Solar-Driven Removal of 1,4-Dioxane Using WO 3 / n γ -Al 2 O 3 Nano-catalyst in Water

: Increasing demand for fresh water in extreme drought regions necessitates potable water reuse. However, current membrane-based water reclamation approaches cannot e ﬀ ectively remove carcinogenic 1,4-dioxane. The current study reports on the solar-driven removal of 1,4-dioxane (50 mg L − 1 ) using a homemade WO 3 / n γ -Al 2 O 3 nano-catalyst. Characterization methods including scanning electron microscope (SEM), X-ray photoelectron spectroscopy (XPS) and X-ray ﬂuorescence (XRF) analyses are used to investigate the surface features of the catalyst. The 1,4-dioxane mineralization performance of this catalyst under various reaction conditions is studied. The e ﬀ ect of the catalyst dosage is tested. The mean oxidation state carbon (MOSC) values of the 1,4-dioxane solution are followed during the reaction. The short chain organic acids after treatment are measured. The results showed that over 75% total organic carbon (TOC) removal was achieved in the presence of 300 mg L − 1 of the catalyst with a simulated solar irradiation intensity of 40 mW cm − 2 . Increasing the dose of the catalyst from 100 to 700 mg L − 1 can improve the treatment e ﬃ ciency to some extent. The TOC reduction curve ﬁts well with an apparent zero-order kinetic model and the corresponding constant rates are within 0.0927 and 0.1059 mg L − 1 s − 1 , respectively. The MOSC values of the 1,4-dioxane solution increase from 1.3 to 3 along the reaction, which is associated with the formation of some short chain acids. The catalyst can be e ﬀ ectively reused 7 times. This work provides an oxidant-free and energy saving approach to achieve e ﬃ cient removal of 1,4-dioxane and thus shows promising potential for potable reuse


Introduction
Rapid population expansion of cities results in increasing water consumption and requires exploitation of alternative water resources for potable purposes, especially in extremely water-scarce regions [1][2][3][4]. Membrane-based water purification techniques including ultrafiltration (UF) and reverse osmosis (RO) can effectively remove the major proportion of salts and organic contaminants in municipal wastewater secondary effluent (MWSE) [5][6][7][8][9]. However, some micropollutants including 1,4-dioxane (1,4-D) present in the MWSE in trace concentrations cannot be easily removed [10]. 1,4-D is a reagent stabilizer in industrial chlorinated solvents, which commonly exists in cosmetics, toiletries, and food addictive [11][12][13]. Uncontrolled exposure to 1,4-D can cause failures of human organs including kidney and liver [14,15]. It can even cause cancer when it exists in drinking water. Thus, it has been classified as a Group 2B human carcinogen [1]. or energy in these systems. These ROS tend to attack organic pollutants and result in their degradation [22][23][24][25][26]. Photolytic AOPs can take advantage of the photons' energy, especially ultraviolet (UV), to activate the oxidants [27]. In order to eliminate 1,4-D from water for potable reuse, recent studies provided several possible solutions based on photolytic AOP approaches [1,28,29]. Patton et al. investigated the performance of UV/H 2 O 2 -based system for 1,4-D removal in the presence of mono-and dichloramines [1]. On that basis, Li  to establish a UV/S 2 O 8 2− system to eliminate 1,4-D [5]. Although these works revealed that monoor dichloramines can promote treatment efficiency to some extent by participating in the reaction processes, these systems inevitably introduce external reagents (the oxidants) into the target solution. This compromises the practical application of those methods for potable water reuse since the effect of the residual reagents on human health needs to be further considered. Photocatalytic oxidation systems are possible candidates to avoid that drawback. In photocatalytic systems, oxidants are not indispensable since ROS can be formed from the activation of dissolved oxygen or even water molecules by the photocatalysts under photoirradiation [30][31][32]. Electrons on the valence band (VB) of the catalyst can be excited to transit to the conduction band (CB) when the energy of photoirradiation reaches the band gap [33]. Electron/hole (h + /e − ) pairs can be formed on the catalysts during this process and lead to formation of ROS from dissolved oxygen and water molecules [30,34]. Photocatalytic approaches have been used to treat 1,4-D contaminated water [35][36][37]. However, these systems were mainly activated by UV light rather than visible light, since the band gap of traditional photocatalysts (like TiO 2 ) is so broad that photons with lower energy (visible light) are not capable of inducing the formation of ROS [38,39].
Tungsten-based catalysts have been developed in recent years since they are sensitive to visible light and thus can improve the photocatalytic efficiency [40][41][42][43]. Meanwhile, γ-Al 2 O 3 supporter showed more desirable stability in AOP systems than conventional supporters like activated carbon due to its high active phase-supporter interaction [16]. Moreover, using nano-size supporters may provide a stronger active phase-supporter link due to their unique features like extremely large specific surface area. Thus, the nano-size supporters have the potential to improve the reusability of the catalyst [44]. However, to the best of our knowledge, no previous study applied nano tungsten-based catalysts with γ-Al 2 O 3 as the supporter for the photocatalytic breakdown of 1,4-D.
The current study aimed at the removal of 1,4-D using an oxidant-free photocatalytic system with a homemade nano-size tungsten-based catalyst (WO 3 /nγ-Al 2 O 3 ) under solar light irradiation. The characteristics of the catalyst were studied using scanning electron microscope (SEM), X-ray photoelectron spectroscopy (XPS) and X-ray fluorescence (XRF). 1,4-D mineralization efficiency under various conditions was considered and the reaction kinetic rates were calculated. The effect of the catalyst dosage was tested. The oxidation state of the effluent solution during the reaction was followed and the formation of short chain organic acids after the reaction was measured.

Characterization of the Catalyst
To have an insight into the features of the WO 3 /nγ-Al 2 O 3 catalyst, characterization was conducted. The SEM image (Figure 1a) shows the appearance of the prepared catalyst. It can be seen that the shape of the catalyst is roughly sphere-like.
The diameter of the catalyst is around 50 nm (Figure 1a). XPS analyses of the WO 3 /nγ-Al 2 O 3 catalyst were also conducted. Figure 1b depicts the W 4f core level XPS spectra. There are two symmetric peaks at binding energies 35.9 and 37.4 eV, which are associated with the W-Al band and WO 3 , respectively [45,46]. The O 1s XPS profile ( Figure 1c) shows two symmetric peaks at 530.8 and 532.7 eV. These two peaks correspond to WO 3 and Al 2 O 3 , respectively [47,48]. Al 2p XPS spectra profile ( Figure 1d) shows two peaks at 74.5 and 75.9 eV, which are associated to the Al-W band and Al 2 O 3 [45,49], respectively. The XPS spectra profiles support the fact that a certain amount of active phase (WO 3 ) has been loaded on the carrier (nγ-Al 2 O 3 ). It can be inferred from the results that the three elements (W, O, and Al) are interconnected in terms of WO 3 and Al 2 O 3 and the former two compounds are probably connected by the W-Al band. The diameter of the catalyst is around 50 nm (Figure 1a). XPS analyses of the WO3/nγ-Al2O3 catalyst were also conducted. Figure 1b depicts the W 4f core level XPS spectra. There are two symmetric peaks at binding energies 35.9 and 37.4 eV, which are associated with the W-Al band and WO3, respectively [45,46]. The O 1s XPS profile ( Figure 1c) shows two symmetric peaks at 530.8 and 532.7 eV. These two peaks correspond to WO3 and Al2O3, respectively [47,48]. Al 2p XPS spectra profile ( Figure 1d) shows two peaks at 74.5 and 75.9 eV, which are associated to the Al-W band and Al2O3 [45,49], respectively. The XPS spectra profiles support the fact that a certain amount of active phase (WO3) has been loaded on the carrier (nγ-Al2O3). It can be inferred from the results that the three elements (W, O, and Al) are interconnected in terms of WO3 and Al2O3 and the former two compounds are probably connected by the W-Al band.
XRF profile provides quantitative proportions of the surface compounds on the catalyst (Table  1) where the weight percentage of WO3 is 4.73% on the catalyst surface, confirming that WO3 was formed and fixed on the surface of the catalysts. XRF profile provides quantitative proportions of the surface compounds on the catalyst (Table 1) where the weight percentage of WO 3 is 4.73% on the catalyst surface, confirming that WO 3 was formed and fixed on the surface of the catalysts.

Solar-Driven Removal of 1,4-D
Solar-driven oxidation of 1,4-D was firstly conducted without the catalyst. As can be seen in Figure 2a, less than 17% of mineralization can be achieved after 4 h. This limited removal is due to the non-photosensitive structure of 1,4-D [15,36]. A comparative experiment using the catalyst in dark conditions was also carried out to check the contribution of adsorption for 1,4-D removal. Figure 2a shows that almost no total organic carbon (TOC) removal can be observed under these conditions, indicating the negligible effect of adsorption. Then, 1,4-D was exposed under solar radiation in the presence of the WO 3 /nγ-Al 2 O 3 nano-catalyst. Under this condition, a significant TOC reduction (over 75%) was obtained following a reaction time of 4 h Figure 2a. The above results indicate that the current WO 3 /nγ-Al 2 O 3 nano-catalyst can effectively mineralize 1,4-D and the effect is mainly due to photocatalytic decomposition. (XRF).

Solar-Driven Removal of 1,4-D
Solar-driven oxidation of 1,4-D was firstly conducted without the catalyst. As can be seen in Figure 2a, less than 17% of mineralization can be achieved after 4 h. This limited removal is due to the non-photosensitive structure of 1,4-D [15,36]. A comparative experiment using the catalyst in dark conditions was also carried out to check the contribution of adsorption for 1,4-D removal. Figure 2a shows that almost no total organic carbon (TOC) removal can be observed under these conditions, indicating the negligible effect of adsorption. Then, 1,4-D was exposed under solar radiation in the presence of the WO3/nγ-Al2O3 nano-catalyst. Under this condition, a significant TOC reduction (over 75%) was obtained following a reaction time of 4 h Figure 2a. The above results indicate that the current WO3/nγ-Al2O3 nano-catalyst can effectively mineralize 1,4-D and the effect is mainly due to photocatalytic decomposition. To check the effect of catalyst dose on mineralization of 1,4-D, various amounts from 100 to 700 mg L −1 were tested (Figure 2b). It can be observed that increasing the catalyst dose can improve the treatment efficiency with the final mineralization extent within 67% and 85%. The TOC evolution was fitted to a zero-order kinetic model which has been used in previous photo-oxidation systems [26]. Table 2 collects the corresponding mineralization rate constants with the correlation coefficients (≥ 0.95). It also includes the rate constants obtained using 300 mg L −1 of the catalyst alone and using solar light alone. It can be seen that the rate constant of the run using 700 mg L −1 of the catalyst reaches up to 0.1059 mg L −1 s −1 , which is over 4 folds and 81 folds the results achieved when using 300 mg L −1 catalyst alone and using solar light alone, respectively. The results confirm the high efficiency of the current WO 3 /nγ-Al 2 O 3 nano-catalyst for improving the solar-driven photocatalytic systems for the breakdown of 1,4-D.

Oxidation State and Formation of Short-Chain Acids
The mean oxidation state carbon (MOSC) of a solution provides an overall oxidation state of all the compounds in the solution in terms of their averaged MOSC values [50]. MOSC value can be calculated considering both chemical oxidation demand (COD) and TOC via the following equation (Equation (1)).
A higher MOSC value indicates a higher oxidation state, whereas negative ones stand for a higher potential for further oxidation of the compounds in the solution. A previous study reported that the MOSC values of a chemical industrial wastewater containing 1,4-D were within the interval of [4,−4] and 1,4-D had an MOSC value of around 1 [36].
In the current study, to better reveal the oxidation state of the effluent during the course of the reaction, the evolution of solution MOSC was followed and the corresponding results are included in Figure 3a. As can be observed in the figure, the MOSC value of the solution increases from around 1.3 to approximately 3 after 4 h of reaction. This trend is in agreement with a previous work [36]. Extension of reaction time can hardly increase the MOSC value, indicating that the solution reaches a relatively stable oxidation state in the current photocatalytic system. Some short-chain organic acids were formed after 4 h, including acetic, formic, and fumaric ones (Figure 3b). Among them, acetic acid shows the highest concentration. These short-chain acids have been frequently reported as the final oxidation byproducts in advanced oxidation systems with low ecotoxicity [16,18,51]. It can be concluded that the high MOSC value after reaction must be associated with the formation of these reaction byproducts.  In the current study, to better reveal the oxidation state of the effluent during the course of the reaction, the evolution of solution MOSC was followed and the corresponding results are included in Figure 3a. As can be observed in the figure, the MOSC value of the solution increases from around 1.3 to approximately 3 after 4 h of reaction. This trend is in agreement with a previous work [36]. Extension of reaction time can hardly increase the MOSC value, indicating that the solution reaches a relatively stable oxidation state in the current photocatalytic system. Some short-chain organic acids were formed after 4 h, including acetic, formic, and fumaric ones (Figure 3b). Among them, acetic acid shows the highest concentration. These short-chain acids have been frequently reported as the final oxidation byproducts in advanced oxidation systems with low ecotoxicity [16,18,51]. It can be concluded that the high MOSC value after reaction must be associated with the formation of these reaction byproducts.

Reusability of the Catalyst
Reusability is an important factor impacting on the potential for practical application of a photocatalyst. In that respect, 300 mg L −1 catalyst was added to the system to lead the solar-driven decomposition of 1,4-D for 4 h. Then, the catalyst was separated and dried at 60 °C after reaction, and the same amount of 1,4-D was added again to repeat the reaction. As is shown in Figure 4, a TOC removal of 72.1% is achieved after 7 times of reuse of the catalyst, showing fairly desirable reusability. The slight decline of mineralization extent may be due to the inactivation of the catalyst by adsorption of some reaction byproducts on the active sites.

Reusability of the Catalyst
Reusability is an important factor impacting on the potential for practical application of a photocatalyst. In that respect, 300 mg L −1 catalyst was added to the system to lead the solar-driven decomposition of 1,4-D for 4 h. Then, the catalyst was separated and dried at 60 • C after reaction, and the same amount of 1,4-D was added again to repeat the reaction. As is shown in Figure 4, a TOC removal of 72.1% is achieved after 7 times of reuse of the catalyst, showing fairly desirable reusability. The slight decline of mineralization extent may be due to the inactivation of the catalyst by adsorption of some reaction byproducts on the active sites.

Postulation of Photocatalytic Mechanisms
The results so far prove that the current WO3/nγ-Al2O3-based photocatalytic system can effectively remove 1,4-D and the byproducts include some short-chain organic acids. HO • and O2 •− /HO2 • radicals must be responsible for the high efficiency of the current system. These reactive free radicals can be generated through the following scheme: the photon from solar light with certain energy excites the electrons, causing them to transit from the valence band (VB) to the conduction band (CB) and thus, holes are formed on VB. The electron-hole pairs can further excite the dissolved oxygen and water to form ROS ( Figure 5). The electrons have a strong reducing ability and can promote the formation of O2 •− /HO2 • radicals from dissolved oxygen. In the meantime, the holes act as positions with high oxidation ability and thus tend to take electrons from water molecules to generate HO • radicals. In addition, the combination of excessive HO • radicals may occur to further form hydrogen peroxide (H2O2) [34,[52][53][54]. The above mechanisms enable the current system for the efficient mineralization of the target pollutant 1,4-D in the absence of oxidant ( Figure 5).

Postulation of Photocatalytic Mechanisms
The results so far prove that the current WO 3 /nγ-Al 2 O 3 -based photocatalytic system can effectively remove 1,4-D and the byproducts include some short-chain organic acids. HO • and O 2 •− /HO 2 • radicals must be responsible for the high efficiency of the current system. These reactive free radicals can be generated through the following scheme: the photon from solar light with certain energy excites the electrons, causing them to transit from the valence band (VB) to the conduction band (CB) and thus, holes are formed on VB. The electron-hole pairs can further excite the dissolved oxygen and water to form ROS ( Figure 5). The electrons have a strong reducing ability and can promote the formation of O 2 •− /HO 2 • radicals from dissolved oxygen. In the meantime, the holes act as positions with high oxidation ability and thus tend to take electrons from water molecules to generate HO • radicals. In addition, the combination of excessive HO • radicals may occur to further form hydrogen peroxide (H 2 O 2 ) [34,[52][53][54]. The above mechanisms enable the current system for the efficient mineralization of the target pollutant 1,4-D in the absence of oxidant ( Figure 5).

Postulation of Photocatalytic Mechanisms
The results so far prove that the current WO3/nγ-Al2O3-based photocatalytic system can effectively remove 1,4-D and the byproducts include some short-chain organic acids. HO • and O2 •− /HO2 • radicals must be responsible for the high efficiency of the current system. These reactive free radicals can be generated through the following scheme: the photon from solar light with certain energy excites the electrons, causing them to transit from the valence band (VB) to the conduction band (CB) and thus, holes are formed on VB. The electron-hole pairs can further excite the dissolved oxygen and water to form ROS ( Figure 5). The electrons have a strong reducing ability and can promote the formation of O2 •− /HO2 • radicals from dissolved oxygen. In the meantime, the holes act as positions with high oxidation ability and thus tend to take electrons from water molecules to generate HO • radicals. In addition, the combination of excessive HO • radicals may occur to further form hydrogen peroxide (H2O2) [34,[52][53][54]. The above mechanisms enable the current system for the efficient mineralization of the target pollutant 1,4-D in the absence of oxidant ( Figure 5).

Materials
Na 2 WO 4 ·2H 2 O was purchased from Sigma-Aldrich company, St. Louis, MO, USA. 1,4-D, nano-size γ-Al 2 O 3 (nγ-Al 2 O 3 , diameter < 20 nm) and other reagents were supplied by Aladdin Reagent company, Shanghai, China. All the reagents were of AR grade and used without further purification. Ultrapure water was used throughout the experiments.
The WO 3 /nγ-Al 2 O 3 catalyst was prepared by incipient wetness impregnation. Na 2 WO 4 ·2H 2 O (0.08 M) was added drop by drop onto nγ-Al 2 O 3 to adjust W load to achieve nominal 5% (w/w). The impregnated nγ-Al 2 O 3 was shaken at room temperature at 60 rpm for 2 h and then dried at 60 • C for 12 h. After that, the sample was calcinated at 300 • C with a heating rate of 4.5 • C min −1 from room temperature. The total calcination time was 8 h.

Characterization of the Catalyst
Surface morphology of the catalyst was observed using SEM (Zeiss Merlin VP Compact, Oberkochen, Germany).
XPS (PHI Quantera SXM, produced by ULVAC-PHI, Chigasaki, Japan) was used to analyze the components on the surface of the catalyst. An Al-Kα X-radiation source was used with a vacuum in the analysis chamber lower than 1.0 × 10 −7 Pa. The X-ray beam spot size was 200 µm at an incident angle of 45 • with pass energy of 55 eV and step length of 0.1 eV. XPS profiles of W 4f, O 1s and Al 2p were recorded and fitted using XPS Peak 4.1 software.
Quantitative analysis of the compounds on the catalyst surface was conducted using XRF (Shimadzu XRF-1800 SEQUENTIAL, Kyoto, Japan). Based on the protocol, the weight percentage of each compound was presented.

Solar-Driven Photocatalytic Processes
The solar-driven photocatalytic experiments were conducted in cylinder quartz reactors with a diameter of 6 cm and a height of 7 cm, which were placed in a solar simulator. For a typical run, 100 mL 1,4-D (50 mg L −1 ) and certain amount of catalyst (100 to 700 mg L −1 ) were added in the reactor with magnetic stirring at 200 rpm. Then, the lamp was turned on to start the reaction. The light source was a Xe lamp with an irradiation intensity of 40 mW cm −2 . The spectrum of the light source is in the range of 190-1100 nm with peaks at 436 nm and 546 nm. The initial pH value of the 1,4-D solution was around 6.8 and not adjusted during the reaction. Solution temperature was controlled by water circulation (< 40 • C after 4 h of reaction), during which no obvious evaporation was observed.

Analytical Methods
TOC of solution was measured using a TOC analyzer (Shimadzu, model 5000A, Kyoto, Japan). COD was determined by a Hach COD reactor (DRB200, Hach, Loveland, CO, USA) equipped with a spectrophotometer. Organic short-chain acids were measured by an ionic chromatography (IC) equipped with a conductivity detector (Thermo ICS5000+, Waltham, MA, USA). 3.2 mM Na 2 CO 3 and 1mM of NaHCO 3 were used as mobile phases at pumping rate of 0.7 mL min −1 .

Conclusions
The current work reports an oxidant-free solar-driven photocatalytic system for the breakdown of 1,4-D based on a homemade WO 3 /nγ-Al 2 O 3 nano-catalyst. More than 75% mineralization extent can be achieved using 300 mg L −1 catalyst at solar light intensity of 40 mW m −2 after 4 h of irradiation. Increasing the dose of catalyst to 700 mg L −1 can improve the TOC removal to 85%. The MOSC values of the solution were followed and it was found that the oxidation state was greatly raised after reaction. The short-chain acids formed during the reaction-including acetic, formic and fumaric acids-were Catalysts 2019, 9, 389 9 of 11 believed to be associated with the high solution oxidation state. A reusability study indicates that the current catalyst can still be efficient after being reused seven times.
The current work provides a promising approach to eliminate 1,4-D from water for potable reuse. It is not only oxidant-free but can also make use of solar light radiation to lead 1,4-D mineralization. However, it should be noticed that around 20% of organic carbons still remain in the solution after treatment. In that sense, future studies should be conducted to either analyze the composition of these organic carbons or evaluate their ecotoxicity.