All chemicals used in this study were analytical grade or higher. Sodium PS (Na₂S₂O₈, >99%), ferrous sulfate (FeSO₄, >99%), Ferric sulfate (Fe₂(SO₄)₃, >99%), HA (>90%), 1,4-benzoquinone (BQ), anhydrous sodium sulfate, silica gel, n-hexane (95%, HPLC grade), dichloromethane and acetone (99.9%, HPLC grade) were purchased from China National Pharmaceutical Group Corporation. The standard PAH samples were purchased from the National Center for Standard Substances. All solutions were prepared using deionized water (18.2 MΩ cm⁻¹).

A series of batch experiments were carried out at room temperature. Four treatments were set up in the experiment, including HA/Fe(II)/PS, HA/PS, Fe(II)/PS and PS alone, respectively. In detail, a set amount of FLU solution (15 mM, 10 ml), 10 ml of catalysts (FeSO₄, 0.5 M; HA, 0.5 M) and a certain amount of PS (0.5 M, 10 ml) were all added into a 100 ml conical flask. Then they were mixed in a constant temperature incubator shaker (25°C) at 150 rpm for 24 h. The samples were collected within a preset time (5 min, 10 min, 20 min, 30 min, 1 h, 2 h, 4 h, 8 h, 16 h and 24 h), and immediately terminated the reaction process by adding anhydrous ethanol in tiny quantities. The samples were collected and filtered through 0.45 μm PTFE membranes at a predetermined time, then they were stored in a refrigerator at 4°C for further analyses. After filtration, extraction, purification and condensation, the concentrations of FLU were determined. Results showed that the recoveries of all treatments were in the range of 89.8%–95.1%. Besides, the relative deviations of FLU concentrations in all parallel treatments were less than 10%. The reproducibility of the measurements was determined in triplicates and then mean value was calculated.

In order to evaluate the feasibility of HA combined with iron in PS activation, the key parameters were systematically analyzed. Benzoquinone (BQ) was choosn as the representative active substance in HA to explore the free radical process of synergistic activation of PS with Fe (Ⅱ). In a typical experiment, the concentration of PS is fixed at 5 mm, the concentration of DMPO is 0.1 mm, and various dosages of BQ, Fe (Ⅱ) and Fe (Ⅲ) were set up. The variables had been combined to examine the free radicals of the PS oxidation system. Briefly, all the reagents or solution were quickly put into the centrifuge tube and determined on site.

The pre-treatment processes, analytical methods and determining techniques for PAHs had been described elsewhere (Li et al., 2019). FLU in the reaction solution was extracted with 10 ml dichloromethane three times at 250 r min⁻¹ for 10 min. The extraction solutions were collected in an eggplant bottle. Subsequently, it was concentrated to 0.5 ml by a rotary evaporator and diluted with n-hexane to constant volume (1 ml). The samples were determined by gas chromatography-mass spectrometry (GC-MS 7890A-5975C, United States) for quantitative analysis. The instrument condition and parameters settings were not reiterated here. Full scan mode of GC-MS was employed and the intermediates generated were distinguished the by the ChemStation library. According to the relative response area compared with the internal standard, the quantitative and semi-quantitative analysis of PAHs were undertaken. The detection limit of FLU was 0.0029 ug ml⁻¹ and the average recovery were between 87.36.

To investigate the variation of functional groups, the Fourier transform infrared (FTIR) spectra of the HA in HA/Fe (Ⅱ)/PS system over time were obtained by an infrared spectrometer (Frontier FT-IR, United States). The test parameters were set as follows: wave number range: 4000–400 cm⁻¹; scan times: 32; resolution ratio: 4 cm⁻¹; sample thickness: 10–50 um.

Short-lived intermediates include a variety of free radical adducts were observed by an electron paramagnetic resonance (EPR) spectrometer (E-500, BRUKER, Germany). The free radicals adduct generated from the capture of 5,5-dimethyl-1-pyrroline N-oxide (DMPO), then they were collected at selected intervals and tested by means of electron spin resonance (ESR) technique.

The instrument conditions of ESR were set as follows: sweep width, 100 G; modulation amplitude, 3.08 G; a resonance frequency, 9.75 GHz; time constant, 10.24 ms; conversion time, 20.48 ms.

To verify the feasibility of PS activated by HA, the performance of various activation ways of PS for FLU degradation was investigated. The elimination of FLU directly demonstrated the PS activation performance. As shown in Figure 1A, the concentrations of FLU presented a continuous downward trend after degradation. The largest loss in FLU yields was observed in HA/Fe (Ⅱ)/PS system, which had reached 98% after 24 h. While separately 66% or 92% of FLU was degraded by Fe (Ⅱ)/PS and HA/PS. The findings revealed that HA had significant synergistic effects on Fe (Ⅱ) activated PS process (Zhang et al., 2022). The EPR spectra of free radicals generated in various systems were presented in Figure 1B. As we know, hydroxyl radical (OH) and SO₄·^- made a decisive contribution to the successful degradation of the organic pollutant in PS oxidation (Tsitonaki et al., 2010; Yuan et al., 2011; Zeng et al., 2015; Zeng et al., 2016; Lee et al., 2020). Apart from the above free radicals, semiquinone radical (SQ·^-) was also detected in the presence of BQ. Besides, the intensity of OH·, SO₄·^-, and SQ·^- were significantly enhanced when introduced BQ into the reaction system, especially for the HA/Fe (Ⅱ)/PS system. The results were consistent with the kinetic experiments. Result had reported that carbon-based catalyst exhibited an excellent performance on persulfate activation, primarily relay on OH and SO₄·^- (Luo et al., 2020; Li et al., 2022). This might be attributed to the rich porous structure, huge surface area and abundant functional groups of HA, which could benefit the formation of countless bonding sites (Hung et al., 2022; Xu et al., 2022; Yao et al., 2022). The existence of functional groups in HA, especially quinones, was considered to be an important factor in PS activation (Fang et al., 2013; Zhang et al., 2021).

To explore the leading functional groups which take responsible for FLU degradation, FTIR spectra of HA were obtained at different reaction time with the region from 4000 to 500 cm⁻¹ (Figure 2). The absorption peaks at 3415 cm⁻¹ gradually decreased along with the prolonged time (0–7 days), that might be ascribed to N-H stretching vibration or O-H stretching (Li et al., 2022; Zhang et al., 2022). The prominent peak at 1612 cm⁻¹ was attributed to asymmetric stretching of ionized carboxyl groups and aromatic C=C, C=O of quinone or H-bonded conjugated ketones (Ghosh et al., 2009; Yao et al., 2022). The peaks of the hydroxyl group showed the same pattern as the quinone group, which indicated HA played an active role in the production of hydroxyl radicals. A band around 1415 cm⁻¹ had been assigned to the C-OH stretching of phenolic hydroxyl. Studies had shown that phenolic hydroxyl was the main group to reduce Cr(VI) into Cr(III) and generate carbonyl and carboxyl groups as well (Hsu et al., 2009). As the decline of carboxylic groups, the peaks at 1380 cm⁻¹ presented decreased, which were assigned to O-H bending deformation of carboxylic groups, C-O stretching of phenolic groups or C-H deformation (Monda et al., 2017; Zhang et al., 2017). The small peak at 615 cm⁻¹ might be caused by the presence of a large number of iron ions in the reaction system (Yao et al., 2022). The results indicated that the carboxylic, phenolic, hydroxyl, quinone groups were the dominant groups which involved in the electron transfer and oxidation-reduction reactions.

The dosages and proportions of various activators were crucial for degrading toxic recalcitrant organic pollutants (Zhang et al., 2022). To further explore the enhancement of BQ on activation, the orthogonal experiments were designed to figure out the optimum conditions in terms of the concentration of BQ and Fe (Ⅱ). The production of free radicals under various conditions were displayed in Figure 3. In general, the response intensity of free radicals was strongly affected by the BQ/Fe (Ⅱ) molar ratio.

Figure 3A indicated that OH intensity significantly enhanced when introduced BQ into Fe (Ⅱ)/PS system. The results demonstrated that the intensity of OH increased as the reaction time extending. There continues to be a gradual decline of OH with the increasing concentration of Fe(II) when BQ concentration was 0 mm. At the given initial concentration of BQ (0.1 and 0.2 mm), the intensity of OH presented a slow increase followed by a rapid decrease. The highest catalytic activity of OH was exhibited when BQ concentration was 0.1 and 0.2 mm, respectively. When the concentration of BQ was 1.0 mm, the intensity of OH increased gradually as Fe(II) concentration rose, and reached the maximum at 1.0 mm. Basically, the yield of OH will decline once the proportion of Fe(II)/BQ exceeds 1. The most OH was produced at the same concentration of BQ and Fe(II).

The intensity of SO₄·^- was described in Figure 3B. It had demonstrated that BQ could coordinate with Fe(II) to promote more SO₄·^- production. Moreover, there was no distinct difference among all the treatments without BQ. When BQ was added, the intensity of SO₄·^- improved obviously. The most significant enhancement generally happened when the BQ concentration was in the range of 0.1–0.2 mm. The catalytic activity of SO₄·^- seemed to decline due to an excessive amount of BQ and Fe (Ⅱ), which might be due to the competitive effect of Fe (II) on SO₄·^- (Dominguez et al., 2019).

The variation of SQ·^- in different systems was exhibited in Figure 3C. There was evidence that SQ·^- could not generate in the absence of BQ. Besides, the intensity of SQ·^- had decreased markedly over time, which differed from the variation of OH and SO₄ ^·-. With the content of BQ fixed, the order of SQ·^- intensity was similar to that of OH·. When the concentration of BQ was in the range of 0.1–0.2 mm, the variation of SQ·^- intensity with Fe(II) concentration increased first, and then decreased. While as the BQ concentration rose to 1.0 mM, the intensity of SQ·^- was significantly strengthened with the increase of Fe(II) concentration. The maximum SQ·^- was usually produced at the same concentration of BQ and Fe(II).

The catalytic performance had been greatly enhanced by quinone groups. Specifically, as the BQ concentration increased, the response intensity of SQ·^- and OH had strengthened rapidly. Without BQ, Fe (II) had an inhibitory effect on hydroxyl radicals, which might be caused by the unproductive consumption of OH from Fe(II) (Wang et al., 2020). The moderate level of quinones could reduce the unproductive consumption of PS, while excessive dosage could grab SO₄·^- and OH from the system. Quinone could affect directly on the production of active species, which was attributed to the electron shuttles action during chemical processes (Aeschbacher et al., 2012; Zhang et al., 2021). The preliminary study noted that there might be Fe(II)/Fe(III) and semiquinone/quinone cycle in the catalyzed oxidation system (Leng et al., 2014; Li et al., 2022). In conclusion, the addition of an appropriate amount of Fe (II) and quinone could promote the decomposition of PS and the generation of free radicals to attack FLU. In this study, we had confirmed the optimizing ratio of Fe (II) and BQ was 1:1, which demonstrated a superior performance on FLU degradation.

To further prove the effect of iron in various valence on the degradation characteristic, the interaction between Fe (Ⅲ) and BQ was investigated (see Figure 4). According to the previous results, the study was conducted with 1.0 mM BQ. In contrast, the intensity of free radicals was restrained in the co-existence of HA and Fe (Ⅲ), revealing that Fe (Ⅲ) had inhibited the activation process. Specifically, the intensity of OH and SQ·^- gradually reduced with the increase of Fe(III) concentration. The intensity of SQ·^- reached the maximum within 2–4 min, then decreased sharply and tended to be flat after 10 min. It might be due to the decomposition of persulfate and the generation of free redicals decelerated. The more Fe (Ⅲ) we added, the effect of antagonism was more abvious. By contrast, there was little variation in SO₄ ^·-. It might be caused by the reversible reaction between Fe(II)/Fe(III) and SQ·^-/Q·^-, which promote the formation of more SO₄ ^·−in the system (Fang et al., 2013). Besides, excessive Fe(III) would diminish the reduction capacity and electron transfer process, resulting in low generation of free radicals and poor degradation efficiency.