Unveiling the environmental significance of acetylperoxyl radical: Reactivity quantification and kinetic modeling

Abstract Acetylperoxyl radical (CH3C(O)OO•) is among highly reactive organic radicals which are known to play crucial roles in atmospheric chemistry, aqueous chemistry and, most recently, peracetic acid (PAA)-based advanced oxidation processes. However, fundamental knowledge for its reactivity is scarce and severely hampers the understanding of relevant environmental processes. Herein, three independent experimental approaches were exploited for revelation and quantification of the reaction rates of acetylperoxyl radical. First, we developed and verified laser flash photolysis of biacetyl, ultraviolet (UV) photolysis of biacetyl, and pulse radiolysis of acetaldehyde, each as a clean source of CH3C(O)OO•. Then, using competition kinetics and selection of suitable probe and competitor compounds, the rate constants between CH3C(O)OO• and compounds of diverse structures were determined. The three experimental approaches complemented in reaction time scale and ease of operation, and provided cross-validation of the rate constants. Moreover, the formation of CH3C(O)OO• was verified by spin-trapped electron paramagnetic resonance, and potential influence of other reactive species in the systems was assessed. Overall, CH3C(O)OO• displays distinctively high reactivity and selectivity, reacting especially favorably with naphthyl and diene compounds (k ∼ 107–108 M−1 s−1) but sluggishly with N- and S-containing groups. Significantly, we demonstrated that incorporating acetylperoxyl radical-oxidation reactions significantly improved the accuracy in modeling the degradation of environmental micropollutants by UV/PAA treatment. This study is among the most comprehensive investigation for peroxyl radical reactivity to date, and establishes a robust methodology for investigating organic radical chemistry. The determined rate constants strengthen kinetic databases and improve modeling accuracy for natural and engineered systems.


Significance Statement
Organic radicals are ubiquitous in the atmosphere, aqueous environments, and living cells, yet their reactivity has been scarcely studied.Herein, we comprehensively investigated the reactivity of acetylperoxyl radical (CH 3 C(O)OO • , which has been suggested among the most oxidative organic radicals), with structurally diverse organic compounds using laser flash photolysis, pulse radiolysis and ultraviolet photoreactor approaches.This study not only unveils the high reactivity and selectivity of acetylperoxyl radical, facilitating more accurate modeling and elucidating their importance in environmental and catalytic oxidation processes, but also establishes a robust strategy for future research on organic radicals.

Introduction
Organic peroxyl radicals, commonly produced by oxygen (O 2 ) addition onto carbon-centered radicals, are ubiquitous in natural environments and living cells (1).Although most peroxyl radicals lack reactivity and undergo fast self-decay, acyl peroxyl radicals (R-C(O)OO • ) are among the most reactive and stable organic radicals (2)(3)(4).For instance, acetylperoxyl radical (CH 3 C(O)OO • ) is among the most long-lived (i.e. a first-order self-decay rate at 1.82 s −1 ) and reactive organic radicals with a standard redox potential at ∼1.6 V (vs normal hydrogen electrode) (5,6) Nonetheless, previous studies have mainly focused on inorganic radicals (e.g.atomic O, • OH, and • Cl), while the kinetic information of those reactive organic radicals has received relatively limited exploration.
To date, however, there is a dearth of information regarding the reactivity of CH 3 C(O)OO • , which hinders the delineation of its contribution in natural geochemical and biochemical cycles, as well as engineered AOP systems.For instance, the contribution of CH 3 C(O)OO • in the PAA-or diketone-based water decontamination processes has only been indirectly inferred or roughly estimated, rather than quantitatively examined due to the inability to distinguish CH 3 C(O)OO • from other radicals and high-valent metal species co-generated in the AOP systems (12-15, 26, 29).Therefore, an unambiguous, comprehensive investigation on the reactivity of CH 3 C(O)OO • is urgently needed.Herein, we developed a laser flash photolysis (LFP) method using 2,3-butanedione (hereinafter referred to as biacetyl) in aerobic solutions as a clean source to predominantly generate CH 3 C(O) OO • , and 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) as the colorimetric probe, to determine the rate constants between CH 3 C(O)OO • and compounds such as naproxen (NPX), trans-cinnamic acid (CINN), and others.Then, NPX and CINN were employed as probes for competition kinetics in a UV/biacetyl reactor, where the rate constants of CH 3 C(O)OO • with various structurally diverse organic compounds were comprehensively investigated by tracking their degradation during the UV/biacetyl AOP (SI Appendix, Table S1).Furthermore, the key rate constants were cross-validated by pulse radiolysis (PR) of acetaldehyde, Fig. 1.Generation pathways of acetylperoxyl radical and the methodology summary for quantifying the reactivity of acetylperoxyl radical (A), second-order rate constants between acetylperoxyl radical and organic compounds (B).
another clean source of CH 3 C(O)OO • (Fig. 1A).Overall, this study not only built a kinetic dataset (Fig. 1B) of CH 3 C(O)OO • , but also established a systematic approach for investigating CH 3 C(O)OO • that could be further utilized for studying other organic radicals.

Reactive species in LFP of biacetyl
The setup of the LFP system is illustrated in Fig. 2A (30).The radical precursor, biacetyl (10 mM), was photolyzed with an excimer laser at λ = 351 nm to generate CH 3 C(O)OO • in the presence of dissolved oxygen (∼ 2.6 × 10 −4 M, measured by a DO meter) (Fig. 2A, see later discussion).ABTS, whose reaction with CH 3 C(O)OO • has been studied previously [Eq. 3;(31,32)], was used as the probe for competition kinetics.
The absorbance spectra of reaction solutions before and after LFP experiments were measured by a UV-vis spectrophotometer.Obvious absorbance change was observed after LFP (Fig. 2B), with the maximum decrease at ∼ 340 nm and maximum increase at ∼ 415 nm, clearly indicating a conversion from ABTS (peak absorbance: ϵ 340 nm = 3.66 × 10 4 M −1 cm −1 ; Fig. 2C) to ABTS •+ radical (peak absorbance: ϵ 415 nm = 3.4 × 10 4 M −1 cm −1 ; SI Appendix, Figure S2).The ABTS •+ once produced was stable in the next few millisecond and was measured at 635 nm versus time (Fig. 2B, SI Appendix).Notably, the laser light at 351 nm did not lead to ABTS photolysis or ABTS •+ generation (control in Fig. 2B), suggesting that ABTS was oxidized by reactive species generated by laseractivated biacetyl.Preliminary experiments showed 248 nm laser not suitable due to photodegradation of ABTS.The photolysis of biacetyl has been extensively studied and the pathway is depicted as Fig. 2A (7,29,33).The triplet biacetyl undergoes C-C bond cleavage between two carbonyl groups to generate two CH 3 C(O) • with a ∼ 90% yield (reported at 313 nm) (7), which reacts with O 2 at 4.1 × 10 8 M −1 s −1 to generate CH 3 C(O) OO • (32).In addition, dissolved O 2 scavenges triplet biacetyl at 4.1 × 10 8 M −1 s −1 , producing either singlet oxygen (14,34).Due to the low reactivity of 1  .Their potential roles are addressed one by one as below.
1 O 2 is a weak oxidant that could not oxidize various susceptible organic compounds [e.g.tetramethyl-p-phenylenediamine (TMPD) (8)], hence its role in oxidizing the recalcitrant contaminants is minor (35,36).Furthermore, the lifetime of 1 O 2 is reported to be ∼ 3.8 µs in H 2 O (37,38), which is definitely too short to be responsible for the much longer term ABTS oxidation that lasted for milliseconds (Fig. 2B).
Recent literature showed the triplet biacetyl also exhibited oxidation capacity for some organic compounds including carbamazepine (CBZ) (29) and β-carotene (33), evidenced by their nonnegligible removal in the absence of oxygen (O 2 ), which is critical for the formation of CH 3 C(O)OO • .However, triplet organic matter is always rapidly quenched by the surrounding solvent and/or O 2 , hence their lifetime and oxidation kinetics typically last for only microseconds in the presence of O 2 (39,40).Previous literature has confirmed that, despite its oxidation contribution in anoxic conditions, triplet biacetyl could not contribute to oxidation in the presence of O 2 due to the fast quenching reaction that leads to the formation of either (41) also found the triplet state acetylacetone (another simple diketone) had a lifetime and oxidation duration at microsecond level.Furthermore, we estimated the lifetime of triplet biacetyl to be only 6.1 µs in aerobic solution due to the quenching by O (14,34), hence triplet biacetyl is not responsible for the millisecond-long oxidation of ABTS.
More importantly, we used PR of acetaldehyde to confirm that HO 2 , as well as CH 3 CO • , could hardly react with ABTS (see later Discussion).Therefore, the only candidate for ABTS oxidation in the LFP system was CH 3 C(O)OO • .

Observation of acetylperoxyl radicals by spin-trapped EPR
In situ EPR was harnessed for direct identification of CH 3 C(O)OO • (Fig. 3A, SI Appendix).We could not detect CH 3 C(O)OO • in water samples due to instability of the adduct of CH 3 C(O)OO • with DMPO (Fig. 3B), similar to previous research (43).Therefore, we employed tert-butyl alcohol (TBA), which has negligible reactivity toward CH 3 C(O)OO • , as the medium to stabilize the DMPO-peroxyl adduct without quenching the radical (43).During the in situ photolysis of biacetyl, DMPO-CH 3 C(O)OO • was accumulated, successfully identified, and simulated with the conditions marked in Fig. 3C , and its oxidation product by 1 O 2 (5,5-dimethylpyrrolidone-2(2)-oxyl-(1) (DMPOX) and DMPO-• OH) (44) were not observed in either deionized (DI) water or TBA solutions, indicating the minor roles of these reactive species during continuous irradiation of biacetyl.

The rate constant between ABTS and acetylperoxyl radical
As ABTS oxidation could be primarily attributed to CH 3 C(O)OO • , the rate constant between ABTS and CH 3 C(O)OO • could be first determined by modeling the pseudo-first-order ABTS •+ buildup rate at different initial ABTS concentrations (Eqs.4 and 5).3), k ABTS can be calculated by Eq. 6. final absorbance at 635 nm (i.e. when ABTS •+ formation reaches the plateau), and A t is the absorbance at time t.Preliminary tests have shown that the photolysis of biacetyl and the selected compounds could not produce signal at 635 nm, hence the A t and A final should be linearly proportional to the corresponding ABTS •+ concentration.
As shown in Fig. 2B and D, the buildup of 635 nm signal (indicating ABTS •+ formation) was faster with increasing ABTS concentration, consistent with the kinetic calculation.On the other hand, A final , indicating [ABTS •+ ] at the plateau and the total amount of photo-generated CH 3 C(O)OO • , decreased with increasing ABTS concentration.This is due to the competitive light absorbance of ABTS with biacetyl, where ABTS dominated the overall light screening at 351 nm despite its much lower concentration (Fig. 2C).According to Eq. 6, k ABTS was determined to be (2.0 ± 0.1) × 10 9 M −1 s −1 , by linear regression of the pseudo-first-order rate constants for signal buildup against initial ABTS concentrations (Figs.2D and 1B, SI Appendix, Table S1).This rate constant is in excellent agreement with the reported values, confirming the accuracy of our analytical methods (31,32).Notably, the oxidation of ABTS lasted for several milliseconds, indicating a relatively long lifetime of CH 3 C(O)OO • and ruling out the contribution of other transient intermediates (e.g.triplet biacetyl, 1 O 2 ) with microsecond-lifetime.

Competition kinetics study in LFP of biacetyl
In the competition kinetics experiments, biacetyl, ABTS, and compound X (at different concentrations) were mixed together and underwent pulse laser activation, where the photo-generated CH 3 C(O)OO • was consumed by either compound X or ABTS (Eqs.7 and 8).
k X is the second-order rate constant between selected compound X and .
A control is the final absorbance (indicating ABTS •+ formation) without compound X; and A X is the final absorbance with compound X.The reactivity of CH 3 C(O)OO • toward ascorbic acid has been investigated in previous studies (31,32).Hence, we tested the reactivity of ascorbic acid by competition kinetics with ABTS in the same LFP setup (Eqs.8 and 9) to further validate our approach.A final , indicating ABTS •+ formation, decreased remarkably with increasing ascorbic acid concentration, suggesting that the ABTS oxidation was inhibited due to the coexistent ascorbic acid competing for CH 3 C(O)OO • (Fig. 2E).Through a linear regression according to Eq. 9, the rate constant between ascorbic acid and CH 3 C(O)OO • was determined to be (9.4 ± 0.2) × 10 8 M −1 s −1 (Fig. 2E, SI Appendix, Table S1), concurring with previous studies and confirming the reliability of our method (32).

Competition kinetics study in a UV reactor with biacetyl
To investigate the rate constants between less reactive compounds with CH 3 C(O)OO • , we compared the degradation of selected organic compounds versus suitable probes during UV/ biacetyl AOP in a bench-scale UV (254 nm) reactor, which enables continuous CH 3 C(O)OO • generation and long-term monitoring of compound degradation.Biacetyl was continuously activated by UV 254 irradiation following the photochemical pathways in Fig. 2A, generating CH 3 C(O)OO • as the sole reactive species for degradation of selected organic compounds (in addition to direct photolysis degradation).Therefore, the difference between degradation rates during direct photolysis (without biacetyl addition) and UV/biacetyl could be attributed to the oxidation by CH 3 C(O) OO • , and the oxidation rate should be linearly proportional to their second-order rate constants with CH 3 C(O)OO • (Eq.10).
k obs,probe and k obs,X are the observed pseudo-first-order rate constants for the degradation of the probe compound (whose reactivity has been quantified) and compound X, respectively (in min −1 ); k UV,probe and k UV,X are their direct photolysis rate constants (in min −1 ); k probe and k X are the second-order rate constants between CH 3 C(O)OO • and the probe and compound X, respectively (in M −1 s −1 ).NPX and CINN, which can be easily measured by highperformance liquid chromatography and have been investigated by LFP, were selected as probes for the competition kinetics study in the UV reactor.Firstly, the method was validated by comparing the degradation of NPX and CINN in the UV reactor (SI Appendix, Figure S7).As calculated by Eq. 10, the observed pseudo-first-order oxidation rate constant (k obs− k UV ) of NPX was about 12 times higher than that of CINN, very close to the ratio of their second-order rate constants with CH 3 C(O)OO • obtained from the LFP system (SI Appendix, Table S1).Subsequently, the rate constants of CH 3 C(O) OO • with more organic compounds were determined by competition kinetics with either NPX or CINN in the UV reactor (SI Appendix, Figure S7).As summarized in Fig. 1B, SI Appendix, and Table S1, CH 3 C(O)OO • exhibited the highest reactivity toward naphthalene and NPX (k > 10 8 M −1 s −1 ), followed by naphthol, 2,4,6-trimethyphenol (TMP), 2,4-HD, CINN, and tryptophan (k ∼ 10 7 M −1 s −1 ).The reactivity with benzoic acid, bisphenol A, phenyl methyl sulfoxide (PMSO), methyl p-tolyl sulfide, ibuprofen (IBP), and diethyltoluamide (DEET) were too low to determine.
Overall, CH 3 C(O)OO • preferentially reacts with naphthyl compounds (k ∼ 10 7 -10 8 M −1 s −1 ), and, to a lesser extent, aromatic, and diene compounds (k ∼ 10 6 -10 7 M −1 s −1 ) (Fig. 1B, SI Appendix, Table S1).Notably, N-and S-containing compounds, that are electron-rich and susceptible to oxidation by • OH ( 46), • Cl (46), chlorine (47), high-valent iron (48), or peroxyacids (49,50) The above reactivity information can shed light on the recent debates about reactive species in PAA-related AOPs.In delineating the reactive species in metal-PAA AOPs, the 100% conversion of sulfoxides (e.g.DMSO and PMSO) to sulfones has been frequently used as evidence for the absence of CH 3 C(O)OO • , and the oxidation of CBZ has been attributed to other species due to its "suspected low reactivity" with CH 3 C(O)OO • (15,23).Herein, we demonstrated that CH 3 C(O)OO • reacts with CBZ at (3.0 ± 0.6) × 10 6 M −1 s −1 , hence holds the potential to degrade CBZ.On the contrary, given the negligible reactivity of CH 3 C(O)OO • toward S-containing compounds, the transformation products of sulfoxides cannot be used as the evidence for the absence of CH 3 C(O)OO • .

Reactive species in PR of acetaldehyde
PR of acetaldehyde was applied as an alternative clean source of CH 3 C(O)OO • to verify the rate constants (32).The absorbance spectrum of ABTS •+ obtained after PR was similar to that of LFP (Fig. 2B), indicating bleaching of the ground-state ABTS peak absorbance (∼340 nm) with the appearance of a transient absorbing species in the vicinity of 415 nm (Fig. 4B).As shown in Fig. 4B, the formation of ABTS •+ at 423 nm built up slowly within 1.0 ms and stayed stable in 3.0 ms.
In brief, the radiolysis of water leads to formation of hydrated electrons (e aq − ), hydroxyl radicals ( • OH), protons (H + ), and a low yield of hydrogen atoms and molecular products ( radicals as another candidate reactive species (Fig. 4A).Overall, the reactive species that may contribute to ABTS oxidation during We first tested PR of acetaldehyde with ABTS in an anoxic environment (N 2 O saturation only) and found that ABTS •+ transient generation was totally inhibited (Fig. 4B).The lack of O 2 suppressed the formation of CH 3 C(O)OO • and HO 2 , but could not reduce the yield of • OH and CH 3 C(O) • .Therefore, the absence of ABTS •+ signal in the anoxic solution confirmed that (i) • OH was completely scavenged by 10 mM of acetaldehyde hence could not lead to ABTS oxidation; and (ii) CH 3 C(O) • was ineffective for ABTS oxidation, thus, its role in the LFP system and UV reactor should also be negligible, which was consistent with previous studies (32).
Furthermore, H 2 O 2 itself is a relatively weak oxidant and we have confirmed that it could not react with ABTS (49).Finally, to evaluate the role of HO 2 • /O 2 •− , we used methanol, instead of acetaldehyde, as a scavenger of • OH, while not affecting HO 2 • /O 2 •− production from the reaction between O 2 and e aq − (Fig. 4A and C).As a result, we found that the ABTS radicals (Fig. 4A), they could not oxidize ABTS in the system (Fig. 4C).These results are consistent with the previous studies that reported the limited reactivity of carbon-centered radicals and alkyl peroxyl radicals, distinguishing CH 3 C(O)OO • from other organic radicals in terms of reactivity (2,4,12).

Competition kinetics study in PR of acetaldehyde
Then, PR of acetaldehyde was harnessed to verify the rate constants for two major probes in this study, i.e.NPX and CINN, using the competition kinetics approach (Eqs.[4][5][6].As shown in Fig. 4D and E, the rate constants for NPX and CINN were determined to be (1.0 ± 0.1) × 10 8 and (9.8 ± 0.6) × 10 6 M −1 s −1 , respectively, by the PR approach (SI Appendix, Table S1).These data are in excellent agreement with those obtained from LFP, confirming the reliability of the LFP approach and the effectiveness of these probes for UV reactor experiments.The rate constants for TMP and cysteine were also studied by PR (Fig. 1B, SI Appendix, Table S1, Figure S6).

Remodeling the reaction kinetics of UV/PAA
This study is among the first to establish a large reactivity dataset for an organic radical (CH 3 C(O)OO • ), which could be incorporated into the kinetic models for natural and engineered systems to enhance modeling accuracy.Herein, we take an established kinetic model on UV/PAA AOP as the example and demonstrate the importance of CH 3 C(O)OO • -related reactions.UV/PAA is an emerging (waste)water treatment technology that has been extensively applied for inactivation of pathogens (20,51,52) and degradation of micropollutants (12,19) (Fig. 5A, SI Appendix, Table S2).S3).These results further demonstrate the environmental significance of CH 3 C(O)OO • and indirectly verify the rate constants generated in this study.

Conclusions
Compared with the common inorganic radicals, the importance of ubiquitous organic radicals in the environment has been scarcely studied.In this study, we established systematic approaches for quantifying the reactivity of an organic peroxyl radical with high reactivity, i.e.CH 3 C(O)OO

Chemicals and reagents
The selected compounds for study are listed in SI Appendix, Table S1.The sources for these compounds and other chemicals are provided in SI Appendix.

Laser flash photolysis
The setup of the LFP system has been illustrated in our previous studies and Fig. 2A (30).The reaction solution containing biacetyl (10 mM), ABTS, and a selected compound was pumped through the reaction cell that received pulse laser activation at λ = 351 nm.Produced ABTS •+ was measured, by a continuous wave laser at 635 nm as the light source combined with a fast photodiode as the detector (SI Appendix), over the reaction time after activation by the laser pulse.

Spin-trapped EPR
EPR study was conducted with a Bruker EMX micro spectrometer and an ER 4103TM resonator to evaluate the radical species during biacetyl photolysis.DMPO was used as the trapping agent.The solution of biacetyl with DMPO (2 mM) was prepared in either water or TBA and irradiated by a 100 W mercury arc lamp that emitted light from 200 to 600 nm (connected by optical fiber, Fig. 3A), and the glass tube cut off any irradiation below 300 nm (SI Appendix).
The reactor was open to the atmosphere with a surface to volume ratio at 0.83 cm −1 to ensure sufficient reaeration during the experiments.Periodically, 0.5-mL aliquots were collected to amber vials for concentration analysis using an Agilent 1100 high-performance liquid chromatography equipped with a diode-array detector.The experiments were conducted in duplicate.

Pulse radiolysis
The PR experiments were performed using an 8 MeV Linear Accelerator at the Notre Dame Radiation Laboratory (SI Appendix) (53).The prepared solutions containing acetaldehyde, ABTS, and one organic compound, at pH 5.   S1 and S2).
the thiocyanate dosimeter.The competition kinetics experiments for each compound were conducted under the identical condition and provided consistent results.The transient kinetic analysis of the produced ABTS •+ was conducted at 423 nm.

Kinetic modeling
The rate constants for reactions between CH 3 C(O)OO • and organic contaminants are incorporated into the kinetic model for UV/PAA (SI Appendix, Table S2) (12), and the contaminant removal by UV/ PAA was modeled using the Kintecus program 4.55.31.
1 O 2 or CH 3 C(O) • + CH 3 C(O)OO • .For example, Mortensen et al. (33) distinguished triplet biacetyl and CH 3 C(O)OO • by their different oxidation products of β-carotene and confirmed the negligible role of triplet biacetyl in aerated solution.Darmanyan et al. (34) observed the rapid quenching of triplet biacetyl by O 2 directly by the transient absorption spectra changes in the solution in equilibrium with air.Jin et al.

k
ABTS is the second-order rate constant between ABTS and CH 3 C(O)OO • (in M −1 s −1 ); t is reaction time after activation by the laser pulse (in s); [ABTS] t and [CH 3 C(O)OO • ] t are the concentrations of ABTS and CH 3 C(O)OO • at the particular reaction time t (in M), respectively; [CH 3 C(O)OO • ] 0 is initial concentration of CH 3 C(O)OO • generated by the pulse laser (in M).As ABTS was in excess of the produced CH 3 C(O)OO • , we could assume the ABTS concentration remained close to the initial concentration.Considering the 1:1 stoichiometry between CH 3 C(O)OO • and ABTS (Eq.
are the concentration changes of X, ABTS, and ABTS •+ , respectively, after the ABTS oxidation reaches the plateau (in M).Note that the change in overall absorbance at 351 nm was negligibly affected by selected compounds, hence the addition of compound X could not affect CH 3 C(O)OO • production by light-shielding or producing additional reactive species.In other words, [CH 3 C(O)OO • ] 0 remained at the same value regardless of the addition of compound X, and the difference between [CH 3 C(O)OO • ] 0 and [ABTS •+ ] final should be totally attributed to the consumption by compound X (i.e.Δ[X]) (Eq.9) , are not efficiently degraded by CH 3 C(O)OO • , distinguishing CH 3 C(O) OO • from other reactive species in environmental processes.According to previous studies, CH 3 C(O)OO • usually react with aromatic and diene compounds via radical addition to C=C double bonds, followed by CH 3 C(O)O • -leaving reaction, resulting in epoxidation reaction (Eq.11) (3, 33, 45).On the other hand, CH 3 C(O)OO • oxidize ABTS and TMPD via an electron transfer pathway, producing PAA and organic radical (ABTS •+ or TMPD •+ ) (8).The oxidation mechanism should be further elucidated by transient spectra or mass spectrometry techniques for oxidation products.
7. Saturation with N 2 O or N 2 O/O 2 mixtures was achieved by bubbling for at least 25 min per 80 mL sample.Mixtures of N 2 O/O 2 were calibrated using flow meters and maintained at the same level in all experiments.Solutions were prepared fresh and protected from adventitious exposure to room light.The radiation dose was 10.4 Gy per 4 ns pulse or 16 Gy per 5 ns pulse, as determined by

Fig. 5 .
Fig. 5. Mechanisms for acetylperoxyl radical generation by UV/PAA AOP (A), relative contribution of UV, • OH, and CH 3 C(O)OO • to organic compound degradation by UV/PAA (B), kinetic modeling of the degradation by UV/PAA (C-G).The experimental data are retrieved from our previous studies (Zhang and Huang; Cai et al.) with permission(12,19).The new UV/PAA model included the reactions from the old model(12) and new reactions between CH 3 C(O) OO • and organic compounds (SI Appendix, TablesS1 and S2).
, further confirming its critical role in the UV/biacetyl system.Notably, our EPR pattern was consistent with Hoshino et al.'s (8) result for aerobic biacetyl photolysis in benzene solution.
(12,20)f, photolysis of PAA by UV cleaves the O-O bond and generate • OH and CH 3 C(O)O • , which in turn produce CH 3 C(O)OO • , • OOCH 3 , and • CH 3 through radical chain-reactions.Unlike CH 3 C(O)O • and• CH 3 , whose reactivity and steady-state concentrations are too low to contribute to decontamination, the steady-state concentration of CH 3 C(O)OO • is ∼3 orders of magnitude higher than • OH, indicating its significant contribution in oxidation(12,20).However, the contribution of CH 3 C(O)OO • could not be included in the kinetic model previously due to lack of reliable reactivity information, leading to an underestimation of the overall oxidation efficiency.Herein, we integrated the reactions between CH 3 C(O)OO • and organic contaminants into the UV/PAA kinetic model to simulate the removal of micropollutants and to quantify the relative contribution of direct UV photolysis, CH 3 C(O)OO • , and • OH in the system (Fig.5B).As shown in Fig.5C-G, the incorporation of CH 3 C(O)OO • oxidation reactions significantly improved the accuracy of modeling micropollutant degradation by UV/PAA.Although its contribution is negligible for IBP and DEET due to low reactivity, CH 3 C(O)OO • accounted for 24.08, 40.78, and 73.04% removal of caffeine, CBZ, and NPX, respectively (Fig.5B, SI Appendix, Table • .LFP of biacetyl, UV reactor with biacetyl, and PR of acetaldehyde under aerobic condition, were verified as clean sources of CH 3 C(O)OO • .The second-order rate constants between CH 3 C(O)OO • and 34 different organic compounds demonstrated high reproducibility during cross-validation and literature comparison.CH 3 C(O)OO • exhibited the highest reactivity toward naphthyl and diene compounds, while showed much less capacity in oxidizing N-, and S-containing functional groups.Even with such structural selectivity, the high reaction rates of CH 3 C(O)OO • render its nonnegligible contribution in environmental processes.Indeed, incorporating the new kinetic information remarkably improved the modeling accuracy for UV/PAA AOP, where CH 3 C(O)OO • contributes significantly to the oxidation of several common organic contaminants in water.The kinetic dataset and experimental approaches established by this study will be useful to facilitate future research on organic peroxyl radicals.