E1, E2, E3, BPA, HA, BSA and SA were purchased from Sigma-Aldrich (USA); EE2 was purchased from Dr. Ehrenstorfer (Germany). The ultraperformance liquid chromatography-mass spectrometry (UPLC/MS/MS) system (UPLC-Quattro Premier XE) was purchased from Waters (USA). The chromatographic columns (BEH C18) were purchased from Waters (USA). The solid-phase extraction chromatography system (Visiprep DL) was purchased from Supelco (USA). An ozone generator (DHX-SS-1G) from Harbin JiuJiu Electrochemical Engineering Technology Co., LTD (China), was used to generate ozone; the shift of the generator was adjusted to I, and the inlet gas was pure oxygen with a flow rate of 1 L· min^-1. The ozone water carried the ozone/oxygen mixture gas to the 4 ℃ ultrapure water to prepare ozone.

Ultrapure water, NaHCO₃, NaCl and CaCl₂ were added to the EDCs stock solution and organic substance stock solution to prepare the simulated secondary effluent working solution that contained the three organic matters. The solutions were prepared such that the total organic carbon (TOC)of each organic matter was 20 mg· L^-1 and the concentration of EDCs was 1 μ mol· L^-1, higher than the concentration in secondary effluent; however, this concentration is typical of small-scale experiments^{[10, 11]}. The working solution (10 mL) was first transferred to a 40 mL sample bottle that contained a magnetic bar, (10-V mL) ultrapure water was then added to the sample bottle, and V mL of ozone water was added to start the reaction. For the DO₃: TOC was 0, 0.2, 0.5, 0.8, 1.0 and 1.2, successively. Then, at 5, 10, 20, 30, 45, 60, 80, 100, 120 and 150 s, 0.2 mL of the above solution was transferred to a potassium indigotrisulfonate solution, and the indigo methodwas immediately used to determine the concentration of the D O3[12]. The reactor was stored overnight, and then the residual ozone was removed for the analysis EDCs concentrations. All experiments were repeated three times at room temperature.

1.3.1 Analysis methods for EDCs

Solid-phase extraction-chromatography-MS was used to quantify E1, E2, E3, EE2 and BPA in the solution simultaneously. The mobile phase consisted of acetonitrile (A) and ultrapure water (B); the analysis conditions are listed in Table 1.

Table 1

Table 1

Table 1 Analysis conditions of UPLC/MS/MS Technique Component Conditions UPLC analysis conditions Chromatographic column (100 mm× 2.1 mm I. D.× 1.7 μ m) 90%A 0 min→ 64%A 0.5 min→ maintained to 6 min→ 10% A 7min→ 0%10 min→ returned to the initial state 10.1 min→ maintained to 12 min Temperature of columns 50 ℃ Injection volume 10 μ L Flow rate of mobile phase 0.3 mL· min-1 MS analysis conditions Ionization mode Temperature of ion source ESI- 120 ℃

Table 1 Analysis conditions of UPLC/MS/MS

The removal efficiency of EDCs was used to evaluate the effect of ozonation degradation of EDCs; R= c0-cc× 100, where R is the removal rate of EDCs by ozone, c₀ is the initial concentration of EDCs, and c is the residual concentration of EDCs in the solution after ozonation degradation.

1.3.2 Spectral analysis

The scanning ranges of the three-dimensional excitation-emission matrix (EEM) fluorescence spectroscope (F-7000, Hitachi, Japan) were excitation wavelength/emission wavelengths (Ex/Em) of 220~420 nm (step of 5 nm)/240~600 nm (step of 2 nm). After obtaining the fluorescence intensity corresponding to each Ex/Em, Origin 8.0 was used to plot the contour map. UV-Vis spectroscopy (UV-2401PC, Shimadzu, Japan) was used to scan the full wavelength range of 200~800 nm.

After converting the concentsation to 10 mg· L^-1 TOC, the impact of the three organic matters on the attenuation of the DO₃ is shown in Fig.1. The difference between the ozone dosage concentration and the concentration of the DO₃ at 60 s is defined as the ozone demand (O3 demand, OD). In H₂O, when DO₃ was 2~12 mg· L^-1, the corresponding OD was 1.58~6.11 mg· L^-1, and more than half of the DO₃ was consumed. In SA, the attenuation of DO₃ was close to that of the H2O system. In HA, DO₃ was rapidly depleted, and the corresponding OD was 0.80~11.49 mg· L^-1. In the BSA medium, the DO₃ attenuation rate was between that of the two systems above. Overall, the organic matters significantly promoted the attenuation rate of the DO₃. Moreover, the effect produced by HA species was substantially higher than those by polysaccharides and proteins, which agrees with past research results^[3].

The effect of the three organic matters on the removal efficiency of the five EDCs was further investigated, and the results are shown in Fig.2. It can be seen from the figure that ozone exhibited excellent removal efficiency for EDCs. In addition, the species of organic matter in the solution yielded a significant effect on the degradation of EDCs. For DO₃: TOC≥ 0.8, the removal efficiency of EDCs reached above 90% in all three media. For DO₃: TOC=0.5, the removal efficiency of EDCs reached 90% in BSA and SA but was only 68%~73% in HA. For DO₃: TOC=0.2, there were significant residues of the five EDCs in HA and BSA; the removal efficiency of EDCs was 68%~76% in the BSA solution, while it was only 26%~39% in the HA solution. These results indicate that the inhibiting effect of the ozonation degradation of EDCs by the three organic matters followed HA> BSA> SA, which agrees with the DO₃ attenuation data. Furthermore, the data sufficiently indicated that the organic matters causeda certain degree of ozone attenuation, which decreased the chance of ozonation degradation of EDCs and resulted in the decreased removal efficiency of EDCs. Moreover, the humus-type organic substances presented the greatest inhibitory effect on the ozonation degradation of EDCs.

Fig.1

	Figure Option ViewDownloadNew Window
	Fig.1 Effect of three kinds of organic matters on dissolved ozone attenuation

Fig.2

	Figure Option ViewDownloadNew Window
	Fig.2 Effect of three kinds of organic matters on EDCs removal rate

From the results above, the three organic matters exhibited different impacts on the ozone attenuation rate and the ozonation degradation of EDCs. Therefore, we analyzed the structure of the three organic matters in depth. The characteristic UV-Vis spectra of the three organic matters are shown in Fig.3. It can be seen from the figure that the response intensity of the three organic matters followed HA> BSA> SA, which complies with the ozone attenuation and EDCsremoval efficiency results, indicating that the active groups in the organic matters that could react with ozone had specific response characteristics in the UV-Vis region.

Fig.3

	Figure Option ViewDownloadNew Window
	Fig.3 UV-Vis spectra of three kinds of organic matters

The EEM fluorescence spectra of the three organic matters are shown in Fig.4. Based on the response characteristics of the spectra for the organic matters and the corresponding substance types^[13], the spectra can be divided into 5 fluores-cence response regions (as shown in Table 2). It can be seen from the table that the response intensity of the three organic matters at Ex/Em=240/396 nm and Ex/Em=345/436 nm followed HA> BSA> SA, which agrees with the ozone attenuation and EDCs removal efficiency data, indicating that the active groups in the organic matters could react with the ozone responses at Ex/Em=240/396 nm and Ex/Em=345/436 nm, i.e., the fulvic acid-type humic acids and quinoline-type structures contributed significantly to ozone attenuation.

Fig.4

	Figure Option ViewDownloadNew Window
	Fig.4 Three-dimensional fluorescence spectrum of three kinds of organic matters

Table 2

Table 2

Table 2 Fluorescence response of the three organic matters Fluorescence peak (Ex/Em)/nm Type of matter Fluorescence intensity of HA Fluorescence intensity of BSA Fluorescence intensity of SA 280/350~370 Soluble microbial products 191 1 544 265.1 230/350 Tryptophan-type aromatic protein 205.2 2 198 374.8 275/306 Aromatic protein 705.3 1 862 1 391 240/396 Fulvic acid-type humic acids 1 142 365 133.4 345/436 Quinoline-type matters 1 174 54.68 313.1

Table 2 Fluorescence response of the three organic matters

The aforementioned results show that HA had the greatest inhibiting effect on the ozonation degradation of EDCs. Therefore, we analyzed the change of UV and fluorescence spectra in the presence of HA.

(1) Ultraviolet absorbance (UVA)

Based on the above results and results from the literature^{[14, 15]}, the changesin UVA (UVA₂₅₄, UVA₂₅₈ and UVA₂₈₀) that represent aromatic and oxygen-containing aromatic structures were chosen to indirectly indicate the degradation effect on EDCs. It can be seen from Fig.5 that the UVA₂₅₄, UVA₂₅₈ and UVA₂₈₀ of HA decreased with increasing ozone dosage and exhibited an excellent linear relationship. Therefore, the UVA₂₅₄, UVA₂₅₈ and UVA₂₈₀ of HA were used as indicator parameters for the ozone attenuation caused by organic matters. Furthermore, UVA₂₈₀ was the most sensitive indicator parameter and was chosen for the detailed exploration as follows.

Fig.5

	Figure Option ViewDownloadNew Window
	Fig.5 The relationship between UVA removal rate of HA and O3 dosage

The relationship between the removal efficiency of HA associated with UVA₂₈₀ and that of EDCs is shown in Fig.6. When the removal efficiency associated with UVA₂₈₀ was less than 15%, the removal efficiency of the five EDCs increased sharply with increasing removal efficiency of HA. When the removal efficiency of HA associated with UVA₂₈₀ was greater than 15%, the change in the removal efficiency of EDCs was steady, and when the removal efficiency of HA associated with UVA₂₈₀ was 18%, the removal efficiency of EDCs reached 100%. Therefore, UVA₂₈₀ can be used as an indicator parameter to predict the degradation efficiency of EDCs.

Fig.6

	Figure Option ViewDownloadNew Window
	Fig.6 The relationship between UVA280 removal rate of HA and EDCs removal rate

(2) Characteristics of the EEM fluorescence spectra

According to the above results, Ex/Em=240/396 nm and Ex/Em=345/436 nm were chosen as references. As shown in Fig.7, the EEM fluorescence spectra show that the removal efficiencyincreased with increasing ozone dosage for the HA solution, exhibitingan excellent linear relationship. Among the paired peaks, Ex/Em=240/396 nm presented a high linear correlation, and thus, was chosen for the detailed exploration.

Fig.7

	Figure Option ViewDownloadNew Window
	Fig.7 The relationship between EEM removal rate of HA and O3 dosage

In the presence of HA, the relationship between the removal efficiency associated with Ex/Em=240/396 nm and the removal efficiency of EDCs had a significant linear correlation, as shown in Fig.8. In addition, the removal efficiency of the five EDCs exhibited a similar trend as that of the EEM. When the removal efficiency associated with Ex/Em=240/396 nm was less than 35%, the removal efficiency of EDCs increased significantly with increasing effective removal efficiency presented in the EEM. When the removal efficiency associated with Ex/Em=240/396 nm was greater than 35%, the removal efficiency of EDCs nearly reached 100%. Therefore, Ex/Em=240/396 nm can be used as an indicator parameter to predict the degradation efficiency of EDCs.

Fig.8

	Figure Option ViewDownloadNew Window
	Fig.8 The relationship between Ex/Em=240/396 nm removal rate of HA and EDCs removal rate