Preparation and Optimization of Photocatalytic Polyacrylic Acid/Guar Gum/TiO2 Hydrogels for Absorption and Removing of Methylene Blue Under Visible and UV Irradiation

In this research, photocatalytic hydrogels of polyacrylic acid/guar gum/Titanium dioxide (PAA/GG/TiO2) were synthesized and optimized by response surface methodology for the absorption and degradation of methylene blue (MB). The variables were the amount of cross-linking agent, the amount of guar gum, the amount of TiO2 nanoparticles, and the pH of the environment, respectively. The optimizing responses were swelling ratio and absorption in two environments with and without UV radiation, respectively. The optimal samples contained 4% (molar) cross-linking agent and 2 wt. % of guar gum. The amount of TiO2 was optimized at 5 and 7 wt. %. To achieve dye absorption rate, kinetic models such as pseudo-first-order, pseudo-second-order, Elovich, and intraparticle diffusion (Weber–Morris), had been investigated in visible light and UV radiation at an initial concentration of MB 20 (ppm) for one day. Also, absorption isotherm equations in visible light and UV irradiation at various concentration of 20–60 ppm were evaluated based on the models of Freundlich, Langmuir (Hanes–Woolf), Temkin, and Dubinin–Radushkevic. The results showed that the hydrogel containing 4% (molar) of crosslinker, 5 wt. % of TiO2, and 2 wt. % guar gum, have good efficiency for removing MB from the aqueous medium and this property is intensified by UV radiation.


Introduction
The water crisis is one of the biggest challenges in the world. Water is polluted by hazardous materials such as heavy oils, dyes, some heavy metals [1][2][3]. Clean water resources are dwindling continuously; therefore, refinement of wastewater operation has become one of the most significant issues for the environment and the industries. Textile Industries use yearly over 150,000 tons of water which can enter over 150 tons of cationic dye wastewater into nature and damage human and marine creatures [4]. The presence of dye in water not only causes pollution and toxicity but also prevents the passage of sunlight and can reduce the photosynthesis of marine plants. Ultimately, lack of oxygen in the water leads to the death of aquatic animals [5]. Therefore, the presentation of a method for removing dye from wastewater is vital before entering into nature. For the removal of dyes in wastewater several methods can be utilized as membrane filtration, coagulation/flocculation, advanced oxidation, ozonation, electrocoagulation, adsorption or absorption by super absorbents, photocatalytic degradation and etc. [6][7][8][9][10][11][12].
Using the absorption process is considered one of the most efficient methods for collecting water pollutants. Hydrogels as superabsorbent polymers (SAPs) can absorb water up to several hundred times of their volume and trap pollutants inside themselves during this process. If anionic polymers such as poly (acrylic acid) (PAA) are used in the construction of these hydrogels, the absorption efficiency of cationic dyes such as methylene blue (MB) will be greatly increased [13]. Based on this, Esmaeildoost et al. used Xanthan Gum/PAA/Cloisite 15A semiinterpenetrating polymer network (semi-IPN) hydrogels to absorb heavy metals [14]. Although hydrogels have significant applications as pollution collectors, the main problem has not disappeared. This means that the pollution has changed from scattered to concentrated, but has not been destroyed. Photocatalyst materials such as n-type semiconductor materials TiO 2 , ZnO, SnO 2 , Fe 2 O 3 , WO 3 , In 2 O 3 and nonmetallic semiconductor C 3 N 4 are used to destroy pollution. Therefore, if the photocatalyst nanoparticles are placed in the hydrogel, they first absorb the pollutants and then destroy them [15,16]. For example, Mota et al. prepared Zn(II)-porphyrin/poly(acrylic acid) (Zn(II) Pr@PAA) microparticles as potential photocatalysts [17]. They suggested that the Zn(II)Pr@PAA microparticles are promise catalysts for photodegradation of aqueous organic pollutants in practical applications.
By placing photocatalyst nanoparticles inside hydrogels, more parameters will be involved in the efficiency of hydrogels, and these parameters should be optimized. Response surface method (RSM) is an experimental design method in which the independent variables are varied at different levels and the dependent variables are examined as their responses [18,19]. In this method, in addition to the main effects between factors, it is possible to estimate interactive effects and interaction between factors (reciprocal effects). Behnken-Box (BBD) and central composite design (CCD) are among the most common designs available in the RSM. The advantage of CCD over BBD is more changes of variables and the possibility of more accurate estimation of the response behaviors [20].
Therefore, in this research, our goal is to prepare photocatalytic semi-IPN hydrogels based on PAA/Guar Gum (GG)/TiO 2 for absorbing and then destroying MB. MB absorption variable parameters such as amounts of gum, crosslinker, nanoparticle, and pH were analyzed in visible light and under UV radiation and swelling ratio for optimizing RSM responses as shown in Scheme 1. Based on the optimization part of the RSM, the optimal samples and a control sample (without TiO 2 ) were synthesized, and XRD, FTIR, and SEM analyses were performed to investigate the structure and presence of nanoparticles in the hydrogels. The absorption capacities of the optimal samples and control were obtained and compared by various kinetic and isotherm models.

Experimental Design
The experimental design was drafted by Design-Expert program 11.0.3 from the central composite design (CCD) subcategory of the response surface methodology (RSM) [21]. The amount of AA was constant at 4 g. Variables in the RSM were the quantities of BDOD (X 1 ), GG (X 2 ), TiO 2 (X 3 ), and pH of absorption medium (X 4 ) which are noted in Table 1S and Scheme 1. Total runs suggested by CCD can be obtained from Eq. (1). which n and n c are the numbers of variables, and the central replicate runs, respectively. The evaluating responses were swelling (Y 1 ), removing without UV radiation (Y 2 ), and removing with UV radiation (Y 3 ). Also, one quadratic model was used for explaining the mathematical relationship between the three independent variables by the following Eq. (2): (1) Overall run numbers = 2n + n 2 + n c Scheme 1 Algorithm for preparation and optimization of PAA/GG/ TiO 2 nanocomposite hydrogels for MB removal where Y j , b 0 , b i , b ii , and b ij are the predicted response, constant-coefficient, linear-coefficient, quadratic equation, and interaction coefficient, respectively. In addition, X i and X j are the values of variables. The CCD proposed 30 runs, as shown in Table 1. These samples were fabricated based on what was described briefly.

Synthesis Method
TiO 2 powder was dispersed into the distilled water. For TiO 2 uniformity, the TiO 2 mixture was put into an ultrasonic device. The sample was irradiated in an ultrasonic cleaning bath (28 kHz, 1000 W) for 30 min at room temperature. GG as a stabilizer was then added to the TiO 2 (2) b ij X i X j beaker and stirred for 3 h. Then AA as a monomer, BDOD as a crosslinker, and KPS as an initiator were added into the beaker (based on experimental design). Eventually, the mixture was entered into a silicon mold and cured in an oven at 90 °C for 30 min. After curing, the hydrogel was settled into a beaker to separate unreacted monomers. Each hydrogel was located in an oven at 50 °C for two days to dry and prepare for further evaluations.

Swelling
The swelling test was done as in previous works [22,23]. In summary, the chopped dry hydrogels were put into the distilled water. At several times, the samples were removed from the water and their surface was dried with tissue paper and their weight was measured. The swelling ratio of samples was obtained by Eq. (3). which m 0 is the weight of the dry hydrogel, and m t is the weight of the swollen hydrogel at the time (t).

Absorption of MB in Aqueous Solution
0.5 g prepared dry hydrogels were generally poured into the petri dish containing 50 ml of aqueous MB solution at a concentration of 20 ppm at different pHs for one day. In experiments, the intensity of MB was evaluated by UV-vis spectroscopy sequential times. The percentages of absorption of MB were obtained by Eq. (4).
which C 0 is the initial concentration of MB and C t is the concentration of the solution after absorption at the variable time (t).

Characterization
In this study, UV-vis spectroscopy model V-770 was used to measure the intensity of MB with an accuracy of 0.1 nm. Burker tensor 27 (made in Germany) FT-IR Spectrometer was used to analyze the bond among the particles of the samples in the wavenumber of 400 to 4000 cm −1 . SEM images using to observe particle size and the presence of TiO 2 nanoparticles in the hydrogel using the AIS2100C device (made in South Korea). The amorphous or crystalline structure of the hydrogels was examined and analyzed by an XRD device with a wavelength of X-ray lamp (1.5 angstroms), model PW1730 (manufactured by Philips, the Netherlands).

UV Chamber
For the degradation of MB under UV radiation, a handmade UV chamber covered by aluminum foil was designed. This device consists of the following parts such as a fan installed for air conditioning, UV-C, and UV-AB lamps (Philips).

Synthesis of Photocatalytic PAA/GG/TiO 2 Hydrogels
Hydrogels were synthesized based on Table 1 and shown in Fig. 1S. As can be seen, all samples were uniform. The samples containing TiO 2 nanoparticles were whiter and the dimensional stability of the samples was improved by increasing the amount of crosslinking agent. For example, run-3 was more transparent than the other samples that did not contain nanoparticles. According to Table 1, some samples contain different amounts of TiO 2 . For this purpose, from the TiO 2 used in this study, SEM images and particle size distribution of TiO 2 were taken in Fig. 1. TiO 2 nanoparticles are spherical with an average diameter of 60 nm. The swelling rate of hydrogels as superabsorbents is a key parameter to absorb pollutants. The swelling of synthesized hydrogels was evaluated and shown in Fig. 2. Run-8, run-19, run-27, and run-30, which had high swelling rates, had a low cross-linking agent. Likewise, run-4 and run-22 have low swelling ratios, which can be due to the higher amount of crosslinks. According to Table 1, run-22 had a crosslinker of 4%, which was not the highest value. Therefore, it seems that the synergistic effect of other variables has been involved challengingly. The analysis of software results is shown in Fig. 2S. In contour diagrams, red and blue regions indicate the highest and lowest response values respectively. Therefore, the color spectrum changes between blue and red. The contour plots are green and the color spectrum does not change significantly. The lower is the crosslinker, the yellower would be the swelling ratio diagrams. The results confirm that increasing the amount of TiO 2 nanoparticles reduces the amount of swelling. Since these nanoparticles can prevent water from entering their structure, they can reduce swelling in hydrogels.
The results show that the swelling rate is maximum at alkaline pH. Previous scholars proved that PAA swells more in alkaline environments [24,25]. In alkaline environments, the hydrogen of the repeating units of acrylic acid is released and the chains take on an anionic charge, and this causes repulsion between the chains and more swelling.

Evaluation of MB Absorption in Visible and UV Light as Y 2 and Y 3
According to Fig. 3, all runs for absorption of MB in the UV chamber are significantly higher than absorption in visible light. The significant increase can also be due to TiO 2 photocatalytic nanoparticles. This proves the successful presence of TiO 2 nanoparticles for MB degradation. Although other factors also affect the absorption percentage in the two environments. According to the contour diagrams (Fig. 3S), the effect of alkaline medium on the adsorption of MB is more favorable than that of acidic medium. By comparing the contour diagram in both visible and UV radiation environments (Figs.3S and 4S), it is obvious that the yellow and red areas are more due to UV radiation. Color spectrum changes of contour diagrams are more noticeable with pH change. Therefore, increasing the pH is the most effective factor in the absorption ratio of MB in visible light. Figure 3S indicates that absorption increases by scale-up the amount of GG from 1 to 2 (wt %) and reduction of BDOD from 8 to 4 (mol %). The minimum amount of TiO 2 and BDOD have a favorable influence on improving the absorption of MB from the aqueous solution. Figure 3S likewise mentions that the improvement of absorption ratio occurs by changing the amount of GG from 1 to 2 (wt %) and TiO 2 from 15 to 5 (wt %). Guar gum, as a natural water-soluble polymer, has increased the swelling rate in hydrogels. As shown in Fig. 4S, yellow and red colors expose in most images, which manifests the efficiency of MB absorption under UV radiation compared with visible light.

Design Expert Results
Base on Tables 2S, 3S, 4S, all responses are significant. As shown in the color spectrum of Fig. 4, when TiO 2 reaches 15 wt%, the graphs approach blue. Of course, the graphs confirm the optimal value of about 5-8 wt. % TiO 2 . When pH = 8 and GG = 2 (wt %); the color spectrum tends to red regularly, so increasing the pH and GG have the most positive effect on the desirability. BDOD has the worst effect on the ratio of desirability.
Based on the results of RSM, two samples were selected as optimal, which were compared by a control sample that did not contain TiO 2 nanoparticles (Table 2).

Morphological study of the fracture surface
SEM images of the cross-section of the optimal and control samples are shown in Fig. 5. In Optimum-1 and Optimum-2, the presence of TiO 2 nanoparticles can be seen. These nanoparticles exhibit as lumps, in the optimal-2 is increased the intensity of aggregations excessively.

XRD and FTIR Analyses
The XRD patterns are for optimal and control samples in the angle range 2θ = 10-80°. Since no peak diffraction in the control sample diagram is apparent in Fig. 6 is amorphous. By incorporation of TiO 2 nanoparticles into the hydrogels; the morphology of the hydrogel alternates from an amorphous status to the crystalline composites, and the peak angles of this nanoparticle are constant in most of the hydrogels. Dispersions of TiO 2 nanoparticles are contributed in Fig. 6a at angles of 2θ = 27°, 38°, 48°, 54°, 56°, 63°, 69°, and 71° [26].
The intensity of peaks in Optimum-2 is higher than in Optimum-1, which indicates a higher amount of TiO 2 nanoparticles. The FTIR of the fabricated samples is definite in the range of 400-4000 cm −1 (Fig. 6). The peak at 3436 cm −1 relates to the stretching bond of the hydroxyl functional group due to intermolecular hydrogen bonding between PAA and GG chains. The wavenumber roughly 1718 cm −1 should pertain to the tenacious stretching vibrations of the carbonyl group of PAA [27]. The wavenumbers 1454 and 1162 cm −1 relates to the medium bending bonds of methyl substations and the medium stretching bonds of the amine group, respectively. Between 900 and 500 cm −1 wavenumber locating in the fingerprint zone; relates to the symmetric and asymmetric links of Ti-O-Ti [28]. The peaks of 1256 and 1240 cm −1 are attributed to the atactic and syndiotactic, respectively, confirming the existence of PAA [25].

Evaluation of MB Absorption by Optimal and Control Hydrogels
The UV-vis spectroscopy reports the intensity of the colors (Abs) based on the index peaks. MB dye has different peak places such as 300, 585, and 665 nm as is shown in Fig. 5S. The most intense considering peak is 665 nm as the base peak to obtain the desired concentration. For this aim, calibration curves at a variety of concentrations   The results of absorption capacity of MB by hydrogels over time are shown in Fig. 7. As it is shown in Fig. 7a, MB was absorbed by the control sample 1780 mg/g in 225 min, but Optimum-1 and Optimum-2 could have absorbed about 1600-1650 mg/g MB for 225 min at visible light condition. Therefore, MB absorption rate is more in control than optimums at without UV radiation. Figure 7b show the MB absorption by Optimum-1, Optimum-2, and control in the UV chamber. The absorption capacity rate of the Optimum-1 is higher than all samples and is approximately 1720 mg/g in 90 min. It is noteworthy that Optimum-2 has an equilibrium absorption capacity of 1896 mg/g in visible light, but when it is placed under UV light, it increases nearly 1996 mg/g after 1500 min, which can effectively infer the photocatalytic hydrogel activity at UV chamber.

Kinetic Absorption Models
For obtaining the absorption pace and the rate-limiting step, several absorption kinetic models are used. The kinetic models of the pseudo-first-order Eq. (6), pseudo-second-order Eq. (7), intraparticle diffusion (Weber-Morris) Eq. (8), and Elovich Eq. (9) based on absorbent capacity: t q e (8) q t = k d t 1∕2 + C   In all kinetic Eq.s of this work, q t and q e are the quantities of pollutants absorbed at t (min) and equilibrium status, respectively. k 1 (min −1 ) in Eq. (6) is the pseudo-first-order absorption rate constant, k 2 (g.mg −1 .min −1 ) in Eq. (7) is the equilibrium constant of the pseudo-second-order; C is a constant of diffusion resistance (mg/g) and a constant rate of intraparticle diffusion, k d (mg.g −1 .min −0.5 ) in Eq. (8). Elovich's two constants are initial absorption rate α (mg. g −1 .min −1 ) and desorption constant β (g/mg) in Eq. (9). Figure 8 and Table 3 show kinetic absorption model diagrams and constants in visible light. As it is shown, the kinetic model of Elovich absorption with an average of R 2 = 0.99 is a favorable kinetic model for absorption in visible light. This is while the q e (cal) corresponds better to the q e (exp) value by pseudo-first-order.
(9) q t = 1 ln ( ) + 1 ln t According to the obtained equations for absorption kinetics, it was found that some models are in good agreement with the results. As shown in Fig. 9 and Table 3, following the table can be found out these results: the pseudofirst-order model with averages of R 2 = 0.92 and χ 2 = 1495 cannot be a preferable model for the absorption kinetics. Figure 9b, the pseudo-second-order model with an average of R 2 = 0.999 and χ 2 = 0.011 should report a perfect model for UV absorption kinetics. Figure 9c, the intraparticle diffusion model cannot expose an acceptable for the UV absorption kinetic model and so on. Since the average of q e (cal) = 2000 mg/g of the pseudo-second-order kinetic model is closer to the average of q e (exp) = 1995 mg/g than other models, the pseudo-second-order model is the best model for absorbing MB in the UV chamber. Based on the Elovich model from Table 3, the initial absorption rate MB under UV light is faster than visible light.

Isotherm Absorption Models
The absorption isotherm is the relation between the absorbed pigment capacity and the different pigment concentrations in an aqueous solution under equilibrium status based on In all Eq.s, q e (mg/g) and C e (mg/L) are Equilibrium capacity and concentration respectively, in Eq. (11), the absorption capacity constant of Freundlich is K F (mg/g) (L/g) n . In Eq. (12), the constant Langmuir capacity K L (L/ mg), maximum monolayer coverage capacities q m (mg/g). In Eq. (13), the equilibrium constant of Temkin is k T (J/mol) , and the heat absorption Constant is b 1 (g.L −1 .K); if b 1 is more than zero, the absorption process would be exothermic [30]. The theoretical saturation capacity q m (mg/g),∈ = RTln(1 + 1 C e ) , and D-R isotherm constant β (11) log q e = log k F + 1 n log C e (12) 1 q e = 1 q m + 1 k L q m C e (13) q e = b 1 ln k T + b 1 ln C e (14) ln q e = ln q m − β ∈ 2 (mol 2 kJ −2 ) in Eq. (14), in the lateral relation, the global constant of gases R is equivalent to 8.314 and the absolute ambient temperature in terms of Kelvin T. There is a significant constant in the Langmuir relation called the Langmuir intrinsic absorption (R L ) which is determined from Eq. (15). If the intrinsic absorption is more than 1, the absorption capacity is undesirable. When it is 0, the reaction is irreversible. When it is 1 to 0 absorption is desirable. Langmuir intrinsic absorption depends on the K L and the initial concentration (C 0 ). K L indicates the porosity and active absorption Sites of an absorbent, so the higher the Langmuir constant, the greater the absorption capacity and can set the Langmuir intrinsic absorption between 1 and 0.
In Fig. 10, the Optimum-1 has the highest equilibrium absorption capacity at different concentrations in both visible light and UV radiation environments; in addition, the Optimum-2 has less absorption capacity in visible light than the control sample, although in the UV chamber due to the activation of the photocatalytic properties of TiO 2 , its absorption capacity is higher than the control sample. The reason is that Optimum-2 contains more TiO 2 nanoparticles than Optimum-1. The presence of these nanoparticles, as shown, reduces the amount of swelling in the hydrogels, and (15) therefore, the absorption rate of Optimum-2 was lower in the environment without UV radiation. From Table 4 and Fig. 11, based on the Freundlich isotherm model (average R 2 = 0.94), the orders of the functions (n) vary 0.8-1. Langmuir isotherm model indicates that only the optimum-1 K L has a positive range, and the intrinsic isotherm absorption R L = 0.5-0.75 is in visible light. Other samples with an R L of more than one cannot demonstrate  desirable absorption at higher concentrations. k T (Temkin equilibrium constant) varies from 1.5 to 0.5, and due to the low R 2 for optimum-1 and control, just can be referred for optimum-2. The average q m = 5830 mg/g of the D-R model reports that with an average of the correlation coefficient, 0.93 for the optimum-1 and optimum-2 samples, which is significant with the obtained results. Freundlich was the best model of isotherm of MB absorption in visible light with χ 2 averages, 89 mg/g. Table 4 and Fig. 11 show that the Optimum-1 based on the Freundlich isotherm model has the highest order of absorption equilibrium capacity (n) equivalent to 2.5. Langmuir model points out that the equilibrium Langmuir constants (K L ) vary from 0.66 to 2. All optimal and control samples have intrinsic isotherm absorption (R L ) between 0.008 and 0.024. The equilibrium constant of the Optimum-1 is higher than other samples. Finally, the significant result is that the best model of absorption isotherm is Freundlich (averages of R 2 = 0.98, and χ 2 = 27.5). According to Table 4, orders of absorption concentration capacities (n) in visible light are approximately first order, while the orders (n) in the UV chamber rise to 2.2 to 2.5 order, which indicates the effect of UV radiation and photocatalytic activity of TiO 2 nanoparticles in the absorption and destruction of MB.

EDS Analysis
Images of the fracture surface of the samples before and after MB absorption were carried out to detect existing elements. The elements carbon (red), oxygen (green), nitrogen (blue), sulfur (purple), and titanium (yellow) were assessed in Fig. 12. It is clear that the control samples lack the element titanium.
Mapping Optimum-1 elements after MB absorption (after immersion in MB solution) in visible light, increasing sulfur, nitrogen, and carbon densities represent that the absorption of MB in Optimum-1. Mapping Optimum-1 elements after MB absorption in UV light, compared to Optimum-1 after absorption in visible light, declining carbon and oxygen justify that degradation of MB appears with the Optimum-1. Radical reactions among oxygen, water, and MB due to the presence of TiO 2 in UV light; degradation and absorption arise simultaneously. Based on MB degradation, CO 2 is produced ideally [31]. Finally, the results of this study are  compared with previous works and the results are summarized in Table 5.
Comparison between the results of other studies confirms the effect of acrylic acid hydrogels containing photocatalytic nanoparticles prepared in this study. For example, Santoso et al. showed that with increasing the pH of the environment, the rate of absorption and removal of MB increased [26]. They attributed this effect to creating negative charges in the alkaline environment on the adsorbent

Conclusions
The examinations of swelling and absorption of MB in visible light and UV radiation present as three test design responses that fundamentally can be reached the conclusions: the swelling assessment was optimal when the amounts of TiO 2 , GG, and BDOD approached the minimum range. The absorption assessment in visible and UV light at the optimal status when pH = 8, GG (2 wt %), BDOD (2 mol %), and TiO 2 (5-7 mol %). Likewise, MB absorption due to photocatalytic activation of TiO 2 was reported favorably at pH = 6 in the UV chamber. SEM, XRD, and FTIR analysis indicated the presence of intensity of nanoparticles in the hydrogel. Superlative equilibrium absorption capacities (q e ) of Optimum-1 were 1934 (mg/g) in visible light and 1996 (mg/g) under UV irradiation. The results of equilibrium absorption capacity (q e ) and absorption capacity (q t ) at different times showed that the optimum-1 had the topmost absorption capacity at various minutes (q t ) and concentrations (C e ). The absorption kinetic results for the optimal samples generally showed that the absorption in visible light was pseudo-first-order and under UV radiation trended to be second-order. Significantly, based on the Elovich, the initial absorption rate (α) of MB under UV light was 3-13 times faster than visible light. The results of the Freundlich absorption isotherm model for the Optimum-1 demonstrated that the order of equilibrium concentration (n) in visible light was 1.07 and under UV radiation was 2.49 relative to the equilibrium absorption capacity. Langmuir model results pointed out that the optimum-1 had the higher equilibrium constant (K L ) than the other samples and put the intrinsic absorption (R L ) between 0 and 1 in the two environments. Ultimately, EDS analysis confirmed the presence of TiO 2 nanoparticles, degradation MB to CO 2 , and more demanded absorption of MB under UV radiation than visible light. Funding The authors did not receive support from any organization for the submitted work.