English

• Recently, the combination of UV light and S(Ⅳ) (HSO₃⁻/SO₃²⁻) has been acknowledged as an efficient advanced reduction method for the aqueous removal of different halogenated micropollutants [1-3]. Studies have uncovered the strong reducing capability of hydrated electrons (e_aq⁻), hydrogen atoms (H^•), and sulfite radicals (SO₃^•−) [4,5]. However, its reductive characteristics lead to its inefficiency in abating non-halogenated electron-rich compounds, which constitute a large portion of environmental contaminants. The introduction of electron-shuttling photosensitizers (e.g., Fe³⁺, FeTiO_X) into the UV/S(Ⅳ) system was recently demonstrated to facilitate SO₃²⁻ conversion into SO₃^•− without producing e_aq⁻ and H^• [6]. Subsequent reactions between SO₃^•− and dissolved oxygen could spark chain reactions, yielding multiple oxidizing species such as SO₅^•−, SO₄^•−, ^•OH, and ¹O₂ (Eqs. 1-6). These highly reactive intermediates surpassed the limited reactivity of reducing species, exhibiting broader oxidizing capabilities against diverse organic micropollutants. Moreover, the addition of photosensitizers might induce an absorption redshift, thereby expanding the catalytic UV wavelength range [7-9].

  $ \mathrm{SO}_3^{\cdot-}+\mathrm{O}_2 \rightarrow \mathrm{SO}_5^{\cdot-} $

  (1)

  $ \mathrm{SO}_5^{\cdot-}+\mathrm{HSO}_3^{-} \rightarrow \mathrm{SO}_4^{\cdot-}+\mathrm{SO}_4^{2-}+\mathrm{H}^{+} $

  (2)

  $ \mathrm{SO}_4^{\cdot-}+\mathrm{H}_2 \mathrm{O} \rightarrow ^{\cdot}\mathrm{OH}+\mathrm{SO}_4^{2-} $

  (3)

  $ \mathrm{HO}_2^{\cdot} \rightarrow \mathrm{O}_2^{\cdot-}+\mathrm{H}^{+} $

  (4)

  $ 2 \mathrm{O}_2 ^{\cdot-}+2 \mathrm{H}^{+} \rightarrow{ }^1 \mathrm{O}_2+\mathrm{H}_2 \mathrm{O}_2 $

  (5)

  $ \mathrm{SO}_5^{\cdot-}+\mathrm{SO}_5^{\cdot-} \rightarrow \mathrm{S}_2 \mathrm{O}_8^{2-}+{ }^1 \mathrm{O}_2 $

  (6)

  As a class of multi-functional metal complexes, iron(Ⅲ) porphyrins have exhibited considerable potential in various fields such as environmental catalysis, C—H functionalization, CO₂ reduction, and biomedical imaging [10-12]. Particularly, iron(Ⅲ) porphyrins have proven to be exceptional electron-shuttling agents for modulating the redox characteristics of the UV/S(Ⅳ) system. This assertion was supported by the following rationales: (1) The near-visible light absorption of iron(Ⅲ) porphyrins facilitates their potential photocatalytic activity [13,14]. (2) iron(Ⅲ) porphyrins have a high oxidation capability (E⁰(Fe³⁺/Fe²⁺) = 0.77 V), making them thermodynamically favorable in SO₃²⁻/SO₃^•− conversion. (3) As a typical pentacoordinate complexes, iron(Ⅲ) porphyrins retain superior coordination activity, facilitating the binding with SO₃²⁻ and O₂ to potentially accelerate photocatalytic oxidation. However, limited information has been gained about the role of these electron-shuttling photosensitizers in adjusting the redox reactivity of the UV/S(Ⅳ) system, the formation mechanisms of oxidative species, and their interactions with micropollutants.

  In this context, TPPFe was integrated into a UV₃₆₅-irradiated photochemical system containing S(Ⅳ) and carbamazepine (CMZ), aiming to explore the photocatalytic degradation process, which utilizes light to drive chemical reactions that break down organic contaminants into harmless or less harmful substances [15,16]. Firstly, morphological and structural characterization was conducted to ensure the structure of TPPFe (Figs. S1 and S2 in Supporting information). As illustrated in Fig. S3 (Supporting information), the incorporation of 0.2 g/L TPPFe into the UV₃₆₅/S(Ⅳ) system enhanced the removal efficiency of CMZ from 18% to 95% in 150 min at pH 7.0. Degradation performance for different pollutants was evaluated, which showed a higher removal efficiency for electron-rich pollutants (Fig. S4 in Supporting information). These findings implied the generation of highly reactive species in the combined system. Furthermore, investigations were conducted to examine the influences of S(Ⅳ) concentrations, TPPFe dosages, initial pH, and coexisting anions on the photocatalytic efficiency of CMZ. The pertinent results and discussions are shown in Figs. S3 and S5 and Text S5 (Supporting information).

  The electron spin resonance technique and various trapping agents were employed to gain deeper insights into the reaction mechanisms. This revealed the primary radicals formed in the system, including ^•OH, SO₄^•−, ¹O₂, and O₂^•− (Fig. S6 in Supporting information). Based on the experimental evidence, a photocatalytic cycle mechanism was postulated for the generation of SO₃^•− and the regeneration of TPPFe (Scheme 1). To validate the mechanistic plausibility, density functional theory (DFT) calculations were conducted to determine the Gibbs energy for each process. Initially, the entry of the TPPFe species into the cycle was facilitated by the coordination of SO₃²⁻ rapidly forming the [TPPFe^ⅢCl(SO₃)]²⁻ (Fig. S9 in Supporting information). Subsequently, this adduct absorbed light under UV₃₆₅ irradiation. The high-energy [TPPFe^ⅢCl(SO₃)]^2−* dissociated to generate SO₃^•− and [TPPFe^ⅡCl]⁻, releasing heat in the process. The first generated adduct was localized as 251.5 kcal/mol lower than the dissociated pair, indicating a significantly exothermic coordinate reaction [17]. The energy-demanding step for SO₃^•− formation was the dissociation of the [TPPFe^ⅢCl(SO₃)]²⁻ complex (Fig. S9) [18,19]. This could be characterized as a homolytic reaction of the Fe–O bond. Thermodynamic analysis, aided by DFT calculation, revealed a Gibbs free energy increase of 55.89 kcal/mol, correlating with an irradiation energy of 512 nm. This suggested that the reaction might be photoactivation-induced. The excited states of [TPPFe^ⅢCl(SO₃)]²⁻ were computed by TDDFT and UV–vis spectroscopy was simulated by the observed intense absorption at 392–566 nm (Fig. S10 in Supporting information). To unveil the electronic excitation properties, excited states with high oscillator strength for absorption were analyzed by hole-electron and interfragment charge transfer (IFCT) analysis. The hole–electron analysis indicated the continuous existence of ligand-to-metal charge transfer (LMCT) from the π orbital of porphyrin to the d orbital of Fe^Ⅲ for each excited state (Fig. S11 in Supporting information) [20,21]. Additional IFCT analysis revealed that the excitation of [TPPFe^ⅢCl(SO₃)]²⁻ belonged to a highly mixed charge transfer (CT) and locally excited (LE) mode, with concrete excitation parameters listed in Table S1 (Supporting information). Collectively, these findings indicated that the LMCT between porphyrin and Fe could induce a redshift in the complex absorption. Consequently, the formed adduct could be excited to a high energy state under UV₃₆₅ light, facilitating the subsequent dissociation reaction.

  Scheme 1

  

  Scheme 1.  Proposed formation mechanisms of reactive species by the UV-induced photocatalytic cycle of TPPFe^ⅡCl-[TPPFe^ⅡCl]⁻.

  DownLoad: Full-Size Img PowerPoint

  During TPPFe regeneration, the produced [TPPFe^ⅡCl]⁻ was coordinated by dissolved oxygen, leading to the formation of an O₂ adduct. Upon absorbing light, this adduct subsequently reacted with H₂O to form TPPFe, HO₂^• and OH⁻. Theoretical analysis revealed that the complexation reaction with dissolved oxygen slightly increased the energy by 3.28 kcal/mol (Fig. S12 in Supporting information). The dissociation of the O₂ adduct was identified as the most energy-demanding step in TPPFe regeneration. Various dissociation mechanisms were postulated and validated through DFT. Initially, a direct dissociation mechanism of the adduct into O₂^•− and TPPFe was considered. However, the process proved to be energetically unfavorable, with an energy difference of 86.96 kcal/mol, indicating its thermodynamical infeasibility. Therefore, a more plausible mechanism emerged: A concerted reaction with H₂O to generate the more stable radicals HO₂^• and OH^–. Energy increased by 41.96 kcal/mol before and after the reaction. Further analysis revealed that the excited states of [TPPFe^Ⅱ(O₂)]⁻ absorbed light near 390 nm (Fig. S13 in Supporting information), possibly induced by a LMCT, as supported by hole-electron and IFCT analysis. Consequently, the most plausible mechanism for TPPFe regeneration involved the excited [TPPFe^Ⅱ(O₂)]^−* reacting with H₂O to revert to TPPFe and generate HO₂^• and OH⁻, releasing heat in the process. The participation of H₂O in this reaction synergistically promoted the breakage of the Fe–O bond.

  The above photocatalytic cycle mechanism fully accounted for the generation of SO₃^•− and HO₂^•. The chain reactions involving the mutual transformation of radicals are well known [22-25]. Dissolved oxygen rapidly oxidized the generated SO₃^•− into SO₅^•− (Eq. 1). The unstable and highly oxidizing SO₅^•− can react with HSO₃⁻ to generate SO₄^•−, which are the primary oxidizing radicals in the PMS-dominated system. Additionally, SO₅^•− can react with H₂O to generate ^•OH ((2), (3)). Furthermore, the ¹O₂ could be generated by O₂^•− and SO₅^•− through different pathways (Eqs. 4-6) [22-24].

  The reaction mechanisms of radicals have been the subject of intensive research for decades [25-27]. However, a knowledge gap persists regarding the detailed reaction mechanisms at the atomic level between radicals and organic micropollutants in water. To overcome this limitation, DFT calculations were employed to analyze the Gibbs free energy variations during CMZ's reactions with various reactive intermediates.

  Initially, CMZ's electronic structure was examined to pinpoint potential reaction sites using Fukui functions. The computed f⁰ value indicated CMZ's reactivity toward radicals. Visualizing the Fukui functions via isosurface plots (Fig. S14 in Supporting information) revealed that the C7 and C8 positions are most susceptible to radical addition reactions [28]. Subsequently, five oxidizing species (i.e., ^•OH, SO₃^•−, SO₄^•−, ¹O₂, and HO₂^•) were selected for analyzing the radical addition process to the C7–C8 double bond. The initial, transition and final state structures of these addition reactions were optimized, and the Gibbs free energy for each structure was computed. The resulting energy barrier diagram (Fig. 1) revealed that ^•OH and SO₄^•− exhibited the highest kinetic and thermodynamic reactivities. In contrast, HO₂^• and SO₃^•− displayed less pronounced reactivity trends with CMZ. Furthermore, given that the steady-state concentration of reactive radicals was typically much lower than the added oxidant amount, the reaction of these two radicals with CMZ was virtually negligible. Notably, the nonradical ¹O₂ demonstrated a unique reaction mechanism distinct from the other four radicals, rendering it inappropriate for classification as a simple radical addition reaction [29,30]. Based on CMZ's structure and the chemistry of ¹O₂, a comprehensive reaction mechanism was proposed and validated using DFT calculations (Fig. 2). Specifically, ¹O₂ preferentially reacts with the double bond, forming an oxidized epoxy intermediate that undergoes further rearrangement reaction to form a quadruple ring structure [29]. Subsequently, direct ring cleavage leads to the formation of a stable dialdehyde compound, consistent with experimental observations. In addition, quenching experiments were carried out to identify key active species (Fig. S15 in Supporting information) [31-33]. The results showed that SO₄^•− played a dominant role in CMZ degradation, which was consistent with calculation results.

  Figure 1

  

  Figure 1.  Free energy diagram for the initial radical addition to C7, C8 double bond of CMZ reaction by ^•OH, SO₃^•−, SO₄^•−, and HO₂^•.

  DownLoad: Full-Size Img PowerPoint

  Figure 2

  

  Figure 2.  Free energy diagram for the reaction between CMZ and ¹O₂.

  DownLoad: Full-Size Img PowerPoint

  Based on the identified transformation products and potential reaction mechanisms, the distinct degradation pathways for CMZ were proposed, involving radicals and ¹O₂ (Fig. 3). As illustrated in Fig. 2, the pathway mediated by ¹O₂ resulted in the generation of TP-268, exhibiting electrophilic characteristics in its aldehyde group. This electrophilicity potentially facilitated subsequent reactions, such as amide attack, yielding the addition products. Specifically, the aldehyde group underwent an attack by the lone pair of electrons on the nitrogen atom, followed by dehydration, ultimately producing TP-250. Despite the amide nitrogen's weak nucleophilicity, nucleophilic addition proved viable due to the conjugate system's stabilizing effect on the molecule. Subsequent oxidation led to the formation of TP-166. Notably, the products resulting from electron-rich phenyl nucleophilic addition, widely documented by other studies in oxidative systems, were not observed [34-36]. We hypothesized that the absence of acid catalysis might have rendered this reaction pathway unfavorable. Additionally, the hydroxyl compound formed lacked the capacity for swift elimination reactions necessary to create a stabilizing conjugate system.

  Figure 3

  

  Figure 3.  The proposed degradation pathways of CMZ during interactions with different reactive species.

  DownLoad: Full-Size Img PowerPoint

  In contrast, radical-mediated degradation pathways exhibit greater complexity, involving the generation of radical intermediates. The selective addition to the C7–C8 double bond was attributed to the stabilization of radicals produced at the benzyl position by benzene conjugation. These radical intermediates had the potential to combine with other radicals (e.g., ^•OH), leading to the formation of vicinal diols. These diols might then undergo dehydration and ring-closing reactions, resulting in epoxy products such as TP-252. The formation of the dialdehyde compound TP-268 could be attributed to a ^•OH attack on the C7–C8 σ bond, which produced a more stable radical species. Subsequent binding with another ^•OH and dehydration led to the production of TP-268. As previously discussed, the production of TP-250 followed a comparable mechanism, encompassing the nucleophilic addition of amide and a dehydration step. The subsequent degradation steps were predominantly governed by radical hydrogen abstraction and radical binding reactions.

  Upon comparing various methods for generating the key compound TP-268, it was discovered that a single molecule of ¹O₂ was sufficient to oxidize CMZ into TP-268. Conversely, ^•OH required interactions with four molecules to initiate the reaction, thereby increasing its complexity. Each molecular interaction consumed a considerable amount of time and energy. Consequently, a higher number of interactions resulted in a longer completion time for the entire reaction. Therefore, from the perspective of molecular interactions and reaction sequences, the ¹O₂–CMZ reaction tended to exhibit a faster rate than the radicals reaction, primarily due to its straightforward nature.

  In summary, TPPFe has demonstrated remarkable photocatalytic efficiency in enhancing the UV/S(Ⅳ) system to generate oxidizing radicals for the degradation of CMZ. Through a combination of experimental and theoretical analyses, this research has elucidated the photocatalytic reaction mechanisms of TPPFe and the reactivity of various radicals toward CMZ. While significant progress has been made in elucidating TPPFe's photocatalytic mechanisms, further exploration is needed. Currently, mechanisms remain largely theoretical, with direct capture of intermediate reaction products not feasible. Advanced characterization methods are necessary to validate these mechanisms and enhance result authenticity. Deeper investigation into the interaction between TPPFe and reactants is also warranted, as it may reveal new aspects of reaction mechanisms and kinetics. Overall, this research offers a novel understanding of photocatalytic reactions modulated by photosensitizer-based systems and demonstrates the value of studying reactions at the molecular level. Future research can expand our knowledge and guide the development of efficient, sustainable water treatment technologies.

  Declaration of competing interests

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  This study was financially supported by the National Key R&D Program of China (No. 2023YFE0112100) and the XMU Training Program of Innovation and Entrepreneurship of Undergraduates (No. 2024Y1328).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110743.

  
  
• 1. [1]

     L. Yang, L.Y. He, J.M. Xue, et al., Chem. Eng. J. 397 (2020) 125412. doi: 10.1016/j.cej.2020.125412

  2. [2]

     D.D. Rao, H.Y. Dong, L.S. Lian, et al., ACS EST Water 1 (2021) 1785–1795. doi: 10.1021/acsestwater.1c00110

  3. [3]

     Y.R. Gu, W.Y. Dong, C. Luo, T.Z. Liu, Environ. Sci. Technol. 50 (2016) 10554–10561. doi: 10.1021/acs.est.6b03261

  4. [4]

     P. Neta, R.E. Huie, Environ. Health. Perspect. 64 (1985) 209–217. doi: 10.1289/ehp.8564209

  5. [5]

     X.C. Li, J. Ma, G.F. Liu, et al., Environ. Sci. Technol. 46 (2012) 7342–7349. doi: 10.1021/es3008535

  6. [6]

     Y.W. Xie, W.Z. Chen, H.R. Li, et al., J. Hazard. Mater. 462 (2024) 132801. doi: 10.1016/j.jhazmat.2023.132801

  7. [7]

     M. Cheng, Y.X. Cui, J. Wang, et al., ACS Appl. Mater. Interfaces 11 (2019) 13158–13167. doi: 10.1021/acsami.9b02695

  8. [8]

     P. Wang, S. Guo, H.J. Wang, et al., Nat. Commun. 10 (2019) 3155. doi: 10.1038/s41467-019-11099-8

  9. [9]

     Q. Xiao, T. Wang, S.L. Yu, P. Yi, L. Li, Water Res. 111 (2017) 288–296. doi: 10.1016/j.watres.2017.01.018

  10. [10]

      K.J. Hock, A. Knorrscheidt, R. Hommelsheim, et al., Angew. Chem. Int. Ed. 58 (2019) 3630–3634. doi: 10.1002/anie.201813631

  11. [11]

      S. Amanullah, P. Gotico, M. Sircoglou, et al., Angew. Chem. Int. Ed. 63 (2024) e202314439. doi: 10.1002/anie.202314439

  12. [12]

      J.J. Chen, Y.F. Zhu, S. Kaskel, Angew. Chem. Int. Ed. 60 (2021) 5010–5035. doi: 10.1002/anie.201909880

  13. [13]

      J. Chen, W.R. Browne, Coordin. Chem. Rev. 374 (2018) 15–35. doi: 10.1016/j.ccr.2018.06.008

  14. [14]

      E. Pugliese, P. Gotico, I. Wehrung, et al., Angew. Chem. Int. Ed. 61 (2022) e202117530. doi: 10.1002/anie.202117530

  15. [15]

      X.D. Zhao, Q. Liu, X.L. Li, H.M. Ji, Z.R. Shen, Chin. Chem. Lett. 34 (2023) 108306. doi: 10.1016/j.cclet.2023.108306

  16. [16]

      Q. Liu, H. Li, H. Zhang, Z.R. Shen, H.M. Ji, Chin. Chem. Lett. 33 (2022) 4756–4760. doi: 10.1016/j.cclet.2021.12.089

  17. [17]

      T. Lu, F.W. Chen, J. Comput. Chem. 33 (2012) 580–592. doi: 10.1002/jcc.22885

  18. [18]

      M.X. Sun, H.J. Song, X.B. Xie, et al., Environ. Sci. Technol. 56 (2022) 3170–3180. doi: 10.1021/acs.est.1c08362

  19. [19]

      L. Chen, X.Z. Peng, J.H. Liu, J.J. Li, F. Wu, Ind. Eng. Chem. Res. 51 (2012) 13632–13638. doi: 10.1021/ie3020389

  20. [20]

      Z.Y. Liu, T. Lu, Q.X. Chen, Carbon 165 (2020) 461–467. doi: 10.1016/j.carbon.2020.05.023

  21. [21]

      T. Lu, Q.X. Chen, Comput. Theor. Chem. 1200 (2021) 113249. doi: 10.1016/j.comptc.2021.113249

  22. [22]

      Y. Jia, Z.H. Hu, Y. Wu, Y. Zhang, Chem. Eng. J. 499 (2024) 156062. doi: 10.1016/j.cej.2024.156062

  23. [23]

      P.C. Xie, Y.Z. Guo, Y.Q. Chen, et al., Chem. Eng. J. 314 (2017) 240–248. doi: 10.1016/j.cej.2016.12.094

  24. [24]

      Y.Y. Chu, L.J. Xu, L. Gan, et al., Water Res. 198 (2021) 117143. doi: 10.1016/j.watres.2021.117143

  25. [25]

      Y. Yang, J.J. Pignatello, J. Ma, W.A. Mitch, Environ. Sci. Technol. 48 (2014) 2344–2351. doi: 10.1021/es404118q

  26. [26]

      H.M. Huang, M. Koy, E. Serrano, P.M. Pflüger, Nat. Catal. 3 (2020) 393–400. doi: 10.1038/s41929-020-0434-0

  27. [27]

      N. Zhang, S.R. Samanta, B.M. Rosen, V. Percec, Chem. Rev. 114 (2014) 5848–5948. doi: 10.1021/cr400689s

  28. [28]

      T. Lu, Q.X. Chen, Realization of conceptual density functional theory and information-theoretic approach in multiwfn program, in: In Conceptual Density Functional Theory, WILEY-VCH GmbH, Weinheim, 2022, pp. 631–647.

  29. [29]

      E.L. Clennan, A. Pace, Tetrahedron 61 (2005) 6665–6691. doi: 10.1016/j.tet.2005.04.017

  30. [30]

      D.B. Min, J.M. Boff, Compr. Rev. Sci. F. 1 (2002) 58–72. doi: 10.1111/j.1541-4337.2002.tb00007.x

  31. [31]

      R.D. Su, Y.X. Gao, L. Chen, et al., Proc. Natl. Acad. Sci. 121 (2024) e2319427121. doi: 10.1073/pnas.2319427121

  32. [32]

      J. Zhang, C. Shan, W.M. Zhang, B.C. Pan, Proc. Natl. Acad. Sci. U. S. A. 120 (2023) e2305255120. doi: 10.1073/pnas.2305255120

  33. [33]

      L.D. Lai, H.Y. Zhou, Y.C. Hong, et al., Chin. Chem. Lett. 35 (2024) 108580. doi: 10.1016/j.cclet.2023.108580

  34. [34]

      Y.C. Hong, H.Y. Zhou, Z.K. Xiong, et al., Chem. Eng. J. 391 (2020) 123604. doi: 10.1016/j.cej.2019.123604

  35. [35]

      L.J. Bu, S.Q. Zhou, S.M. Zhu, et al., Water Res. 146 (2018) 288–297. doi: 10.1016/j.watres.2018.09.036

  36. [36]

      R.N. Guo, L.C. Nengzi, Y. Chen, et al., Chin. Chem. Lett. 31 (2020) 2661–2667. doi: 10.1016/j.cclet.2020.03.068

• 

  Scheme 1  Proposed formation mechanisms of reactive species by the UV-induced photocatalytic cycle of TPPFe^ⅡCl-[TPPFe^ⅡCl]⁻.

  下载: 全尺寸图片 幻灯片
  

  Figure 1  Free energy diagram for the initial radical addition to C7, C8 double bond of CMZ reaction by ^•OH, SO₃^•−, SO₄^•−, and HO₂^•.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Free energy diagram for the reaction between CMZ and ¹O₂.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  The proposed degradation pathways of CMZ during interactions with different reactive species.

  下载: 全尺寸图片 幻灯片
•