Heterogeneous photocatalytic degradation of anthraquinone dye Reactive Blue 19: optimization, comparison between processes and identification of intermediate products

Treatment of textile wastewater using heterogeneous photocatalysis began in the the last decade and attracted the attention of researchers due to its versatile application. The variety of applications of TiO2 as a photocatalyst was due toits numerous positive properties, such as low operating temperature, biologically inert nature, low energy consumption, water insolubility, availability and photoactivity, low toxicity, high chemical stability, suitable flat band potential, narrow bandgap and the fact that it is environmentally benign. Heterogeneous UV-TiO2 photocatalysis is capable of removing organic pollutants from textile wastewater; this has been widely studied, with the technology also having been commercialized in many developing countries. Decolorization of anthraquinone dye Reactive Blue 19 (RB 19) by heterogeneous advanced oxidation processes TiO2/UV/H2O2, TiO2/UV/KBrO3 and TiO2/UV/(NH4)2S2O8 was studied under different conditions and in the presence of electron acceptors such as hydrogen peroxide (H2O2), potassium bromate (KBrO3) and ammonium persulphate ((NH4)2S2O8). Decolorization was very fast for all three processes, and complete dye decolorization was achieved in 10 min. The effect of various ions (Cl–, SO4 2– and HCO3 –) on RB 19 decolorization was also studied. The optimal condition for the decolorization of the dye were determined to be: TiO2 concentration 1 g∙dm–3, electron acceptor concentration 30.0 mmol∙dm–3, dye concentration 50.0 mg∙dm–3, UV intensity 1 950 μW∙cm–2, at temperature 25 ± 0.5°C. In addition, experiments were performed and compared in three different matrices. In the surface water and dyebath effluent water, the removal efficiency for RB 19 was lower than that achieved in the deionized water because of the interference of complex constituents in the surface water and effluent. LC-MS analysis was carried out and the detected intermediates were compared with the previously published data for anthraquinone dyes.


INTRODUCTION
Textile industries play a vital role in the economic development of many developing countries and therefore also in increasing the gross domestic products of these countries (Masum, 2016). These industries use different raw materials, such as cotton, synthetic and woollen fibres, and chemicals including dyes. Approximately 10 000 different synthetic dyes are available in the market and worldwide annual production of these dyes is over 700 000 t. Nearly 200 000 t of synthetic dyes are lost into the environment because of the inefficient dyeing process used in textile industries. According to World Bank estimates, about 17-20% of industrial wastewater is generated from textile dyeing and finishing treatments (Holkar et al., 2016;Hossain et al., 2018;Ribeiro et al., 2017).
Thus, though the textile industry provides significant economic benefits, it also faces the environmental and social impacts associated with the generation of toxic wastewaters from its processing operations, such as de-sizing, sizing, scouring, bleaching, mercerizing, dyeing, printing, finishing and other processes (Masum, 2016;Miguel et al., 2002;Ledakowicz et al., 2001;Punzi et al., 2015). In order to meet the colour requirement, reactive and azo dyes are highly water-soluble and therefore around 10-20% of the used dye is washed out with water as effluents which are hazardous (carcinogenic or mutagenic) and toxic to the environment (Ganesh et al., 1994;Weber and Adams, 1995;Zhang et al., 1998).
Semiconductor TiO 2 -based photocatalysis has received much attention because of its properties such as non-toxicity, chemical and biological stability, low cost and higher photocatalytic activity. However, for the technical and commercial feasibility of the process, extensive investigations are needed to overcome some problems with TiO 2 -based photocatalysts (Sharotri and Sud, 2017).
One of the major disadvantages of heterogeneous photocatalysis is the recombination of the photo-generated electron (e cb − ) and hole (h vb + ). This step decreases the quantum yield and causes energy wasting. These could be overcome by using electron acceptors or hole scavengers. The addition of the electron acceptor, such as KBrO 3 , H 2 O 2 and (NH 4 ) 2 S 2 O 8 enhanced the degradation rate by (i) preventing the electron-hole recombination by accepting the conduction band electron; (ii) increasing the hydroxyl radical concentration; and (iii) generating other oxidizing species (SO 4 • -) to improve the efficiency of intermediate compounds (Wei et al., 2009).
The objective of the present study was to evaluate the effectiveness of heterogeneous advanced oxidation processes TiO 2 /UV/KBrO 3 , TiO 2 /UV/H 2 O 2 and TiO 2 /UV/(NH 4 ) 2 S 2 O 8 at degrading the anthraquinone dye Reactive Blue 19 (RB 19). For all three processes, the influence of background ions Cl -, SO 4 2and HCO 3 -, which can compete with the target contaminant for reaction with radicals and holes, was examined. The effect of these ions has not been reported in the literature. Therefore, the influence of the aqueous matrix should also be considered when applying TiO 2 /UV/KBrO 3 , TiO 2 /UV/H 2 O 2 and TiO 2 / UV/(NH 4 ) 2 S 2 O 8 in practice. For this reason, experiments were performed in three different matrixes (laboratory deionized water, surface water collected from the local Nišava River and dyebath effluent water from a local cotton dyeing facility).
Most investigators have provided information on dyes removal during the degradation process and not much information has been provided about the degradation pathway or intermediate compound formation by heterogeneous advanced oxidation processes TiO 2 /UV/KBrO 3 , TiO 2 /UV/H 2 O 2 and TiO 2 /UV/ (NH 4 ) 2 S 2 O 8 . Hence in the present study, an attempt was made to identify the intermediate compound formed during the dye photocatalytic degradation by using LC-MS analysis.

Photoreactor
Photochemical experiments were carried out in a batch photoreactor handmade in our laboratory (Mitrovic et al., 2012). The UV lamps were turned on 10 min before performing each experiment. The intensity of UV radiation was measured by a UV radiometer Solarmeter model 8.0 UVC (Solartech, USA). The total UV intensity was controlled by turning on different numbers of UV lamps and the maximum intensity was 1 950 μW•cm -2 (with all ten UV lamps on) at a distance of 220 mm from the working solution surface.

Procedures
A stock solution of RB 19 was prepared by dissolving 1.0 g dye in 1 000.0 cm -3 of deionized water. Working solutions were freshly prepared before irradiation, by diluting the stock to the desired concentration with deionized water. The pH of the solutions was adjusted by addition of NaOH or HCl (0.1/0.01 mol•dm -3 ) with pH/ISE meter (Orion Star A214, Thermo scientific, USA). The suspensions of dye and TiO 2 were magnetically stirred in the dark for 30 min to attain adsorption-desorption equilibrium between dye and TiO 2 , and the dye solutions were then treated in the UV reactor.
During irradiation, the solution was magnetically stirred (Are, Velp Scientifica, Italy) at a constant rate, and temperature was maintained at 25 ± 0.5°C by thermostatting. At required time intervals, 4.0 cm -3 of samples were withdrawn, centrifuged (3 000 r•min -1 , 15 min) and filtered through a 0.20 μm regenerated cellulose membrane filter (Agilent Technologies, Germany) to separate the catalyst. Absorbance at 592 nm was measured using a UV-vis spectrophotometer Shimadzu UV-1800 PC (Shimadzu, Japan) to determine the degree of decolorization of the solution. The removal (%) of RB 19 dye was calculated as: where c 0 and c t are the concentration values of the dye solution before and after UV irradiation, respectively.
Identifying the degradation product of the RB 19 solution was carried out by LC-MS system. After treatment, the samples of RB 19 were processed using LCQ Fleet mass spectrometer (Thermo Fisher Scientific, USA) with orthogonal electrospray (ESI) source and ion trap (IT) as an analyser. The LCQ Fleet mass spectrometer was linked to the HPLC system (Ultimate 3000, Thermo Fisher Scientific, USA). Thermo Scientific column Dionex Hipersil GOLD C18 was used for separation in a liquid chromatograph. The samples of anthraquinone dye RB 19 were analysed in a negative mode of the mass spectrometer.
To ensure the accuracy, reliability, and reproducibility of the collected data, all experiments were carried out in triplicate, and mean values are recorded. OriginPro 2016 (OriginLab Corporation) software was used for statistical analysis and calculation of the data.

Influence of experimental parameters on removal efficiency of RB 19
The preliminary experiments were carried out in order to investigate the effect of UV radiation only, TiO 2 without UV https://doi.org/10.17159/wsa/2020.v46.i2.8245 radiation, electron acceptors (KBrO 3 , H 2 O 2 and (NH 4 ) 2 S 2 O 8) without UV radiation and UV irradiation in the presence of TiO 2 and electron acceptors. The solution of RB 19 dye (initial dye concentration was 50.0 mg•dm -3 ) was irradiated for 180 min to examine the effect of UV light radiation alone, and there was no observable decrease in residual dye concentration. This indicated that the direct photolysis of RB 19 dye by UV irradiation was slow. Experiments with only electron acceptors were done for 180 min in the dark. The dye removal efficiency, in that case, was also negligible. No decolorization was observed for the dye solution with TiO 2 without UV radiation. But if electron acceptors are applied in combination with UV radiation and TiO 2 , residual dye concentration rapidly decreases (Fig. 1).
Complete decolorization was obtained in less than 15 min, with an initial dye concentration of 50.0 mg•dm -3 in the presence of 1.0 g•dm -3 TiO 2 and 30.0 mmol•dm -3 electron acceptors and under 1 950 μW•cm -2 light intensity.
Titanium dioxide has drawn much attention from the industry as a good candidate for a large band-gap semiconducting oxide for photodecomposition processes in pollutant treatment, because of its favourable physical/chemical properties, low cost, ease of availability, and high stability Yang et al., 2013;Xu et al., 2013;Cetinkaya et al., 2013). Numerous studies in the literature show that TiO 2 and other catalysts remain unchanged after the photocatalytic process, confirming their stability even after several cycles of use. (Wan et al., 2012;Li and Wu, 2017;Zhu et al., 2011). Because of all the abovementioned factors, TiO 2 was selected as a photocatalyst in the processes examined in this manuscript. The process optimization for the best dye removal efficiency is presented below.

Effect of electron acceptors
BrO 3 ion is an efficient electron acceptor and is used to enhance photocatalytic decolorization rate (Poulios and Tsachpinis, 1999;Gratzel et al., 1990;Sanchez et al., 1998). The effect of the addition of KBrO 3 (10.0-100.0 mmol•dm -3 ) on removal efficiency of RB 19 is shown in Fig. 2. It can be seen that adding a small amount of KBrO 3 , from 10.0 to 30.0 mmol•dm -3 ,increases the decolorization from 88.62% to 96.38% during the time period of 10 min. The enhancement of the removal efficiencyis due to the reaction between BrO 3 ion and conduction band electron (Eq. 2) (Wei et al., 2009;Muneera and Bahnemannb, 2002;San et al. 2001) (2) With the further increase in KBrO 3 concentration from 30.0 to 100.0 mmol•dm -3 the removal efficiency is almost constant. This is due to the adsorption effect of Brions on the TiO 2 surface, which affects the catalytic activity of TiO 2 (San et al., 2001). So the optimum concentration of bromate ion is 30.0 mmol•dm -3 for photocatalytic decolorization of RB 19.
The photocatalytic decolorization of dye occurs on the surface of TiO 2 , and O 2 and H 2 O 2 are necessary for photocatalytic decolorization (Vesely et al., 1991;Chen and Liu, 2007;Dionysiou et al., 2000;Houas et al., 2001): H 2 O 2 may be photolyzed or react with superoxide anion to form hydroxyl radical directly: H 2 O 2 is also a powerful hole scavenger. In excess, it may react with holes to produce oxygen and protons. In photocatalytic decolorization, the hole directly oxidizes the dye and with water produces a hydroxyl radical. Hence, the removal efficiency for the dye decreases due to the removal of holes (Pichat et al., 1995).
The effect of the addition of S 2 O 8 2on the photolytic oxidation of RB 19 (50.0 mg•dm -3 ) was investigated by varying the amount of (NH 4 ) 2 S 2 O 8 from 10.0 to 100.0 mmol•dm -3 . The results are shown in Fig. 2. The addition of (NH 4 ) 2 S 2 O 8 from 10.0 to 30.0 mmol•dm -3 increases the removal efficiency from 93.63% to 100.00% within 10 min. These results are in agreement with earlier research (Poulios andTsachpinis, 1999, Sanchez et al., 1998). With the further increase of (NH 4 ) 2 S 2 O 8 concentration from 30.0 to 100.0 mmol•dm -3 the removal efficiency is almost constant. Addition of persulfate to photocatalytic processes enhances the decolorization rate in the following ways: (10) In the reaction with a photogenerated electron and with a water molecule, the sulfate radical anion SO 4 •can generate a hydroxyl radical. The sulfate radical anion is a strong oxidant and participates in the decolorization process. At a high dosage of S 2 O 8 -2 , inhibition of reaction occurs due to the increase in the concentration of the SO 4 2ion. On the surface of the TiO 2 catalyst there is absorption of the SO 4 2ions. Therefore, catalytic activity is reduced. On the other hand, the adsorbed SO 4 -2 ion also reacted with photogenerated holes (Eq. 11) and hydroxyl radical (Eq. 12).

Effect of initial RB 19 concentration
Pollutant concentration is a very important parameter in wastewater treatment. The effect of initial dye concentration on decolorization was investigated over a range of 10.0 to 100.0 mg•dm -3 . The removal efficiency of RB 19 by TiO 2 /UV/KBrO 3 after 10 min of treatment is shown in Table 1. The results indicate that whileincreasing the initial concentration of dye from 10.0 mg•dm -3 to 100.0 mg•dm -3 removal efficiency decreased from 100.0% to 82.55%.
The results after 10 min of treatment by TiO 2 /UV/H 2 O 2 process are shown in Table 1. With the increase in the initial concentration of dye from 10.0 mg•dm -3 to 100.0 mg•dm -3 , removal efficiency decreases from 95.49% to 39.14%. At high concentrations the penetration of photons entering into the solution decreases, consequently lowering the hydroxyl radical concentration (Ghodbane and Hamdaoui, 2010;Muruganandham and Swaminathan, 2004b;Behnajady et al., 2004;Galindo and Kalt1998). It can be seen that removal efficiency decreased as initial dye concentration increased at the same concentration of (NH 4 ) 2 S 2 O 8 .

Effect of UV radiation intensity
By increasing UV radiation intensity, the efficiency of dye decolorization increases considerably. Based on the obtained results, which are shown in Fig. 3, it can be concluded that removal efficiency of dye increases with the increase in radiation intensity from 730 μW•cm -2 to 1 950 μW•cm -2 . The lowest difference within the process efficiency is between the radiation intensity 1 750 μW•cm -2 and 1 950 μW•cm -2 , from which follows that the back lamps in photoreactor have the least contribution to dye decolorization. Results have also shown that the UV intensity tested in the study lies in the linear range and all the photons produced are effectively used (Fig. 3).
The increase in light intensity from 730 μW•cm -2 to 1 950 μW•cm -2 increases the decolorization by TiO 2 /UV/H 2 O 2 process from 29.70 to 87.79% within 10 min. The investigation is consistent with previous studies which generally observed an increase in decolorization rate with increasing UV intensity (Mills et al., 1993;Lea and Adesina, 1998). This is a consequence of a higher quantity of generated •OH radicals, which make oxidative decolorization of anthraquinone dye more efficient.
The influence of UV light intensity on the decolorization of RB 19 by TiO 2 /UV/KBrO 3 and TiO 2 /UV/(NH 4 ) 2 S 2 O 8 processes has been monitored by varying the UV radiation intensity as in previous experiments, and similar results were obtained.

Comparison of decolorization by TiO 2 /UV/KBrO 3 , TiO 2 /UV/ H 2 O 2 and TiO 2 /UV/(NH 4 ) 2 S 2 O 8
In order to optimize the process, a comparison was made between the three heterogeneous oxidation processes after  10 min of treatment under given conditions. The initial electron acceptor concentrations in these experiments were 30.0 mmol•dm -3 , and the amount of TiO 2 was 1.0 g•dm -3 . In the case of photocatalytic decolorization, S 2 O 8 -2 is the most effective for the photodecolorization of RB 19 among the additives studied in this paper. The decolorization efficiencies of RB 19 are in the following order TiO 2 /UV/(NH 4 ) 2 S 2 O 8 (100%) > TiO 2 /UV/KBrO 3 (96.38%) > TiO 2 /UV/H 2 O 2 (87.79%).
Electron acceptors such as hydrogen peroxide, potassium bromate and ammonium persulfate were added into the solution in order to enhance the decolorization (Poulios and Tsachpinis, 1999;Gratzel et al., 1990;Sanchez et al., 1998). All the additives showed a beneficial effect on the decolorization of the dye, whereas S 2 O 8 -2 has been found to remarkably enhance the decolorization of pollutant. The efficiency of the TiO 2 /UV/H 2 O 2 process is comparable to the TiO 2 /UV/KBrO 3 process.
TiO 2 -based photocatalysts also offer advantages such as high physical and chemical stability, low cost, availability, low toxicity, and excellent photoactivity (Banerjee et al., 2014). However, purification of water and wastewater using the TiO 2 / UV/KBrO 3 process leads to the formation of bromide ions (Lv et al., 2008). Although bromide ion are not harmful to the human body, they can be converted to bromated and other brominated pollutants (Haag and Holgne, 1983;Gunten and Oliveras, 1998) which have suspected carcinogenic potential. Therefore, it is necessary and significant to remove the DBP (disinfection by-product) precursor bromide.
A traditional precursor removal strategy (enhanced coagulation) and novel precursor removal strategy (anion exchange such as activated carbon adsorption processes) are two areas of active research for controlling DBP formation (Johnson and Singer, 2004;Boyer and Singer, 2005). Also, Brions are absorbed on the surface of TiO 2 and the ability to convert them into BrO 3 ions is reduced. In the case of TiO 2 /UV/H 2 O 2 , the final products of dyes degradation are carbon dioxide, water and inert salts (Muruganandham et al., 2004b, Sharma et al., 2016. The persulfates have high solubility and stability at ambient temperature, while the sulfate ions, which are the major products of persulfate reduction, are relatively harmless and considered to be environmentally friendly (Peternel et al., 2012, Olmez-Hanci et al., 2014. Therefore, these processes are a promising environmental engineering technique.

Effect of salt addition and decolorization test of RB 19 in surface water and dyebath effluent water
Starting from the assumption that the typical constituents of natural water and wastewater (CO 3 2-, HCO 3 -, SO 4 2-, Cl -, NO 3 -, HPO 4 2-, H 2 PO 4 -) can influence the rate of decolorization of the tested substrates, the effects of different concentrations of some ions were studied. Decolorization experiments were performed by dissolving 50.0 mg•dm -3 of dye in deionized water. The added amount of catalyst was 1.0 g•dm -3 and the initial pH was 7.0. The obtained results are shown in Table 2.
The decrease in decolorization efficiency of the dye is due to the hole scavenging and hydroxyl radical scavenging properties of chloride and sulfate ions (Wei et al., 2009;Wenhua et al., 2000). The presence of bicarbonate increased the decolorization efficiency. Bicarbonate ions react with hydroxyl radical and produce carbonate radical, CO 3 • - (Aleboyeh et al., 2012). The carbonate radical is a strong oxidant and very selective for organic compounds.
After 10 min irradiation, the TiO 2 /UV/KBrO 3 process achieved 96.38%, 89.63% and 79.99% RB 19 removal for the DW, SW, and DEW, respectively;the TiO 2 /UV/H 2 O 2 process achieved 81.84%, 77.34% and 69.56% for the DW, SW, and DEW, respectively; and the TiO 2 /UV/(NH 4 ) 2 S 2 O 8 process 100.0%, 96.56% and 86.35% for the DW, SW, and DEW, respectively. As shown in Fig. 4, in the surface water and dyebath effluent the efficiency of removal of RB 19 was lower than that achieved in the deionized water because of the interference of complex constituents in the surface water and dyebath effluent.

LS-MS analyses
On the mass spectra obtained after the applied heterogeneous oxidation processes, signals are observed at similar m/z values, so it can be assumed that the degradation of the RB 19 dye by the applied processes is probably carried out by a similar mechanism.
After the preliminary fragmentation of RB 19, the samples obtained during treatment with selected heterogeneous advanced oxidation processes were analysed. In Fig. 5b it can be seen from the mass spectrum of the sample after 2 min of treatment that a new ion at m/z of 499.1 was reported, compared to the spectrum of untreated dye sample. MS 2 ion analysis on m/z 499.1 gave ions on m/z 435.1 and m/z 408.0. These ions are found at m/z values greater than Δm/z 16 of the fragments at m/z 419.1 and 393.0 of the untreated RB 19 dye sample. Based on these facts, it can be assumed that the formation of monohydroxylated products has occurred. In the mass spectrum of the sample after 4 min of treatment, the peak intensity at m/z 602.9 was significantly reduced, and in the spectrum there was a peak at m/z 317.1 (Fig.5c). The presence of this ion indicates one of the possible mechanisms of degradation where there is a breakdown of the relationship between the carbon of the aromatic nucleus and nitrogen. MS 2 fragmentation of the peak at m/z 317.1 gave an ion on m/z 253.2.
After 6 min of treatment, in the mass spectrum of the dye RB 19, no further signal is available on m/z 602.9 (Fig. 5 (d)). A new signal is an output at m/z 515.1. MS 2 analysis of the ion on m/z 515.1 gave a peak at m/z 451.1. The ion on m/z 451.1 is found at m/z values greater than Δm/z 16 from ion to m/z 435.1 identified by the MS 2 analysis of the ion at m/z 499.1 after 2 min of colour treatment, which is probably due to the attachment of another •OH radical to an anthraquinone nucleus and the formation of a di-hydroxylated degradation product. MS 3 ion analysis on m/z 451.1 gave ions on m/z 424.9, 377.2 and 360.2, while MS 2 ion analysis on m/z 499.1 obtained peaks at 435.1 and 408.1.
After 10 min of treatment, no new signals were detected, and the intensities of all previously detected peaks were significantly reduced. After a longer treatment time (60 min), the signals of all detected ions have disappeared, indicating further oxidative degradation of intermediate products. Further degradation leads to the formation of low molecular weight aldehydes, organic acids, nitrate and sulfate that cannot be detected by this technique (Amorisco et al., 2011(Amorisco et al., , 2013. These results are consistent with the results for the change in RB 19 dye concentration over time (Fig. 1), which show a significant drop in dye concentration at the same time. Based on the structure of the intermediate degradation products identified by the ESI/IT technique, a possible mechanism of degradation of the anthraquinone dye RB 19 can be predicted (Fig. 6).

CONCLUSIONS
The decolorization of the RB 19 solutions by TiO 2 /UV/KBrO 3 , TiO 2 /UV/H 2 O 2 and TiO 2 /UV/(NH 4 ) 2 S 2 O 8 processes strongly depends on the system parameters, such as electron acceptors, dye initial concentration and radiation intensity. From an economic point of view, the TiO 2 /UV/(NH 4 ) 2 S 2 O 8 process emerges as the most attractive oxidation system for reactive dye effluents in terms of complete decolorization (100.00% in less than 10 min), very closely followed by the TiO 2 /UV/KBrO 3 process (96.44% after 10 min) and TiO 2 /UV/H 2 O 2 process (87.79% after 10 min). The presence of chloride and sulfate ions decreased the photocatalytic decolorization, while the presence of bicarbonate increased the decolorization efficiency. All three oxidation processes were carried out in three matrices (laboratory deionized water, surface water collected from the Nišava River and dyebath effluent water from a local cotton dyeing facility).
In the surface water and dyebath effluent, the removal efficiency of RB 19 was lower than that achieved in the deionized water because of the interference of complex constituents in the surface water and the dyebath effluent. Lastly, LS-MS analyses were carried out to identify the intermediates produced during dye degradation. At longer treatment times no organic by-products were identified by LS-MS.