1 Introductions

The molecular composition of dissolved organic matter (DOM) in coastal seawater is primarily affected by anthropogenic activities. Recently, the proportion of sulfur (S)-containing DOM molecules in the surface waters of the Bengal Bay in the Indian Ocean is reported to be generally higher than that for the deep ocean (Liu et al. 2023b). In the Dalian coastal seawaters of China, the S-containing DOM molecules were partially derived from the effluent and transformation of sulfa antibiotics (e.g., sulfachloropyridazine) by land-based sources or mariculture activities (Guo et al. 2021; Wang et al. 2018b). Tokyo Bay has been reported to receive a large amount of municipal effluent from wastewater treatment plants in the Tokyo metropolitan area, which is abundant in CHOS compounds (Phungsai et al. 2016, 2018). For example, the proportion of CHOS compounds in the DOM from the natural coastal seawater of Tokyo Bay was higher than that in the natural freshwater upstream of Tama River (Fu et al. 2020a), suggesting that the coastal seawater quality was susceptible to anthropogenic activities. Most S-containing substances are derived from various productive activities, such as the extensive use of surfactants (Liu et al. 2023a). The S-containing DOM compounds in wastewater have a potential endocrine-disrupting effect, consequently reducing the water quality and promoting the production of neurotoxic methylmercury (Ksionzek et al. 2016; Michael et al. 2015). Therefore, it is urgent to investigate the environmental behavior of CHOS substances in natural coastal seawater.

Photochemical halogenation is capable of producing organohalogen compounds in seawater. For example, the phenol moiety in seawater can be transformed into organic halogenated compounds (OHCs) after illumination (Calza et al. 2008). Various OHCs result from the nonspecific halogenation of DOM by reactive halogen species (RHSs) photochemically generated at the natural surface seawater (Méndez-Díaz et al. 2014) and artificial seawater containing Suwannee River natural organic matter (SRNOM) (Hao et al. 2017, 2018; Méndez-Díaz et al. 2014). The S-containing compounds (e.g., sulfadimethoxine) will undergo a photochemical transformation, such as direct or indirect photolysis accompanied by loss of SO2, in natural sunlit surface aquaculture waters (Guerard et al. 2009). Photolysis is another important degradation pathway for S-containing compounds (sulfa drugs including sulfadiazine, sulfamethazine, and sulfamerazine) in surface waters, which contributes to the increase of SO2 extrusion photoproduct (Boreen et al. 2005). Furthermore, S-containing OHCs will be photoconverted to other OHCs species in seawater via photonucleophilic substitution with chloride ions (Liu et al. 2019). However, the photochemical halogenation of anthropogenically derived organic compounds in coastal seawater remains unclear. Many brominated disinfection byproducts containing S have been documented in aquaculture seawater (Wang et al. 2018a) and coastal ballast water (Gonsior et al. 2015). Therefore, in this study, we proposed the hypothesis that photochemical halogenation may affect the molecular composition and transformation of anthropogenically derived S-containing organic compounds in natural coastal seawater.

Recently, the Fourier transform ion cyclotron resonance mass spectrometry coupled with electrospray ionization (ESI-FT-ICR MS) has been extensively employed to characterize the molecular composition of DOM (Baluha et al. 2013; Stenson et al. 2003), halogenated disinfection byproducts (Xu et al. 2013; Zhang et al. 2014), and OHCs (Hao et al. 2017, 2018). Therefore, the main objective of this study was to (1) investigate the effects of photochemical reactions on the chemodiversity of natural coastal seawater DOM by using FT-ICR MS; (2) reveal the molecular composition characteristics of photochemically generated organic brominated compounds (OBCs) and organic iodinated compounds (OICs); (3) propose a new molecular index to indicate the unsaturated degree of carbon skeleton for DOM molecules. For the first time, the present study explored the chemodiversity and photochemical halogenation of DOM molecules in coastal seawater under irradiation.

2 Materials and methods

2.1 Photochemical formation experiments of OBCs and OICs

Approximately 5 L coastal seawater sample was collected by a bucket from the Aomi Minami Terminal Park at Tokyo Bay, Japan (35.61607° N, 139.77564° E; red point in Fig. S1), and immediately transferred and filtered with 0.45 μm membrane in the laboratory at Tokyo Institute of Technology, followed by photohalogenation experiments. The experimental conditions of photohalogenation were detailed in Content S1. All samples were treated as follows: (i) the original coastal seawater sample without any treatment (CSW); (ii) CSW under sunlight (CSW + Sun); (iii) CSW with 50 mg/L Br− under sunlight (CSW + Sun + Br); (iv) CSW with 5.0 mg/L I− under sunlight (CSW + Sun + I); (v) CSW under UV light for 12 h (CSW + UV); (vi) CSW with 5.0 mg/L I− under UV light for 12 h (CSW + UV + I); (vii) CSW with 50 mg/L Br− under UV light for 12 h (CSW + UV + Br-1); (viii) CSW with 50 mg/L Br− under UV light for 24 h (CSW + UV + Br-2). It is worth noting that seawater samples in the glass bottles were subjected to sunlight treatment to examine the effects of visible and partially penetrated UVA (310 nm—400 nm) lights on the DOM transformation (Table S1). Moreover, the quartz beaker (500 mL) was employed to examine the DOM transformation induced by the UVA and UVC irradiation.

2.2 Sample preparation

Before and after irradiation, all seawater samples were filtered through a 0.45 μm membrane and subjected to purification and extraction using a solid-phase extraction (SPE) method described elsewhere (Dittmar et al. 2008). Briefly, the filtered seawater samples were acidified with concentrated HCl to pH ≈ 2 and then passed through a Bond Elut PPL cartridge (6 mL and 1 g, Agilent, US) activated with 6 mL of methanol (LC–MS grade). Ultrapure water (≥ 18.25 MΩ) was prepared by a Milli-Q system (Millipore, Billerica, MA). The cartridges were rinsed with 20 mL of HCl-acidified ultrapure water (pH ≈ 2), followed by complete drying using high-purity N2 gas (≥ 99.9%). The DOM was then gravitationally eluted with 6 mL methanol.

The aliquot (0.2 mL) of each SPE-extracted DOM was completely dried with a centrifugal concentrator (CVE-2200, EYELA, Japan) and redissolved in ultrapure water. The total organic carbon (TOC) of all filtered seawater and resolved DOM samples was determined using a TOC analyzer (TOC 4100, Shimadzu, Japan). The SPE recovery of DOM was averaged at 59.70% (Table S3). All SPE-extracted samples were diluted twofold with ultrapure water for FT-ICR MS measurement.

2.3 FT-ICR MS analysis

The spectra of 8 seawater samples were obtained from a 9.4 T FT-ICR MS instrument equipped with an ESI ionization source (SolariX XR, Bruker Daltonics Inc., USA) at the Tohoku University, Japan, using the following instrumental parameters: negative ion mode, 150 µL/h direct infusion rate, -4.5 kV capillary voltage, 1 ms ion accumulation time, 450 average scans, two megaword of time-domain data size, and 150–1,000 mass-to-charge ratio (m/z) range (Fu et al. 2020a, 2020b). All FT-ICR MS spectra were externally calibrated with NaI clusters before measurement.

2.4 Data analysis

The formula assignment was performed by our FTMSDeu algorithm (Fu et al. 2022b) according to the following calculation conditions: signal-to-noise ratio ≥ 6 and ≥ 10 for nonhalogenated and halogenated monoisotopic formula, respectively; 0.3 ≤ (H + Cl + Br + I)/C ≤ 2.25 and 0 < O/C ≤ 1.2 for molecules with C ≥ 5; (H + Cl + Br + I)/C ≤ 4 and 0 ≤ O/C ≤ 1.2 for molecules with C ≤ 4; an integer value ≥ 0 for double bond equivalent (DBE); -10 ≤ double bond equivalents minus oxygen (DBE-O) ≤ 10; 1 ≤ 12C ≤ 50; 13C ≤ 2; 18O ≤ 2; 14N ≤ 5; 32S ≤ 3; 33S ≤ 1; 34S ≤ 1; P ≤ 1; 35Cl ≤ 5; 37Cl ≤ 5; 79Br ≤ 5; 81Br ≤ 5; and I ≤ 3 for all samples. The pathways (electrophilic substitution reaction, electrophilic addition reaction, and both or others) for OHC formation were established based on stoichiometric formula changes using our early method (Fu et al. 2022b).

The DBE, double bond equivalents minus oxygen atom number (DBE-O), double bond equivalents minus the half of oxygen atom number (DBE-0.5O), the nominal oxygen state of carbon (NOSC), and the modified aromaticity index (AImod) values were calculated using the equations described in Content S2 (Fu et al. 2022b; Gonsior et al. 2009). The assigned molecular formulae were classified into nine biochemical groups (Table S4) based on their spaces in the van Krevelen diagram (Zhou et al. 2023).

For a given parameter, if all values did not follow a normal distribution, the non-parametric Kruskal–Wallis one-way of variance was employed for statistically comparing the difference among the compared groups at the significance level of p < 0.05. Otherwise, the one-way analysis of variance would be used in this study. Statistical analysis was performed by using Origin software (version 2023 OriginLab Corporation, USA).

3 Results and discussion

3.1 New index for DOM unsaturation degree

Generally, the DBE-O is considered a better indicator to quantify the unsaturation degree of the carbon skeleton than the DBE because most oxygen atoms in DOM molecules are assumed in the carboxyl and hydroxyl functional groups (Gonsior et al. 2009). However, it should be noted that only half of the oxygen atoms in the carboxyl functional group (an oxygen atom on the C = O double bond) contribute to the DBE value (Koch and Dittmar 2006). Eventually, the DBE-O will underestimate the unsaturation degree of the carbon skeleton. Inspired by the modification of the aromaticity index, we have proposed a new and more reasonable index, DBE-0.5O, to evaluate the unsaturation degree of the carbon skeleton for DOM molecules in this study. For the carboxyl and hydroxyl functional groups, the unsaturation of the carbon skeleton of DOM will be underestimated by an additional 1 and 0.5 DBE using the DBE-O compared to those using the DBE-0.5O. Moreover, two oxygen atoms in the two S = O double bonds of the sulpho functional group (-SO3) contribute to 2 DBE. Therefore, DBE-O and DBE-0.5O will underestimate sulpho-containing DOM molecules (such as the surfactants frequently detected in municipal wastewater effluent) by 1 DBE and overestimate them by 0.5 DBE. Similarly, the DBE-0.5O will exhibit less error than the DBE-O for the quantification of DBE in the phosphate functional group (1 DBE versus 3 DBE). The superiority of the index of DBE-0.5O over DBE-O is further supported by the photolysis-sensitive CHOS1 formulae, which are detailed in Content S3. For example, the C13H18O5S1 formula in the CSW (Fig. S2) is most likely to be the most common isomer for the co-products and metabolites (e.g., sulfophenyl carboxylic acid, C14H20O5S1) of linear alkyl benzene sulfonates (Gonsior et al. 2011). Compared with the 20 increment units of DBE-O (Fig. 1a), ranging from -10 to 10 for the ESI(-)-FT-ICR MS spectra of aquatic DOM (Herzsprung et al. 2014), 37 increment units were observed for the index of DBE-0.5O (-5 to 13, Fig. 1b). The increment range was enlarged by the index of DBE-0.5O with a factor of approximately 2, suggesting a more sensitive potential in revealing the unsaturation degree of the carbon skeleton in DOM molecules than the index of DBE-O. Therefore, it is more reasonable to indicate the unsaturated carbon density of DOM molecules using the index of (DBE-0.5O)/C than the conventional (DBE-O)/C (Phungsai et al. 2018).

Fig. 1
figure 1

The frequency histograms of DBE-O (a) and DBE-0.5O (b) values for the number of photolysis-sensitive CHOS1 formulae detected in CSW, CSW + Sun, and CSW + UV

Full size image

3.2 Effect of photohalogenation on the DOM molecular composition of raw coastal seawater

3.2.1 Chemodiversity of raw coastal seawater DOM

DOM molecules in the CSW were divided into two categories, namely common and other DOM formulae (defined as the formulae identified in all eight samples and partially identified in the other seven treated samples, respectively), to investigate the photosensitivity of DOM molecules in the raw coastal seawater. There were 9159 nonhalogenated formulae (excluding isotopic formula) identified in the CSW DOM, including 2249 common and 6910 other DOM formulae. As illustrated in Figs. 2a and S3a, lignin-like compounds were the most abundant molecular classes (accounting for 63% and 68% of common DOM and other DOM formulae, respectively), followed by nitrogen (N)-free saturated compounds (15% versus 11%) and tannin-like compounds (9% versus 6%). The plot (DBE-0.5O)/C against NOSC (Figs. 2b and S3b) was employed to indicate the extent of unsaturation and oxidation of DOM molecules in this study. The larger positive values of (DBE-0.5O)/C and NOSC generally represent a greater extent of unsaturation and oxidation, respectively. Compared with common DOM molecules, the other DOM molecules typically exhibited higher unsaturation [(DBE-0.5O)/C averaged at: 0.22 versus 0.13] and oxidation [NOSC averaged at: -0.18 versus -0.26] (Figs. 2b and S3b), indicating that other DOM molecules were more susceptible to light irradiation. These results could be attributed to the highly aromatic molecules of other DOM molecules, which were rich in electrons and absorbed more light (Sharpless et al. 2014).

Fig. 2
figure 2

The van Krevelen diagram (a) and plot of (DBE-0.5O)/C vs. NOSC (b) for common formulae in CSW

Full size image

The coastal seawater in Tokyo Bay was rich in S-containing substances, accounting for 33% of the total number and 24% of the summed intensity of all identified formulas in the CSW, respectively. The most abundant CHOS and CHONS formulae contributed to 24% (1655) and 19% (1342) of the total number of the other DOM molecules (Fig. 2), respectively, but only 5% (120) and 1% (30) for the common DOM molecules (Fig. S3). As illustrated in Fig. S4, the abundance of S-containing classes decreased in the order of S1O3-18 > N1S1O3-14 > O6-11 for the other DOM molecules. The DBE-0.5O of O6-11 class for CHOS and CHONS compounds in other DOM ranged from -4.5 to 15, suggesting that these highly aromatic compounds contained less S and were generally derived from terrestrial substances (Guerard et al. 2009). Overall, the coastal seawater DOM, especially S-containing compounds with high unsaturation, are more prone to photodegradation or phototransformation during irradiation.

3.2.2 Photochemical transformation of DOM

The overall profile of FT-ICR MS spectra in the sunlight-irradiated samples with and without halide ions (CSW + Sun, CSW + Sun + Br, and CSW + Sun + I) was close to that for the CSW before irradiation (Fig. S5), indicating the insignificant DOM transformation of coastal seawater in glass bottles under the sunlight irradiation. This observation is attributed to the fact that only the low-energy UVA and visible lights in the sunlight could penetrate the glass bottles to yield an obvious molecular transformation (Table S1). However, the intensities of peaks at a low m/z (< 300) were increased obviously by the UV (Fig. S5). For example, the intensities of C8H15O3Cl1 peaks increased by more than one order of magnitude for the UV treatment than those for the sunlight treatment and CSW sample, indicating the UV-induced chlorination of DOM in the seawater (Fig. S6). Notably, compared to the CSW, there were ~ 44%—~ 67% and ~ 81%—~ 91% number of CHOS and CHONS compounds decreased by the UV treatment, respectively (Fig. 3a). In this study, the compounds present in the CSW but disappeared in the samples treated with sunlight or UV light irradiation were defined as the “transformed” formulae. Instead, the “produced” compounds were the formulae in the samples treated with sunlight or UV light irradiation that disappeared in the CSW.

Fig. 3
figure 3

The number and proportion of DOM composition of the samples according to (a and b) elemental composition and (c and d) molecular composition

Full size image

The number of transformed CHOS and CHONS compounds by sunlight and UV light irradiation was significantly more (p < 0.05) than their produced number (Fig. S7). The number of transformed CHOS and CHONS formulae by UV light irradiation was significantly larger (p < 0.05) than those for sunlight irradiation. These results suggest that the desulfonation pathway contributes more to the photodegradation of most CHOS and CHONS molecules than the decarboxylation (Fig. S8). In general, the significantly higher (p < 0.05) m/z value but lower (p < 0.05) AImod and (DBE-0.5O)/C values of the produced CHOS and CHONS molecules by irradiation than those transformed molecules indicated that S-containing molecules had higher molecular weight, aromaticity, and unsaturation (Fig. S9). However, the significantly higher (p < 0.05) S/C value of the produced CHONS molecules (Fig. S9) than that for the transformed CHONS molecules during UV light might result from the decreased proportion of CHON molecule number because the transformed CHON molecule number was significantly larger (p < 0.05) than the transformed CHONS molecule number during UV light. Moreover, the transformed and produced CHOS/CHONS molecules were mostly lignin-like compounds, followed by saturated compounds (Fig. S10).

For transformed or produced compounds, no significant difference (p > 0.05) was observed between the numbers of CHOS and CHO molecules in the same irradiation treatment (Fig. S7). Despite the molecular transformation, the CHO, CHON, and lignin-like compounds were the predominant components identified for coastal seawater in all treatments (Fig. 3). These results were further supported by the intensity-weighted values of 12 molecular parameters (Fig. S11). The H/Ciw and O/Ciw typically increase during sunlight and UV light irradiation (Gonsior et al. 2009, 2013) due to extensive partial photooxidation (i.e., oxygen is incorporated into DOM molecules). Therefore, the coastal seawater DOM with higher molecular weight, aromaticity, and unsaturation are more readily transformed to lower mass molecules through desulfonation, decarboxylation, and photooxidation (partial oxidation or mineralization) (Boreen et al. 2005; Waggoner et al. 2017). The irradiation-induced increase of Cl/Ciw and Br/Ciw is attributed to the halogenation (e.g., substitution and addition reactions) of coastal seawater DOM molecules.

3.3 Photobromination of coastal seawater DOM

There were 23 OBCs identified in the CSW, including 10 OBCs-Br (OBCs containing one bromine) and 13 OBCs-2Br (OBCs containing two bromines) species. As exemplified in Fig. S12, the exclusive presence of C9H9O8S1Br1 in the CSW suggests that the C9H9O8S1Br1 may have originated from the anthropogenic input such as the effluent from the wastewater treatment plant in Tokyo Bay (Andersson et al. 2019; Fu et al. 2022b). Therefore, the effect of abiotic bromination of CHOS molecules in the CSW was limited under visual and UV light irradiation. The OBCs species generally resulted from abiotic reactions between reactive bromine species (RBrSs) and DOM precursors through four main mechanisms: substitution reactions (SR) and addition reactions (AR), reactions involving both substitution and addition (SR + AR), and other reactions (OR) (Fu et al. 2022b). According to the stoichiometric changes based on these mechanisms, the majority (19, 86%) of OBCs identified in the CSW were saturated CHOX-class compounds (CHOX indicates formula composited of C, H, O and halogen atom) formed via the OR reaction (Fig. 4a). Compared to CSW, the decreased number of OBCs by sunlight irradiation resulted from photodegradation (Abusallout and Hua 2016). In comparison, the increased number of OBCs by UV light irradiation could be attributed to the photohalogenation of UV light, which provides higher energy to generate RHSs. In addition, the number of OBCs in the CSW + UV + Br-2 (83) was approximately 1.4 folds of that in the CSW + UV + Br-1 (60), suggesting that the number of newly formed OBCs is positively correlated with the UV irradiation time within a reasonable range. The prolonged UV light irradiation will result in more RBrS and thus promote the generation of OBCs by RHS reacting with DOM (Méndez-Díaz et al. 2014).

Fig. 4
figure 4

The molecular composition of OBCs of photobrominated CSW samples with/without Br− after sunlight and UV light irradiation. The van Krevelen diagrams are depicted based on different bromine numbers (a and b) and reaction types (c and d). Note that dots might be overlapping at the same position

Full size image

These OBCs identified in CSW + UV, CSW + UV + Br-1, and CSW + UV + Br-2 were mostly CHOX-class OBCs-Br formed via SR + AR reaction and mainly composed of saturated compounds, followed by lignin-like and other compounds (Fig. 4). The UV light irradiation-induced OBCs were primarily composed of unsaturated and saturated reduction compounds, which was supported by the observation that these molecules were in the quadrants II and III of the (DBE-0.5O)/C vs NOSC plots (Fig. S13).

3.4 Photoiodination of coastal seawater DOM

In total, 10 OICs-I (OICs containing one iodine) identified in the CSW before irradiation were mainly assigned to saturated compounds. Several OICs species (e.g., C8H5O5I1) in the CSW were easily photodegraded under visual and UV light irradiation (Fig. S14). Further, 80% of OICs in the CSW were saturated CHOX-class compounds and might be formed via the OR reaction. Compared with the OBCs, UV light irradiation did not contribute significantly to the formation of OICs because of the slightly decreased number of OICs. However, 70 OICs were additionally detected in the CSW + Sun + I compared with the CSW. These results suggest that the penetrated UVA lights from glass (1.00—1.31 mW/cm2, Table S1) are of enough energy to trigger the iodination of DOM. OICs are formed but subsequently transformed into iodine-free compounds for the UV light treatment because of the high energy and luminous intensity (0.34 mW/cm2) of UVC light used in this study. Over 73% of OICs identified in the CSW + Sun + I sample were CHOX-class and assigned to lignin-like compounds via SR + AR reaction (Fig. 5). The high unsaturation and oxidation state of OICs (Figs. S15 and S16) again support our observation that few OICs were identified in the UV light treatment because highly unsaturated molecules are more prone to photolysis, such as most commonly photodecarboxylation (Fu et al. 2022a; Gonsior et al. 2014; Ward and Cory 2016).

Fig. 5
figure 5

The molecular composition of OICs of photoiodinated CSW samples with/without I− by sunlight and UV light irradiation. The van Krevelen diagrams are depicted based on different iodine numbers (a and b) and reaction types (c and d). Note that dots might be overlapping at the same position

Full size image

3.5 Conceptual mode of photohalogenation for coastal seawater DOM

The different OBCs and OICs formation observed in this study are attributed to the discrepancy in the redox potentials of bromine and iodine. Indeed, the formation of OBCs is not favorable in dark environments and even in natural sunlight irradiation environments (Hao et al. 2017) due to the high redox potential of bromine (Eθ (Br/Br−) ~ 1.92 V). While the generation of OICs is easier because of the lower redox potential of iodine (Eθ (I/I−) ~ 1.30 V) (Stanbury 1989). Thus, sufficient UV light irradiation will form abundant ROS. ROS can promote the formation of RBrS by oxidizing Br− due to its relatively higher redox potential in comparison with the bromine (Hao et al. 2018), which is confirmed by the result that CSW + UV + Br-2 generates more OBCs than CSW + UV + Br-1. Although UV light produces sufficient ROS and can facilitate the formation of OICs, OICs are more prone to photodegradation compared to OBCs because the bond dissociation energies of C–I (209 kJ/mol) are lower than that for C–Br (305 kJ/mol) (Bernardes et al. 2007; Wong and Cheng 2001). This discrepancy also supports our observation in this study that the OICs have higher aromaticity and unsaturation than OBCs.

The photochemical transformation of DOM in surface seawater is conceptually depicted in Fig. 6 with an emphasis on the formation of OBCs and OICs through the following critical reactions: (1) the generation of reactive oxygen species (ROS), such as hydroxyl radical (HO·), hydrogen peroxide (H2O2), singlet oxygen (1O2), superoxide (O2−·), ozone (O3), and DOM triplet state (3DOM*), by the absorption of sunlight by DOM in surface seawater (Hao et al. 2018; Marchisio et al. 2015; Richard and Canonica 2005; Zhang et al. 2022); (2) the generation of RHSs and or nonradical halogenated species (non-RHSs) from the DOM-photosensitized reduction of iodate and the oxidation of halides by DOM-induced ROS and photoexcited DOM or chlorophyll (Jammoul et al. 2009; Reeser et al. 2009; Saunders et al. 2012); (3) the addition of unsaturated C–C bonds and recombination with carbon-centered radicals during the reaction of RHS with DOM molecules (Alegre et al. 2000; Grebel et al. 2010); (4) electrophilic substitution and/or addition of non-RHSs with DOM (Deborde and von Gunten 2008; Heeb et al. 2014).

Fig. 6
figure 6

The mechanism diagram of coastal seawater DOM during photohalogenation

Full size image

4 Environmental Implications

DOM is a ubiquitous component of natural coastal seawaters and can be an essential precursor for natural organohalogen compounds (Liu et al. 2023b; Méndez-Díaz et al. 2014). The natural coastal seawater is highly complex and chemically diverse and is highly susceptible to anthropogenic activities, such as the extensive use of detergents, surfactants, and sulfa drugs (Boreen et al. 2005; Guo et al. 2021; Liu et al. 2023a; Wang et al. 2018b). Our newly proposed index, DBE-0.5O, is of great importance in accurately revealing the unsaturation degree of the carbon skeleton in DOM molecules for natural and engineered systems. Moreover, the molecular composition of natural coastal seawater DOM is affected by its photochemical transformation. The decarboxylation and desulfonation are expected to contribute mainly to the photodegradation of DOM in surface coastal seawater. Results of this study reveal that S-containing DOM molecules with high molecular weight, aromaticity, and unsaturation are preferentially photodegraded, highlighting the importance of the photochemical transformation of DOM in affecting the carbon and sulfur cycles in natural coastal seawaters.

The formation of OBCs and OICs in coastal seawater is primarily affected by light wavelength. Specifically, the UVA is the primary light for halogenation in natural sunlight. The higher light energy can generate more ROSs through the absorption of sunlight by DOM in surface seawater, which facilitates the generation of RHSs and results in the formation of more OHCs with highly unsaturated DOM molecules. Furthermore, OICs are expected to form more easily than OBCs under natural sunlight due to the higher redox potential of bromine than iodine. The formed OICs will be degraded preferentially by high energy UV light than OBCs because of the lower bond dissociation energies of C–I compared with C–Br. For the first time, the individual molecular compositions and characteristics of OBCs and OICs formed in coastal seawater under natural sunlight and UV light irradiation are elucidated by FT-ICR MS in this study. Compared with the environmentally relevant levels of Br− (50.0 mg/L used in this study versus ~ 65 mg/L in seawater), the I− concentration used in this study (5.0 mg/L) was approximately 100 folds of the total dissolved iodine concentration (66.1 ± 6.0 μg/L) in the surface seawater in Japan (Satoh et al. 2019). Therefore, photobromination is expected to be more prevalent than photoiodination in the surface coastal seawater in Japan. Moreover, the presence of stable emerging OHCs species such as C16H17O4S1I1 detected in all treatments of this study highlights the importance of investigating their source, structure, and toxicity.

5 Conclusions

The molecular composition, photobromination, and photoiodination of coastal seawater DOM (especially for S-containing DOM) are primarily governed by the light wavelength. The CHOS- and CHONS-class DOM molecules are more likely to be photodegraded through decarboxylation and desulfonation reactions. Although the coastal seawater in Tokyo Bay is abundant in S-containing DOM, the results of this study indicate that photochemical halogenation will not significantly yield S-containing OBCs and OICs species. Furthermore, the irradiation of UV light and penetrated sunlight preferentially facilitate the formation and accumulation of OBCs and OICs, respectively. In this study, the OBCs and OICs are predominantly formed by SR + AR and belong to lignin-like compounds, implying that the levels of OBCs and OICs and the terrigenous DOM precursors in natural coastal seawater should not be underestimated. It is essential to highlight that the proposed DBE-0.5O parameter is of great significance to more accurately reflect the unsaturation of the carbon skeleton of natural aquatic DOM.