Journal of Hazardous Materials 285 (2015) 127–136

Contents lists available at ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Photodegradation of malachite green under simulated and natural irradiation: Kinetics, products, and pathways Li Yong a , Gao Zhanqi b , Ji Yuefei a , Hu Xiaobin c , Sun Cheng a,∗ , Yang Shaogui a , Wang Lianhong a , Wang Qingeng a , Fang Die a a

State Key Laboratory of Pollution Control and Resource Reuse, School of the Environment, Nanjing University, Nanjing 210023, China State Environmental Protection Key Laboratory of Monitoring and Analysis for Organic Pollutants in Surface Water, Jiangsu Provincial Environmental Monitoring Center, Nanjing 210036, China c School of Life Science, Huzhou University, Huzhou 313000, China b

h i g h l i g h t s • • • • •

Photofate of malachite green was studied under simulated and natural irradiation. Favorable conditions for degradation were optimized by the orthogonal array design. Main ROS for the decomposition were determined by free radical quenchers. Fifty-three products were determined by LC–MS and GC–MS. Pathways were proposed with the aid of theoretical calculation.

a r t i c l e

i n f o

Article history: Received 24 August 2014 Received in revised form 24 November 2014 Accepted 30 November 2014 Available online 3 December 2014 Keywords: Photodegradation Malachite green Simulated/natural conditions Products Pathways

a b s t r a c t In this work photodegradation rates and pathways of malachite green were studied under simulated and solar irradiation with the goal of assessing the potential of photolysis as a removal mechanism in real aquatic environment. Factors influencing the photodegradation process were investigated, including pH, humic acid, Fe2+ , Ca2+ , HCO3 − , and NO3 − , of which favorable conditions were optimized by the orthogonal array design under simulated sunlight irradiation in the presence of dissolved oxygen. The degradation processes of malachite green conformed to pseudo first-order kinetics and their degradation rate constants were between 0.0062 and 0.4012 h−1 . Under solar irradiation, the decolorization efficiency of most tests can reach almost 100%, and relatively thorough mineralization could be observed. Forty degradation products were detected by liquid chromatography–mass spectrometry, and thirteen small molecular products were identified by gas chromatography–mass spectrometry. Based on the analyses of the degradation products and calculation of the frontier electron density, the pathways were proposed: decomposition of conjugated structure, N-demethylation reactions, hydroxyl addition reactions, the removal of benzene ring, and the ring-opening reaction. This study has provided a reference, both for photodegradation of malachite green and future safety applications and predictions of decontamination of related triphenylmethane dyes under real conditions. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Traditionally, malachite green (MG) is an organic dye for materials such as silk, leather, and paper. In recent decades, however,

Abbreviations: MG, malachite green; LMG, leucomalachite green; NOM, natural organic matter; DLBP, 4-(dimethylamino)benzophenone; FED, frontier electron density. ∗ Corresponding author. Tel.: +86 25 89680258; fax: +86 25 89680580. E-mail address: [email protected] (S. Cheng). http://dx.doi.org/10.1016/j.jhazmat.2014.11.041 0304-3894/© 2014 Elsevier B.V. All rights reserved.

it has emerged as a controversial agent in aquaculture due to its role as an antibacterial and parasiticide agent, and its mutagenicity in Salmonella typhimurium TA98 reflected marked cytotoxicity and induced cell transformation and lipid peroxidation [1]. MG cannot undergo efficient biodegradation, because its BOD5 is almost equal to zero. However, in aqueous solution, MG and MG leucocarbinol can convert into each other, the equilibrium being pH-dependent [2]. MG leucocarbinol not only spreads quickly across cell membranes, but also metabolizes into leucomalachite green (LMG) in fish muscles (Fig. 1). Being non-polar, LMG is found

128

L. Yong et al. / Journal of Hazardous Materials 285 (2015) 127–136

Fig. 1. Conversions between MG (left), MG leucocarbinol (middle), and LMG (right).

to be retained in catfish muscle longer (t1/2 = 10 days) [3] than MG (t1/2 = 2.8 days). In fishes’ intestines, MG can be reduced to LMG metabolites, remaining in catfish muscle with a half-life of about 10 days and even longer in fat and organs [3]. The potential carcinogenic, genotoxic, mutagenic, and teratogenic properties of its metabolites were demonstrated in many animal species [4–7]. Consequently, it is not permitted in the USA and EU. Nevertheless, it is still widely used for fishery disinfection because of its low price and high efficiency, but few studies have been involved in its degradation in natural conditions. It is still indispensable to conduct thorough research on its transportation and transformation in natural and simulated waters. Previous literature on degradation of MG indicated that pure oxidization, electrochemical treatment, and biodegradation had weaknesses, such as low degradation rate, high investment, etc. [8]. A number of studies focused on the photodegradation of MG by photocatalysis with TiO2 , K3 PW12 O40 , etc. [9–13]. Some researchers also studied the adsorption by different materials to remove MG from aqueous solutions [14,15]. However, the photodegradation of MG in the actual or simulated environment was rarely reported, except for one case that MG was photodegraded under merely natural sunlight [16]. Although MG possesses relatively low bioaccumulation and hydrolysis, it can be degraded by photolysis in water, which is the primary feature that draws our attention. Photolysis is an important removal pathway of organic compounds besides hydrolysis, adsorption, volatilization, and biodegradation owing to the ubiquitousness of light in natural processes. Direct photolysis occurs when the absorption spectrum of one compound overlaps with the solar emitting spectrum. Indirect photolysis takes place in the presence of reactive oxygen species generated from photosensitizers such as nitrate and dissolved organic matter. For those organic compounds unable to undergo direct photolysis, indirect photolysis plays a particularly important role. Several important reactive oxygen species are generated as a consequence of light absorption in the aquatic environment. HO• , 1 O2 , HO2 • /O2 •− , 3 NOM* , and eaq − are prevalent and generally respond to oxidization of organic pollutants in the environment. It is expected that the photodegradation of MG could be accelerated, and it could be mineralized into CO2 and H2 O under simulated and/or natural conditions. Considering MG as a pollutant in dye wastewater and a bactericide for fish, it is crucial to investigate its transportation and transformation process under natural conditions. Never before has photodegradation of MG under simulated or natural conditions been reported. In this study, photodegradation of MG was explored under simulated and natural conditions. Experimental parameters such as pH, humic acid, cationic, and anionic ions were investigated by the orthogonal array design in an attempt to better understand the transformation process in water matrix. The photodegradation intermediates of MG have been identified by LC-ESI–MS and GC–MS. Possible mechanism of the solar photodegradation was

proposed via kinetics, theoretical calculation, and identification of degradation products. 2. Experiment 2.1. Reagents Malachite Green chloride, humic acid, isopropanol 1,4diazabicyclo[2.2.2]octane (DABCO), 1,4-benzoquinone (BQ), formic acid, ammonium acetate, and acetonitrile were used. For detailed information, please refer to the Supplementary material. 2.2. Orthogonal array design In this work, the orthogonal array design was employed to investigate the extent of MG degradation under simulated natural environment. Moreover, the influencing factors and their importance for MG degradation under simulated natural conditions were also studied according to the orthogonal test results. The ambient pH of natural waters varies from about 5 to 9 and pKa of MG is 6.9 [2]. In the pH range of 5–9, the main species of ions in surface water are ferrous, bicarbonate, and nitrate [17]. The concentrations of the main ions were set based on literature and the data of main waters of China, i.e., the concentration of humic acid is between 0.3 and 30 mg/L [18]; Fe2+ is about 0.1 mM [19–21]; Ca2+ is 32.2 (the Yellow River) to 48.1 mg/L (the Yangtze River), HCO3 − is 118 (the Yellow River) to 228 mg/L (the Yangtze River) [22], and NO3 − is 0.1 to 1 mM [23,24]. Therefore, one L25 (56 ) orthogonal array design with six factors at five levels was performed to optimize the favorable conditions of photodegradation of MG under simulated and natural irradiation as in Table S1. The five levels for controlling factors are 5–8 for pH, 0–40 mg/L for humic acid, 0–0.12 mM for Fe2+ , 0–60 mg/L for Ca2+ , 0–200 mg/L for HCO3 − , and 0–1.2 mM for NO3 − . And the statistical software Minitab 16.0 was executed to process the data analysis. Humic acid was considered because it is one of the most important light absorbers that may induce indirect photolytic transformations of organic pollutants in natural waters; iron in different speciations could be the source of HO• ; Ca2+ represents the hardness of water and it is responsible for complexation and chelating; NO3 − is an important ion affecting indirect photolysis and induces production of reactive radicals such as HO• , NO• , and NO2 • ; HCO3 − was considered one scavenger of HO• . 2.3. Degradation studies Under the dark condition, the hydrolysis test was conducted as follows: a test tube with glass stopper containing 50 mL of 10 mg/L MG solution entirely enclosed by tinfoil was placed in a temperature-controlled reactor. The absorbance variation was measured by UV–vis spectrometer at regular time intervals, and

L. Yong et al. / Journal of Hazardous Materials 285 (2015) 127–136

degradation intermediates were separated and identified by LCESI–MS and GC–MS. The simulated experiment was carried out on a XPA-7 photoreactor (Nanjing Xujiang Electromechanical Plant, China) containing a Xenon lamp. This apparatus was set near permanently-open windows to keep the temperature close to the atmospheric environment. The intensity of the artificial sunlight source was measured with a radiometer (400 <  < 1000 nm, Photoelectric Instrument Factory of Beijing Normal University), which covered the main part of sunlight spectrum. The light intensity was continuous at 400 W/m2 during the process. The corresponding light dose of 1 h’s irradiation was 1440 kJ/m2 . The relative intensity of the light of Xenon lamp was measured by an S3000-VIS spectrometer (Hangzhou Seeman Technology Co., Ltd., China), and the UV–vis absorption spectrum of MG and emission spectrum of Xenon lamp were demonstrated in Fig. S1. The Xenon lamp was placed inside a quartz cooling well with constant water circulation, keeping the temperature fluctuation less than 3 ◦ C. Capped quartz tubes containing 50 ml of 10 mg/L reaction solution were set in a rotating unit at a fixed distance of 5 cm to the light source. Filters ( > 290 nm) were set between the light source and tubes to guarantee the spectrum approximate to the wavelength of natural sunlight. All 25 tests were divided into five groups according to the orthogonal array design. As soon as the solutions of each group were prepared, the irradiation started. Each supernatant photoreaction aliquot (3 ml) was sampled at different irradiation time points to measure the UV–vis absorptions. The experiment under the solar irradiation was carried out from March 6 to 9, 2013 for four days. All 25 test tubes were installed on the roof top of the building of School of the Environment, Nanjing University, China. After each day’s irradiation, 3 ml supernatant samples of photoreaction aliquots were collected to measure on the UV–vis spectrometer. The intensity of solar radiation was measured by a pyranometer (285 <  < 2800 nm, CMP11, Kipp and Zonen, Netherlands). This apparatus operated uninterruptedly only 10 meters away from the experiment location. The total accumulated irradiation was 52.14 MJ/m2 , the mean duration of sunlight was 12.5 h with the maximum solar radiation 716 W/m2 and the mean sunlight intensity 176 W/m2 , during the daytime. The temperature was monitored by a probe (Vaisala HUMICAP® Humidity and Temperature Probe HMP155A, Finland) and the average temperature during the process was about 22.5 ◦ C. To determine the reactive oxygen species generated and reacted in the photolysis process, free radical scavengers IPA, DABCO, and BQ were added respectively, in contrast with the control group. The conditions for simulated light experiment and solar irradiation were identical in accordance with the orthogonal array design. Influencing factors such as pH, humic acid, and ions were investigated and favorable conditions were optimized by the orthogonal array design. The concentration of samples was tested by a UV–vis spectrometer.

129

The pre-treatment process for GC–MS was as follows: 10 mL filtered solution was extracted with 10 mL n-hexane three consecutive times, and the extracted solutions were dehydrated using anhydrous sodium sulphate. Afterward, the dehydrated solution was concentrated to 1 mL and blown dry with nitrogen flow. Before GC–MS analysis, trimethylsilylation was carried out at 50 ◦ C for 30 min using 0.5 mL bis(trimethylsilyl)trifluoroacetamide (BSTFA). 2.5. Analytical methods The pH determination and adjustment of the solutions were carried out using Shanghai Leici ZDJ-5 automatic titrator. It was calibrated with standard buffers (pH 4.00, 6.86, 9.18) every time before measurements. UV–vis spectrometer (Shimazu UV-1800) was applied to determine the absorptions to calculate the kinetic rate constants and residue concentrations of MG during the degradation process. LC-ESI–MS analysis was performed by a Thermo Finnigan Surveyor Modular HPLC system equipped with a Finnigan LCQ Advantage MAX ion trap mass spectrometer and HPLC column (Agilent Zorbax SB-C18, 5 ␮m, 2.1 × 150 mm) to detect and identify polar degradation products, such as MG and N-demethylated intermediates. The mobile phase was a mixture of water (A) and acetonitrile (B) at 0.2 mL/min of the flow rate, and the gradient was set as follows: t = 0 min, A:B = 5:95 (V/V); t = 20 min, A:B = 95:5 (V/V). The column temperature was 30 ◦ C and the injection volume was 10.00 ␮l using the autosampler. The mass fragments to monitor were from m/z 50 to m/z 400. GC–MS analysis was conducted by a Thermo Trace gas chromatography Ultra interfaced with an ISQ mass spectrometer equipped with DB-5 fused-silica capillary column (30 m × 0.25 mm i.d., 0.25 ␮m film thickness) to detect small molecular products during the ring-opening reaction. The final trimethylsilylated sample (1.0 ␮l) was automatically injected into the column with splitless mode. The oven temperature was programmed as follows: the initial temperature was 60 ◦ C, then 60–300 ◦ C at ramp rate of 10 ◦ C min−1 , held 10 min. The MS was operated with 70 eV electron impact (EI) and positive ion modes. 2.6. Theoretical calculation Molecular orbital calculations were carried out at B3LYP/6311 + G* level with the optimal conformation having a minimum energy obtained in the Gaussian 09 program. The frontier electron densities (FEDs) of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) were determined. Values of FED2 HOMO + FED2 LUMO were calculated to predict the reaction sites for hydroxyl addition. 3. Results and discussion 3.1. Kinetics

2.4. Sample preparation for the degradation products At first, the SPE cartridges (Waters Oasis MCX column, 6 ml/150 mg) were equilibrated by 3 ml acetonitrile and 3 ml 2% (V/V) formic acid solution. Next, 10 ml final reaction solutions were loaded through the cartridges at 0.2 ml/min. Subsequently, 2 ml 2% (V/V) formic acid/acetonitrile solution and 6 ml acetonitrile were added in turn at 2 ml/min to perform washing process. Afterwards, 4 ml 5% (V/V) ammonium acetate (5 mol/L, pH 7.0)/methanol solution was added at 1 ml/min to complete the elution. Finally, the eluents were transferred into rotary evaporator flasks and evaporated at 40 ◦ C until almost dry. The final product was prepared into 1 ml acetonitrile solution and analyzed by LC-ESI–MS.

3.1.1. Hydrolysis of MG After 40 h hydrolysis, 81.20% of MG was still remaining. Hydrolysis was fit by first-order kinetics model and its apparent rate constant kh was 0.0192 h−1 (R2 = 0.9409) as shown in Fig. S2. The spectra variation (Fig. 2) shows no blue shift of the major peaks at about 315 nm, 425 nm, and 618 nm, but the peaks exhibit a slight decrease trend. Meanwhile a slight increase trend between 200 nm and 280 nm is observable, indicating the slight increase trend of benzene rings and conjugated structure with double bonds. This observation is consistent with the LC–MS results (Fig. S9). Based on the structure of MG, the triarylmethyl cation is stabilized by the conjugation that delocalizes the positive charge, and Cl− is a good leaving group. Therefore, the hydrolysis mechanism

130

L. Yong et al. / Journal of Hazardous Materials 285 (2015) 127–136

Fig. 2. Spectra variation of MG during hydrolysis.

should be SN 1 mechanism and the reaction rate is unrelated to the concentration of the nucleophile H2 O as Eqs. (1) and (2). The pH varied from 8.98 to 7.87 in the reaction process, which conformed to the presumption above.

MG+ Cl− → MG+ + Cl−

(1)

MG+ + H2 O → MGOH + H+

(2)

3.1.2. Direct photolysis of MG In the direct photolysis test, 83.40% of MG was still remaining after 40 h. The direct photolysis process conformed to first-order kinetics with the rate constant kdp of 0.0145 h−1 (Fig. S3). The spectra (Fig. 3) show the same trends as those in hydrolysis. Compared with hydrolysis under dark condition, the direct photolysis rate constant kdp of 0.0145 h−1 (R2 = 0.9989) was slightly smaller than hydrolysis rate constant kh of 0.0192 h−1 (R2 = 0.9409), and a slight increase trend of the spectra between 340 nm and 380 nm was observed, indicating the formation of DLBP (4-(dimethylamino)

Fig. 3. Spectra variation of MG during direct photolysis.

L. Yong et al. / Journal of Hazardous Materials 285 (2015) 127–136

benzophenone). This observation was explained by the LC–MS results (Fig. S12). 3.1.3. Indirect photolysis of MG It is possible that MG could undergo indirect photolysis through reactions with reactive oxygen or transient species (e.g., 1 O2 , HO• , and 3 DOM*) produced by photosensitizers which widely exist in natural waters (e.g., NO3 − , DOM, and Fe3+ ). The indirect photolysis under simulated sunlight in laboratory included tests from 2 to 25 in Table S2. The degradation processes followed pseudo first-order kinetics, and the rate constants of degradation kobs were between 0.0062 and 0.4012 h−1 , with half-life 111.62–1.73 h (Table S2). It can be seen that as the humic acid concentration increased, the decolorization rate did not increase correspondingly, but have a maximum value of Test 2–5, and Test 4 represented the favorable condition. This could be explained by the fact that humic acid can induce organic pollutant photodegradation through direct electron energy transfer or through indirect oxidation with photo-generated reactive oxygen species [25]. Meanwhile, humic acid may also play a negative role in photodegradation by inhibiting the transformation of oxidation intermediates, competing with organic contaminants to absorb the photons [26]. 3.1.4. Reactive oxygen species during indirect photolysis Radical scavengers were employed as diagnostic tools to verify the indirect photolysis mechanism and investigate the importance of certain kinds of radicals since these scavengers selectively react with radicals to form stable or persistent intermediates [27,28]. This experiment was carried out in the presence of IPA as the hydroxyl radical (HO• ) quencher, DABCO as the singlet oxygen (1 O2 ) quencher and BQ as the superoxide anion (O2 •− ) quencher, respectively. According to the consequences of indirect photolysis, Test 4 represents the favorable condition of degradation after irradiation of 40 h. Therefore, keeping the MG concentration 10 mg/L and taking Test 4 as the control group, three tubes were added with 5 ml IPA, DABCO (100 mg/L), and BQ (100 mg/L), respectively. The result is illustrated in Fig. 4. On the basis of these results, we conclude that HO• and O2 •− existed in the photolysis process while 1 O2 did not participate in the oxidation process. As a powerful oxidant, HO• is capable to unselectively react with most of the aquatic organic compounds with a secondary rate constant of 107 –1010 M−1 s−1 [29]. O2 •− can be generated via the reaction of eaq − in the presence of oxygen.

Fig. 4. Quenching effects for MG.

131

At acidic condition, O2 •− undergoes rapid protonation to produce H2 O2 through the intermediate hydroperoxyl radical (HO2 • ), which is capable of forming a stronger oxidant HO• upon irradiation. Previous studies [16,30] have also pointed out the photo-induced production of these two radicals upon MG in solutions containing dissolved oxygen as Eqs. (3)–(7). MG∗ + O2 → MG O2

•−

•+

+ O2

•−

(3)

+ H+  HO2 •

•

(4)

•

HO2 + HO2 → H2 O2 + O2 O2

•−

+ H2 O2 →

H2 O2 + hv →

HO•

(5)

−

+ OH + O2

(6)

HO•

(7)

3.2. Results of orthogonal array design and data analysis As indicated by our results, the photodegradation efficiencies of MG depended on factors that can be controlled by variations of pH, humic acid, cationic, and anionic ions in different conditions. Therefore, data analysis of the orthogonal array design by Minitab 16.0 was employed to evaluate the effect of various parameters on the photodegradation. Range analysis is a visual method of orthogonal experiment by calculating the maximum and minimum gap (range) to determine the primary and secondary indicators of various factors on the experiment. The value of the range reflects the importance of the factor; the larger the range is, the more important the factor is. Table 1 Products determined by LC–MS. Compound

Tentative structure

RT (min)

m/z

ESI/MS

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

MG MG − CH2 MG − 2CH2 MG − 3CH2 MG − 4CH2 MG − 4CH2 − NH LMG LMG − CH2 MG + OH MG + OH − CH2 MG + OH − 2CH2 MG + OH − 3CH2 MG + OH − 4CH2 MG + OH − 4CH2 − NH MG + 2OH MG + 2OH − CH2 MG + 2OH − 2CH2 MG + 2OH − 3CH2 MG + 2OH − 4CH2 MG + 2OH − 4CH2 − NH MG + 3OH MG + 3OH − CH2 MG + 3OH − 2CH2 MG + 3OH − 3CH2 MG + 3OH − 4CH2 MG + 3OH − 4CH2 − NH DLBP DLBP − CH2 DLBP − 2CH2 DLBP − 2CH2 − NH DLBP + OH DLBP + OH − CH2 DLBP + OH − 2CH2 DLBP + OH − 2CH2 − NH DLBP + 2OH DLBP + 2OH − CH2 DLBP + 2OH − 2CH2 DLBP + 2OH − 2CH2 − NH CH3 COOH DLBP − 2H + O

8.68 2.47 2.35 2.16 1.58 1.79 7.64 3.44 2.15 7.39 3.44/9.19 2.60 2.27 2.35 2.32/8.73 2.92/3.74 2.37 2.14 2.03 2.19 1.82/2.26 2.30/2.76 2.28 2.26 2.20 2.19 2.48 2.18 2.01 1.86 2.00 1.81 1.75 2.19 2.24 2.11 1.89 1.92 1.87 2.24

329 315 301 287 273 258 330 316 345 331 317 303 289 274 361 347 333 319 305 290 377 363 349 335 321 306 225 211 197 182 241 227 213 199 257 243 229 214 60 239

329.4 315.3 301.3 287.2 273.2 258.2 331.4 317.3 345.1 331.2 317.2/317.3 303.3 289.4 274.3 361.2 347.2 333.2 319.2 305.1 290.4 377.3/377.1 363.3/363.1 349.0 335.1 321.0 306.2 226.2 212.3 198.2 183.6 242.2 228.3 214.0 200.5 258.2 244.1 230.3 215.2 60.1 240.3

132

L. Yong et al. / Journal of Hazardous Materials 285 (2015) 127–136

Table 2 Products determined by GC–MS. Compound

Retention time(min)

Structure of derivatives

Derivatives’ fragments (m/z, abundance)

1

5.06

147(999) 73(546) 191(240)

2

5.82

73(999) 117(696) 147(691)

3

5.86

140(999) 73(895) 174(869)

4

5.92

147(999) 73(924) 148(165)

5

6.05

75(999) 73(823) 173(578)

6

7.04

147(999) 73(785) 117(513)

7

7.28

73(999) 174(744) 59(293)

8

8.14

117(999) 73(750) 147(560)

9

8.27

179(999) 105(974) 77(710)

10

8.51

73(999) 147(495) 205(384)

11

8.92

102(999) 147(715) 73(485)

12

18.12

120(999) 197(564) 92(146)

L. Yong et al. / Journal of Hazardous Materials 285 (2015) 127–136

133

Table 2 (Continued) Compound

Retention time(min)

13

19.43

Structure of derivatives

This study has selected signal/noise (S/N) values (larger is better) to quantify the range. The quantification results are listed in Table S3, and it can be speculated that the sequence of the influencing factors for the photodegradation of MG is pH, HCO3 − , humic acid, Ca2+ , NO3 − , and Fe2+ . The results indicated that the optimal pH was 5, and pH values of 6–8 were also beneficial to the reaction; however, pH of no adaptation was less beneficial to the reaction, indicating the variation of pH was favorable to the degradation of MG. As the concentration of HCO3 − increased, the reactions were less favorable; this phenomenon coincides with the fact that HCO3 − is one scavenger of free radicals reducing reaction rate. Humic acid could induce organic pollutant photodegradation through direct electron energy transfer or through indirect oxidation with photo-generated reactive oxygen species; meanwhile, the increasing trend of humic acid reflected an inhibiting effect in the reaction process [26]. Generally, the existence of Ca2+ did not significantly affect the reaction efficiency, and it was believed that Ca2+ was responsible for complexation and chelating with humic acid, indicating in the actual environment, the existence of Ca2+ may influence the reaction. NO3 − was not favorable to the degradation to a large extent, and the variation of concentration did not evidently affect the reaction. This can be explained by the competition for light between NO3 − and MG and the formation of HO• by photo-excited NO3 − . HO• can promote MG degradation, however, the reduced fraction of light due to NO3 − absorption may suppress MG direct photolysis. During the reaction processes, Fe2+ was oxidized to Fe3+ , and precipitations can be observed. The competition for oxidizers or free radicals unquestionably reduces the oxidation rate. 3.3. Data analyses of UV–vis spectra during solar photolysis Figs. S4–S8 describes the variation of absorption spectra of MG during the solar photodegradation process. During the 72 h irradiation, the spectra exhibit a decrease trend with slight or intense blue shift of the major peak at about 618 nm of MG, which indicated the process of N-demethylation, i.e., the non-selective attack of reactive oxygen species on C N bond as reported [31]. The absorbance peaks at 425 and 315 nm decline (Figs. S4–S8), which evidently indicate the whole conjugated chromophore structure of MG has been destroyed. Fig. S5 illustrates the spectra between about 200 nm and 240 nm is remarkably lower than others (Figs. S4, S6–S8). It can be concluded that the existence of NO3 − leads to the increasing inclination of spectra between about 200 nm and 240 nm, which represents the increasing trend of benzene rings and conjugated structure with double bonds. In Fig. S6, it can be observed that a new absorbance peak appears at about 360 nm. Based on previous research [31,32], the attack on central carbon could generate DLBP as one of the main products with obvious absorbance at about 360 nm. Fig. S8 illustrates a noticeable increase on peak area between about 200 nm and 420 nm, representing different conjugated

Derivatives’ fragments (m/z, abundance)

148(999) 225(641) 77(281)

structures, benzene rings or other chromophores. It can be speculated that the high concentrations of Fe2+ and NO3 − were responsible for the increase of peak area. From the analyses above, competition between the cleavage of the central carbon reaction and N-demethylation reaction or other unclear reactions might exist, which constitutes the main part of the degradation mechanism. 3.4. Degradation products Unlike the experiment under simulated sunlight, the experiment under natural solar irradiation was for identification of degradation products and speculation of degradation mechanism afterwards. Most products during photodegradation processes were easily recognized in TIC, however, species with low concentrations can be found only by searching typical ion fragments according to common molecule moieties. This can help us identify minor products in selected ion chromatography. 3.4.1. Results from LC–MS identification MG hydrolysis occurred during dark reaction in water matrix. Intermediates (Fig. S10) were detected as follows: (1) 2.88 min, m/z = 329.4, MG; m/z = 347.3, MG leucocarbinol. (2) 3.47 min, m/z = 315.4, MG − CH2 ; 329.4, MG. (3) 4.37 min, m/z = 331.3, LMG. (4) 4.80 min, m/z = 329.4, MG; m/z = 361.2, MG + 2OH. These results conformed to the fact that MG, MG leucocarbinol, and LMG can transform naturally to each other in water matrix. Under natural solar irradiation, MG in most tests has attained almost 100% decolorizations as in Fig. S11, and the rest were also close to being colorless. Although the decolorization of all experiments under simulated sunlight could not reached 100%, those under solar irradiation did that. Test 1 represented direct photolysis while Test 2–25 experienced indirect photolysis. As illustrated in Figs. S12 and S13, the main differences between direct and indirect photolysis on TIC can be seen from retention time 4 min to 7 min. The intermediates of direct photolysis were: (1) 4.30 min, m/z = 226.2, DLBP. (2) 4.70 min, m/z = 333.2, MG + 2OH − 2CH2 ; m/z = 315.4, MG − CH2 . (3) 5.39 min, m/z = 334.2, unknown. (4) 5.72 min, m/z = 301.3, MG − 2CH2 ; m/z = 333.2 MG + 2OH − 2CH2 ; (5) 6.15 min, m/z = 329.4, MG. In order to identify intermediates as many as possible, a mixture sample composed of indirect photolysis was analyzed by LC–MS. The total ion chromatography was illustrated in Fig. S13 and forty products determined were displayed in Table 1. Compared with direct photolysis, indirect photolysis detected many more products indicating indirect photolysis was more effective in degradation of MG. 3.4.2. Results from GC–MS identification As same as LC–MS detection, a mixture sample composed of indirect photolysis was analyzed by GC–MS, and the structures of derivatives conformed by GC–MS were listed in Table 2. As illustrated in Figs. S15 and S16, the products of direct and indirect

134

L. Yong et al. / Journal of Hazardous Materials 285 (2015) 127–136

Fig. 5. Degradation pathways of MG under solar radiation.

L. Yong et al. / Journal of Hazardous Materials 285 (2015) 127–136

photolysis with low molecular mass were basically identical; however, the TIC of hydrolysis differed (Fig. S14). EI full-scan provided adequate structural information for unequivocal identifications. Products were identified by an identification program of the U.S. National Institute of Standards and Technology (NIST) library. The results of GC–MS detection indicated that most species were carboxylic acid, alcohol, and amine, and these species are not likely to cause severe hazards to the environment. This proved that MG can be degraded into small innocuous molecules under compatible conditions. However, benzophenones, such as DLBP and its derivatives engendered from the degradation process, which have a sweet aroma and were used in additives of soap or perfumes. These chemicals are also environmental endocrines and might induce male infertility.

3.4.3. Calculation of FED In order to correctly characterize positions of hydroxylation, a simple method is theoretical calculations. In the present study, the frontier electron densities (FEDs) of MG and DLBP were calculated to predict the reaction sites for HO• attack as summarized in Tables S4 and S5. According to Frontier Orbital Theory, positions with higher values of 2FED2 HOMO are more easily subject to electron extraction, while positions with higher value of FED2 HOMO + FED2 LUMO are more susceptible to HO• attack. As shown in Tables S4 and S5, for MG (Fig. S17), 10C, 12C, 14C, 15C, 17C, and 19C have the relatively highest FED2 HOMO + FED2 LUMO value (in bold), while 7C, 8C, 9C, 11C, 13C, 16C, and 18C have the relatively higher FED2 HOMO + FED2 LUMO value (underlined); for DLBP (Fig. S18), 4C, 5C, and 6C have the relatively highest FED2 HOMO + FED2 LUMO value (in bold), while 8C, 11C, and 14C have the relatively higher FED2 HOMO + FED2 LUMO value (underlined), which is indicative of the high possibility of hydroxylation formation with HO• attack occurring at aromatic ring. These results are favorable to the speculation of degradation pathways.

3.5. Pathways Main degradation pathways of MG during natural photolysis under different conditions were the decomposition of conjugated structures, N-demethylation reactions and HO• addition reactions, which are illustrated in Fig. 5. N-Demethylation reactions: the HO• radicals attacked the N,N-dimethyl position, inducing the reactions such as demethylation and even deamination, which reduced the polarity of molecules. Unfortunately, this process could not effectively induce the decolorization because the chromophore structure could not be destructed in this way. HO• addition reactions: due to the non-selectivity of this kind of radical, HO• could be added to MG and MG-derivative structure in 1:1–1:3 ratio. With the help of calculations of FEDs of MG and DLBP, the reaction sites for HO• attack could be predicted. Decomposition of conjugated structure: the HO• radicals attacked the central carbon, resulting in the cleavage of central carbon structure, which led to the decomposition of conjugated structure and the generation of DLBP. In this way, the substance MG in the solution was diminished rapidly, which is the decreasing trend of the major peak in UV–vis spectra. Afterwards, the removal of benzene ring and the ring-opening reaction took place (Fig. 5), which engendered small organic molecules, and expectedly thorough mineralization into CO2 and H2 O. Also, all the products from N-demethylation reactions were degraded by decomposition of conjugated structure, the removal of benzene ring and ring-opening reaction into small molecules.

135

4. Conclusion In different conditions under simulated sunlight, the degradation of MG could be significant in some processes such as Test 4, leading to up to 95.62% removal of MG at most. The degradation processes conformed to pseudo first-order kinetics. Under solar irradiation, the decolorization efficiencies of 17 tests can reach 100% after 72 h irradiation and relatively thorough mineralizations could be observed, indicating that solar irradiation was much more effective than our artificial sunlight source in the degradation process of MG. Forty intermediates and products were detected by LC-ESI–MS and thirteen kinds by GC–MS for MG photodegradation, of which N-demethylation and hydroxyl addition products were prevalent, and a pivotal intermediate DLBP decomposed into carboxylic acids, alcohols, and amines. On the basis of calculations of FEDs of MG and DLBP predicting the reaction sites for HO• attacks, possible pathways were speculated as the decomposition of conjugated structure, N-demethylation reactions, hydroxyl addition reactions, the removal of benzene ring, and the ring-opening reaction. This study has illustrated that photodegradation under natural conditions is an important way to remove MG. Suitable conditions accelerate the degradation process of MG in natural waters so as to reduce dye pollution and potential toxicity. Further investigation by these means could be performed to test and verify the decontamination of other dyes. Acknowledgements This work was supported by National Natural Science Foundation of China (No. 21177055) and Jiangsu Provincial Science and Technology Supporting Program (No. BE2012116). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhazmat. 2014.11.041. References [1] S. Srivastava, R. Sinha, D. Roy, Toxicological effects of malachite green, Aquat. Toxicol. 66 (2004) 319–329. [2] A.R. Fischer, P. Werner, K.U. Goss, Photodegradation of malachite green and malachite green carbinol under irradiation with different wavelength ranges, Chemosphere 82 (2011) 210–214. [3] G.Y. Chen, S.I. Miao, HPLC Determination and MS confirmation of malachite green, gentian violet, and their leuco metabolite residues in channel catfish muscle, J. Agric. Food Chem. 58 (2010) 7109–7114. [4] S.J. Culp, F.A. Beland, R.H. Heflich, R.W. Benson, L.R. Blankenship, P.J. Webb, P.W. Mellick, R.W. Trotter, S.D. Shelton, K.J. Greenlees, M.G. Manjanatha, Mutagenicity and carcinogenicity in relation to DNA adduct formation in rats fed leucomalachite green, Mutat. Res. Fundam. Mol. Mech. Mutagen. 506–507 (2002) 55–63. [5] A.A. Bergwerff, P. Scherpenisse, Determination of residues of malachite green in aquatic animals, J. Chromatogr. B 788 (2003) 351–359. [6] R.A. Mittelstaedt, N. Mei, P.J. Webb, J.G. Shaddock, V.N. Dobrovolsky, L.J. McGarrity, S.M. Morris, T. Chen, F.A. Beland, K.J. Greenlees, R.H. Heflich, Genotoxicity of malachite green and leucomalachite green in female Big Blue B6C3F(1) mice, Mutat. Res. Genet. Toxicol. Environ. Mutagen. 561 (2004) 127–138. [7] S.J. Culp, P.W. Mellick, R.W. Trotter, K.J. Greenlees, R.L. Kodell, F.A. Beland, Carcinogenicity of malachite green chloride and leucomalachite green in B6C3F1 mice and F344 rats, Food Chem. Toxicol. 44 (2006) 1204–1212. [8] K.S. Wang, C.L. Lin, M.C. Wei, H.H. Liang, H.C. Li, C.H. Chang, Y.T. Fang, S.H. Chang, Effects of dissolved oxygen on dye removal by zero-valent iron, J. Hazard. Mater. 182 (2010) 886–895. [9] Y. Bessekhouad, D. Robert, J. Weber, Bi2 S3 /TiO2 and CdS/TiO2 heterojunctions as an available configuration for photocatalytic degradation of organic pollutant, J. Photochem. Photobiol. A 163 (2004) 569–580. [10] C.C. Chen, C.S. Lu, Y.C. Chung, J.L. Jan, UV light induced photodegradation of malachite green on TiO2 nanoparticles, J. Hazard. Mater. 141 (2007) 520–528.

136

L. Yong et al. / Journal of Hazardous Materials 285 (2015) 127–136

[11] S. Malato, P. Fernandez-Ibanez, M.I. Maldonado, J. Blanco, W. Gernjak, Decontamination and disinfection of water by solar photocatalysis: recent overview and trends, Catal. Today 147 (2009) 1–59. [12] K. Natarajan, T.S. Natarajan, H.C. Bajaj, R.J. Tayade, Photocatalytic reactor based on UV-LED/TiO2 coated quartz tube for degradation of dyes, Chem. Eng. J. 178 (2011) 40–49. [13] T. Ochiai, A. Fujishima, Photoelectrochemical properties of TiO2 photocatalyst and its applications for environmental purification, J. Photochem. Photobiol. C 13 (2012) 247–262. [14] P. Janos, H. Buchtova, M. Ryznarova, Sorption of dyes from aqueous solutions onto fly ash, Water Res. 37 (2003) 4938–4944. [15] A. Salima, B. Benaouda, B. Noureddine, L. Duclaux, Application of Ulva lactuca and Systoceira stricta algae-based activated carbons to hazardous cationic dyes removal from industrial effluents, Water Res. 47 (2013) 3375–3388. [16] L.A. Perez-Estrada, A. Aguera, M.D. Hernando, S. Malato, A.R. Fernandez-Alba, Photodegradation of malachite green under natural sunlight irradiation: kinetic and toxicity of the transformation products, Chemosphere 70 (2008) 2068–2075. [17] W. Stumm, J.J. Morgan, Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, John Wiley & Sons, 2012. [18] Y. Zuo, R.D. Jones, Photochemistry of natural dissolved organic matter in lake and wetland water—production of carbon monoxide, Water Res. 31 (1997) 850–858. [19] W. Xinying, Effects of humic acid and iron on photodegradation of atrazine, Chin. J. Environ. Eng. 6 (2012). [20] D. Xu, W. Wu, S. Ding, Q. Sun, C. Zhang, A high-resolution dialysis technique for rapid determination of dissolved reactive phosphate and ferrous iron in pore water of sediments, Sci. Total Environ. 421–422 (2012) 245–252. [21] S. Schnell, S. Ratering, K.H. Jansen, Simultaneous determination of iron(III), iron(II), and manganese(II) in environmental samples by ion chromatography, Environ. Sci. Technol. 32 (1998) 1530–1537. [22] W. Jiasheng, Experimental study on effect of positively charged ion in river on the velocity of sediment particles, Adv. Water Sci. 16 (2005).

[23] Y. Zuo, C. Wang, T. Van, Simultaneous determination of nitrite and nitrate in dew rain, snow, and lake water samples by ion-pair high-performance liquid chromatography, Talanta 70 (2006) 281–285. [24] Y. Zuo, K. Zhang, Y. Deng, Occurrence and photochemical degradation of 17␣-ethinylestradiol in Acushnet River Estuary, Chemosphere 63 (2006) 1583–1590. [25] Y. Chen, C. Hu, X. Hu, J. Qu, Indirect photodegradation of amine drugs in aqueous solution under simulated sunlight, Environ. Sci. Technol. 43 (2009) 2760–2765. [26] J. Wenk, U. von Gunten, S. Canonica, Effect of dissolved organic matter on the transformation of contaminants induced by excited triplet states and the hydroxyl radical, Environ. Sci. Technol. 45 (2011) 1334–1340. [27] Y. Chen, S. Yang, K. Wang, L. Lou, Role of primary active species and TiO2 surface characteristic in UV-illuminated photodegradation of Acid Orange 7, J. Photochem. Photobiol. A 172 (2005) 47–54. [28] A.C. Herath, R.M.G. Rajapakse, A. Wicramasinghe, V. Karunaratne, Photodegradation of triphenylamino methane (magenta) by photosensitizer in oxygenated solutions, Environ. Sci. Technol. 43 (2009) 176–180. [29] G.V. Buxton, C.L. Greenstock, W.P. Helman, A.B. Ross, Critical-review of rate constants for reactions of hydrated electrons, hydrogen-atoms and hydroxyl radicals (OH/O− ) in aqueous-solution, J. Phys. Chem. Ref. Data 17 (1988) 513–886. [30] V. Brezova, J. Pigosova, B. Havlinova, D. Dvoranova, M. Durovioc, EPR study of photochemical transformations of triarylmethane dyes, Dyes Pigm. 61 (2004) 177–198. [31] Y.M. Ju, S.G. Yang, Y.C. Ding, C. Sun, A.Q. Zhang, L.H. Wang, Microwave-assisted rapid photocatalytic degradation of malachite green in TIO2 suspensions: mechanism and pathways, J. Phys. Chem. A 112 (2008) 11172–11177. [32] Y.M. Ju, S.G. Yang, Y.C. Ding, C. Sun, C.G. Gu, Z. He, C. Qin, H. He, B. Xu, Microwave-enhanced H2 O2 − based process for treating aqueous malachite green solutions: intermediates and degradation mechanism, J. Hazard. Mater. 171 (2009) 123–132.