%) of membranes to MO was calculated by the Equation (1): where A0 and At are the original and equilibrium absorbances of the MO solutions.
The batch experiments of MO onto the fiber membranes were carried out as functions of TiO2 concentration (0-5 wt %), pH (4-9), contact time (0-85 min), and initial concentration (10-80 mg/L) at room temperature. The adsorption and photocatalysis ability was computed as follows: where qt (mg/g) represented the removal amount at time t, C0 (mg/L) and Ct (mg/L) are the original and equilibrium concentrations of the MO solutions, V (L) is the volume of MO solutions, m (g) is the quantity of fiber membrane.
In the reusability experiment, the used adsorbent was washed by deionized water. Then the supernatant of the solution, after the reaction was replaced by fresh MO solution. The photodegradation reaction was carried out under UV illumination with magnetic stirring for 1 h. This experiment was cycled five times.

Morphology Observation
The morphology and diameter distributions of the CA substrate are displayed in Figure 2. The surfaces of the CA fibers were smooth without beads. The mean diameter of CA fibers was 0.55 ± 0.24 μm (shown in Figure 2b). The morphology and the diameter distributions of the CS microspheres with different additions of TiO2 are shown in Figure 3. Pure CS showed concave hemisphere morphology as shown in Figure  3a. With the content of 1, 2, 3 wt % TiO2, CS still maintained the concave hemisphere morphology, and TiO2 nanoparticles were evenly distributed on the surface of CS hemispheres (see Figures 3b-d). When TiO2 content was 4 wt % and 5 wt %, TiO2 nanoparticles begun to accumulate, and the original uniform morphology of CS hemispheres gradually disappeared (Figures 2e,f).
The particle size of pure CS hemispheres was 1.07 ± 0.24 μm. With the increase of TiO2 content from 1 wt % to 3 wt % (Figures 3b1-d1), the diameters of CS hemispheres were 1.35 ± 0.26 μm, 1.40 ± 0.28 μm and 1.46 ± 0.34 μm, showing an upward trend. This was because the increase of viscosity was more important than the conductivity of the electrosprayed solution ( Table 1). As the content of The morphology and the diameter distributions of the CS microspheres with different additions of TiO 2 are shown in Figure 3. Pure CS showed concave hemisphere morphology as shown in Figure 3a. With the content of 1, 2, 3 wt % TiO 2 , CS still maintained the concave hemisphere morphology, and TiO 2 nanoparticles were evenly distributed on the surface of CS hemispheres (see Figure 3b-d). When TiO 2 content was 4 wt % and 5 wt %, TiO 2 nanoparticles begun to accumulate, and the original uniform morphology of CS hemispheres gradually disappeared (Figure 2e,f).
The particle size of pure CS hemispheres was 1.07 ± 0.24 µm. With the increase of TiO 2 content from 1 wt % to 3 wt % ( 1.40 ± 0.28 µm and 1.46 ± 0.34 µm, showing an upward trend. This was because the increase of viscosity was more important than the conductivity of the electrosprayed solution ( Table 1). As the content of TiO 2 continued to increase, the diameters of CS hemispheres were 1.10 ± 0.24 µm and 1.11 ± 0.30 µm, respectively. The quick increase in the viscosity of the solutions led to the agglomeration of TiO 2 on the surface of CS hemispheres, as shown in Figure 3e 1 ,f 1 . The size of the decrease in the CS hemispheres was due to an increase in the conductivity of the electrospray solution. TiO2 continued to increase, the diameters of CS hemispheres were 1.10 ± 0.24 μm and 1.11 ± 0.30 μm, respectively. The quick increase in the viscosity of the solutions led to the agglomeration of TiO2 on the surface of CS hemispheres, as shown in Figures 3e1,f1. The size of the decrease in the CS hemispheres was due to an increase in the conductivity of the electrospray solution.

XRD Characterization
The XRD patterns of all the samples are illuminated in Figure 4. The broad peak at 20.10° and 22.01° were characteristic peaks of CS and CA, respectively [47,48]. TiO2 powder showed three sharp feature peaks at 48.06°, 37.86° and 25.41° of 2θ values [45]. Compared with XRD of CA powders, the

XRD Characterization
The XRD patterns of all the samples are illuminated in Figure 4. The broad peak at 20.10 • and 22.01 • were characteristic peaks of CS and CA, respectively [47,48]. TiO 2 powder showed three sharp feature peaks at 48.06 • , 37.86 • and 25.41 • of 2θ values [45]. Compared with XRD of CA powders, the peak at 22.01 • of CS/CA was weakened due to the addition of amorphous CS which inhibited crystallization of CA. Compared with XRD of CS and CA powders, the peaks at 20.10 • and 22.01 • of C-T-3/CA were weakened due to the addition of TiO 2 which inhibited crystallization of CS and CA. It indicated that there were interactions between TiO 2 and CS/CA, which might stabilize TiO 2 on the membranes [45].

Effect of TiO2 Content on MO Removal
The effect of the amount of TiO2 on the removal of MO under visible light is demonstrated in Figure 5. Before visible light irradiation, the absorbance of MO solutions for all samples decreased due to adsorption of MO by the CS based membrane. Under visible light, the absorbance of all solutions with different fibrous membranes decreased as the reaction time increased (shown in Figure  5a). There was a significantly faster removal trend for the MO in visible light than in darkness because of the effects of both adsorption and photocatalysis. The removal of MO consisted of two stages for all samples. In the first stage, as the reaction time increased, the removal rate of MO was fast. In the second stage, the removal rate declined and got close to an equilibrium state. All samples could reach saturation state at 120 min. In combination with Figure 5b, the removal amount increased at first and then decreased with a further increase of TiO2 content. When the TiO2 content was higher than 3 wt %, the MO removal amount declined. This might be attributed to the reduction of the reaction sites by the TiO2 agglomeration. Therefore, the best adsorption and photocatalysis performance was obtained at 3 wt % content of TiO2 for CS based adsorbents.

Effect of TiO 2 Content on MO Removal
The effect of the amount of TiO 2 on the removal of MO under visible light is demonstrated in Figure 5. Before visible light irradiation, the absorbance of MO solutions for all samples decreased due to adsorption of MO by the CS based membrane. Under visible light, the absorbance of all solutions with different fibrous membranes decreased as the reaction time increased (shown in Figure 5a). There was a significantly faster removal trend for the MO in visible light than in darkness because of the effects of both adsorption and photocatalysis. The removal of MO consisted of two stages for all samples. In the first stage, as the reaction time increased, the removal rate of MO was fast. In the second stage, the removal rate declined and got close to an equilibrium state. All samples could reach saturation state at 120 min. In combination with Figure 5b, the removal amount increased at first and then decreased with a further increase of TiO 2 content. When the TiO 2 content was higher than 3 wt %, the MO removal amount declined. This might be attributed to the reduction of the reaction sites by the TiO 2 agglomeration. Therefore, the best adsorption and photocatalysis performance was obtained at 3 wt % content of TiO 2 for CS based adsorbents. The effect of TiO2 contents on the MO removal under UV light is demonstrated in Figure 6. Under UV light, the absorbance of all solutions decreased faster than samples in darkness because of concurrent adsorption and photocatalytic degradation as the reaction time increased (shown in Figure 6a). A similar situation was found as in the visible light, where the removal of MO consisted of two stages for all samples. In the first stage, the removal rate of MO was fast as reaction time increased. In the second stage, the removal rate declined and then got close to an equilibrium state. All samples could reach saturation state at 30 min [20]. In Figure 6b, it can be seen that the removal amount increased at first and then decreased with further increase of TiO2 content. When the TiO2 content was higher than 3 wt %, the MO removal amount declined. This might be attributed to the reduction of the reaction sites by the TiO2 agglomeration. The removal amount of MO by adsorption and photocatalytic degradation under UV light were superior to that in visible light because of more energy for electronic transitions [23]. Therefore, C-T-3/CA was selected for further MO removal studies by UV irradiation in the following tests.

Effect of pH Values on the MO Removal
The pH value, as a crucial parameter, controlled the MO removal ability by affecting the degree of ionization of hydrogen ions, hydroxide ions in solutions and structural stability. Figure 7 shows the removal amount of MO in the pH value range from 4 to 9 under UV light. Both samples exhibited high removal amount with pH value in the range of 4-6, and MO removal amount gradually declined  Figure 6. Under UV light, the absorbance of all solutions decreased faster than samples in darkness because of concurrent adsorption and photocatalytic degradation as the reaction time increased (shown in Figure 6a). A similar situation was found as in the visible light, where the removal of MO consisted of two stages for all samples. In the first stage, the removal rate of MO was fast as reaction time increased. In the second stage, the removal rate declined and then got close to an equilibrium state. All samples could reach saturation state at 30 min [20]. In Figure 6b, it can be seen that the removal amount increased at first and then decreased with further increase of TiO 2 content. When the TiO 2 content was higher than 3 wt %, the MO removal amount declined. This might be attributed to the reduction of the reaction sites by the TiO 2 agglomeration. The removal amount of MO by adsorption and photocatalytic degradation under UV light were superior to that in visible light because of more energy for electronic transitions [23]. Therefore, C-T-3/CA was selected for further MO removal studies by UV irradiation in the following tests. The effect of TiO2 contents on the MO removal under UV light is demonstrated in Figure 6. Under UV light, the absorbance of all solutions decreased faster than samples in darkness because of concurrent adsorption and photocatalytic degradation as the reaction time increased (shown in Figure 6a). A similar situation was found as in the visible light, where the removal of MO consisted of two stages for all samples. In the first stage, the removal rate of MO was fast as reaction time increased. In the second stage, the removal rate declined and then got close to an equilibrium state. All samples could reach saturation state at 30 min [20]. In Figure 6b, it can be seen that the removal amount increased at first and then decreased with further increase of TiO2 content. When the TiO2 content was higher than 3 wt %, the MO removal amount declined. This might be attributed to the reduction of the reaction sites by the TiO2 agglomeration. The removal amount of MO by adsorption and photocatalytic degradation under UV light were superior to that in visible light because of more energy for electronic transitions [23]. Therefore, C-T-3/CA was selected for further MO removal studies by UV irradiation in the following tests.

Effect of pH Values on the MO Removal
The pH value, as a crucial parameter, controlled the MO removal ability by affecting the degree of ionization of hydrogen ions, hydroxide ions in solutions and structural stability. Figure 7 shows the removal amount of MO in the pH value range from 4 to 9 under UV light. Both samples exhibited high removal amount with pH value in the range of 4-6, and MO removal amount gradually declined

Effect of pH Values on the MO Removal
The pH value, as a crucial parameter, controlled the MO removal ability by affecting the degree of ionization of hydrogen ions, hydroxide ions in solutions and structural stability. Figure 7 shows the removal amount of MO in the pH value range from 4 to 9 under UV light. Both samples exhibited high removal amount with pH value in the range of 4-6, and MO removal amount gradually declined as the pH value increased. C-T-3/CA had a higher removal amount than CS/CA due to both adsorption and photocatalysis for C-T-3/CA. The removal amount was greatly improved by the addition of TiO 2 (that allowed for simultaneous adsorption and photocatalysis) under acidic conditions. This was because under acidic conditions, the -NH 2 on the CS molecule was protonated, and TiO 2 existed in the form of TiOH 2 + , resulting in the electrostatic attraction among-SO 3 − in MO, -NH 3 + and TiOH 2 + [48,49]. Under alkaline conditions, the -NH 2 on the CS molecule was deprotonated, TiO 2 generated the TiO − , and an electrostatic repulsion appeared between fibrous membranes and MO, which resulted in the decrease of removal amount [47]. The maximum removal amount of MO for both samples was obtained at pH = 4. as the pH value increased. C-T-3/CA had a higher removal amount than CS/CA due to both adsorption and photocatalysis for C-T-3/CA. The removal amount was greatly improved by the addition of TiO2 (that allowed for simultaneous adsorption and photocatalysis) under acidic conditions. This was because under acidic conditions, the -NH2 on the CS molecule was protonated, and TiO2 existed in the form of TiOH2 + , resulting in the electrostatic attraction among-SO3 − in MO, -NH3 + and TiOH2 + [48,49]. Under alkaline conditions, the -NH2 on the CS molecule was deprotonated, TiO2 generated the TiO − , and an electrostatic repulsion appeared between fibrous membranes and MO, which resulted in the decrease of removal amount [47]. The maximum removal amount of MO for both samples was obtained at pH = 4.

Effect of Initial MO Concentration on the MO Removal
The effect of initial MO concentration on MO removal is indicated in Figure 8. The MO removal ratio of CS/CA and C-T-3/CA decreased with the increase of the initial concentration of MO. At a lower initial concentration of MO, fiber membranes CS/CA and C-T-3/CA provided sufficient reactive sites for removal of MO. With the increase of initial concentration of MO, the total amount of MO molecules increased. However, the amount of fiber membranes remained unchanged, meaning that the amount of reactive sites did not change. Therefore, the removal ratio of MO decreased at the higher initial concentration of MO. C-T-3/CA showed a higher removal ratio of 98.4% than CS/CA at initial concentration of 10 mg/L due to simultaneous adsorption and photocatalysis for the MO removal by C-T-3/CA.

Effect of Initial MO Concentration on the MO Removal
The effect of initial MO concentration on MO removal is indicated in Figure 8. The MO removal ratio of CS/CA and C-T-3/CA decreased with the increase of the initial concentration of MO. At a lower initial concentration of MO, fiber membranes CS/CA and C-T-3/CA provided sufficient reactive sites for removal of MO. With the increase of initial concentration of MO, the total amount of MO molecules increased. However, the amount of fiber membranes remained unchanged, meaning that the amount of reactive sites did not change. Therefore, the removal ratio of MO decreased at the higher initial concentration of MO. C-T-3/CA showed a higher removal ratio of 98.4% than CS/CA at initial concentration of 10 mg/L due to simultaneous adsorption and photocatalysis for the MO removal by C-T-3/CA.

Effect of Contact Time and Kinetic Study
The effect of time on the removal capacity of MO by CS/CA and C-T-3/CA are shown in Figure  9a. As time increased, the removal amount of MO by CS/CA and C-T-3/CA increased and then gradually reached the equilibrium state. In the initial stage (t < 25 min), due to the large number of active sites in both adsorbents, the removal amount of MO increased rapidly with the increase of contact time. The removal amount of MO increased slowly and finally reached the equilibrium 25 min later [49]. For CS/CA, the adsorption equilibrium appeared at 40 min, while for C-T-3/CA, equilibrium was reached at 60 min. Due to the addition of TiO2 nanoparticles, which increased the specific surface area of the fiber membrane, more time was token for the MO to penetrate the fiber membranes. Therefore, the contact time of 60 min was used for subsequent tests to ensure sufficient adsorption and photocatalytic degradation treatment. The pseudo-first-order and pseudo-secondorder kinetic models were used to study the behavior of MO removal of fiber membranes [46], which can be expressed as follows: Pseudo-first-order kinetic model: Pseudo-second-order kinetic model: where qe and qt (mg/g) represent the equilibrium removal amount and the removal amount at time t; k1 and k2 are the constants pseudo-first-order and pseudo-second-order rate, respectively. The experimental kinetic curves of MO removal are shown in Figures 9b,c. The relevant kinetic parameters calculated from the curves are listed in Table 2. The correlation coefficients of the pseudofirst-order model and the pseudo-second-order model for MO removal by CS/CA were nearly equal to and greater than 0.95, indicating that physical adsorption and chemical adsorption all contributed to the removal of MO. The R 2 of the pseudo-second-order model (0.9952) was higher than that of the pseudo-first-order model (0.9878) for the removal of MO by C-T-3/CA. This suggested that fiber membranes exhibited mainly chemical adsorption accompanied by physical adsorption via the addition of TiO2 [50].

Effect of Contact Time and Kinetic Study
The effect of time on the removal capacity of MO by CS/CA and C-T-3/CA are shown in Figure 9a. As time increased, the removal amount of MO by CS/CA and C-T-3/CA increased and then gradually reached the equilibrium state. In the initial stage (t < 25 min), due to the large number of active sites in both adsorbents, the removal amount of MO increased rapidly with the increase of contact time. The removal amount of MO increased slowly and finally reached the equilibrium 25 min later [49]. For CS/CA, the adsorption equilibrium appeared at 40 min, while for C-T-3/CA, equilibrium was reached at 60 min. Due to the addition of TiO 2 nanoparticles, which increased the specific surface area of the fiber membrane, more time was token for the MO to penetrate the fiber membranes. Therefore, the contact time of 60 min was used for subsequent tests to ensure sufficient adsorption and photocatalytic degradation treatment. The pseudo-first-order and pseudo-second-order kinetic models were used to study the behavior of MO removal of fiber membranes [46], which can be expressed as follows: Pseudo-first-order kinetic model: Pseudo-second-order kinetic model: where q e and q t (mg/g) represent the equilibrium removal amount and the removal amount at time t; k 1 and k 2 are the constants pseudo-first-order and pseudo-second-order rate, respectively. The experimental kinetic curves of MO removal are shown in Figure 9b,c. The relevant kinetic parameters calculated from the curves are listed in Table 2. The correlation coefficients of the pseudo-first-order model and the pseudo-second-order model for MO removal by CS/CA were nearly equal to and greater than 0.95, indicating that physical adsorption and chemical adsorption all contributed to the removal of MO. The R 2 of the pseudo-second-order model (0.9952) was higher than that of the pseudo-first-order model (0.9878) for the removal of MO by C-T-3/CA. This suggested that fiber membranes exhibited mainly chemical adsorption accompanied by physical adsorption via the addition of TiO 2 [50].

Isotherm Study
The Langmuir model and Freundlich model were used to describe the isotherm equilibrium data of MO removal by CS/CA and C-T-3/CA fiber membranes [50]. It can be expressed as Equations (5) and (6) The Langmuir model: The Freundlich model: where qe (mg/g) represents the equilibrium removal amount, Ce (mg/L) is the equilibrium concentration of MO in aqueous solution and qm (mg/g) represents the maximum removal amount, KL is the Langmuir isotherm constant related to free energy and KF and n are the Freundlich isotherm constant which expressed removal capacity and removal intensity, respectively. The experimental isotherm curves of MO removal are shown in Figures 10a,b. The relevant isotherm parameters calculated from the curves are listed in Table 3. On the basis of the correlation

Isotherm Study
The Langmuir model and Freundlich model were used to describe the isotherm equilibrium data of MO removal by CS/CA and C-T-3/CA fiber membranes [50]. It can be expressed as Equations (5) and (6) The Langmuir model: The Freundlich model: where q e (mg/g) represents the equilibrium removal amount, C e (mg/L) is the equilibrium concentration of MO in aqueous solution and q m (mg/g) represents the maximum removal amount, K L is the Langmuir isotherm constant related to free energy and K F and n are the Freundlich isotherm constant which expressed removal capacity and removal intensity, respectively.
The experimental isotherm curves of MO removal are shown in Figure 10a,b. The relevant isotherm parameters calculated from the curves are listed in Table 3. On the basis of the correlation coefficients of MO removal by CS/CA, the Freundlich isotherm model (Figure 9b) was better fitted than the Langmuir isotherm model (Figure 9a). Therefore, the CS/CA mainly exhibited multiple layers of adsorption. After the addition of TiO 2 , the correlation coefficients of the Langmuir model and Freundlich model were approximately equal, indicating that C-T-3/CA exhibited multiadsorption behavior. This might be due to both the effects of photocatalytic degradation and adsorption [50]. coefficients of MO removal by CS/CA, the Freundlich isotherm model (Figure 9b) was better fitted than the Langmuir isotherm model (Figure 9a). Therefore, the CS/CA mainly exhibited multiple layers of adsorption. After the addition of TiO2, the correlation coefficients of the Langmuir model and Freundlich model were approximately equal, indicating that C-T-3/CA exhibited multiadsorption behavior. This might be due to both the effects of photocatalytic degradation and adsorption [50].

FTIR Spectra Results
The FTIR spectra of C-T-3/CA fibers before and after removal of MO are demonstrated in Figure  11. Before removal of MO, the peaks at 3100-3500 cm −1 were ascribed to -OH and -NH2, 1638 and 1572 cm −1 were attributed to -NH3 + and -NH2 of CA and CS, respectively. The peak appearing at 1154 cm −1 was attributed to the stretching of the C-O-C band of CA. The peaks in the range of 400-800 cm −1 were ascribed to the characteristic absorption of Ti-O [47,51]. After the adsorption and photocatalytic degradation of MO, the peak of the -NH2 group at 1572 cm −1 decreased and the -NH3 + at 1638 cm −1 increased [49]. It demonstrated that -NH2 was protonated as -NH3 + in the process of MO removal. Several new characteristic peaks that appeared were listed as follows: 1400-1450 cm −1 were ascribed to -N=N-group, and 1039 cm −1 was attributed to -SO3 − of MO. It showed that MO was adsorbed on surfaces of CS molecules during the reaction process. The peaks at 3100-3500 cm −1 increased, which might be due to the formation of abundant active hydroxyl radicals (·OH) during the photodegradation process (see Equations (9) and (10)).

FTIR Spectra Results
The FTIR spectra of C-T-3/CA fibers before and after removal of MO are demonstrated in Figure 11. Before removal of MO, the peaks at 3100-3500 cm −1 were ascribed to -OH and -NH 2 , 1638 and 1572 cm −1 were attributed to -NH 3 + and -NH 2 of CA and CS, respectively. The peak appearing at 1154 cm −1 was attributed to the stretching of the C-O-C band of CA. The peaks in the range of 400-800 cm −1 were ascribed to the characteristic absorption of Ti-O [47,51]. After the adsorption and photocatalytic degradation of MO, the peak of the -NH 2 group at 1572 cm −1 decreased and the -NH 3 + at 1638 cm −1 increased [49]. It demonstrated that -NH 2 was protonated as -NH 3 + in the process of MO removal. Several new characteristic peaks that appeared were listed as follows: 1400-1450 cm −1 were ascribed to -N=N-group, and 1039 cm −1 was attributed to -SO 3 − of MO. It showed that MO was adsorbed on surfaces of CS molecules during the reaction process. The peaks at 3100-3500 cm −1 increased, which might be due to the formation of abundant active hydroxyl radicals (·OH) during the photodegradation process (see Equations (9) and (10)). Polymers 2019, 11, x FOR PEER REVIEW 12 of 19 Figure 11. FTIR spectra of C-T-3/CA before and after MO removal.

XPS Spectra Results
As shown in Figure 12, XPS was employed to analyze the chemical element of the membrane surface before and after the removal (that allowed for coinstantaneous adsorption and photocatalysis) of MO with full-range and magnified spectra of S, N, O and Ti. It can be seen in Figure  12a that C-T-3/CA membranes before MO removal showed peaks of binding energy at 285.5, 397.7, 531.2 and 457.1 eV, which were assigned to carbon atom (C 1s), oxygen atom (O 1s), nitrogen atom (N 1s) and titanium atom (Ti 2p), respectively. As seen in Figure 12b, the peak of S 2p appeared at 167.7 eV after removal of MO, indicating MO adsorbed on the surface fiber membranes. For the N 1s spectrum in Figure 12c, the peak at 399.1 eV was assigned to -NH2 group before MO removal. After MO removal, a new peak appeared at 401.7 eV and was attributed to -NH3 + , and the peak of -NH2 at 399.1 eV shifted to 399.8 eV. This indicated that -NH2 was protonated as -NH3 + , and electron transfer occurred between -NH2 and MO molecules, respectively [22]. As seen in Figures 12d,e, the O 1s and Ti 2p binding energy increased after the removal of MO. Ti could exist in the form of TiOH2 + in acid solutions. After the reaction with MO molecules, the groups of -OH2 + as electron acceptors could interact with -SO3 − of MO due to electrostatic attraction, which could influence charge density of Ti and O [20].

XPS Spectra Results
As shown in Figure 12, XPS was employed to analyze the chemical element of the membrane surface before and after the removal (that allowed for coinstantaneous adsorption and photocatalysis) of MO with full-range and magnified spectra of S, N, O and Ti. It can be seen in Figure 12a that C-T-3/CA membranes before MO removal showed peaks of binding energy at 285.5, 397.7, 531.2 and 457.1 eV, which were assigned to carbon atom (C 1s), oxygen atom (O 1s), nitrogen atom (N 1s) and titanium atom (Ti 2p), respectively. As seen in Figure 12b, the peak of S 2p appeared at 167.7 eV after removal of MO, indicating MO adsorbed on the surface fiber membranes. For the N 1s spectrum in Figure 12c, the peak at 399.1 eV was assigned to -NH 2 group before MO removal. After MO removal, a new peak appeared at 401.7 eV and was attributed to -NH 3 + , and the peak of -NH 2 at 399.1 eV shifted to 399.8 eV. This indicated that -NH 2 was protonated as -NH 3 + , and electron transfer occurred between -NH 2 and MO molecules, respectively [22]. As seen in Figure 12d,e, the O 1s and Ti 2p binding energy increased after the removal of MO. Ti could exist in the form of TiOH 2 + in acid solutions. After the reaction with MO molecules, the groups of -OH 2 + as electron acceptors could interact with -SO 3 − of MO due to electrostatic attraction, which could influence charge density of Ti and O [20].

BET Analysis
The BET results of CS/CA and C-T-3/CA membranes were measured by nitrogen adsorption method (see Table 4). After the addition of TiO 2 , the fiber membrane attained higher specific surface area and pore volume owing to uniform distribution of TiO 2 on CS hemispheres and an increase of the distance among C-T hemispheres (see Figure 3d). Therefore, the relative high adsorption capacity of MO on C-T-3/CA was ascribed to the high specific surface area, which should be good for the photodegradation of MO. Polymers 2019, 11, x FOR PEER REVIEW 13 of 19 Figure 12. XPS spectra of C-T-3/CA before and after MO removal.

BET Analysis
The BET results of CS/CA and C-T-3/CA membranes were measured by nitrogen adsorption method (see Table 4). After the addition of TiO2, the fiber membrane attained higher specific surface area and pore volume owing to uniform distribution of TiO2 on CS hemispheres and an increase of the distance among C-T hemispheres (see Figure 3d). Therefore, the relative high adsorption capacity of MO on C-T-3/CA was ascribed to the high specific surface area, which should be good for the photodegradation of MO.

Mechanism Analysis
According to the above analysis, the high removal capacity of C-T-3/CA was mainly due to the effective dispersion of TiO 2 in combination with CS, which improved the specific surface area of fiber membranes and increased the sites of photocatalytic degradation and adsorption. MO molecules could be adsorbed to the vicinity of photodegradation sites by electrostatic attraction between CS and MO molecules, promoting the contact probability of TiO 2 and MO in the following photocatalytic degradation process. The reaction mechanism between MO and C-T-3/CA is illustrated in Figure 13. The mechanism of photocatalysis is summarized from Equation (7) to Equation (12). The MO removal mechanism included: (1) Adsorption, the electrostatic attraction between -NH 3 + group of CS and -SO 3 − of MO molecules; (2) photocatalysis, the photocatalytic degradation of the MO by UV illumination began with photoexcitation of the TiO 2 and then formed an electron-hole pair (Equation (7)). High oxidation valence-band holes (h + ) directly oxidized MO degradation (Equation (8)). Water decomposition produced ·OH (Equation (9)) or a reaction of h + with OH − (Equation (10)). Meanwhile, the reaction between conduction-band electrons (e − ) and proper electron acceptors (such as O 2 ) yielded oxidative radicals as described by Equation (11). The generated hydroxyl radicals easily degraded MO to form inorganic small molecules (Equation (12)) [6]. These results agreed with BET, FTIR and XPS analyses.

Mechanism Analysis
According to the above analysis, the high removal capacity of C-T-3/CA was mainly due to the effective dispersion of TiO2 in combination with CS, which improved the specific surface area of fiber membranes and increased the sites of photocatalytic degradation and adsorption. MO molecules could be adsorbed to the vicinity of photodegradation sites by electrostatic attraction between CS and MO molecules, promoting the contact probability of TiO2 and MO in the following photocatalytic degradation process. The reaction mechanism between MO and C-T-3/CA is illustrated in Figure 13. The mechanism of photocatalysis is summarized from Equation (7) to Equation (12). The MO removal mechanism included: (1) Adsorption, the electrostatic attraction between -NH3 + group of CS and -SO3 − of MO molecules; (2) photocatalysis, the photocatalytic degradation of the MO by UV illumination began with photoexcitation of the TiO2 and then formed an electron-hole pair (Equation (7)). High oxidation valence-band holes (h + ) directly oxidized MO degradation (Equation (8)). Water decomposition produced ·OH (Equation (9)) or a reaction of h + with OH − (Equation (10)). Meanwhile, the reaction between conduction-band electrons (e − ) and proper electron acceptors (such as O2) yielded oxidative radicals as described by Equation (11). The generated hydroxyl radicals easily degraded MO to form inorganic small molecules (Equation (12) Figure 13. The mechanism of MO removal by C-T-3/CA. Figure 13. The mechanism of MO removal by C-T-3/CA.

Recycling Ability
To research the recycling ability of the composite membranes, the absorbance changes were measured in five cycles of MO removal. As demonstrated in Figure 14, the MO removal of fiber membranes was almost invariable for five consecutive cycles. The membranes still had a removal ratio of 98% at the end of the fifth cycle, testifying that the membranes maintained good performance of adsorption and photocatalytic degradation. This was because the stability of CS in aqueous solutions was improved by the intermolecular hydrogen bonds between CS and CA, and TiO 2 could be uniformly distributed via the interactions with CS molecules for effective photocatalytic degradation of MO [45]. membranes was almost invariable for five consecutive cycles. The membranes still had a removal ratio of 98% at the end of the fifth cycle, testifying that the membranes maintained good performance of adsorption and photocatalytic degradation. This was because the stability of CS in aqueous solutions was improved by the intermolecular hydrogen bonds between CS and CA, and TiO2 could be uniformly distributed via the interactions with CS molecules for effective photocatalytic degradation of MO [45].

Conclusions
In this work, the C-T/CA fiber membranes were prepared via electrospinning and electrospraying. After the addition of TiO2, the removal capacity of MO was improved, as a result of concurrent adsorption and photocatalytic degradation. When TiO2 content was 3 wt %, the highest MO removal amount for fiber membranes (C-T-3/CA) was reached, 98% at pH value of 4 and MO concentration of 40 mg/L. The kinetic and isotherm model for the removal of MO by fiber membranes were well fitted pseudo-second-order kinetic and Freundlich isotherm models. Especially, C-T-3/CA membranes exhibited multiple layer adsorption with the addition of TiO2. The involved mechanisms including adsorption and photocatalytic reaction occurred in the removal process which was demonstrated by FTIR, XPS and BET results. The recycling experiment showed fiber membrane had excellent stability and reusability.
Funding: This research was funded by the National Natural Science Foundation of China (No. 51103058).