Introduction

Sea spray is a natural phenomenon that forms when seawater mixes with organic matter, particulate matter, and salt1. Usually, under the action of waves, wind, and other factors, the spray in the water is brought to its surface. Certain organic substances in seawater with surfactant properties, such as marine organisms and anthropogenic pollution emissions, can facilitate spray formation; these emissions stabilise by reducing the surface tension of the water, ultimately forming a sea spray2. Prolonged contact with marine foam may induce allergic reactions in human skin and cause respiratory infections if inhaled3. In particular, research on the toxic mechanisms of sea spray is limited, mainly focusing on the direct harm to the human body caused by toxic algae and bacteria4.

Aerosol particle matters (PM) are a mixture of solid or liquid particles dispersed in gaseous media5. Toxicology and human health-related studies have shown that PM inhaled into the lungs can induce diseases of the human respiratory and cardiovascular systems and can even reduce the human immune capacity and increase the risk of cancer6. Respirable suspended particulates (RSP) and their surface organic toxic components have been the focus of atmospheric toxicity research in the past7.

However, our current knowledge of known toxic compounds does not fully explain the toxicity observed in sea spray and PM. Consequently, there is an immediate need to investigate previously unidentified toxic substances and the mechanisms underlying their toxicity.

Environmentally persistent free radicals (EPFRs), a unique category of free radicals, are capable of long-term stability within environmental matrices, such as airborne particles, soil, sediments, and water. They have lifespans that extend over several days8. Free radicals containing unpaired electrons exhibit high reactivity with other molecules. Based on the g-factor values, the predominant types of EPFRs present in the environment are C-centred-, O-, and C-centred radicals in proximity to oxygen9. In addition to natural sources such as volcanic eruptions, sandstorms, and forest fires10, the proportion of anthropogenic sources, such as fossil fuel combustion, garbage incineration, industrial waste gas, and automobile exhaust emissions, is gradually increasing11,12. The production of secondary EPFRs via ultraviolet irradiation cannot be neglected13. EPFRs are inhaled or enter the human body through the food chain, and these substances may cause oxidative stress and damage human cells and tissues14. EPFRs can react with chemicals in the atmosphere, potentially exacerbating air pollution and acid rain problems15. In addition, EPFRs can react with building materials and metal structures, leading to material ageing and corrosion16.

Toxic and harmful substances adsorbed onto the PM surface can directly react with cell components to generate active oxygen species (ROS) or induce ROS generation through indirect pathways17,18. However, the relationship between sea spray and ROS requires further study. ROS, including hydrogen peroxide (H2O2), superoxide free radicals (·O2−), hydroxyl free radicals (·OH) and organic free radicals (·R), is a type of oxygen species with a short life span but active chemical properties19. Although ROS generally exist in organisms, when the amount of ROS produced exceeds the antioxidant defence capacity of an organism, it causes oxidative stress and eventually leads to tissue damage and disease20. Current research indicates that EPFRs can trigger or be directly involved in ROS production21,22,23. Zhu et al. discovered that the production of various ROS, including superoxide and hydroxyl radicals, within polystyrene microplastics (PS-MP) was strongly correlated with the presence of EPFRs. Moreover, Li et al. discovered that ROS produced by EPFRs not only increased oxidative stress but also promoted gene expression associated with biological transformation in C. Furthermore, Ai et al. found that the rate of hydroxyl radical generation in secondary EPFRs was 4.4 times that in primary EPFRs, indicating that EPFRs in atmospheric secondary processes have a stronger oxidative potential.

Oxidative potential (OP) is frequently employed as a metric to evaluate the potential toxicity of airborne particulate matter24,25. OP is the ability of a substance to cause oxidative stress, that is, the ability to generate or promote free radical reactions26. Substances with high OP values are more likely to damage cells and biomacromolecules27. The development of acellular assays, including dithiothreitol (OPDTT), glutathione (OPGSH), and ascorbic acid (OPAA) assays, has made OP a useful tool for measuring future health28. Current research indicates that quinones and transition metals are the primary factors contributing to OP29,30. Wang et al. found that OP is a measure of the health effects of PM2.5 and is associated with various toxic substances, especially organic quinones. Wu et al. believed that the synergistic and antagonistic effects of metal and organic matter mixing in PM2.5 may be the main cause of OP. In addition, transition metals can not only react with aromatic compounds to produce EPFRs but also generate excess ROS by inducing REDOX reactions9,31. This trend not only indicates that metals contribute to the production of EPFRs and ROS but also implies that the high contribution of metals to OP may be related to the promotion of EPFR production by metals32.

In the present study, based on the amount of sea spray deposition, distance from the coast, and the degree of influence of wave movement, the sea spray sample collection area was divided into five blocks. The concentration, type, and lifetime of EPFRs in sea spray and size-resolved PM samples were analysed and compared using electron paramagnetic resonance (EPR) spectroscopy. The OP of the sea spray and PM samples, both water-soluble (OPWS) and water-insoluble (OPWIS), were determined using OPDTT to explore the contribution of different components to OPTotal. In addition, 5,5-dimethyl-1-pyrroline N-oxide (DMPO) combined with electron paramagnetic resonance (EPR) was used to analyse the ability of sea spray and PM samples to generate ROS. Finally, based on the two-film theory, the mass transfer process of EPFRs between the gas and liquid phases was characterised and calculated theoretically.

Results and discussions

The concentration of EPFRs in sea spray and size-resolved particle

Figure 1 summarises the EPFR concentrations found in sea spray samples collected from the coastal city and compares these concentrations with those observed in other samples. Analysis of variance (one-way ANOVA) showed no significant difference in the mass per unit area of the 64 sea spray samples collected from the sampling area. The average mass of each sample is 4.26 μg/cm2. As per the EPR analysis, the spin volume concentration of EPFRs in sea spray ranges from 0.3 × 1013 spins/m3 to 7.5 × 1013 spins/m3. To examine the spatial variability of EPFR concentrations in sea spray, the sampling area was divided into five regions (DA, RA, MA, GA, and SA) according to the amount of sea spray deposition, distance from the coast, and the degree of influence of wave movement. The study revealed that more sea spray was deposited in locations closer to the coast, and the concentration of EPFRs was higher in the collected samples (Fig. 1a). The EPFRs concentration of sea spray samples collected in the DA region was the highest, with an average value of 5.51 × 1013 spins/m3. The concentrations of EPFRs in sea spray samples collected from RA and MA regions were 4.15 × 1013 spins/m3 and 3.01 × 1013 spins/m3, respectively. The concentrations of EPFRs in sea spray samples collected from GA and SA regions were the lowest, with average values of only 0.97 × 1013 spins/m3 and 0.49 × 1013 spins/m3. Seawater is dense because it contains large amounts of salts and impurities. When water lashes against the shore under the action of tides, many bubbles gather and eventually form a sea spray. Closer to the coast, more sea spray is deposited and remains longer. According to previous studies, the concentration of EPFRs in samples is closely related to the PM mass and the presence of organic matter33,34. Sea spray closer to the coast deposited more PM and organic matter over a long period, resulting in a higher concentration of EPFRs in this part of the samples.

Fig. 1: The concentration of EPFRs in sea spray samples and comparison with the concentration of EPFRs in different samples.
figure 1

a Spatial distribution of EPFRs concentration in sea spray samples. b Comparison of EPFRs concentration in different samples and sea spray samples in the DA region, where the concentration of EPFRs in combustion, illuminated combustion (I-C), traffic PM and urban PM are from other studies (Guo et al.13; and Zhang et al.11) and the orange front represents sea spray samples. c Comparison of EPFR concentration in size-resolved atmospheric PM, and the orange front represents sea spray samples.

Full size image

Previous studies have shown that the concentrations of EPFRs vary in different media. EPFR concentrations in combustion-related PM, illuminated combustion-related PM (I-C), traffic-related PM and city-related PM ranged from 0.666 × 1015 spins/m3 to 8.730 × 1015 spins/m3, 0.779 × 1015 spins/m3 to 9.027 × 1015 spins/m3, 0.699 × 1014 spins/m3 to 3.99 × 1014 spins/m3, 9.8 × 1011 spins/m3 to 6.9 × 1014 spins/m3, respectively13,35. The concentration of EPFRs in sea spray samples collected from the DA region ranged from 3.90 × 1013 spins/m3 to 7.47 × 1013 spins/m3 (Fig. 1b), indicating that EPFRs in the samples from this region may be similar to the sources of EPFRs in PM samples related to traffic and cities. To obtain a more comprehensive understanding of the concentration of EPFRs in sea spray, 72 size-resolved PM samples were simultaneously collected along the coast of the sea spray collection area. The results showed that the concentration of EPFRs in size-resolved PM samples ranged from 4.57 × 1013 spins/m3 to 11.46 × 1013 spins/m3, higher than most sea spray samples (Fig. 1c). Because the particle size of sea spray is larger than that of fine particles, the surface area is smaller and the surface contains fewer porous structures, leading to a reduction in the number of sites for free radical formation, consequently resulting in lower concentrations of EPFRs32. Sea spray can be disseminated over a broad area owing to external factors such as wave motion, which enhances the potential for residents in coastal regions to be exposed to EPFRs when engaging in outdoor pursuits. To intuitively assess the exposure risk of EPFRs in sea spray, we equated the exposure level of EPFRs from sea spray to an equivalent dose from cigarette inhalation36, revealing that coastal residents inhale the equivalent of 0.91 cigarettes daily or 332 cigarettes annually per individual. This underscores the need to be vigilant about the potential health hazards associated with sea spray, particularly in regions where there is significant sea spray deposition.

Types and lifetime of EPFRs in sea spray and size-resolved particle

Free radicals are categorised as carbon-centred, oxygen-centred, nitrogen-centred, or metal-centred (International Union of Pure and Applied Chemistry, 2014). The predominant EPFRs in the environment are characterized by C-centred radical species (e.g. polycyclic aromatic radicals), C-centred radicals near oxygen (e.g. 3,4-methoxylphenacyl), and O-centred radicals (e.g. hemiquinone radicals)37,38. The important parameters for identifying EPFR type are g-factor and line width (△Hp-p). The closer the free electron is to the oxygen atom, the greater the value of the g-factor9. The g-factor of EPFRs in sea spray samples ranged from 2.0019 to 2.0032, with an average value of 2.0026 ± 0.0003, and mainly concentrated in 2.0025–2.0030, which accounted for 61% of the total 64 samples (Fig. 2a). The results show that the types of EPFRs in sea spray are mainly C-centred free radicals. The g-factor of EPFRs in size-resolved PM ranged from 2.0040 to 2.0053, with an average value of 2.0044 ± 0.0002, and mainly concentrated in 2.0041 to 2.0047, which accounted for 54% of the total 72 samples (Fig. 2c). The results showed that the types of EPFRs in the PM were mainly O-centre free radicals. Central carbon radicals are easily oxidised in the atmosphere, whereas central oxygen radicals are stable in the atmosphere39. Therefore, EPFRs in sea spray are more reactive than those in PM. As another characteristic information of EPFRs, line width (△Hp-p) is a standard method for measuring the spectrum width of EPR, which can be used to reflect the structure of precursor molecules32. The average △Hp-p of EPFRs in sea spray was 4.28 ± 0.6 G, and mainly concentrated in the range of 3–5 G, which accounted for 92% of the total 64 samples (Fig. 2a). The average △Hp-p of EPFRs in PM was 5.56 ± 0.4 G, and mainly concentrated in the range of 3–7 G, which accounted for 65% of the total 72 samples (Fig. 2c). In addition, the type and concentration of free radicals will affect the value of △Hp-p38. In the present study, EPFRs in sea spray samples and PM samples showed a trend of △Hp-p increasing with the increase of free radical concentration (Fig. 2b, d) (R2 = 0.76, R2 = 0.63, respectively), which was consistent with Kibet’s findings.

Fig. 2: The distribution of EPFR concentration in sea spray and PM with g-factor and line width (△Hp-p).
figure 2

a The number in the grid represents the number of sea spray samples in the range of g-factor and △Hp-p. b The number in the grid multiplied by 1013 represents the average concentration of EPFRs in all sea spray samples in this grid. c The number in the grid represents the number of PM samples in the range of g-factor and △Hp-p. d The number in the grid multiplied by 1013 represents the average concentration of EPFRs in all PM samples in this grid.

Full size image

To study the lifetime of EPFRs in sea spray and PM samples after 120 days of exposure at ambient temperature, we conducted a second EPR analysis of the samples and observed that the EPFRs within them were subject to decay. The concentration of EPFRs in over 90% of the sea spray and PM samples exhibited a range of decreasing concentrations (Fig. 3). According to the different types of EPFRs in the sea spray and PM samples, the EPR spectrum was divided into C-centred- and O-centred regions for attenuation analysis (Fig. 3b, d). The results showed that the signal intensities of both types of EPFRs decreased; however, the degrees of attenuation were different. For the same decay time, the C-centred EPFRs decayed faster than the O-centred EPFRs, indicating that the O-centred EPFRs were more stable. The attenuation of the EPFRs in the sea spray and PM samples satisfied the first-order attenuation model. It was calculated that the average half-life of EPFRs in sea spray samples was 0.32 years, and that in PM samples was 0.51 years, which was shorter than that of EPFRs in road dust collected in Xi ‘an32. According to the definition rules of decay types by Chen et al.40, most EPFRs in the sea spray samples belong to the fast decay type, whereas most EPFRs in the PM samples belong to the slow decay type. However, the concentrations of EPFRs in some sea spray and PM samples increased instead of decreasing (blue arrow in Fig. 3a and orange arrow in Fig. 3c). These samples were collected at the Circular Island Road Beach, a very dense area of human and traffic flow, which not only produces a large number of primary EPFRs but also potentially large amounts of precursor substances, such as polycyclic aromatic hydrocarbons (PAHs), that can produce secondary EPFRs. Zhao et al. 12 found that EPFRs can be formed on the surface of Cu (II)-montmorillonite containing PAHs. Guo et al.13 showed that the important precursors of non-extractable and extractable PM components that form secondary EPFRs are metals and quinones, respectively. In addition, the mean g-factor of the EPFRs in the sea spray samples showed an increase of 0.0012 post-decay relative to the initial measurement, indicating the C-centred EPFRs in these samples might gradually convert into O-centred EPFRs following prolonged exposure to atmospheric oxygen. This also implies that EPFRs in the gas and liquid phases can transform into each other.

Fig. 3: Concentration changes of EPFRs in samples before and after decay and EPR spectra.
figure 3

a Blue arrows represent sea spray samples with increased concentrations of EPFRs after decay. b Signals of EPFRs in sea spray samples before and after decay. c Yellow arrows represent PM samples with increased concentrations of EPFRs after decay. d Signals of EPFRs in PM samples before and after decay.

Full size image

OP of sea spray and size-resolved particle

The oxidative potential is a measure that can quantify the ability of PM oxidative target molecules and is used to represent the ability of PM to produce ROS41. The consumption of the reducing substance dithiothitol over time (OPDTT) is a common method used to analyse the oxidative potential capacity of samples42. The results show that the average OPTotal of sea spray samples in the five regions is 3.49 ± 0.16 nmol/min/μg, and the average OPTotal of size-resolved PM samples is 3.23 ± 0.12 nmol/min/μg. After the OP analysis of the five sampling areas, we found that the closer to the coast, the more sea spray was deposited and the higher the OPTotal (Fig. 4a). The average OPTotal values of DA and RA were 4.37 nmol/min/μg and 3.76 nmol/min/μg, respectively. We also conducted an OP analysis on size-resolved PM samples, and the results showed that the OPTotal of samples with smaller aerodynamic diameter were higher (Fig. 4b), which is consistent with the results of Simonetti et al.43. Chirizzi and colleagues44 similarly observed that the rate of DTT consumption for PM10 (3.1 nmol/min/μg) was significantly less than that for PM2.5 (9.2 nmol/min/μg). To delve deeper into the OP characteristics of the sea spray and PM samples, we differentiated the OP into water-soluble (OPWS) and water-insoluble (OPWIS) fractions. It was found that the average OPWS and OPWIS of sea spray samples were 2.89 nmol/min/μg and 0.60 nmol/min/μg, respectively, of which OPWS accounted for 82.6% of the OPTotal (Fig. 4a). The average OPWS and OPWIS of PM samples were 2.35 nmol/min/μg and 0.88 nmol/min/μg, respectively, of which OPWS accounted for 73.2% of the OPTotal (Fig. 4b). These findings showed that water-soluble constituents were primarily responsible for the oxidative toxicity observed in the sea spray and PM samples.

Fig. 4: DTT consumption rates and percentages of different components to DTT consumption for sea spray samples and size-resolved PM samples.
figure 4

a DTT consumption rate and percentages of different components to DTT consumption for sea spray samples. b DTT consumption rates and percentages of different components to DTT consumption for size-resolved PM samples.

Full size image

The transition metals and quinones in the samples can lead to the oxidation of DTT through a catalytic REDOX reaction, that is, the transition metals and quinones can contribute to the OP of the samples45. To further explore the mechanism of the toxic release of OPWS sea spray and PM samples, the concentrations of water-soluble transition metals in the samples and their contribution to OPTotal were analysed in this study. The concentration distribution of transition metals in the size-resolved PM samples is shown in Fig. S3a, where the blue dots represent the average transition metal concentration in the samples with each aerodynamic diameter, and the red dots represent the average transition metal concentration in all PM samples. The concentration distributions of transition metals in the sea spray samples are shown in Fig. S3b, where the red dots represent the average transition metal concentration of the samples in region DA, and the blue dots represent the average transition metal concentration of the samples in the other four regions. The results showed that Fe, V and Zn were the most concentrated in the sea spray and PM samples. Fe and Zn are important metals released by the wear and tear of cars and ships, and V may be widely derived from marine organisms. The contribution of each component to OPTotal was evaluated by establishing a regression equation between the transition metal concentration and DTT consumption rate (Table 1). Transition metals and other unknown factors were identified as major contributors to the OPTotal of the sea spray and PM samples (Fig. S4).

Table 1 Regression equations for different components and DTT losses in sea spray and PM samplesa
Full size table

The results showed that transition metals contributed an average of 66.2% to the OPTotal of the size-resolved PM samples and 51.3% to the OPTotal of the sea spray samples. The contribution of transition metals to the OPTotal of the PM<4.7 samples showed little change, with an average value of 57.9%, whereas the OPTotal contribution to the PM4.7–10 samples was greater, reaching 82.7%. This part of the PM samples had a relatively low OPTotal value but contained a higher concentration of transition metals, resulting in a higher contribution rate. Fe, Cu and Mn were the main contributors to the OPTotal of the sea spray and PM samples, with average contribution rates of 12.7%, 15.5%, and 10.1%, respectively. The average contribution rate of Co, Zn, and V to OPTotal was 22.5%. These results indicate that transition metals play a significant role in the overall oxidative potential (OPTotal) of the sea spray and PM samples. Lyu et al.46 found that transition metals (mainly copper and manganese) contributed 55 ± 13% to the DTT activity of size-resolved PM, while quinone only accounted for 8 ± 3%. Li et al.32 reported that the final contribution of transition metals to the OPWS was 51%. This discussion is consistent with our findings. In addition, transition metals can also participate in the formation of EPFRs. Precursor molecules such as aromatic compounds can form stable persistent free radicals by transferring electrons to the transition metal surface39. Simultaneously, transition metals can generate excess ROS by inducing REDOX reactions31. This outcome suggests that transition metals not only enhance the OP of the sample but also play a vital role in the production of EPFRs and reactive oxygen species (ROS).

ROS of sea spray and size-resolved particle

When PM is inhaled and deposited by respiration, ROS such as ·OH and ·O2− will be produced, resulting in changes in the REDOX state of cells, and eventually causing oxidative stress hazards to the human body47. Capturing these free radicals with DMPO is a common method used to analyse the reactive oxygen generation capacity of samples. By observing the EPR spectrum (black) of the aqueous solution of sea spray and PM samples after adding DMPO, it was found that the spectrum presented a multi-peak alternating state, which indicated that a variety of free radicals were produced in the sea spray samples. After deconvolving the observed spectrum into a four-wire signal of DMPO-OH (red) and a six-wire signal of DMPO-R (blue), we found that the generated analogue spectrum (green) fitted the original spectrum well (Fig. 5). The above results showed that the ·OH and ·R were generated after sea spray and PM samples were exposed to water (Tables 2 and 3). After EPR quantitative analysis, the proportion of ·OH and ·R in sea spray samples was about 54 ± 3% and 46 ± 4%, respectively. Size-resolved PM samples generate 36 ± 7% ·OH and 64 ± 5% ·R after contact with water. This difference may be due to the chemical compositions of the sea spray and PM samples. Tang et al.48 detected the production of hydroxyl radicals (·OH, ranging from 20% to 50%) and organic radicals (·R, ranging from 35% to 80%) in various urban PM samples. This suggests that both sea spray and PM samples play a significant role in inducing ROS production; however, more experimental evidence is needed to derive a more detailed explanation. In addition, in the present study, the mean ROS concentration in sea spray was found to be 9.4 pmol/m3, surpassing the ROS levels measured in urban PM (ranging from 3 pmol/m3 to 7.5 pmol/m3) studied by Tang et al.48, but lower than the ROS value of PM near highways (14 pmol/m3) studied by us before11. This trend suggests that the health risks posed to humans by the formation of ROS from sea spray may be somewhere between PM near highways and urban PM.

Fig. 5: Spin-capture process of reactive oxygen species generated from sea spray samples after contact with water.
figure 5

The black line is the observed spectrum of the actual EPR, the red line is the standard curve of DMPO-OH, and the blue line is the standard curve of DMPO-R. The green line is a fitting curve formed by the addition of the red and blue lines. The vertical dashed lines indicate the positions of each peak of the DMPO-OH and DMPO-R adducts.

Full size image
Table 2 The g-factor and hyperfine split constant (hsc) of size-resolved PM in the spin-capture experiment
Full size table
Table 3 The g-factor and hyperfine split constant (hsc) of sea spray in the spin-capture experiment
Full size table

EPFRs may catalyse or even directly participate in ROS production. To explore the relationship between EPFRs and ROS in sea spray samples, the correlation between the concentration of EPFRs and the chemical components that may have corresponding associations with EPFRs was studied (Fig. S5). In particular, the results revealed a positive correlation between EPFR concentrations and ROS concentrations, and the correlation between EPFR concentrations and ·OH was significantly higher than that between R radicals. The concentrations of EPFRs were positively correlated with OPTotal. The OPDTT can be used as an indicator of ROS generation49. The above results may indicate that EPFRs in sea spray samples participate in the generation of ROS and contribute more to the generation of ·OH. Recent studies have also shown that EPFRs can induce ROS production and cause oxidative stress in humans12,50. In addition, transition metal (Fe, Cu, and Mn) concentrations were positively correlated with EPFR concentrations, as well as ROS concentrations, especially ·OH. Previous studies have shown that transition metals can transfer electrons through precursor substances such as aromatic compounds to form stable persistent free radicals51. Similarly, transition-metal-induced REDOX reactions can generate excess ROS31. This finding not only indicates that metals contribute to the production of EPFRs and ROS but also implies that the high share of metal contribution to OPDTT may be related to metal promotion of EPFR production.

EPFRs mass transfer process between gas and liquid phase

EPFRs have been confirmed to exist widely in the atmosphere, water, and soil52,53. However, the transport and transformation processes of EPFRs between the gas and liquid phases have not been fully studied. The two-film theory has been widely used in the theoretical analysis of the mass transfer of gaseous pollutants between the gas and liquid phases54,55. Based on the two-film theory, the mass transfer process of the EPFRs was assumed to be as follows: EPFRs of the gas phase entered the liquid phase through the gas and liquid film layers via molecular diffusion. Similarly, a transfer process from the liquid to gas phase also occurs. In the gas and liquid phases, there is no concentration gradient of EPFRs from a macroscopic perspective, owing to fluid motion. The molecular transport process was concentrated in the two film layers; that is, the concentration gradient was concentrated in the two film layers. In the steady-state absorption process, the flux of EPFRs transferred from the gas phase to the phase interface is equal to the flux of EPFRs transferred from the phase interface to the liquid phase; that is, there is no absorption resistance at the phase interface and no accumulation or consumption of EPFRs (Fig. 6).

Fig. 6: Two-film mass transfer process of EPFRs between gas and liquid phase.
figure 6

C1 and C2 represent the average concentration of EPFRs in the gas phase and the liquid phase, respectively. C1* and C2* represent the concentration of EPFRs at the phase interface. NA and NB represent the mass transfer rate equations of EPFRs in the gas phase and the liquid phase, respectively.

Full size image

Based on the above concepts, the mass transfer process of EPFRs between the gas and liquid phases was calculated theoretically. The average concentration of EPFRs in the gas phase (C1) was 6.57 × 1013 spins/m3 and that in the liquid phase (C2) was 2.82 × 1013 spins/m3. The gas mass transfer rate equation NA = 120.41 × (C1 − C1*), liquid mass transfer rate equation NB = 95.32 × (C1 − C1*). After theoretical calculation, the interfacial concentrations of C1* and C2* were 4.92 × 1013 spins/m3 and 4.91 × 1013 spins/m3, respectively. The difference in the interfacial concentration was within the error range and conformed to the theoretical concept. The above results show that the mass transfer resistance of the EPFRs in the gas and liquid phases cannot be ignored and that the mass transfer rate is controlled by the gas and liquid films.

Methods

Experimental materials

A 100 µg/ml standard solution of Mn2+ and Cr3+ was utilised for the calibration of the absolute spin concentration and g-factor of the samples before and following experimentation32. Dithiothreitol (DTT, 1 mM), phosphate-buffered saline (PBS, 0.1 M, pH 7.4), tricarboxylic acid (TCA, 10%, w/v), (hydroxymethyl) aminomethane hydrochloride (Tris, 0.4 M), 5,5’-dithio bis-(2-nitrobenzoic acid) (DTNB, 10 mM), metal standard solution (Cu, Fe, Zn, Mn, Co, and V, 100 μg/ml) were used in oxidative potential experiments. 5,5-dimethyl-1-pyrroline N-oxide (DMPO, Dojindo), diethylenetriaminepentaacetic acid (DTPA, 98%), hydrogen peroxide (H2O2, 1 mM), horseradish peroxidase (HRP, 100 μM) were used in spin-capture experiments.

Division of sea spray sampling area

To completely characterize the potential risks of sea spray, the collection area was divided into five blocks (Fig. 7a) according to the amount of sea spray deposition, distance from the coast, and the degree of influence of wave movement. Ι. Depositional area (DA): the junction of seawater and the coast, where the sea spray deposition was the largest, remained stable for a long time, with the least influence from the movement of waves. II. Rich area (RA): this area has an increasingly stable sea spray deposition, with less influence from wave movement. III. Moderate area (MA): this area has moderate sea spray deposition, which is affected by wave movement. IV. Generator area (GA): the area where waves hit the sea surface and formed sea spray, which generated in this area was fresher. V. Sea area (SA): this area is far from the coast, mainly composed of clear seawater, with no obvious sea spray deposition on the surface of the seawater.

Fig. 7: Sampling diagram of sea spray and size-resolved PM.
figure 7

a Sampling diagram of sea spray. b Sampling diagram of size-resolved PM.

Full size image

Collection of sea spray and PM samples

Sea spray samples were collected from major coastal cities in southeastern China and a key hub in the local maritime transport system. According to data released by the local statistics bureau, Sea, with an area of approximately 390 square kilometres, with several natural goods ports, has been vital to foreign trade on the southeast coast of China. We performed sampling within 20–31 August 2023, with sampling locations in the Circular Island Road Beach (24°51′N 118°21′E) and Rainbow Beach (24°64′N 118°16′E). Based on a predefined sampling area (Fig. 7a), In the unit area (50 cm2), the sea spray was first separated from the seawater by glass plates, then the upper sea spray was extracted by folding pump (S-82, Xiaomi, China), and finally collected into brown glass bottles. To explore the relationship between the sea spray and PM samples, an Anderson Stage Eight sampler (TE-20-800, Tisch, USA) was set up at a flow rate of 28.3 L/min adjacent to the sampling site to capture the size-distributed PM samples (Fig. 7b). The PM samples were collected on pre-baked quartz filters, and these filters were heated to 450 degrees Celsius for 4 h before the sampling process. Ultimately, 64 sea spray samples and 72 PM samples were preserved at a temperature of −20 °C in the laboratory for subsequent analysis.

Analysis of EPFRs

Electron paramagnetic resonance (EPR) spectroscopy (EMXnano, Bruker, Germany) was employed to quantify the concentrations, identify the types, and determine the lifetimes of EPFRs in the sea spray and size-resolved PM samples. The instrument settings were as follows: magnetic field intensity 340–345 mT, receiver gain 42 dB, microwave power 20 dB, sweep width 20 mT, scans 5; scan time 50 s11.

The concentration of the EPFRs was expressed as the spin volume concentration (spins/m3). The classification of EPFR types was mainly based on g-factor and line width (△Hp-p). To ascertain the decay patterns of EPFRs in sea spray and size-resolved PM, the samples were placed at ambient temperature (28 ± 5 °C) and relative humidity (75 ± 5 °C) after EPR detection and were tested again 120 days later. The decay kinetics of samples are commonly analysed using a first-order decay model to determine the lifespan of free radicals56. The lifetime of the free radicals was calculated based on the difference in the signal strengths generated by the two tests of the samples. In addition, to evaluate the potential health hazards of EPFRs in sea sprays for coastal inhabitants, a comparable number of cigarettes per person was calculated based on daily EPFR exposure (Text S1).

Analysis of OP

The rate of DTT consumption over time (OPDTT) was used to assess OP produced by sea spray and PM. The experimental steps for the determination of OPTotal are as follows: (1) after the samples were fully dissolved with ultra-pure water, 100 µl mixture was removed and transferred to three centrifuge tubes. (2) One hundred microliter DTT solution and 800 µl PBS solution were added to each centrifuge tube. (3) The centrifuge tube was incubated in a water bath environment at 37 °C for 10 min, 20 min, and 30 min, respectively. (4) One hundred microliter TCA solution was added to each centrifuge tube to terminate the reaction. (5) After the solution was filtered, 200 µl filtrate was collected from each centrifuge tube and transferred to the reaction dish four times, and subsequently, 100 µl Tris solution and 20 µl DTNB solution were added successively. (6) A spectrophotometer (MOLECULAR SPECTRAMAX M2; Molecular Devices, San Jose, CA, USA) was used to determine the absorbance of the samples at 412 nm. (7) OPDTT was calculated according to the remaining concentration of DTT across various durations of reaction, represented as the DTT activity per unit mass of the sample (DTTm, nmol/min/µg). The experimental protocol for the determination of OPWS differs from that of OPTotal in that the sample is filtered immediately after being dissolved in ultrapure water to obtain water-soluble samples. The average OPTotal minus the average OPWS was the average OPWIS.

To assess the contribution of water-soluble metal components to the OPTotal of the samples, the levels of water-soluble metals in sea spray- and size-separated PM samples were quantified using an inductively coupled plasma mass spectrometer (ICPMS-2030LF, Shimadzu Corporation, Japan). Six common metals (Cu, Fe, Mn, Zn, Co, and V) were selected from the 51 metals measured, because they have been shown to contain high concentrations and REDOX activity to produce ROS. The metal concentration was divided into several gradient ranges, and a regression curve between the water-soluble transition metal concentration and the DTT consumption rate was established (Figs. S1 and S2). To mitigate the influence of extraneous factors on the OPDTT, the DTT consumption rate of the blank control group (KBD) was measured. The actual DTT consumption rate (KAD) of the sample was equal to the theoretical DTT consumption rate (KTD) derived from the standard curve minus KBD (Table S1).

Analysis of ROS

The ROS generated in the samples were captured using the spin-trapping agent DMPO (Dojindo, China) and analysed using EPR. The experimental procedure is as follows: 1) After the sample was fully dissolved with ultra-pure water, 50 μl sample liquid was added to the centrifuge tube. (2) One-hundred thirty microliter PBS solution containing DTPA was added to the centrifuge tube as a transition metal chelating agent. (3) Five microliter DMPO was added to the centrifuge tube for trapping free radicals. (4) Ten microliter H2O2 solution was introduced as a scavenger. (5) Twenty microliter HRP solution was added to initiate the reaction. (6) Twenty microliter solution was extracted for EPR analysis. The EPR instrument settings were consistent with those employed for the analysis of the EPFRs; only the microwave attenuation was adjusted to 10 dB. The SpinFit module in the Xenon software was used to scan the observation curve of the samples. Deconvolution of the observed curve and comparison with the standard curve can be used to identify the free radical species. The SpinCount module enabled the quantitative analysis of free radicals. The concentration of generated ROS was expressed in terms of the sampling volume concentration (pmol/m3).

Statistical analysis

The statistical relationships between EPFRs and other chemicals in sea spray were studied using Pearson correlation analysis. After the normality and variance homogeneity tests, the data were entered into a statistical software platform (IBM SPSS Statistics, version 27.0, USA) for analysis.

Two-film mass transfer process of EPFRs

Based on the two-film theory, a detailed mass transfer process of the EPFRs between the gas and liquid phases was assumed. Owing to the sufficient turbulence of the fluid, the concentration of EPFRs in the gas and liquid phases is uniform from a macro point of view; that is, there is no concentration gradient in the gas and liquid phase environments. The concentration of EPFRs in the gas phase C1 is the average concentration of EPFRs in the 72 size-resolved PM samples. The concentration of C2 of the EPFRs in the liquid phase was the average concentration of the EPFRs in the 64 sea spray samples. Through molecular diffusion, EPFRs in the gas phase continuously enter the liquid phase through gas and liquid films. Similarly, the reverse process also exists. We assumed that the phase interface experienced no absorption resistance; that is, the concentration gradient was concentrated in the two film layers.

$${\rm{Gas}}\; \; {\rm{transfer}}\; {\rm{rate}}\; {\rm{equation}}:\;{N}_{A}={k}_{a}\,\cdot\,({C}_{1}-{C}_{1}^{*})$$
(1)
$${\rm{Gas}}\; {\rm{mass}}\; {\rm{transfer}}\; {\rm{rate}}\; {\rm{equation}}:\;{N}_{A}={k}_{a}\,\cdot\,({C}_{1}-{C}_{1}^{*})$$
(2)

In a stable absorption process, there is no accumulation or consumption of substances at the phase interface; therefore, the absorption rate of EPFRs through the gas film is equal to that through the liquid film:

$${N}_{A}={k}_{a}\,\cdot\,({C}_{1}-{C}_{1}^{* })={N}_{B}={k}_{b}\,\cdot\,({C}_{2}^{*}-{C}_{2})$$
(3)

According to Henry’s law at the interface:

$${C}_{1}^{* }=m\,\cdot\,{C}_{2}^{* }$$
(4)

the aforementioned formula is solved simultaneously:

$${{{C}}}_{1}^{* }=\frac{{k}_{a}\,\cdot\,{C}_{1}+{k}_{b}\,\cdot\,{C}_{2}}{\frac{{k}_{b}}{m}+{k}_{a}}\,$$
(5)
$${{{C}}}_{2}^{* }=\frac{{k}_{a}\,\cdot\,{C}_{1}+{k}_{b}\,\cdot\,{C}_{2}}{{k}_{b}\,+\,m\,\cdot\,{k}_{a}}$$
(6)

where NA, NB are the mass transfer rate, spins/(m2 s); ka, kb are mass transfer coefficients, m/s. Because it cannot be measured directly experimentally, the mass transfer coefficient is derived from the common empirical formula of the mass transfer coefficient48,57; C1 and C1* are the concentrations of EPFRs at the gas phase and phase interface, spins/m3; C2 and C2* are the concentrations of EPFRs in the liquid phase and phase interface, spins/m3; and m is the phase equilibrium constant.