3.1 Photobromination of phenol in bromide solutions
Phenol’s bromination was first studied in aqueous bromide solutions. The concentration of bromide, [Br−], in seawater is at an average concentration of 0.8 mmol L-1, while it can be enriched in the sea-spray aerosols due to the evaporation of water and can reach dozens of mmol L-1 (Edebeli et al. 2019). Considering the wide [Br−] range in marine environment and the analytical convenience, [Br−] adopted in this work was ranged from 0.8 to 80 mmol L-1. Fig. 1 shows that ~10 ng L-1 bromophenols was generated with 8 mmol L-1 Br−, whereas no detectable bromophenols was formed in dark within 8 h. The formation of bromophenols was obviously enhanced with the increasing of [Br−], with bromophenols concentration reaching ~21 ngL-1 in the presence of 80 mmol L-1 Br−.
The production of bromophenols in aqueous systems has be supposed to occur mainly via bromine radicals (Br2•−/ Br•) (De Laurentiis et al. 2012; Liu et al. 2018). One formation path of bromine radicals is by oxidation of •OH (eqs. 1-3, k+ and k− represent the forward and backward rate constant, respectively, and the unit of k is M-1s-1) (Zhang and Parker 2018).
•OH+ Br− →HBrO•− k+ = 1.1×1010 M-1s-1, k− = 3.3×107 M-1s-1 (1)
HBrO•− + H+ →Br• + H2O k+ = 4.4×1010 M-1s-1, k− = 1.36 M-1s-1 (2)
Br• + Br− → Br2•− k+ = 1.2×1010 M-1s-1, k− = 6.6×103 M-1s-1 (3)
Phenol bromination takes place in two steps, involving a phenoxyl radical formation and bromine radical substitution reaction (eqs. 4 and 5) (Vione et al. 2008). High concentration of bromide helped for Br2•− formation, consequently led to more generation of bromophenols.
ArOH + Br2•− →ArO• + 2 Br− (4)
ArO• + Br2•−→BrArOH +2 Br− (5)
In the reaction system of this work, •OH could be originated from the photolysis of phenol. Jiang et al. (2020) have demonstrated that photolysis of phenolic compounds resulted free radicals including •OH based on the electron paramagnetic resonance investigation. The formation path of •OH from phenol was shown in Fig. S2.
3.2 Effect of AQ2S and BP on the photobromination of phenol
Fig. 2 reports the time evolutions of 2-bromophenol (A) and 4-bromophenol (B) upon irradiation of 2 mg L-1 phenol in the presence of 1 and 10 μmol L-1 AQ2S or BP. AQ2S and BP considerably enhanced the photobromination of phenol, and the concentrations of bromophenols increased with increasing of [AQ2S] and [BP] obviously. The promotion effect of AQ2S was more significant that BP, bromophenols reached 110~150 ng L-1 in the presence of 10 μmol L-1 AQ2S, which was about twice of that in the presence of 10 μmol L-1 BP. Phenol removal was ~ 5% within 8 h irradiation (Fig. S3), and the convert rate of phenol to bromophenols with BP or AQ2S was ~10-3. It is not high but may be of significance, for the universal coexistence of DOM and bromides in marine environment.
Apart from oxidation by •OH, triplet state of aromatic ketones and quinones can oxidize bromide to Br2•− as well (Yang and Pignatello 2017). Brigante et al. (2014) and De Laurentiis et al. (2012) previously showed the possibility of sensitized photochemical production of dihalogen radical anions Cl2•− and Br2•− from Cl− and Br−, using AQ2S as a sensitizer. Jammoul et al. (2009) found that the triplet excited state of benzophenone can oxidize bromide ions to Br2•−. The oxidation of bromide by excited triplet state (3S) of aromatic ketones and quinones is considered to via charge-transfer interactions, and can be confined to the reduction potential of 3S. Excited triplet state of AQ2S (3AQ2S*) is a powerful oxidant (E3S/S- = 2.28 VNHE), and aromatic ketones possess triplet state reduction potentials of 1.10~1.69 V NHE (McNeill and Canonica 2016). Therefore, it is reasonable to propose that 3AQ2S* and 3BP* could oxidize Br− to Br2•− (eq. 6).
3AQ2S* or 3BP* + 2Br− → AQ2S •− or BP•− + Br2•− (6)
In addition, 3AQ2S* and 3BP* can oxidize phenol via an electron transfer reaction, forming phenoxyl radical, which is also an important process for bromophenol generation. Jammoul et al. (2009) have demonstrated that the quenching of triplet state of benzophenone, 3(Ar1Ar2-C=O)*, with phenol occurred by electron transfer. The electron transfer reaction is usually described by a mechanism involving (i) formation of a precursor complex, [PhOH…3(Ar1Ar2-C=O)*], (ii) electron transfer to form a charge transfer (CT) complex, [PhOH•+…Ar1Ar2-C•-O−], and (iii) separation of the oxidized donor and reduced acceptor, yielding a protonated phenoxyl radical (PhOH•+) and a deprotonated ketyl radical (Ar1Ar2-C•-O−), eq 7. PhOH•+ is a strong acid with pKa of -2.0 (Canonica et al. 2000), and is liable to give out one proton forming phenoxy radical (eq. 8).
PhOH + 3(Ar1Ar2-C=O)* ⇄ [PhOH…3(Ar1Ar2-C=O)*] ⇄ [PhOH•+…Ar1Ar2-C•-O−] → PhOH•+ + Ar1Ar2-C•-O− (7)
PhOH•+ → PhOH• + H+ (8)
The rate constants for eq. 6 and eq.7 follows the order AQ2S > BP, which is consistent with the order of their reduction potential. The reduction potentials, E3S/S-, for BP and AQ2S are 1.67 VNHE and 2.28 VNHE, respectively. Therefore, the promotion effect of AQ2S on phenol bromination was more obvious than that of BP.
In addition, although there ensued a long-lasting controversy about the possibility for excited AQ2S to generate •OH upon oxidation of water, some studies have indicated the possibility that free •OH was being formed with significant amounts (Alegía et al. 1999; Maurino et al. 2008). Therefore, AQ2S promoted phenol bromination via generating •OH as well as 3AQ2S*, which are helpful for bromine radicals and phenoxyl radical production.
3.3 Effect of bromide and chloride concentrations
Formation of bromophenols in the presence of 10 μmol L-1 BP or AQ2S with different concentration of bromide is shown in Fig. 3. It is clear that bromophenols increased obviously with [Br−] ranging from 0.8 to 80 mmol L-1. Take 2-bromophenols for example, [2-bromophenol] was around 10, 35 and 65 ng L-1 in the presence of BP, while it was about 13, 85 and 110 ng L-1 in the presence of AQ2S, with [Br−] ranging from 0.8 to 80 mmol L-1.
Bromine radical was the dominant precursor for the formation of bromophenols. According to the model proposed by Loeff et al. (1993), halide reacts with the triplet excited state (3S) to form a charge-transfer binary exciplex, 3(S•−-- X•), and ternary exciplex, 3(S•−-- Br•-- Br−), both of which can decay to the ground state or dissociate to the radical pairs, as shown in Fig. S4. The ternary exciplex has a lower tendency than the binary exciplex to decay to the ground state since it has weaker spin-orbit coupling of the incipient radical. Consequently, ternary exciplex dissociates to the radical products, Br2•−, more favorably than the binary exciplex. High concentrations of halide could favor ternary exciplex formation, thus increase the generation of bromophenols.
Considering chloride is the main anion in seawater, the effect of chloride on phenol bromination was studied next. Fig. 4 shows addition of 0.5 mol L-1 chloride enhanced the formation of bromophenols obviously, especially in the presence of AQ2S (Fig. 4C and D). As mentioned above, high concentration of chloride helped to generate ternary exciplex, 3(S•−-- Br•-- Cl−), which then formed the mixed dihalogen radical anion, BrCl•−. Subsequently, BrCl•− reacted with Br− to produce Br2•− (eq. 9). On the other hand, •OH can oxidize X− to produce reactive radical species, X•/ X2•−, through HXO•−, where X=Br, Cl (Zhang and Parker 2018; Dong et al. 2020). Although Cl− is an ineffective •OH scavenger since the intermediate HClO•− primarily reverts to •OH and Cl− not to Cl• (eq. 10), HClO•− is liable to react with Br− to form BrCl•− (eq. 11). Reactions related to BrCl•− are expected to be predominant in marine environment, since bromide is susceptibly oxidized while chloride is abundant. Consequently, chloride enhanced bromophenols production due to the generation of BrCl•− and the subsequent formation of Br2•−.
BrCl•− + Br−→Br2•− + Cl− k+ = 8.0 ×109 M-1s-1, k- = 4.3×106 M-1s-1 (9)
•OH+ Cl−⇄HClO•− k+ = 4.0×109 M-1s-1, k− = 6.0×109 M-1s-1 (10)
HClO•− + Br−→BrCl•− + OH− k+ = 1.0 ×109 M-1s-1, k- = 3.0×106 M-1s-1 (11)
Fig. 4 C and D shows that bromophenols’ concentration in the presence of AQ2S reached approximate 5000~10000 ng L-1 after adding chloride, which was almost 100 times higher than that without chloride. Brigante et al. (2014) have demonstrated that 3AQ2S* could oxidize chloride to generate chlorine radical. E3S/S- of AQ2S is 2.28 VNHE, thus it can oxidize Cl− and Br− to generate mixed-halogen radical, BrCl•− (eq. 12) and subsequently generate Br2•− (eq. 9) (Zhang and Parker 2018). Therefore, the promotion effect of chloride on the bromophenols’ formation was really significant with AQ2S. In contrast, the promotion effect of chloride in the presence of BP was not so obvious (Fig. 4 A and B), which should be attributed to the lower oxidizing ability of 3BP*. Although no data are available for the reduction potential of BrCl•−, it is typically assumed to present reduction potential between or similar to Br2•− (E Br2•−/Br- = 1.63 VNHE) and Cl2•− (E Cl2•−/Cl- = 2.20 VNHE). E3S/S- for BP is 1.67 VNHE, so 3BP* did not efficiently produce BrCl•−, consequently the promotion effect of chloride in the presence of BP was much lower than that of AQ2S.
3AQ2S* + Br− + Cl− → 3[AQ2S−---Br•---Cl−] → 3AQ2S•− + BrCl•− (12)
It is noticeable that the concentrations of 4-bromophenol were higher that 2-bromophenol in almost all cases, especially in Fig. 4C and D. The similar result has been observed by Vione et al (2008), which was attributed to the higher formation rate of 4-bromophenol compared to 2-bromophenol. The higher formation rate of the para isomer as compared to the ortho isomer might be explained by a lower steric hindrance at the para position, which is more distant from the oxygen atom, that could allow easier addition of the bulky Br atom to the aromatic ring (Vione et al. 2008). In addition, Zakon et al. (2013) found that 4-bromophenol degraded much slower than 2-bromophenol during photolysis. Consequently, 4-bromophenol was easier to be accumulated in the solutions.
3.4 Formation of bromophenols in the presence of SRNOM
The effects of SRNOM on phenol bromination were shown in Fig. 5. Bromophenols reached more than 30 ng L-1 with 1 mg L-1 SRNOM, which was about three times of that without DOM, indicating that SRNOM enhanced bromophenols formation. However, when SRNOM increased to 5 mg L-1, bromophenols decreased, which could be attributed to the competitive consumption of RBS by DOM forming halogenated DOM or light shielding effect (Hao et al. 2017; Dong et al. 2020).
Irradiation of DOM with solar light generates 3DOM* and •OH, which both act as oxidants to oxidize halide ions. The photogeneration of bromine radicals in marine environment can take place upon bromide photooxidation of •OH (reaction 1~3) and 3DOM* (eq. 13 and 14). Although ClBr•− is predicted to be the dominant RHS formed initially via halide oxidation by 3DOM* in seawater, Parker and Mitch et al. (2016) found that Br2•− concentrations may exceed those of ClBr•− by ∼2.5-fold, because the dominance of Br2•− arises from further reactions of ClBr•− (e.g., with Br−). Therefore, ClBr•− was an important intermediate for bromination reaction.
3DOM* + 2 Br− → DOM•− + Br2•− (13)
3DOM* + Br− + Cl− → 3DOM•− + BrCl•− (14)
As far as eq. 13 and 14 are concerned, it is decided that the halogen radicals generation depends on the oxidation capability of 3DOM* which is determined by DOM structure. Although the relationship between the oxidation capability and the structure of DOM needs further investigation, it is undeniable that organic compounds with quinone and aromatic ketone structure can accelerate the bromination reaction. As we know, river DOM mainly contains humic- or fulvic-like components while marine DOM mainly contains protein- or amino acid-like component, whereas river DOM exists in estuarine and coastal aqueous environment due to the mixture of river and sea water. Nevertheless, this result justified the enhancement of DOM for the photochemical bromination reaction in the saline water that is relevant to the marine environment.
Although this research focused on aqueous conditions, RHS formation from direct halide oxidation by 3DOM* likely also applies to marine aerosols, where halide concentrations can exceed those in seawater (Finlayson-Pitts 2003). Since non-radical dihalogen species form as products of radical RHS reactions, e.g., 2 ClBr•− → ClBr + Br− + Cl−, this pathway may contribute to the release of halogens to the atmosphere, with implications for tropospheric ozone degradation.