Distinguishing homogeneous advanced oxidation processes in bulk water from heterogeneous surface reactions in organic oxidation

Significance Heterogeneous catalytic oxidation technologies for water purification, such as Fenton and Fenton-like catalytic oxidation, involve intricate interfacial reactions at the solid catalyst surface and in bulk solution. To date, the difference in the reaction pathways on the catalyst surface and in bulk water solution has not been recognized. In this work, we reveal a widespread surface-dependent reaction pathway that is fundamentally different from the previously accepted pathway. We further elucidate the changes in reaction pathways as oxidizing species (e.g., Mn(III), •OH) detach from the catalyst surface to the aqueous solution. Our study provides new insights on heterogeneous chemical oxidation and catalytic oxidation reactions, potentially leading to the design of more efficient nanocatalysts.

Heterogeneous chemical oxidation reactions are ubiquitous in nature and play important roles in various chemical reaction systems associated with energy conversion, environmental remediation, and catalysis (1)(2)(3)(4). Rapid advances in nanocatalysts have significantly improved the activity and stability of such reaction systems, opening the door for expanded applications. Nevertheless, one fundamental question remains: "How do the chemical reactions at the solid-water interface differ from those in bulk water?" (5).
Recent studies suggest that reactions occurring at the air-water interface and in bulk water solution differ drastically in reaction kinetics, with reaction rates based on the same reaction pathway varying greatly (5)(6)(7). We surmise that altered reaction kinetics compared to that in bulk water solution would also be expected at the solid-water interface of solid catalysts. However, since the nanocatalyst surface is greatly affected by factors such as morphology, atomic composition, atomic vacancies, crystal facets, and surface stress (8)(9)(10), the research blind spot for reaction kinetics is no longer the reaction rate, which is easy to measure, but rather the reaction pathway.
To date, the difference in reaction pathways between the solid-water interface and bulk water solution has not been fully revealed. For example, various Fenton and Fenton-like catalytic systems for organic pollutant oxidation have been reported-from the homogeneous Fe 2+ /hydrogen peroxide and Co 2+ /persulfate systems to the heterogeneous Fe-or Co-based solid catalyst/oxidant systems. In these reaction systems, the ions in solution were thought to play the same catalytic role as their counterparts on the heterogeneous catalyst surface and the removal of pollutants was considered to follow similar degradation/mineralization pathways (11)(12)(13)(14)(15)(16)(17). However, we recently revealed three unexpected functions, i.e., activation, stabilization, and accumulation, of catalyst surfaces for persulfate-triggered catalysis and organic pollutant oxidation (18), which inspire us to unravel the differences in reaction pathways between the solid-water interface and the bulk water solution. Clarifying the reaction pathways is of utmost importance for the design and optimization of heterogeneous catalytic oxidation processes and technologies.
In this work, we shed light on the different reaction pathways of organic oxidation at the solid catalyst surface and in aqueous solution. The reaction systems were constructed by using phenol (PhOH), a common compound in the chemical industry and a ubiquitous environmental pollutant (19)(20)(21), as a model organic compound, and a series of metal ions or metal oxide nanomaterials as the oxidants/catalysts. Manganese (Mn) is the second most abundant redox-active transition metal on earth with multiple valences [i.e., Mn(II), Mn(III), and Mn(IV)] in natural environments, such as the ocean floor, soils and sediments, and freshwater bodies (4,22). Manganese oxides (MnO X ) are one of the strongest naturally occurring oxidants with important roles in biogeochemical elemental cycles, soils, and water treatment (23)(24)(25)(26). First, by using high-valent MnO X solids (Mn 3 O 4 , Mn 2 O 3 , and MnO 2 ) and high-valent Mn ions (Mn 3+ ) as oxidants, the removal behavior, reaction pathways, and reaction sites of PhOH in heterogeneous and homogeneous oxidation systems were thoroughly investigated. We revealed an interface-dependent organic oxidation pathway, which has been considered an aqueous advanced oxidation process (AOP) in previous studies. Second, similar reaction pathway differences were extended to other heterogeneous catalytic oxidation systems (i.e., FeOCl catalyzing hydrogen peroxide and Co 3 O 4 catalyzing persulfate) and the corresponding homogeneous ion catalytic oxidation (i.e., Fe 2+ catalyzing hydrogen peroxide and Co 2+ catalyzing persulfate) systems. Overall, our work clarified the reaction pathways in diverse homogeneous oxidation systems and their corresponding heterogeneous oxidation systems. These findings will inform and direct future studies on heterogeneous chemical (catalytic) oxidation processes, which are conventionally considered to proceed only by aqueous solution AOPs.

PhOH Oxidation Pathway in MnO X Heterogeneous Oxidation
Systems. We first investigated the chemistry of heterogeneous PhOH oxidation by various MnO X , including MnO, Mn 3 O 4 , Mn 2 O 3 , and MnO 2 (SI Appendix, Figs. S1-S6). All the MnO X , except for MnO (which lacks high-valent Mn species), exhibited considerable activity toward PhOH oxidation (Fig. 1A) /PhOH, and MnO 2 /PhOH systems, the removal efficiencies of total organic carbon (TOC) in these reaction processes were evaluated. The aqueous TOC was synchronously removed with PhOH at high efficiency in all the test groups ( Fig. 1 B-D), which differs from the PhOH degradation behavior in AOPs (20). To clarify such an unusual phenomenon of aqueous TOC removal, we examined the surface characteristics of MnOx before and after the reaction. The transmission electron microscopy (TEM) images showed lighter contrast layers on the MnO X nanosurface ( Fig. 1E and SI Appendix, Figs. S7 and S8). In addition, the measurement of the mass loss of organics by thermal gravimetric analysis (TGA) also suggests that the majority of PhOH in the aqueous solution was transferred and accumulated onto the MnO X nanosurface in the oxidation processes ( Fig. 1F and SI Appendix,  (Table 1). These results suggest that a direct oxidative transfer process (DOTP), rather than AOP, may predominate in the high-valent MnO X /PhOH oxidation systems, during which organics are not decomposed, but oxidized and accumulate on catalyst surfaces.
To further decipher the pathways of PhOH oxidation at the MnO X surface, we identified the molecular structures of the surface-accumulated products. Taking the Mn 2 O 3 /PhOH system as an example, we tried to elute the reacted Mn 2 O 3 with ethanol and toluene to dissolve the surface-accumulated product but failed to do so (SI Appendix, Fig. S10), implying that the product might be in a cross-linked state. Having three active sites in its molecular structure (i.e., the ortho-and para-positions of the hydroxyl group), PhOH is readily oxidized and forms a network-like cross-linked polymer that is insoluble in organic solvents (18). If this is the case, the formation of the cross-linked product would be prevented when two of the three active sites in the reactant are sheltered. For validation, we conducted the hypothetico-deductive experiment using 2,6-dimethyl-phenol (2,6-M-PhOH) as the reactant, which has only one active hydrogen site preserved in its molecular structure. Hypothesis. The 2,6-M-PhOH oxidation on the Mn2O3 surface would proceed by the same pathway as that of PhOH but form a noncross-linked product that can be readily dissolved in organic solvents. To provide direct evidence of the DOTP pathway, we identified the dissolved 2,6-M-PhOH oxidation products by liquid chromatography mass spectrometry (LCMS), gel permeation chromatography (GPC), and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS). The results indicate that the substances dissolved in ethanol were biphenyl quinone compounds (i.e., 3,3′,5,5′-tetramethyl-diphenoquinone) ( Fig. 2 A and B), while those dissolved in toluene were identified as chain-like polyphenyl ethers with a peak molecular weight of ~4,635 Da ( Fig. 2 D and E). The former could be generated by the C-C coupling of 2,6-M-PhOH (i.e., Fig. 2C) (27,28), and the latter may be derived from the C-O polymerization of 2,6-M-PhOH ( Fig. 2F) (29,30). Therefore, the PhOH oxidation in the high-valent MnO X systems proceeds by both the coupling reaction (CR) and polymerization reaction (PR) pathways, two types of DOTP pathways.

Dependence of the DOTP Pathway on the Solid-Water Interface.
Next, we performed potassium iodide (KI) oxidation experiments to further identify the PhOH reaction sites in the MnO X oxidation system (Mn 3 O 4 , Mn 2 O 3 , and MnO 2 ). We surmised that dissolved high-valent Mn ions from MnO X under acidic conditions, if present, would efficiently oxidize KI, and this chromogenic reaction can be easily detected by Ultraviolet visible (UV-Vis) absorption spectroscopy. However, no optical or spectral changes in the supernatant solution (i.e., the filtrate of the MnO X / H + suspension) were observed after mixing with KI ( Fig. 3A), suggesting that no dissolution of free high-valent Mn ions occurred in the acidic Mn 3 O 4 , Mn 2 O 3 , and MnO 2 nanoparticle suspensions. Interestingly, when KI was directly added to these acidic MnO X suspensions, the chromogenic reaction occurred immediately (Fig. 3A). Apparently, PhOH oxidation occurred predominantly on the nanosurface of high-valent MnO X (i.e., at the solid-water interface) instead of in the bulk water solution.
Since Mn 3 O 4 , Mn 2 O 3 , and MnO 2 do not dissolve in an acidic environment without redox reactions (SI Appendix, Tables S1 and S2), the released Mn ions in the high-valent MnO X /PhOH reaction systems should result from MnO X  Fig. 3E.

PhOH Oxidation Pathway in Mn 3+ Homogeneous Oxidation
System. To further confirm that the DOTP reaction of PhOH and solid Mn(III, IV) (the reactive species in high-valent MnO X ) is indeed a surface-dependent pathway and to elucidate the underlying mechanisms, we constructed a homogeneous PhOH oxidation system with dissolved-state Mn(III) (i.e., free Mn 3+ in water) for comparison. Here, free Mn 3+ was obtained by two consecutive steps: coordination dissolution and acidification dissociation (Fig. 4A). First, Mn 2 O 3 was dissolved into bulk water with the help of pyrophosphate (PP), a strong but nonredoxactive ligand (11,31), to form a homogeneous Mn(III)-PP bulk solution ( Fig. 4B and SI Appendix, Fig. S13A). The oxidizing ability of Mn(III) in this Mn(III)-PP complex is restricted by the ligands, as evidenced by its inability to oxidize KI (SI Appendix, Fig. S13B). Then, Mn(III)-PP was acidified with H 2 SO 4 , by which process Mn(III) was freed to bulk water via proton (H + ) substitution ( Fig. 4B and SI Appendix, Fig. S13B). The amount of    obtained free Mn 3+ was quantified by KI oxidation chromogenic reaction (a typical one-electron transfer reaction) (SI Appendix, Fig. S14A) following an analogous procedure used for persulfate [including peroxymonosulfate (PMS) and peroxydisulfate (PDS)] concentration determination (24,32). The PhOH removal behavior in the Mn 3+ homogeneous oxidation system was explored qualitatively and quantitatively. At low Mn 3+ dosages (i.e., no more than twice the molar amount of PhOH), the TOC in bulk water remained almost unchanged, although PhOH was rapidly and largely removed (Fig. 4 C and D  and SI Appendix, Fig. S14B). This result indicates that the oxidation products remained in the bulk water, which is consistent with the PhOH removal behavior in AOPs (20). Subsequently, we separated the Mn 3+ /PhOH reaction product from the solution by ultrahigh-performance liquid chromatography (UHPLC). High-resolution mass spectral analysis of the separated and recovered product showed the formation of both dimers and trimers of PhOH (Fig. 4 E and F and SI Appendix,  Fig. S15), which should result from the oligomerization of phenoxy radicals (23,33). Therefore, the PhOH reaction pathways in the Mn 3+ homogeneous oxidation system (dominated by radical polymerization reaction, a typical AOP) differed fundamentally from those in the DOTP-based heterogeneous MnO X reaction system. These findings clarify the long-standing misunderstandings about AOPs in MnO X oxidation systems (17,25,34,35) and identify the change of reaction pathways as oxidizing

Universality of the Pathways in Fenton and Fenton-Like Catalytic
Oxidation Systems. Similar to the homogeneous Mn 3+ oxidation system, the classical Fe 2+ /H 2 O 2 Fenton catalytic oxidation system is dominated by a radical pathway in organic pollutant degradation/mineralization (7,36,37). As expected, the removal behavior of aqueous PhOH and TOC in the homogeneous Fe 2+ / H 2 O 2 reaction system was analogous to that in the Mn 3+ oxidation system: Rapid PhOH removal but no obvious TOC removal from the bulk water was observed at low H 2 O 2 dosage (Fig. 5 A and B). The reaction products in the Fe 2+ /H 2 O 2 reaction system were mainly benzenediols and benzenetriols [the free-radical adducts of PhOH (16,20)] and dimers of PhOH (oligomerization of phenoxy radicals), which resulted from the radical degradation and polymerization of PhOH, respectively (23, 33) (both belong to AOPs) (Fig. 5 C and D and SI Appendix, Fig. S16). Compared with the Mn 3+ oxidation system, the occurrence of the PhOH radical degradation pathway in the Fe 2+ /H 2 O 2 catalytic oxidation system might be attributed to the higher oxidizing ability of hydroxyl radicals (standard oxidation potential 2.7 V) than that of Mn 3+ (oxidation potential 1.54 V) (38)(39)(40).
The results from the homogeneous reaction systems of Mn 3+ / PhOH and hydroxyl radical/PhOH in bulk water indicate that the oxidation pathway of PhOH was an indirect oxidation process. In these systems, the oxidizing species of Mn 3+ and hydroxyl radicals were first generated from high-valent MnO X and H 2 O 2 , which involved a one-electron-transfer intermediate step. Then, the generated oxidizing species underwent radical degradation and radical polymerization reactions, depending on their oxidation potential. The elementary reaction pathways of PhOH oxidized by Mn 3+ and hydroxyl radicals are detailed in Fig. 5E. Such AOP reaction pathways differ fundamentally from the abovementioned DOTP in heterogeneous oxidation systems.
Corresponding to the homogeneous Fenton system in bulk water, the oxidation pathways of PhOH in the heterogeneous Fenton system were also explored by using FeOCl [the most efficient Fe-based nanocatalyst (14,41)  to PhOH (i.e., 2:1), the TOC and chemical oxygen demand (COD) removal efficiencies were still at high levels similar to those in the MnO X oxidation system (SI Appendix, Fig. S18  C and D), suggesting that the DOTP pathway might also dominate the PhOH oxidation in the FeOCl/H 2 O 2 heterogeneous Fenton system. This was verified by product analysis, which showed the accumulation of PhOH oxidizing on the reacted FeOCl surface (formed from surface-dependent coupling and polymerization pathways) (SI Appendix, Figs. S19 and S20). Similar results were found for the PhOH oxidation pathways in the Fenton-like process of Co 2+ /PMS (homogeneous catalytic system) and Co 3 O 4 /PMS (heterogeneous catalytic system) (SI Appendix, Figs. S21 A and B, S22, and S23). Together, these results indicate that the difference in the organic oxidation pathways between the solid-water interface and bulk water solution exists universally in diverse chemical (catalytic) oxidation systems (Fig. 6).

Discussion
Our systematic analysis of PhOH oxidation in homogeneous catalytic oxidation systems and their counterpart heterogeneous catalytic oxidation systems clarified the changes in the reaction pathways as oxidizing species (i.e., high-valent Mn, hydroxyl radical, and sulfate radical) detach from the heterogeneous surface to aqueous solution. In the homogeneous Mn 3+ oxidation system as well as Fe 2+ /H 2 O 2 (Fenton) and Co 2+ /PMS (Fenton-like) catalytic oxidation systems, organics are mainly removed by AOPs via radical degradation and radical polymerization pathways (one-electron indirect oxidation). In these reaction systems, the   oxidants must first be activated to produce highly reactive intermediates (such as ROS) via one-electron pathway, with the latter primarily responsible for the pollutant degradation or polymerization. Such radical-mediated reactions are inevitably energy and chemical intensive. These reactions also suffer from incomplete pollutant removal from water and generation of hazardous by-products, which limit their application in water purification. In contrast, the heterogeneous reaction systems were dominated by surface-dependent coupling and polymerization reactions via two-electron direct oxidation pathway (i.e., DOTP reaction), which allows nondegradative yet highly efficient, complete removal of organic pollutants from water. Specifically, the oxidants on nanocatalyst surfaces directly oxidize the organic pollutants via a two-electron reaction pathway without an ROS intermediate step. In such processes, the oxidized pollutant intermediates are stabilized by the catalyst surface and spontaneously undergo surface coupling and polymerization reactions with other intermediates or pollutant molecules. Their products accumulate in situ on the catalyst surface and the aqueous pollutants are completely removed. The entire process requires much less oxidants and leaves almost no residual by-products in water. Such surface-dependent DOTP reactions were universally found in the high-valent MnO X (i.e., Mn 3 O 4 , Mn 2 O 3 , and MnO 2 ) oxidation systems as well as solid FeOCl/H 2 O 2 (heterogeneous Fenton) and Co 3 O 4 /persulfate (heterogeneous Fenton-like) catalytic oxidation systems, suggesting a great potential for water purification applications.
Distinguishing the difference between homogeneous and heterogeneous oxidation (i.e., the vast difference in pollutant removal effect and kinetics caused by different reaction pathways) would allow us to develop more efficient, economically affordable water treatment nanotechnologies. In the long term, environmentally benign, simple, efficient, and low-cost heterogeneous chemical oxidation (catalytic oxidation) technologies will be highly desired for the purification of organic-polluted wastewaters. Therefore, our findings shed light on a long-lasting misunderstanding of chemical oxidation/catalytic oxidation processes at solid-water interfaces, which may guide the development of more efficient and sustainable water purification systems.  (11,42). Specifically, 2 g Mn 2 O 3 was first added to 100 mL 40 mM Na 4 P 2 O 7 ·10H 2 O (PP) solution. Then, the pH of the suspension was adjusted to 6.5 with H 2 SO 4 . After continuous stirring at ambient temperature for 5 d, the suspension was centrifuged and filtered to obtain the Mn(III)-PP stock solution. Nano FeOCl was also synthesized according to a reported procedure (14,43). Briefly, 2 g FeCl 3 ·6H 2 O powder was first heated and annealed at 220 °C for 2 h. Then, the obtained FeOCl powder was washed with acetone and water. The pH adjustment and acidification of the reaction solution were all performed by using H 2 SO 4 (nonredox activity and no coordination ability). Unless otherwise specified, all chemicals were used as received without further purification. In the above reaction systems, the concentrations of PhOH, TOC, and COD over time were measured to monitor the reaction progress. For the reaction systems requiring TEM and TGA characterizations, the concentration of reactants was usually increased 10 ~ 20 times. All experiments were conducted in triplicate, and error bars are expressed as the arithmetic mean ± SD.

Materials and Methods
Validation of Surface-Dependent Effect. The surface-dependent effect of the DOTP pathway was validated by KI oxidation and chromogenic reactions in high-valent MnOX/PhOH reaction systems. Specifically, a certain amount of Mn 3 O 4 , Mn 2 O 3 , and MnO 2 powders was first added to the dilute sulfuric acid solution (pH = 2.0) and stirred with ultrasonication for 60 min. Then, the suspension was divided into two portions. One portion of the suspension was centrifuged and filtrated to obtain the bulk solution, which was subsequently mixed with KI solution and detected with UV-Vis absorption spectroscopy (2450, SHIMADZU Co., Japan). Another portion was directly mixed with KI solution, and the supernatant was detected with the same UV-Vis absorption spectroscopy. 3+ Concentration. The Mn 3+ concentration in the Mn 3+ stock solution was quantified by KI oxidation and chromogenic reaction. Specifically, 1 mL Mn(III)-PP stock solution (diluted if necessary) was first mixed with 1 mL KI solution (100 g L −1 , containing 5 g L −1 NaHCO 3 ). Then, a certain amount of H 2 SO 4 was added to the mixture to free Mn 3+ and initiate KI oxidation and chromogenic reactions. After 30 min of reaction, the mixture was analyzed by a UV-Vis absorption spectrometer (2450, SHIMADZU Co., Japan) at 396 nm. For absolute quantification of Mn 3+ (the one-electron transfer oxidant), the standard substance PDS (the two-electron transfer oxidant) was used to calibrate the standard curve of oxidant concentration versus absorbance intensity. The procedure used in the PDS/KI oxidation and chromogenic reaction was the same as that used for Mn 3+ /KI.

Solid-water interface reaction
Bulk-water solution reaction  Analytical Methods. The qualitative and quantitative analyses, including PhOH, TOC, and COD concentrations, hypothetico-deductive experiments, and characterization techniques (X-ray powder diffraction (XRD), TEM, TGA, XPS, inductively coupled plasma-mass spectrometer (ICP-MS), LC-MS, GPC, MALDI-TOF-MS), are available in SI Appendix. Data, Materials, and Software Availability. All study data are included in the article and/or SI Appendix.