Photoinduced formation of persistent free radicals, hydrogen radicals, and hydroxyl radicals from catechol on atmospheric particulate matter.

Catechol is speculated to be a potential precursor of environmentally persistent free radicals (EPFRs) in the atmosphere. EPFRs absorbed on PM2.5 have attracted public attention because their toxicity is similar to cigarette smoke. In this study, we found that catechol could produce EPFRs, which were oxygen-centered phenoxy and semiquinone radicals. These free radical species had half-lives of up to 382 days. CaO, CuO, and Fe2O3 markedly promoted EPFR formation from catechol. The valence states of Cu and Fe changed during the photochemical reactions of catechol but no valence state changed for Ca. Alkaline nature of CaO is possibly the key for promoting the free radical formations through acid-base reactions with catechol. In addition to hydroxyl free radicals, hydrogen free radicals and superoxide anions formed from the photochemical reactions of catechol were first discovered. This is of concern because of the adverse effects of these free radicals on human health.

Introduction

Catechol is an important organic chemical that is intentionally manufactured and used as a chemical intermediate, antibacterial agent, and industrial additive in rubber production, adhesive, galvanizing, preservatives, pesticides, and production of epinephrine(Bhuiyan et al., 2020; Esguerra, 2017; Sedo et al., 2013). Catechol can be unintentionally produced and released from full-scale industrial activities such as metallurgy, waste incineration, and chemical manufacture. Catechol is also a natural polyphenolic compound that is present in a wide range of plants such as tea, vegetables, and fruits (Sun Yugang et al., 2000). Plants and lignin decomposition are important sources of catechol in the atmosphere and other environment compartments. Moreover, catechol can be formed by daily human activities such as cooking and household heating by combustion of coal or biomass (Kibet et al., 2015; Sedo et al., 2013). Catechol emissions to the atmosphere through biomass burning were investigated and the average emission ratio of benzenediols (catechol, resorcinol) to CO was 0.37 mmol mol/CO by using the predominant vegetation at two California facilities as fuels (Veres et al., 2010). The total concentrations of phenols in the rainwater ranged from 0 to 1383 μg/L in the east of France (Schummer et al., 2009). Thus, catechol is a ubiquitous organic precursor in the environment, and it can arise from various anthropogenic activities.

Phenols are important precursors for the formation of many highly toxic organic pollutants, including the notorious carcinogenic dioxins (Evans and Dellinger, 2003, 2005; Nishinaga et al., 1977). Environmentally persistent free radicals (EPFRs) are emerging environmental pollutants that could potentially be formed from phenols (Burcat et al., 2003; Cook et al., 1956; Steelink, 1965). EPFRs can contribute to the formation of reactive oxygen species in simulated lung fluid(Tong et al., 2017). The health risks arising from EPFRs in fine atmospheric particulate matter (PM2.5) are reportedly similar to that of cigarette smoking and could explain why non-smokers develop cancers similar to those seen in smokers (Dellinger, 2008; Pryor et al., 1983). It has also been confirmed that EPFRs cause DNA damage (Gehling et al., 2014; Kelley et al., 2013; Khachatryan et al., 2011; Valavanidis et al., 2013). Thus, EPFRs are an emerging concern because of health risks associated with their inhalation along with PM2.5 or dietary intake.

EPFRs have been detected in multiple environmental samples, including PM2.5, soil organic matter, and industrial fly ash samples(Gehling and Dellinger, 2013; Gehling et al., 2014; Jia et al., 2017; Kiruri et al., 2013; Yang et al., 2017a, 2017b). The two major sources of EPFRs are primary release from various anthropogenic activities and secondary formation from precursors in the environment (Yang et al., 2017a). For primary sources, EPFRs are formed and released by waste incineration, metallurgical processes, and combustion of coal and biomass (Yang et al., 2017a). Although secondary formation of EPFRs in the atmosphere is hypothesized to be likely, it is not clear if EPFRs and highly active free radicals can be formed from catechol in the atmosphere, and key factors affecting their formation are not fully understood. Organic precursors are essential for EPFR formation (Lomnicki et al., 2008; Mahmood et al., 2018; Mas-Torrent et al., 2012; Vejerano et al., 2011). Catechol is a widespread organic precursor in the environment, and the formation potential of EPFRs from catechol under natural conditions and the mechanism and influencing factors need to be clarified. Light irradiation is important for triggering the formation of EPFRs and highly active free radicals in natural environments, and especially in the atmosphere (Awwad et al., 2020; Krapf et al., 2016). It is essential to consider the effect of light irradiation on EPFR formation because it could affect secondary formation of EPFRs in the atmosphere.

In this study, catechol was selected as a typical organic precursor because it is ubiquitous in the environment. The formation potentials of EPFRs and highly active free radicals from catechol were evaluated under light irradiation and taking into consideration secondary formation of EPFRs. The results of this study could improve the understanding of the formation of EPFRs and highly active free radicals from catechol under natural conditions.

Results and discussion

Electron spin levels of free radicals formed from catechol by photochemical reactions

Photoinduced electron transfer is an important pathway for initiating oxidation–reduction reactions of organic precursors, including phenols. Free radicals could be involved in and formed by the photochemical reactions of organic precursors; however, the electron spin levels of free radicals formed from catechol through photochemical reactions with different sorts of metal oxide loaded need to be accurately quantified. In this study, the electron spin levels of free radicals formed from catechol under UV and visible light irradiation were evaluated taking into considering the percentages of UV (7%) and visible light (50%) in sunlight(Xu et al., 2016).

The electron spin levels of free radicals formed from catechol under UV irradiation are shown in Figure 1. Distinct signals for free radicals were detected in the experiments conducted with UV irradiation compared with without UV irradiation (Figure 1A). UV irradiation greatly increased the levels of the free radicals. These results showed that free radicals were formed from catechol through photochemical reactions under UV irradiation.

Figure 1

The formation of EPFRs in CT/SiO2 system under UV irradiation

(A) Concentrations of EPFRs on the surface of SiO2 loaded with catechol (CT) and a metal oxide (Fe2O3, CaO, or CuO) or without a metal oxide. Changes in the EPFRs with time under UV irradiation.

(B) EPR spectra for SiO2 loaded with catechol (CT) and a metal oxide (Fe2O3, CaO, or CuO) or without a metal oxide after UV irradiation for 5500 s.

(C) The g values and concentrations of organic free radicals in reaction system samples after UV irradiation for 5500 s. Blank conditions: 5% Fe2O3 + 5% CuO +5% CaO +85% SiO2, with UV irradiation. CT-Dark conditions: 5% catechol +5% Fe2O3 + 90% SiO2, without UV irradiation.

The effects of normal metal oxides, including Fe2O3, CuO, and CaO, on the generation of free radicals from catechol were studied under UV irradiation (Figure 1A). Addition of only Fe2O3 to the reaction system of catechol on SiO2 did not markedly promote the formation of free radicals. However, addition of CaO and CuO to the reaction system greatly increased the electron paramagnetic resonance spectroscopy (EPR) signal. In a previous study, the higher oxidation potential of Fe2O3 resulted in greater decomposition of the adsorbate, which led to lower EPFR yields (Vejerano et al., 2011). This supported our observations in the present study. In thermochemical reaction system, CuO may act as an important surface catalyst and could mediate the EPFR formation (Lomnicki et al., 2008). When CaO was added to the reaction system, the phenol decomposition rate markedly increased, and the reaction time decreased by 2.1–2.6 times (Shin et al., 2020). CaO could absorb CO2 and H2O produced during pyrolysis of biomass and reduce the activation energy required for the reaction, which would promote the pyrolysis (Wang et al., 2020). In this study, CaO also greatly enhanced the reactivity of catechol.

There are distinct differences between UV and visible light, with the energy of visible light being lower than that of UV. Therefore, we also evaluated the formation of free radicals from catechol under visible light irradiation. Under visible light irradiation, free radicals were produced from catechol on Fe2O3/SiO2 (Figure S1). More free radicals were produced as the visible light irradiation period increased. Continual UV irradiation for 5000 s gave an EPFR yield in catechol/Fe2O3/SiO2 that was three times higher than that from continual visible light irradiation for 5000 s. The g values of organic free radicals in catechol/Fe2O3/SiO2 after UV irradiation and visible light irradiation for 5000 s are 2.00459 and 2.00453, respectively. Therefore, different optical wavelengths have different abilities to induce formation of EPFRs.

Formation of free radicals from catechol on PM2.5 as a reaction matrix under light irradiation

The solid matrix is the key factor influencing the formation of free radicals and their stability. Under atmospheric conditions, PM2.5 is a very important matrix for EPFR formation. To improve our understanding of the formation potentials of free radicals in the atmosphere with PM2.5 as the reaction matrix, EPR experiments were conducted with real PM2.5 samples spiked with catechol. The electron spin levels of free radicals formed on the PM2.5 matrix are shown in Figure 2 and the morphologies and elemental compositions of PM2.5 are shown in Figure S2. Under UV irradiation, free radicals were clearly formed from catechol on the PM2.5 matrix. The amount of EPFRs formed from catechol on PM2.5 was markedly higher than that formed from catechol without PM2.5 as a reaction matrix. This suggests that EPFRs in the atmosphere are mainly formed in the solid PM phase. Furthermore, the electron spin levels for the PM2.5 spiked with catechol were higher than those for PM2.5 without catechol. These results indicate that catechol in the atmosphere could contribute greatly to the formation of free radicals on PM2.5.

Figure 2

The formation of EPFRs from catechol on PM2.5 under UV irradiation

Changes in the EPFR concentrations over time under continuous UV irradiation for PM2.5 loaded with catechol (CT, 10%). PM2.5/CT–UV-1 and PM2.5/CT–UV-2 were conducted in parallel.

BFM: blank filter membrane.

It is widely recognized that CaO is a possible inhibitor of organic pollutant formation (Liu et al., 2005). However, it is unclear whether CaO can inhibit free radical formation or not. Thus, we also evaluated the electron spin levels of free radicals after addition of CaO to the PM2.5/catechol/UV irradiation system. The electron spin levels of the free radicals unexpectedly increased after addition of CaO to the reaction system (Figure 2). Therefore, instead of inhibiting free radical formation, CaO promotes the generation of free radicals in the catechol/PM2.5/UV irradiation system. This is important knowledge for those considering the potential of applying CaO to inhibition of free radical formation.

XPS characterization and the influencing mechanisms of metal oxides on free radical formations from catechol

To improve our understanding of the role of metal oxides in free radical formation by photochemical reactions, X-ray photoelectron spectroscopy (XPS) was used to characterize changes in the metal oxides before and after UV irradiation. For catechol/Fe2O3/SiO2, XPS of the Fe 2P region showed a change in low binding energy peak (710 eV–712.5 eV) after UV irradiation (Figure 3A). Furthermore, comparison of the characteristic Fe 2p peaks of Fe2O3 and FeO to those in the catechol/Fe2O3/SiO2 sample after UV irradiation (Figure 3B) showed that Fe3+ accounted for 91.57% and Fe2+ for 8.43% of the total Fe. The XPS spectrum of Cu 2p for catechol/CuO/SiO2 is shown in Figure 3C. After deconvolution of the Cu 2p spectrum of catechol/CuO/SiO2 after UV irradiation (Figure 3D), the characteristic Cu 2p peaks of CuO and Cu2O showed that Cu2+ accounted for 80.3% and Cu+ for 19.7% of the total Cu. The changes in the chemical states characterized by XPS indicated that Fe(III) was reduced to Fe(II) and Cu(II) to Cu(I). This suggested that electron transfer from catechol to Fe2O3 and CuO occurred during the photochemical reactions, and that Fe2O3 and CuO participated in the EPFR formation reactions. In the Ca 2p3/2 spectrum of CaO/catechol/SiO2 before and after the photochemical reactions, peaks were observed for Ca2+O at 347.41 eV and Ca2+ (OH)2 at 346.91 eV (Figures 3E, 3F, and S3).(Sugama et al., 1989; Vandoveren and Verhoeven, 1980) The proportion of Ca(OH)2 increased from 7% before the photochemical reactions to 15% after the photochemical reactions, which suggested that the EPFRs formed via elimination of H2O in a photochemical reaction of catechol with a metal oxide(Lomnicki et al., 2008). It has been reported that CaO can reduce the activation energy required for a reaction in the reaction system and restored system activity, which might also contribute to its ability to promote formation of EPFRs (Wang et al., 2020; Wiechen et al., 2012).

Figure 3

The XPS results

XPS of (A) catechol (CT)/Fe2O3/SiO2 Fe 2p, (B) deconvolution of the Fe 2p spectrum of CT/Fe2O3/SiO2 after UV irradiation. (C) CT/CuO/SiO2 Cu 2p spectrum. (D) deconvolution of the Cu 2p spectrum of CT/CuO/SiO2 after UV irradiation. (E) CT/CaO/SiO2 Ca 2p spectrum, and (F) deconvolution of the Ca 2p spectrum of CT/CaO/SiO2 after UV irradiation.

The XPS O 1s spectrum of the catechol/Fe2O3/SiO2, catechol/CuO/SiO2, and catechol/CaO/SiO2 reaction systems are shown in Figure 4. The XPS O 1s spectrum of the catechol/Fe2O3/SiO2 system after UV irradiation could be deconvoluted into five peaks at 533.14 eV (Si-O-Si of SiO2), 532.20 eV (C=O of benzoquinone), 531.54 eV (Fe-O-C), 530.23 eV (Fe2O3), and 533.68 eV (C-O-H of phenolic hydroxyl) for (Figure 4A) according to the literature(Barr, 1983; Kishi and Ikeda, 1973; Lhoest et al., 1995; Trinh et al., 2018; Xie et al., 2020). The XPS O 1s spectrum of the catechol/CuO/SiO2 system could be deconvoluted into five peaks at 533.13 eV (Si-O-Si of SiO2), 532.13 eV (C=O of benzoquinone), 532.70 eV (Cu-O-C), 531.27eV (CuO), and 534.03 eV (C-O-H of phenolic hydroxyl) (Figure 4B).(Barr, 1983; Cebula et al., 2013; Lhoest et al., 1995; Nefedov et al., 1982; Trinh et al., 2018) The peaks at 531.54 for Fe-O-C and 532.84 eV for Cu-O-C correspond to hydroxyl species from catechol on the Fe2O3 and CuO surfaces. Therefore, complete electron transfer occurred and Fe-O-C and Cu-O-C bonds formed during the photochemical reactions for EPFR formation. The XPS O 1s spectrum of catechol/CaO/SiO2 could be deconvoluted into five peaks at 533.11 eV (Si-O-Si of SiO2), 532.50 eV (C=O of benzoquinone), 533.77 eV (C-O-H of phenolic hydroxyl), and 531.4 eV for CaO, Ca(OH)2, CaCO3, and Ca-O-C because of their similar binding energies (Figure 4C).(Barr, 1983; Lhoest et al., 1995; Sugama et al., 1989; Trinh et al., 2018)

Figure 4

The XPS spectra and deconvolution of the O 1s spectra

The O 1s spectra of (A) catechol (CT)/Fe2O3/SiO2, (B) CT/CuO/SiO2, and (C) CT/CaO/SiO2 after UV irradiation

Identification of highly active free radicals formed from photochemical reactions of catechol

Besides EPFRs, highly active free radicals were also investigated in this study. Free radical trapping was used to identify the active free radicals produced from catechol. DMPO in aqueous solution was used as a trapping agent to detect any active free radicals. Hydroxyl free radicals can reportedly form from EPFRs in PM2.5 (Gehling et al., 2014). In this study, besides hydroxyl free radicals, the formation and occurrence of hydrogen free radicals from the photochemical reactions of catechol and PM2.5 were discovered for the first time. The EPR spectra are shown in Figure 5 and the fitting results are shown in Figures 6 and S4.

Figure 5

EPR spectra of spin-trapped radicals detected in the experiments with the following conditions

(A) 2 μL, 3 M DMPO +30 μL, PBS +2.5 mg of catechol (CT)/SiO2 with UV irradiation for 20 s, control: 2 μL, 3 M DMPO/water +30 μL, PBS +2.5 mg of SiO2 with UV irradiation;

(B) 2 μL, 3 M DMPO +30 uL, PBS +2.5 mg of CT/CaO/SiO2 with UV irradiation, control: 2 μL, 3 M DMPO/water +30 μL, PBS +2.5 mg of CaO/SiO2 with UV irradiation;

(C) 2 μL, 3 M DMPO/water +30 μL, PBS +2.5 mg of CT/SiO2/metal oxide (Fe2O3, CuO, CaO) with UV irradiation;

(D) 2 μL, 3 M DMPO/water +30 μL, PBS +1 mg of PM2.5 with UV irradiation, control: 2 μL, 3 M DMPO +30 μL, PBS +1 mg of BFM with UV irradiation;

(E) 2 μL, 3 M DMPO +2.5 mg of CT/SiO2/metal oxide (Fe2O3, CuO, CaO) with UV irradiation, control: 2 μL, 3 M DMPO +2.5 mg of Fe2O3/CuO/CaO/SiO2 with UV irradiation.

Symbols: ◯, DMPO-H; ▼, DMPO-OH; ▪, DMPO-OOH; ★, EPFR.

Figure 6

Fitting of the EPR spectra of spin-trapped radicals

(A–C)(A) spin-trapped radicals detected in catechol (CT)/SiO2 under UV irradiation. (B) spin-trapped radicals detected in CT/CaO/SiO2 under UV irradiation. (C) a suspension of CT/CaO/SiO2.

(a) Simulated DMPO-OH adducts. (b) Simulated DMPO-H adducts. (c) Fitting of experimental (green line) and simulated (black line) EPR spectra. (d) EPFR signal.

DMPO−OH adducts with a splitting of αN = αH = 14.7 G (literature data αN = αH = 14.9 G) (Rosen and Rauckman, 1981) were detected in the catechol/SiO2 sample after UV irradiation (Figure 5). DMPO−H adducts with splitting of αN = 16.5 G and αH = 22.5 G (literature data αN = 16.6 G and αH = 22.5 G) (Makino et al., 1982) were detected in the catechol/SiO2 EPFRs system after UV irradiation. The direct scission of chemical bonds by UV irradiation was responsible for EPFR photochemical formation. Elimination of ⋅H means more oxygen-containing functional groups will be produced in the catechol–SiO2 system, which might contribute to the formation of phenoxy and semiquinone radicals. Density functional theory calculation was used to verify EPFRs formed by hydrogen (H) abstraction of phenolic precursor in silico study (Dellinger et al., 2007; Yang et al., 2017b). This also supported our observations in the present study.

After the addition of CaO to the catechol–SiO2 system (Figure 5C), more ⋅H and ⋅OH were produced and this suggested that CaO may promote the production of ⋅H and ⋅OH in this system under light irradiation. In addition, the yield of EPFRs from the catechol/CaO/SiO2 system was much higher than that from the catechol/SiO2 system. These results indicate that CaO possible promotes the formation of EPFRs by facilitating the eliminating of ⋅H and ⋅OH from catechol. The XPS results for Ca(OH)2 after UV irradiation also provided further evidence for this mechanism. Interestingly, during the detection of ⋅H and ⋅OH by EPR, a hyperfine signal was observed with a splitting of αH1 = 0. 77 G and αH2 = −3.69 G (literature data: αH1 = 0. 96 G and αH2 = −3.50 G)(Gerson and Huber W, 2003), which indicated the existence of o-benzosemiquinone in the catechol/CaO/SiO2 system in accordance with the g2 type radicals in a Xenon fitting experiment. Oxygen free radicals (O2⋅−) with splitting of αN = 14.3 G, αHβ = 12.5 G and αHβ1 = 0.8 G (literature data: αN = 14.3 G, αHβ = 11.7 G and αHβ1 = 1.2 G) (Makino et al., 1982) was also detected in the solid phase of catechol/SiO2 both with and without Fe2O3 and CuO (Figures 5E and S4). The O2⋅− in the catechol/CaO/SiO2 system was markedly lower than in the other systems, which indicated that the EPFRs formed in the catechol/CaO/SiO2 system were not easy to oxidize with O2. Generally, ⋅H and reactive oxygen species, including O2⋅− and ⋅OH, can be formed in a catechol/SiO2 EPFR system under UV irradiation. The free radicals ⋅H and ⋅OH were mostly formed when the system was spiked with 30 μL of a DMPO aqueous solution, which was considered a liquid-phase system. By contrast, O2⋅− was mostly formed when the system was spiked with only 2 μL of DMPO solution and was considered a solid-phase system. A previous study reported the abilities of the EPFRs in PM2.5 to generate ⋅OH (Gehling et al., 2014). In this study, ⋅OH was also detected in PM2.5 after UV irradiation (Figures 5D and S5). The ⋅OH formed in PM2.5 after UV irradiation may be generated through photochemical reactions of organic pollutants like phenols. This is the first report of ⋅H and O2⋅− in such a reaction system and this knowledge is important for understanding their potential mechanisms and health effects.

Structural identification and mechanism of EPFRs and their stabilities

The photochemical reaction of catechol on Fe2O3/SiO2 could generate persistent radicals under visible light or UV irradiation (Figure 7). In the mixture of catechol and Fe2O3/SiO2 under visible light irradiation, the g value changed from 2.0045 before irradiation to 2.00485 after irradiation. On the fourth day after irradiation, the g value remained at 2.00487 and the peak pattern did not change markedly, which indicated that oxygen-centered radicals were formed during visible light irradiation.

Figure 7

Fitting of the EPR spectra

(A–D)(A) catechol (CT)/SiO2, (B) CT/Fe2O3/SiO2, (C) CT/CuO/SiO2, and (D) CT/CaO/SiO2 during photochemical reactions after spectral deconvolution.

The spectra were complex, indicating the presence of more than one type of radical. Three different radicals were suggested when Xenon software was used to mathematically deconvolute the EPR spectra (Figure 7 and Table S1), which were denoted as g1, g2, and g3 with g-values of approximately 2.0010–2.0020, 2.0040–2.0050, and >2.0050, respectively. The EPR spectral deconvolution suggested there were metal F-centers (g1) in the reaction systems spiked with Fe2O3, CuO, and CaO after UV irradiation and the F-center formed in CaO was similar to that for alkali halides (Henderson et al., 1969) and different from those for heavy metal oxides (CuO and Fe2O3) (Figure S6). Two types of free radicals formed on the surfaces of silica (g2 and g3), which were mainly oxygen-centered phenoxy and semiquinone radicals.

The possible mechanism of the formation of EPFR under UV irradiation was shown in Figure 8. Through taking off ⋅H and ⋅OH from catechol, the resonance stabilized radicals including o-semiquinone, phenoxyl and carbon-centered radicals were formed. The concentration of ⋅H was higher than ⋅OH, indicating that more o-semiquinone were formed comparing to phenoxyl and carbon-centered radicals which were consistent with the deconvolution results of EPFR spectrum (Pathway 1). Apart from the direct and sequential electron transfer from the O 2p orbitals of the phenolic oxygen in catechol to Fe(III) and Cu(II) (Pathway 2)(Lomnicki et al., 2008; Vejerano et al., 2011), ⋅H reacted with electron acceptor – Fe(III) and Cu(II) and the electron transfer from ⋅H to Fe(III) and Cu(II) which was indirect electron transfer from phenolic oxygen in catechol to Fe(III) and Cu(II) through ⋅H (Pathway 1). The lower concentration of ⋅H in Fe and Cu containing sample also supported our hypothesis. Moreover, O2 reacted with o-semiquinone type EPFRs on the surface of solid and electron transferred from EPFRs to O2 generated superoxide radical(Khachatryan et al., 2011). However, the possible mechanism for CaO to promote the formation of EPFR was differed from CuO and Fe2O3 and no electron transfer occurred from CT to Ca2+. CaO belonged to strong alkaline oxide, and catechol was weak acid in which proton hydrogen on phenolic hydroxyl was easily removed. Although CaO reacted with proton hydrogen, a lot of energy could be released to significantly promote the reaction process followed by more ⋅H elimination, EPFR formation and H2O production.

Figure 8

Proposed mechanism for EPFR formation

The proposed mechanism for EPFR formation accompanied by highly active free radicals participated in CT/SiO2 system after UV irradiation.

Half-life is a key parameter for evaluating the stability of EPFRs. In this study, the 1/e half-lives of the EPFRs in the samples were calculated using dynamic changes in their spin levels. The half-lives of EPFRs formed on the surfaces of SiO2 and PM2.5 spiked with catechol or a metal oxide ranged from days to months (Figure S7), which indicated that these radicals were persistent. The EPFRs in the PM2.5, catechol/PM2.5, and catechol/BFM systems displayed fast decay with 1/e lifetimes of 9.98 hr, 3.22hr, and 12.17 hr, respectively. This was followed by slower decay with 1/e lifetimes of 52.17 days, 115.74 days, and 25.85 days for the PM2.5, catechol/PM2.5, and catechol/BFM systems, respectively. Compared with catechol/PM2.5 and catechol/BFM, catechol/PM2.5 showed slower decay of the EPFRs, which suggested that absorption of EPFRs onto the PM2.5 surface increased their stability. In the catechol/CaO/PM2.5 system, the concentration of EPFR first decreased (1/e = 8.37 hr) and then increased slowly without further decay. This might result from continuous formation of the EPFRs from reactions of catechol, CaO, and other complex compounds in PM2.5.

Decay processes for EPFRs generated in catechol/SiO2 with or without metal oxides are shown in Figure S7 e–f. The decay characteristics of the catechol/CuO/SiO2, catechol/CaO/SiO2, and catechol/SiO2 systems were almost the same with a fast decay followed by a slower decay. A fast decay followed by no decay was observed for the catechol/Fe2O3/SiO2 system, which supported our previous research where we found that Fe2O3 stabilized EPFRs formed in this system (Vejerano et al., 2011). The faster decay might be caused by decomposition of a phenoxy-type radical and the slow decay by decomposition of a semiquinone-type radical (Gehling and Dellinger, 2013).

Conclusions

Catechol is a common potential precursor of EPFRs in the atmosphere and anthropogenic emissions. EPFRs are recognized internationally as an important component of PM2.5 when evaluating its health risks. We found that catechol could produce large amounts of EPFRs under light irradiation. Metal oxides, including Fe2O3, CuO, and CaO, markedly promoted the formation of EPFRs from catechol, with their effects on free radical formation in the order CaO > CuO > Fe2O3. The EPFRs produced from catechol underwent relatively fast decay (1/e = 3.2–32.3 hr) followed by slower decay (25.9–382 days), which indicates that they are stable and could have potential adverse effects on human health after inhalation or dietary intake. The photochemical formation mechanism of EPFRs clarified in this study could aid understanding of their secondary formation in the atmosphere. The formation and occurrence of hydrogen free radicals from the photochemical reactions of catechol and real PM2.5 were discovered for the first time. This is concerning because of their adverse effects on human health.

Limitations of the study

Photochemical reactions of organic precursors are extremely complex in atmospheric environment. This study clarified the formation of EPFRs from catechol through photochemical reactions as well as their influencing factors. However, there are multiple precursors in the airborne particle matters. Moreover, the nature of the complexity of airborne particle matters makes it a long way to fully understand the numerous mechanisms of EPFRs from multiple potential precursors. The various factors influencing EPFR formations in atmospheric environment cannot be fully recognized in this study. Therefore, much more studies should be conducted to better understand the formation and control of EPFRs in atmospheric environment in the future.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Guorui Liu (grliu@rcees.ac.cn).

Materials availability

This study did not generate new unique reagents.

Data and code availability

This study did not generate/analyze data sets/code.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.