Degradation of Diclofenac by Bisulfite Coupled with Iron and Manganous Ions: Dual Mechanism, DFT-Assisted Pathway Studies, and Toxicity Assessment

Abstract

Diclofenac (DCF) is often detected in diverse aquatic bodies, and ineffective management can lead to detrimental effects on human health and ecosystems. In this study, degradation of DCF by Fe(III) and Mn(II) activating bisulfite (BS) was investigated. In the Fe(III)/Mn(II)/BS system, 93.4% DCF was degraded at 200 μM BS within 120 s, and additional research on 1000 μM BS achieved 88.4% degradation efficacy. Moreover, kinetics fitting of DCF degradation with the different BS concentrations was studied to find the two highest reaction rates (200 and 1000 μM, k_obs = 0.0297 and 0.0317 s⁻¹, respectively). Whereafter, SO₄^•− and Mn(III) were identified as the main active species at these two concentrations, respectively. Density functional theory (DFT) calculations, molecular frontier orbital theory, and surface electrostatic potential (ESP) forecast electrophilic attack sites. DCF degradation pathways by radical and non-radical ways were proposed by attack site prediction and thirteen intermediates identified by UPLC-QTOF-MS. ECOSAR software 2.2 was used for toxicity assessment. This work studied DCF degradation by the Fe(III)/Mn(II)/BS process in the presence of different concentrations of BS, providing a new insight into water purification.

1. Introduction

Pharmaceuticals and personal care products (PPCPs), recognized as a vital class of “emerging pollutants”, have garnered considerable attention from the scientific community due to their potential implications for human health and aquatic ecosystems [1,2,3,4]. Diclofenac (DCF), a non-steroidal anti-inflammatory drug (NSAID) commonly used to treat fever, pain, and inflammation, was paradoxically often detected in sewage, surface water, groundwater, and even potable water [5,6,7]. Additionally, DCF might spawn biotoxic effects on aquatic organisms, as well as resistance and cross-resistance in microorganisms, and even on ecosystems [8,9]. However, conventional water treatment processes exhibited limited efficacy in managing DCF [10,11,12]. Hence, developing budget-friendly and environmentally benign water treatment technologies to remove DCF is crucial.

Advanced oxidation technologies (AOTs) have been consistently developed for refractory contaminants and water decontamination [13,14,15]. Sulfate radical (SO₄^•−)-based AOTs, namely SR-AOTs, are diffusely approved all the time, because SO₄^•− (2.5–3.1 V) is superior to traditional HO^• (1.8–2.7 V) in terms of higher redox potential or selectivity [16,17,18]. Bisulfite (BS) is commonly identified as an ideal precursor of SO₄^•−, due to its price, harmless products (SO₄²⁻), and non-toxic advantages [19,20]. Transition metals like Fe(II), Cu(II), Co(II), Cr(VI), and Ce(IV) for BS activation were reported, verifying the feasibility of transition metal activation [21,22,23]. Among them, iron is extensively utilized due to its widespread availability, cost-effectiveness, and minimal ecological impact [24,25]. In the meantime, Mn(II), often coupled with Fe(II) or Fe(III) in united catalysts, enhanced the reactivity of substrates for pollutant degradation. For example, Zhao et al. augmented benzoic acid oxidation by using the Fe(III)/H₂O₂ system integrating Mn(II) [26]. Li’s team created biochar-loaded iron and manganese oxides to activate persulfate for soil remediation [27].

Currently, there is a relative paucity of research on the application of Fe(III) and Mn(II) catalysis to BS, especially its application for DCF degradation. It was reported that both Fe(III) and Mn(II) could catalyze BS [17,28]. The extra introduction of either may enhance BS activation. Meanwhile, Fe(II)/Fe(III) and Mn(II)/Mn(III) loops can be facilitated in the presence of BS [29]. Our previous work investigated DCF degradation by the Fe(II)/BS system, whose reaction rate was affected by the BS concentration due to BS being the precursor of SO₄^•− [30]. But excess BS can capture active species because of its strong reducing capacity. Will the catalytic performance of Fe(III) and Mn(II) be affected by BS concentration? Will the main active species transform with BS concentration? These questions enrich the value of our research.

In this study, DCF degradation by the Fe(III)/Mn(II)/BS system was investigated systematically. Herein, the specific targets of this work are to (1) evaluate the DCF degradation efficacy; (2) identify the dominant active species under different BS concentrations (200 and 1000 μM); (3) present DCF degradation pathways via SO₄^•− and Mn(III); (4) assess toxicity in the Fe(III)/Mn(II)/BS system.

2. Materials and Methods

2.1. Materials

Diclofenac sodium (DCF, C₁₄H₁₀C_l2NO₂Na), sulfamethoxazole (SMX, C₁₀H₁₁N₃O₃S), and sodium bisulfite (BS, NaHSO₃) were procured from Aladdin Reagent Co, Shanghai, China. Sodium hydroxide (NaOH) and sulfate acid (H₂SO₄) were obtained from Tianjin Kermel Chemical Reagent Co, Ltd., Tianjin, China. Methanol (MeOH) was purchased from Thermo Fisher Scientific (Shanghai, China) and reached HPLC grade. Tert-butyl alcohol (TBA, C₄H₁₀O), manganese sulfate (MnSO₄), ferric nitrate (Fe(NO₃)₃·9H₂O), 1-hexanol (C₆H₁₄O), and sodium thiosulphate (Na₂S₂O₃) were supplied by Chengdu Kolon Chemical Co, Chengdu, China. Except for MeOH, all chemicals used for the experiments were analytically pure, and the water was deionized water (18.2 MΩ⋅cm) throughout all experiments.

2.2. Experimental Methods

All experiments were conducted at a temperature of 25 °C in 250 mL beakers placed on a magnetic stirrer (600 r·min⁻¹). The prepared DCF was adjusted to the desired pH by 0.2 M NaOH and 0.2 M H₂SO₄. After Fe(NO₃)₃ and MnSO₄ were added, the solution was adjusted again to the preset pH. Finally, BS was added to initiate the reaction. With 1 mL sample collected regularly, 50 μL of 0.4 M Na₂S₂O₃ was added to stop the reaction quickly. In the radical-scavenging experiments, different concentrations of tert-butyl alcohol (TBA), methyl alcohol (MeOH), or 1-hexanol were added to the solution before the addition of Fe(NO₃)₃. All experiments were performed in duplicate except for product analysis, taking the average to calculate the removal rate.

2.3. Chemical Analysis

Waters’s HPLC (Waters 2996-2695, Waters, Milford, MA, USA) was utilized to measure the concentrations of DCF and SMX. For DCF, a mobile phase comprising a 25:75 ratio of 1‰ acetic acid aqueous solution to methanol was employed at a flow rate of 1 mL·min⁻¹, with detection at a wavelength of 276 nm. SMX analysis involved a mobile phase with a 40:60 ratio of the same acetic acid solution to methanol, at a reduced flow rate of 0.8 mL·min⁻¹, and detection at 270 nm. The intermediates of DCF were detected by a UPLC-QTOF-MS (ACQUITY UPLC, Quattro Premier XE, Waters, Milford, MA, USA) with a 40,000 resolution. The stationary phase consisted of a reversed C18 column (100 × 2.1 mm, 1.7 mm), with 0.1% formic acid aqueous solution (A) and acetonitrile (B) at a flow rate of 0.3 mL·min⁻¹ as the mobile phases. The gradient was 10% B for 0.5 min, increasing linearly to 100% B for 7 min, and then decreasing back to 10% B for 3 min. The sample was 0.5 mL, and the column temperature was 40 °C. The solution pH was monitored by a pH meter (PHS-3C, Shanghai Rex, Shanghai, China).

The UV-Vis absorbance spectrum of the Mn(III)-PP (pyrophosphate) complex was measured using a UV-4802 spectrophotometer (manufactured by Unico in Shanghai, China), whose scan range was set from 200 to 600 nm, with distilled and deionized water serving as the blank control. The concentration of Mn(III)-PP was detected by taking the absorbance reading at 260 nm, unless otherwise specified.

2.4. Computational Analysis

The DCF model was visualized and optimized by using the GaussView 5.0 software and the Gaussian 09 program package [31]. The probable electrophilic attack sites on DCF were determined by density functional theory (DFT) at the B3LYP/6-31+G(d) level. To be specific, Multiwfn 3.8 was used to derive the Condensed Fukui Function by Equations (1)–(3), which is based on the Hirshfeld charges [32,33].

$$Nucleophilic attack:\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}{f}_k^{+}=q_k^N-q_k^{N-1}$$

(1)

$$Electrophilic attack:\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}{f}_k^{-}=q_k^{N-1}-q_k^N$$

(2)

$$Radical attack:\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}{f}_0=\frac{1}{2}(f_k^{+}+f_k^{-})$$

(3)

In addition, the highest occupied molecule orbital (HOMO) of the DCF molecule was visualized in the Gaussian 09 program combined with Multiwfn 3.8 and VMD 1.9.3 [34]. Surface electrostatic potential (ESP) was visualized with GaussView 5.0 and the Gaussian 09 program.

2.5. Toxicity Assessment

Software (ECOSAR 2.2) invented by the US Environmental Protection Agency was used to evaluate the aquatic toxicity of DCF and its intermediate products, which is based on the QSAR model [35]. Acute and chronic toxicity to fish, daphnids, and green algae were forecast, covering the half-lethal concentration (LC₅₀), half-effective concentration (EC₅₀), and chronic toxicity value (ChV).

3. Results and Discussion

3.1. Evaluation of DCF Degradation

To evaluate the performance of the Fe(III)/Mn(II)/BS system, different systems for DCF degradation performance were investigated (Figure 1a). DCF degradation was inhibited in the Mn(II)/BS system, while the Fe(III)/BS and Fe(III)/Mn(II)/BS systems could degrade DCF with favorable removal rates within 120 s: 76.4% and 93.3%, respectively. The reasons were as follows: (1) In acidic solutions, Mn(II) was less likely to be oxidized to Mn(III), thus failing to trigger the chain reaction [36,37,38,39]. (2) The great performance of the Fe(III)/Mn(II)/BS system resulted from the generated Mn(III) by Equation (4), which degraded DCF directly, accelerated the Fe(III)/Fe(II) cycle, or produced more active radicals by initiating the Mn(II)/BS system [37,40]. In Figure 1b, DCF degradation in the Fe(III)/BS and Fe(III)/Mn(II)/BS systems accurately adhered to pseudo-first-order kinetic models, and their kinetic equations as follows: y = −0.0118x − 0.0239, R² = 0.99 and y = −0.0297x − 0.0212, R² = 0.99, respectively.

Figure 1c shows the DCF degradation rate in Fe(III)/Mn(II)/BS with different pH, and the optimal pH was 4.0. To better understand the role of pH, Fe(III)/BS was introduced (Figure 1d). By comparison, at pH 4.0, Fe(III)/BS might be the main reaction to degrade DCF, while Mn(II) did not play much of a role in the Fe(III)/Mn(II)/BS system. When the pH was 5.0, DCF was almost not degraded in the Fe(III)/BS system, but it presented remarkable degradation in the Fe(III)/Mn(II)/BS system. Hence, the primary reaction might be the Mn(II)/BS system, and Fe(III) acted as an inducer. At pH 6.0, the DCF removal rate in both systems was negligible, indicating inhibition of the Mn(II)/BS system due to the complete precipitation of Fe(III).

In addition, different Mn(II)/Fe(III) molar concentration ratios were investigated (Figure 1e). When the Mn(II)/Fe(III) concentration ratio was less than 1:2, a higher DCF removal rate was achieved by increasing the Mn(II) concentration due to the promotion of the Fe(II)/Fe(III) cycle and the start-up of the Mn(II)/BS system. When the concentration ratio was 1:1, the DCF degradation rate rapidly increased to 80% within the first 40 s, reaching a plateau afterwards because of the depleted BS [41]. When the ratio was 2:1, BS was consumed rapidly, resulting in a low removal rate.

Finally, BS concentration was investigated in the Fe(III)/Mn(II)/BS system (Figure 1f). For the Fe(III)/Mn(II)/BS system, insufficient BS was consumed rapidly, whereas excess BS captured SO₄^•− [42,43]. When the BS concentration increased from 50 μM to 200 μM, the DCF removal rate gradually increased, which could be explained as follows: (1) The formation of FeOHSO₃H⁺ ligands was facilitated, resulting in more production of active radicals [44]. (2) The reaction of Mn(III)OMn(II) and BS was enhanced, generating more SO₃^•− according to Equation (5) [37]. (3) Fe(II)/Fe(III) and Mn(II)/Mn(III) cycling were accelerated to promote the formation of active radicals. Nevertheless, at 800 µM BS, DCF degradation experienced inhibition due to competition for radicals between excess BS and DCF (Equations (6) and (7)) [42,43,45]. In terms of the above tendency, the more BS, the more severe the inhibition of DCF might be. Meanwhile, high BS concentration might cause the transformation of active species. This has not been studied well. Therefore, 1000 µM BS was investigated additionally, surprisingly achieving 88.4% within 120 s.

Moreover, the results of extra sequential batch experiments are given in S4.

Mn²⁺ + SO₅^•− + H⁺ → Mn³⁺ + HSO₅⁻      K = 1 × 10⁸ M⁻¹ s⁻¹

(4)

Mn(III)OMn(II) + HSO₃⁻ → 2Mn(II) + SO₃^•− + OH⁻   K = 1.5 × 10⁶ M⁻¹ s⁻¹

(5)

HO^• + HSO₃⁻ → SO₃^•− + H₂O      K = 4.5 × 10⁹ M⁻¹ s⁻¹

(6)

SO₄^•− + HSO₃⁻ → SO₃^•− + SO₄²⁻ + H⁺      K = 1.3 × 10⁸~2.5 × 10⁹ M⁻¹ s⁻¹

(7)

3.2. Dual Mechanism

3.2.1. Kinetics Analysis of Different BS Concentrations

Section 3.1 found that 1000 µM BS could also degrade DCF with a favorable removal rate. Therefore, kinetic fitting of DCF degradation at different BS concentrations is shown in Figure 2a. Their k_obs are shown in Figure 2b. As can be seen, with the increase in BS concentration, the k_obs successively increased, decreased, and increased, creating two peaks at 200 and 1000 μM (k_obs = 0.0297 and 0.0317 s⁻¹, respectively). Intriguingly, 1000 μM exhibited a higher k_obs, suggesting the possible transformation of the main active species. Therefore, this phenomenon was subject to further investigation.

3.2.2. Identification of SO₄^•−

The presence of SO₃^•−, SO₄^•−, SO₅^•−, and HO^• should be considered when BS experiences metal-catalyzed autoxidation, but Mn(III) might also be present in the Fe(III)/Mn(II)/BS system [23,46]. Regarding the Fe(III)/Mn(II)/BS (200 μM) system, according to our previous work, SO₄^•− might be the main active species [30]. To identify the main active species, alcohol inhibitory tests were conducted, and CK values are summarized in Table S1 to investigate the competitive relationship of alcohols and DCF towards radicals [30].

Herein, tert-butyl alcohol (TBA) acted as a HO^• quenching agent (K _{TBA, HO•} = 6.0 × 10⁸ M⁻¹ s⁻¹) but not towards SO₄^•− (K _{TBA, SO4•−} = 8.4 × 10⁵ M⁻¹ s⁻¹). Methanol (MeOH) was employed as a scavenger for both HO^• and SO₄^•− (K _{MeOH, HO•} = 9.7 × 10⁸ M⁻¹ s⁻¹; K _{MeOH, SO4•−} = 1.0 × 10⁷ M⁻¹ s⁻¹). Meanwhile, the influence on SO₃^•−/SO₅^•−/Mn(III) caused by addition of TBA or Me can be ignored due to the low rate constants (<10³ M⁻¹ s⁻¹) [47,48]. As shown in Figure 3, in the presence of 10 mM TBA, the DCF removal rate decreased by 5% compared to the blank control. This could be explained by the CK value (CK _{TBA, HO}• = 6.0 × 10⁶ s⁻¹ >> CK _{DCF, HO}• = 7.5 × 10⁴ s⁻¹; CK _{TBA, SO4}•− = 8.4 × 10³ s⁻¹ << CK _{DCF, SO4}•− = 9.2 × 10⁴ s⁻¹), indicating that HO^• was not the main active species. With 100 mM and 500 mM TBA, DCF degradation was obviously retarded because of the scavenging of both HO^• and SO₄^•− by TBA (CK _{TBA, SO4•−} = 8.4 × 10⁴ s⁻¹, 4.2 × 10⁵ s⁻¹ ≈> CK _{DCF, SO4•−} = 9.2 × 10⁴ s⁻¹). Hence, the main active species might be SO₄^•−. Compared with 100 mM TBA, 100 mM MeOH displayed a decrease in DCF removal rate from 67.5% to 19.2%. This drastic inhibition might be due to the greater CK value of MeOH towards SO₄^•− (CK _{MeOH, SO4•−} = 1.0 × 10⁶ s⁻¹ >> CK _{TBA, SO4•−} = 8.4 × 10⁴ s⁻¹). Therefore, the fact that SO₄^•− was the main active species could be confirmed again. In the presence of 500 mM Me, both HO^• and SO₄^•− were quenched, owing to the high CK value of Me towards HO^• and SO₄^•− (CK _{MeOH, HO•} = 4.9 × 10⁸ s⁻¹ > CK _{MeOH, SO4•−} = 5.0 × 10⁶ s⁻¹ >> CK _{DCF, HO• or SO4•−}). However, the DCF removal rate was still 11.4%, proving the existence of SO₅^•− or Mn(III). It was found that the contribution order was SO₄^•− > SO₅^•− or Mn(III). In conclusion, SO₄^•− was the main active species in the Fe(III)/Mn(II)/BS system.

3.2.3. Identification of Mn(III)

Regarding the Fe(III)/Mn(II)/BS (1000 μM) system, Mn(III) was highly likely to be present. Therefore, the alcohol inhibitory experiments were conducted similarly. Based on the CK value theory, 10 mM TBA could totally quench HO^•, while 10 mM 1-hexanol enabled the quenching of both HO^• and SO₄^•− entirely rather than SO₅^•− (K _{1-hexanol, OH•−} = 3.0 × 10⁹ M⁻¹ s⁻¹ > K _{1-hexanol, SO4•−} = 1.6 × 10⁸ M⁻¹ s⁻¹ >> 10³ M⁻¹ s⁻¹ > K _{1-hexanol, SO5•−}) [43,49,50]. As depicted in Figure 4a, with 10 mM TBA, the DCF removal rate almost remained unchanged, suggesting that HO^• was not the main active species. With the addition of 10 mM 1-hexanol, we surprisingly observed the phenomenon that the DCF inhibition rate reached a remarkable 85%, suggesting that SO₅^•− was not the dominant active species. Additionally, competition with sulfamethoxazole (SMX) experiments were conducted to identify whether SO₄^•− played a leading role. The rate constant of SMX towards SO₄^•− approximated that of DCF (K _{SO4•−, SMX} = 8.8 × 10⁹ M⁻¹ s⁻¹; K _{SO4•−, DCF} = 9.2 × 10⁹ M⁻¹ s⁻¹), suggesting that both SMX and DCF could be well degraded by SO₄^•− [51]. Also, the low rate constant between SMX and Mn(III) has been reported by a recent study [46]. Figure 4b shows that the DCF removal rate was favorable, whereas SMX was barely degraded. It was possible that SO₄^•− was not the major active substance. The values of $\frac{K_{ {SO}_4^{•-}/DCF}}{{K }_{{SO}_4^{•-}/SMX}}$ and $\frac{{K }_{obs/DCF}}{{K }_{obs/SMX}}$ were calculated to further verify. The result displayed the values of $\frac{K_{ {SO}_4^{•-}/DCF}}{{K }_{{SO}_4^{•-}/SMX}}=\frac{9.2\times {10}^9}{8.8\times {10}^9}=1.04$ and $\frac{{K }_{obs/DCF}}{{K }_{obs/SMX}}=\frac{0.0263}{0.0016}=16.4$. Undoubtedly, the values of $\frac{K_{ {SO}_4^{•-}/DCF}}{{K }_{{SO}_4^{•-}/SMX}}$ and $\frac{{K }_{obs/DCF}}{{K }_{obs/SMX}}$ did not match, powerfully demonstrating that SO₄^•− was not the primary active substance.

As shown above, radicals did not play a major role, so Mn(III) might have worked. It was necessary to detect it directly. Hence, to stabilize Mn(III) and prevent its disproportionation, pyrophosphate (PP) was employed to form Mn(III)-PP complexes, characterized by a strong 258 nm absorption peak [52,53]. Previously, Michalkiewicz et al. found that the maximum absorption peak of this complex was 260 nm when one electron in the Mn(III)-PP compound was transferred to Mn(III) [54]. Therefore, with the presence of DCF and PP, the full-wavelength scanning curves of Mn²⁺, Fe²⁺, Fe³⁺, Fe³⁺/Mn²⁺, and Fe³⁺/Mn²⁺/BS were investigated (Figure 4c). Compared to the faint absorption peaks of other curves, the Fe³⁺/Mn²⁺/BS curve showed a significant peak at 260 nm, suggesting the existence of Mn(III).

To sum up, SO₄^•−, SO₅^•−, HO^•−, etc., were not the active species for DCF degradation, and the Mn(III)-PP compound confirmed the existence of Mn(III). Hence, Mn(III) very likely served as the predominant active substance.

3.3. Attack Points Prediction by DFT

Figure 5a shows the optimized DCF molecule and Figure 5b lists the Fukui function ($f_0$ and $f_k^{-}$), which forecasts the reaction mechanism mediated by SO₄^•− and Mn(III), electrophilic reagents. $f_0$ is often used to predict radical attacks and $f_k^{-}$ for electrophilic attacks. The molecular frontier orbital theory indicates that HOMO could estimate electrophilic sites, and LUMO could estimate the nucleophilic sites [55]. The orbital energies of HOMO and LUMO demonstrated that DCF is an electron donor: −5.745651 eV and −0.677524 eV, respectively. The HOMO of DCF is shown in Figure 5c.

Due to the electronegativity of the oxygen atoms (O₂₈ and O₂₉), there is a loose electron distribution in the C₂₇, evidenced by ESP (Figure 5d). Meanwhile, the π-π stacking of the aromatic ring protects it from attack. All these allow C₂₄ to be a favorable site for attack, and the moderate electron density in ESP also proves it. In addition, unpaired electrons determine that electron-rich sites are susceptible to be attacked by SO₄^•−, and Mn(III) selectively attacks electron-rich groups, facilitating its bond formation with nitrogen-containing functional groups with lone pair electrons [52]. N₁₂ in the DCF molecule exhibits a significant electron density that can be seen in ESP, and it is observed to have a higher Fukui index ($f_0$ = 0.0610; $f_k^{-}$ = 0.1063, respectively). This indicates that N₁₂ was vulnerable to attack, leading to cleavage of the C−N bond. C₂ is a special site, because dechlorination is accompanied by a reduction reaction. OH^•− converted by SO₄^•− can serve as either an electron donor or acceptor, but only as a donor for Mn(III). Hence, only radical attack is considered ($f_0$ = 0.0253). Finally, C₄ and C₂₀ can also be probable sites for hydroxylation to occur ($f_0$ = 0.0771 and 0.0669; $f_k^{-}$ = 0.0635 and 0.0802, respectively).

3.4. DCF Degradation Pathways

SO₄^•− and Mn(III) were the main active species in the Fe(III)/Mn(II)/BS system with different BS concentrations, and it was meaningful to propose and compare their degradation pathways. Thirteen intermediates were detected and are listed in Table S3. Based on Section 3.3 and intermediates, seven degradation pathways are shown in Figure 6, including dehydrogenation, decarboxylation, hydroxylation, formylation, decarbonylation, cleavage of C−N bonds, and dechlorination.

Pathways 1 to 3 belonged to SO₄^•−. In pathway 1, DCF was attacked at C₄ or C₂₀, with hydroxylation occurring to form m/z 312-1 or m/z 312-2, further hydroxylating to form m/z 328 [56]. In pathway 2, C₂₄ of DCF was the reaction site. DCF went through decarboxylation to form m/z 252, further transforming into m/z 266-1 by formylation [57]. In pathway 3, DCF underwent a series of dechlorination reactions at C₂ and breakage of C₁−N₁₂ bonds to produce a low molecular weight of m/z 114. Researchers also detected m/z 262 and m/z 114 [58,59].

Pathways 4 to 6 belonged to Mn(III). In pathway 4, m/z 268 was produced by decarbonylation, further proceeding through dehydrogenation to form m/z 266-2 (quinoid substance), which detached a hydroxyl group to form m/z 282-2 [59]. In pathway 5, the first two steps of the reaction were identical to pathway 2, but m/z 266-1 could further convert to m/z 282-1 by removing a hydroxyl group at C₂₀ [60]. In pathway 6, breakage of the C₁₄−N₁₂ bond could produce m/z 162, with further hydroxylation at C₄ to form m/z 178 [61].

Pathways led by two active species displayed unique characteristics. For SO₄^•−, pathway 1 could produce a high-molecular-weight substance of m/z 328 for DCF, but not for Mn(III). Meanwhile, pathway 3 could also produce the lowest molecular weight m/z 114. Additionally, dechlorination was a unique reaction to radical attack. For Mn(III), pathways 4 and 6 could generate a quinoid substance and aniline derivative, respectively.

3.5. Toxicity Assessment

The DCF degradation process did not alleviate the environmental risk. Therefore, it was essential to carry out a toxicity assessment. In the Fe(III)/Mn(II)/BS system, different active species had different pathways. This section will analyze the toxicity of different pathways (Figure 7a,b).

The initial toxicity of DCF to fish, daphnids, and green algae, whether acute or chronic, was identified to be “Harmful” or “Toxicity”. Through pathways of SO₄^•−, m/z 252 was a destructive product, demonstrating both acute and chronic “Very Toxicity”. The toxicity of m/z 262 and m/z 312 converted from DCF presented a decreasing tendency to some extent. However, overall, they still harmed aquatic organisms, except for m/z 312, which was “Not Harmful” for acute toxicity. With further degradation, m/z 328 was “Not Harmful” for most acute and chronic toxicity, but the toxicity of m/z 266-1 and m/z 114 was not optimistic. Through pathways of Mn(III), m/z 282-1 presented further acute and chronic toxicity reductions in the richer pathway from m/z 252 to m/z 282-1, but all products of this pathway were located in hazardous regions. Unluckily, the acute and chronic toxicity of m/z 268 and m/z 162 mostly increased compared to DCF, even to “Very Toxicity”. As the degradation was prolonged, the formation of m/z 178 was “Not Harmful” for acute toxicity, but “Harmful” for chronic toxicity. From m/z 268 to m/z 282-2, acute and chronic toxicity both experienced an increase at first and then a decrease, but the ultimate toxicity remained almost unchanged compared to the beginning.

Unfortunately, DCF had a higher potential to produce more toxic intermediate products. The same phenomenon has been reported that NSAIDs’ transformation products often exhibit higher toxicity than the original compounds [62]. Hence, various high-toxicity byproducts were generated during DCF degradation by the Fe(III)/Mn(II)/BS system, so monitoring the fluctuation of product toxicity was of great importance. Meanwhile, the prediction authenticity of ECOSAR software should be verified further.

4. Conclusions

A novel attempt at DCF degradation using the Fe(III)/Mn(II)/BS process was proposed and discussed in this work. SO₄^•− or Mn(III) played a primary role at 200 or 1000 μM BS, respectively. At 200 μM BS, 93.4% DCF was degraded in the Fe(III)/Mn(II)/BS system within 120 s, which was higher than the removal of 76.4% and almost 0% in the Fe(III)/BS and Mn(II)/BS systems. Mn(II) improved the Fe(III)/BS system due to the probable formation of Mn(III). At 1000 μM BS, within 120 s, an 88.4% DCF degradation rate was reached in the Fe(III)/Mn(II)/BS system. Under different BS concentrations, kinetic fitting of DCF degradation showed two peaks of k_obs: 0.0297 s⁻¹ at 200 μM and 0.0317 s⁻¹ at 1000 μM. Then, with 200 μM BS, SO₄^•− was identified by alcohol inhibitory tests. In the presence of 1000 μM BS, Mn(III) was validated by alcohol inhibitory experiments, competitive tests with SMX, and detection of absorption peaks. Further, according to DFT-assisted attack site prediction and thirteen intermediates detected by UPLC-QTOF-MS, DCF degradation pathways were proposed, including dehydrogenation, decarboxylation, hydroxylation, formylation, decarbonylation, cleavage of C−N bond, and dichlorination. Most DCF intermediates appeared to have higher toxicity, and the indispensability of supervision for toxicity fluctuation was revealed.