Skip to main content
NCBI home page
Environment & Health logoLink to Environment & Health
. 2025 Aug 21;3(11):1299–1310. doi: 10.1021/envhealth.5c00157

Are “Clean” Days Really Clean? Single Particle Elemental Signatures of PM2.5 Unveil Their Pulmonary Health Risks

Mengyuan Wang †, Zuoshun Niu †, Xiaojing Yang †, Miao Xu †, Bo Peng †, Zhiqiang Shi †, Zhiyan Wang †, Songda Li ‡, Yunqi Zhang ‡, Hui Zhao ‡, Xinran Ma §, Yi Yang †,∥,⊥,*
PMCID: PMC12645310  PMID: 41306298

Abstract

Ambient PM2.5 mass concentrations inadequately reflect health risks due to compositional heterogeneity. This study utilized single-particle inductively coupled plasma time-of-flight mass spectrometry (spICP-TOF-MS) to characterize high-resolution elemental signatures of metal-containing fine particles (MCFPs) in PM2.5 from an urban area with an intensive anthropogenic influence during different pollution levels in the winter and spring. Al-, Si-, Fe-, Mn-, and Pb-containing FPs accounted for approximately 80% of total MCFPs, with higher number concentrations in spring than in winter, increasing with pollution levels. Unlike Al- and Si-containing FPs, Fe-containing FPs were predominantly multimetal (mm)­FPs (48–87%), with higher proportions in winter than spring and increasing with pollution levels. Notably, a larger fraction of mmFPs, particularly Fe-rich FPs, were associated with toxic metals (e.g., Mn and Pb) on clean days than on polluted days. Lung cytotoxic potencies, including oxidative stress and cytotoxicity, were up to 8.1 and 6.3 times greater on clean days than on polluted days. Fe-rich FPs and their associated toxic metals were identified as first-tier factors in regulating cytotoxic potency, playing a more critical role than organic/elemental carbon and dissolved metals. Machine learning-based source apportionment indicated that anthropogenic-sourced MCFPs, especially Fe-rich FPs, contributed more during winter than in spring, with peak contributions on clean days.

Keywords: PM2.5 , metal-containing fine particles, spICP-TOF-MS, source apportionment, toxic potency

graphic file with name eh5c00157_0008.jpg

graphic file with name eh5c00157_0006.jpg

1. Introduction

Atmospheric fine particulate matter (PM2.5), characterized by its complex chemical heterogeneity, has been extensively documented as a critical environmental pollutant associated with substantial public health burdens. − Over six billion people worldwide still breathe substandard air (24 h average PM2.5 concentration >15 μg/m3 according to WHO). Studies also suggest that even in areas with annual average PM2.5 concentrations below the latest WHO recommendation of 5 μg/m3, these fine particles (FPs) are linked to increased mortality in a superlinear concentration response. Chronic low-concentration exposure to PM2.5 has been associated with progressive health deterioration, which can be attributed to highly toxic species present in the particles. , Exclusive reliance on bulk PM2.5 mass concentration as an air quality metric, without accounting for source-specific chemical heterogeneity, may substantially underestimate the associated health risks.

Sulfate, nitrate, black carbon (BC), organic carbon (OC), and metals are the major components of PM2.5. , Among these, OC and metals are considered the more important toxic components, and they are capable of inducing oxidative stress and inflammatory responses, which may progress to cellular dysfunction and carcinogenesis. − Recent studies have investigated the contribution of polycyclic aromatic hydrocarbons (PAHs) and metals to PM2.5-derived toxicity, with metals being primarily responsible for over 80% of the oxidative stress activity. , Furthermore, inhaled metal-containing FPs (MCFPs) can penetrate biological barriers, including the air-blood and blood-brain barriers, and enter the systemic circulation and critical organs such as the brain − and heart. , Notably, Maher et al. revealed that magnetite nanoparticles, prevalent in urban atmospheric PM, can be transported into the human brain and may be linked to neurodegenerative diseases such as Alzheimer’s disease, while toxic metals like Pb in PM2.5 can cause severe cardiovascular damage. The prolonged biological persistence of MCFPs in these target organs, coupled with their limited clearance mechanisms, , suggests a potential for bioaccumulation and progressive toxicity even at low ambient exposure levels.

MCFPs originate from both natural and anthropogenic sources. Natural emissions, including volcanic eruptions, wildfires, and aeolian processes, contribute substantial quantities of MCFPs dominated by lithogenic elements (e.g., Fe, Al, Si). Anthropogenic activities represent diverse emission profiles; for example, iron and steel manufacturing releases Fe-, Cr-, and Mn-enriched particles, coal combustion generates Pb-, Fe-, and Ti-bearing particles, and vehicular emissions produce particles rich in Mn, Cu, Ba, and Zn. − However, traditional analytical approaches have predominantly relied on bulk metal quantification through acid digestion methods, which obscure particle-specific chemical signatures. Recent advances in analytical techniques, particularly single-particle inductively coupled plasma time-of-flight mass spectrometry (spICP-TOF-MS) revolutionized MCFP characterization by enabling high-throughput, multielemental analysis at the single-particle level. − Although aqueous dispersion pretreatments may induce modifications to ambient PM2.5, the processed samples retain health-relevant properties as the liquid-phase treatment better reflects the actual particle states (e.g., composition and surface chemistry) encountered during respiratory deposition and lung fluid interaction. , This cutting-edge technique has revealed that coal combustion-derived particles, particularly iron in Al-rich FPs and Fe-rich FPs, exhibit distinct reactive oxygen species (ROS) generation capacities in lung cells. Furthermore, size-dependent compositional variations in coal-derived magnetite nanoparticles (NPs) were identified, which further modulate their pulmonary cytotoxicity profiles. The integration of machine learning algorithms with spICP-TOF-MS data has further enhanced source apportionment accuracy, enabling precise identification of MCFP origins. Despite these advancements, significant knowledge gaps persist in understanding the source-specific toxicological profiles of individual MCFPs. High-resolution characterization of MCFPs at the single-particle level remains crucial for elucidating their source contributions and identifying key toxic components, yet comprehensive data sets are still lacking.

This study aims to elucidate the high-resolution elemental signatures of MCFPs at different PM2.5 levels, with particular emphasis on identifying key toxic components and their emission sources. The investigation was conducted in Shanghai, a Chinese megacity with intensive anthropogenic activities and complex air quality challenges. PM2.5 samples were systematically collected during distinct pollution scenarios, specifically targeting winter haze episodes and spring dust storm events. For comparison, background samples were obtained during periods meeting both WHO (PM2.5 < 15 μg/m3) and Chinese air quality standards (PM2.5 < 35 μg/m3). The specific objectives of this study are (1) to quantitatively characterize the distributions and characteristics of MCFPs in PM2.5 at different pollution levels, (2) to assess the toxic potency of PM2.5 through lung cytotoxicity tests and identify the key MCFP components driving cellular responses, and (3) to identify the principal sources of both MCFPs and their key toxic components through advanced single-particle analysis. The data obtained in this study can provide insights into PM2.5-associated health risks and contribute to the development of source-oriented mitigation strategies.

2. Materials and Methods

2.1. Sample Collection

The sampling campaign was conducted at an urban monitoring station situated in Shanghai’s Minhang District (121.5°E, 31.0°N; Figure S1). The monitoring site was strategically located in a mixed urban environment characterized by high-density anthropogenic pollution sources. Situated within 2 km of elevated expressways and arterial roads, the sampling area falls within a 10 km radius of multiple health-relevant emission sources: a coal-fired power station (primary stationary source), concentrated high-tech manufacturing zones, tertiary commercial districts, and high-rise residential clusters. The sampling platform was positioned 2–3 m above ground level on a rooftop, ensuring unobstructed airflow with no proximate buildings or vegetation. For comprehensive PM2.5 characterization, two complementary sampling systems were deployed: (1) a medium-flow sampler (Model KC-120H, Qingdao Laoshan Electronic Instrument Factory, China) equipped with polytetrafluoroethylene (PTFE) filters (Whatman) operating at 100 L/min for trace metal analysis, and (2) a high-volume air sampler (Model TE 6070, Tisch Environmental Inc., USA) utilizing quartz fiber filters (Whatman) with a flow rate of 1.13 m3/min for organic carbon (OC) and elemental carbon (EC) quantification. The sampling protocol encompassed 46 days during the winter and 44 days in the spring, capturing seasonal variations in atmospheric composition and pollution characteristics. According to PM2.5 concentration and PM2.5/PM10 ratio, the sampling periods in this study were classified into four distinct categories: (1) WHO days (PM2.5 < 15 μg/m3), where PM2.5 levels met the World Health Organization’s recommended limit; (2) GB days, (PM2.5: 15–35 μg/m3), meeting China’s grade I environmental air quality standard; (3) haze days (PM2.5 > 35 μg/m3 with PM2.5/PM10 > 0.5), and (4) dust events (PM2.5 > 35 μg/m3 with PM2.5/PM10 < 0.5). This classification scheme was applied to both winter and spring sampling campaigns (Figure S2), allowing for the detailed analysis of MCFP characteristics across varying PM2.5 levels.

Source-specific reference samples were collected from five representative emission categories to establish a comprehensive MCFP source profile: (1) crustal-derived desert dust, (2) coal combustion particulates, (3) diesel vehicle emissions, (4) gasoline/electric vehicle emissions, and (5) biomass combustion products. Detailed information about the sampling protocols is provided in Text S1.

2.2. OC/EC, Electron Microscopy Analysis, and Meteorological Parameters

The OC and EC concentrations in PM2.5 samples collected on quartz fiber filters were measured using a thermal/optical carbon analyzer (DRI Model 2015, Atmoslytic Inc., CA, USA). The morphology, composition, and crystal structures of MCFPs were characterized on a 300-mesh gold grid (Electron Microscopy Sciences) using transmission electron microscopy (TEM, JEOL 2100 TEM), coupled to energy dispersive X-ray spectroscopy (EDXS). During the investigation period, data on PM10 concentrations, relative humidity (RH), wind speed (WS), and temperature (T) were obtained from the Minhang Pujiang Atmospheric Monitoring State Control Point near the sampling station (Figure S3).

2.3. spICP-TOF-MS Analysis

All PM2.5 samples were processed using standardized pretreatment protocols. Briefly, a PM2.5 sampled PTFE filter segment (approximately 0.5–1 cm2) was cut using ceramic scissors for each sample and placed into a 10 mL centrifuge tube with 5 mL of 0.2% SDS solution. The FPs were dispersed using an ultrasonic probe (Scientz-IID Ultrasonic Cell Rrusher, Ningbo Xinzhi) in a water bath for 20 min at 285 W (15–25 °C). These solutions were subsequently diluted 100–5000 times and dispersed before being introduced into the spICP-TOF-MS (icpTOF R, TOFWEAK, Switzerland) equipped with an autosampling system (ESI, microFAST SC, USA) for analysis in standard mode. 32 metals, including Na, Mg, Al, Si, K, Ca, Ti, V, Cr, Fe, Mn, Co, Ni, Cu, Zn, As, Se, Rb, Sr, Zr, Nb, Mo, Ag, Cd, Sn, Sb, Ba, La, Ce, W, Tl, Pb, were determined in individual MCFPs simultaneously. Dissolved calibration standards were prepared from a multielement ICP certified reference mixture (0, 0.1, 0.5, 1, 2, 5, 10 μg/L, diluted in 1% HNO3) to establish the elemental-specific mass response of particles. Based on the spICP-TOF-MS technique, both MCFP number concentrations and dissolved metal concentrations in all samples were simultaneously analyzed. Transport efficiency was calculated using 50 nm Au-NPs as a reference and was 38%. Recovery was calculated based on particle number concentration via standard addition, with triplicate measurements showing 73 ± 5% (mean ± SD) for 50 nm Au-NPs. All PM2.5 samples were analyzed in a single analytical batch to eliminate interbatch variability. Randomly selected samples were measured in triplicate to evaluate reproducibility. Additional details on sample pretreatment, instrument operating conditions, detection limits, and measurement reproducibility are provided in Text S2 and Tables S1–S3.

2.4. Cell Exposure

BEAS-2B human bronchial epithelial cells were selected for PM2.5 cytotoxicity assessment due to their physiological relevance and primary role in PM exposure. − BEAS-2B cells, obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China), were maintained in a humidified 5% CO2 incubator at 37 °C and cultured in DMEM supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. After 24 h of incubation, lung cells were first exposed to PM2.5 samples at varying concentrations (0, 30, 300, 600, and 1200 μg/mL). Concentration–response relationships were established through curve fitting, with toxic potency quantified using two established metrics: EC1.5 (concentration inducing 1.5-fold ROS induction relative to control) and IC20 (concentration causing 20% cell viability reduction). Detailed methodological specifications are provided in Text S3.

2.5. Machine Learning Model for MCFP Source Apportionment

A machine learning-based source discrimination model was developed using Python’s Scikit-learn package, employing the Random Forest algorithm as the primary classifier. The model was trained on high-resolution single-particle signatures acquired through spICP-TOF-MS, incorporating comprehensive elemental composition data and quantitative metal mass fractions in the individual MCFPs. The training data set comprised 12,171 individual MCFPs derived from five distinct emission sources, with source-specific representation maintained through balanced sampling (dust: 2349 particles; coal combustion: 2476 particles; diesel vehicle emissions: 2372 particles; gasoline/electric vehicle emissions: 2,526 particles; biomass combustion: 2448 particles). The data set was partitioned using an 80:20 split for model training and validation, respectively. Model optimization involved the iterative adjustment of detection thresholds for source-specific elemental combinations. Validation comprised: (1) confusion matrix evaluation using held-out test particles and (2 linear regression analysis between predicted versus actual source contributions in blended samples. More details about the model are provided in Text S4.

2.6. Statistical Analysis

Statistical analyses were performed using SPSS 23.0 software (SPSS Inc., Chicago, IL, USA). A Spearman’s rank correlation test was conducted to assess the correlations between the high-resolution elemental signatures of MCFPs and cytotoxicity potencies. Key factors influencing cytotoxicity were identified through significance analysis using R (RandomForest package 4.7–1.1). Differences between PM2.5 pollution periods were analyzed using one-way ANOVA, with p < 0.05 considered statistically significant.

3. Results and Discussion

During the observation period, PM2.5 concentrations exhibited substantial variability, ranging from 8.9 to 91.4 μg/m3 with a mean concentration of 34.3 μg/m3. This average concentration approximates China’s Class I 24-h air quality standard (35 μg/m3) but substantially exceeds the WHO’s revised 24-h guideline value (15 μg/m3). Distinct seasonal patterns were observed: winter concentrations ranged from 8.9 to 70.3 μg/m3, while spring levels were elevated to 13.2–91.4 μg/m3. The PM2.5/PM10 ratio demonstrated significant seasonal variation (p < 0.05), with winter values (0.35–0.95) consistently higher than spring measurements (0.22–0.87; Figure S3b). This ratio pattern reflects the characteristic pollution regimes of each season, with winter dominated by fine-particle-rich haze events and spring influenced by coarse-particle-dominated dust episodes, consistent with previous regional studies. ,

3.1. Dominant MCFPs and Their Distributions

Single-particle analysis using spICP-TOF-MS identified 32 distinct MCFP types in PM2.5 (Figures S4 and S5). The five most abundant MCFP species, comprising Al-, Si-, Fe-, Mn-, and Pb-containing FPs, collectively represented approximately 80% of the total MCFP number. Figure shows the distributions of these five MCFP number concentrations across different PM2.5 levels in both seasons. Quantitative analysis revealed that Al-containing FPs exhibited number concentrations ranging from 9.2 × 105 to 1.5 × 108 particles/m3, while Si-containing FPs ranged from 1.0 × 106 to 4.8 × 108 particles/m3. The concentration ranges for other dominant MCFPs were: 5.2 × 105–5.8 × 107 particles/m3 for Fe-containing FPs, 9.9 × 105–5.5 × 107 particles/m3 for Mn-containing FPs, and 5.5 × 105–4.3 × 107 particles/m3 for Pb-containing FPs. These measurements demonstrate substantial variability in the MCFP abundance across different pollution conditions and particle types.

1.

1

Open in a new tab

(a) Distributions of the number concentrations of Al-, Si-, Fe-, Mn-, and Pb-containing FPs during the observation. At each coordinate point, the outer circle represents the total number concentration of MCFPs, while the inner circle indicates the number concentration of mmFPs. Both the color and size of the circles reflect these number concentrations. Comparison of average number concentrations of Al- (b), Si- (c), Fe- (d), Mn- (e), and Pb-(f) containing FPs (both smFPs and mmFPs) at different periods in both seasons. (Note: the relative standard deviation for the triplicate MCFP number concentrations ranged from 1.9 to 23.4%).

As shown in Figure b–f, the number concentrations of these five MCFPs were generally higher in spring than in winter, particularly for Si-containing FPs (p < 0.01). In winter, the number concentrations of these MCFPs were highest on dust days, followed by haze, GB, and WHO days. For instance, Si-containing FPs reached their maximum concentrations on winter dust days, with values ranging from 2.5 × 107 to 4.7 × 107 particles/m3 with an average of 3.6 × 107 particles/m3. These concentrations were 1.7, 2.6, and 3.0 times higher than those recorded on haze, GB, and WHO days, respectively. A similar distribution pattern was evident in spring, where the number concentrations of these five MCFPs were also highest on dust days and lowest on WHO days. Specifically, the number concentrations of Si-containing FPs on spring dust days ranged from 1.7 × 108 to 4.8 × 108 particles/m3 with an average of 3.7 × 108 particles/m3, while on WHO days, they ranged from 1.1 × 107 to 3.8 × 107 particles/m3 with an average of 2.5 × 107 particles/m3 (Figure c).

According to spICP-TOF-MS analysis, MCFPs can be categorized into two distinct classes: (1) single-metal FPs (smFPs) containing only one detectable metal above the instrument’s particle detection limit (Table S2), and (2) multimetal FPs (mmFPs). For the five MCFPs, Si-smFPs accounted for 58–96% of Si-containing FPs in number, with an average of 85%. Al-smFPs accounted for 29–63% (average of 48%) of Al-containing FPs. Whereas Fe-smFPs, Mn-smFPs, and Pb-smFPs accounted for 13–52% (average: 28%), 15–54% (average: 31%), and 6–47% (average: 21%) of their respective total FPs (Figures and S6). Seasonal analysis revealed that the contributions of smFPs to the five MCFPs were generally higher in the winter than in the spring, with the exception of Fe-smFPs (Figure S7). Specifically, the average contributions of Fe-mmFPs to Fe-containing FPs in winter were 75% (range: 48–87%), significantly higher than the 68% (49–84%; p < 0.05) observed in spring. Notably, as PM2.5 pollution levels increased, a greater proportion of MCFPs were detected as mmFPs. For instance, the contribution of mmFPs to Fe-containing FPs increased progressively from 57% on WHO days to 74% on GB days, 81% on haze days, and 84% on dust days during the winter. In spring, mmFP contributions increased from 67% on WHO days to 76% on dust days (Figure S6h). The contributions of mmFPs to Pb-containing FPs were from 58% on WHO days to 83% on dust days in winter and from 78% on WHO days to 89% on dust days in spring (Figure S6j). These findings suggest that individual MCFPs exhibit more complex elemental compositions during polluted periods compared to clean days.

3.2. Dominant Elemental Composition in Individual mmFPs

Based on quantitative single-particle analysis, metal-rich FPs were classified according to their dominant metal composition, defined as the metal species exhibiting the highest mass fraction (typically >0.4). Metals with mass fractions <0.1 were assumed to be in a surface-adsorbed state or incorporated in FPs with trace amounts, while those ranging between 0.1 and 0.4 were classified as particle-associated impurities. Figures and S8 present the mass fraction distributions of Al, Si, Fe, Mn, and Pb in their respective mmFPs across different PM2.5 pollution levels. The analysis revealed distinct compositional patterns. Al-mmFPs were predominantly characterized by Al mass fractions exceeding 0.1, with a notable seasonal variation. During winter, Al-rich-FPs (Al mass fraction > 0.4) dominated, while 0.1–0.4 fractions dominated in spring (Figure a). Si- and Fe-mmFPs were primarily characterized by mass fractions > 0.4, followed by intermediate (0.1–0.4) and trace (<0.1) levels (Figure b,c). Mn- and Pb-mmFPs showed contrasting patterns, with most particles containing these metals at trace levels (<0.1), followed by intermediate fractions (0.1–0.4), while a minor proportion exhibited mass fractions > 0.4 (Figure d,e).

2.

2

Open in a new tab

Number proportions of Al- (a), Si- (b), Fe- (c), Mn- (d), and Pb-mmFPs (e) base on the corresponding metal mass fraction in individual mmFPs. “>0.4”, “0.1–0.4”, and “<0.1” indicate the mass fractions of a metal in individual mmFPs (e.g., for Fe-mmFPs on winter haze days, 62% had a mass fraction of Fe > 0.4, 33% had 0.1–0.4, and 5% had <0.1). (f) Number proportions of Al-, Si-, Fe-, and other metal-rich FPs among all mmFPs during different periods.

Seasonal analysis revealed significant variations in the metal mass fractions within individual mmFPs. Specifically, the number contributions of Al- and Fe-rich FPs (mass fractions > 0.4) among Al- and Fe-mmFPs were significantly higher in winter than in spring (p < 0.05; Figure S9). Conversely, Si-rich FPs exhibited an inverse seasonal pattern, with greater contributions during spring (p < 0.01), potentially reflecting the influx of mineral particles, such as aluminosilicates predominantly composed of Si, associated with the intrusion of dust-laden air masses during spring , For Mn and Pb, the proportion of particles with mass fractions <0.1 was higher in spring than in winter (p < 0.01; Figure S9).

Regarding the distribution of metal mass fractions in individual mmFPs across varying PM2.5 levels, the number contributions of Fe-rich FPs to Fe-mmFPs showed a decreasing trend from WHO and GB days to haze and dust days, and this trend was also shown for Si-rich FPs in the winter. In contrast, Mn- and Pb-mm FPs had more particles with Mn and Pb mass fractions <0.1, as the PM2.5 pollution level increased. For instance, during winter, the average number proportions of Fe-rich FPs within Fe-mmFPs were 85% (range: 76–88%) on WHO days, 75% (69–88%) on GB days, 62% (52–73%) on haze days, and decreased to 42% (40–44%) on dust days (Figure S8c); Whereas Pb-mmFPs with Pb mass fraction <0.1 contributed an average of 67% (52–78%) of the total Pb-mmFPs on WHO days, 73% (52–86%) on GB days, 84% (75–91%) on haze days, and reached 91% (90–94%) on dust days, by number (Figure S8e).

Based on spICP-TOF-MS analysis, Al-, Si-, and Fe-rich FPs, particles with Al, Si, and Fe mass fractions exceeding 0.4 within individual mmFPs, were identified as the predominant matrices for mmFPs. 63–93% of mmFPs were composed of Al-, Si-, and Fe-rich FPs during the observation period, with an average contribution of 86% (Figure S10). The remaining mmFPs included Mn-, Ti-, Sn-, Mg-, Cr-, Ba-, and Pb-rich FPs. Specifically, as depicted in Figure f, in winter, Al-rich FPs averagely constituted 29, 36, 48, and 57% of all mmFPs on WHO, GB, haze, and dust days, respectively; while these contributions decreased to 13, 15, 16, and 19% in spring. Conversely, these values for Si-rich FPs were only 12, 9, 8, and 10% in winter but increased significantly to 44, 56, 59, and 60% in spring. Fe-rich FPs accounted for 28, 28, 26, and 20% in winter and 30, 19, 14, and 13% in spring. Notably, the proportions of Al- and Si-rich FPs within all mmFPs exhibited an increasing trend with rising PM2.5 levels, whereas the proportion of Fe-rich FPs displayed a decreasing trend. This observation suggests a shift in the particulate matrix composition across different PM2.5 pollution periods. Furthermore, spICP-TOF-MS analysis can provide approximate size estimations of metal-containing particles based on their dominant elemental mass, albeit with inherent uncertainty due to particle compositional complexity. Despite these limitations, the method enabled a comparative assessment of size variations across different PM2.5 levels. Our results revealed a consistent trend: particle sizes increased with pollution levels, following the sequence WHO days < GB days < haze days < dust days (Table S4). Moreover, Fe-rich FPs showed smaller particle sizes compared to Al- and Si-rich FPs. For example, during winter, the sizes of Fe-rich FPs ranged from 67 to 421 nm (average: 112 nm) on WHO days, while they were in the range of 68–443 nm (average: 147 nm) on dust days. This trend was also verified through TEM analysis (Figure S11). As for Al- and Si-rich FPs, their sizes were 62–634 nm (average: 186 nm), 291–966 nm (average: 353 nm) on winter WHO days, and 62–685 nm (average: 237 nm), 292–935 nm (average: 435 nm) on winter dust days, respectively. These size differences indicate that Si/Al-rich FPs in PM2.5, typically characterized by larger diameters, are preferentially removed from the atmosphere through wet and dry deposition processes during precipitation or high-wind events. , In contrast, smaller particulate fractions, e.g., Fe-rich FPs, demonstrate greater atmospheric persistence due to their reduced deposition velocities. This size-dependent screening pattern accounts for the relative enrichment of Fe-rich particles in PM2.5 during clean periods and their consequently extended atmospheric residence times.

3.3. Toxic Metals Associated with mmFPs

Fifteen toxic metals, including V, Cr, Mn, Co, Ni, Cu, Zn, As, Se, Mo, Cd, Sn, Sb, Tl, and Pb, were identified to be associated with Al-, Si-, and Fe-rich FPs, which serve as the dominant matrices of mmFPs across different periods (Figure S12). Among these, Mn and Pb were the most abundant toxic metals in mmFPs, followed by Sn, Cr, and Zn. Si-rich FPs associated with toxic metals exhibited the highest concentrations, particularly in spring, surpassing those of Fe- and Al-rich FPs (Figure a–c). There was, in general, no significant difference in the number concentrations of Al- and Fe-rich FPs associated with toxic metals in different seasons. However, the number concentrations of Al-rich FPs associated with toxic metals were significantly higher on haze days in winter than in spring. Conversely, Fe-rich FPs demonstrated higher concentrations on WHO and dust days in spring than in winter. Si-rich FPs associated with toxic metals consistently exhibited higher concentrations in spring than in winter (p < 0.01; Figure S13). Furthermore, the number concentrations of Al-, Si-, and Fe-rich FPs associated with toxic metals increased with rising PM2.5 levels, peaking during dust periods. For instance, the number concentrations of Fe-rich FPs associated with toxic metals in winter increased from 2.7 × 105–1.2 × 106 particles/m3 on WHO days, to 9.5 × 105–3.7 × 106 particles/m3 on GB days, 2.5 × 106–3.9 × 106 particles/m3 on haze days, and reached 3.0 × 106–6.0 × 106 particles/m3 on dust days. In spring, these values were 1.3 × 106–5.0 × 106 particles/m3 on WHO days and increased to 1.9 × 106–1.9 × 107 particles/m3 on dust days. Figure d shows a typical magnetite aggregate, with identifying characteristic d-spacings of 4.81, 2.93, and 2.51 Å, corresponding to the (1 1 1), (2 2 0), and (3 1 1) lattice spacings of magnetite, respectively. These Fe-rich FPs were observed to be associated with toxic metals, including Mn, Cu, and Zn.

3.

3

Open in a new tab

Comparison of number concentrations (yellow boxes) of Al- (a), Si- (b), and Fe-rich FPs (c) associated with toxic metals, along with their number proportions relative to the total number of their mmFPs (blue boxes) at different periods in both seasons. (d) Typical TEM image of a magnetite aggregate in PM2.5 collected on a winter WHO day, showing associations with Al, Si, Ti, Mn, Cu and Zn.

Interestingly, the number contribution of Al-, Si-, and Fe-rich FPs associated with toxic metals to their respective total mmFPs exhibited distribution patterns distinct from those based on number concentrations. Fe-rich FPs associated with toxic metals displayed the highest number proportions in Fe-mmFPs, followed by Al- and Si-rich FPs, with all three showing significantly higher proportions in winter than in spring (p < 0.01; Figure S14). In addition, the number proportions of Al-, Si-, and Fe-rich FPs associated with toxic metals showed a decreasing trend with elevated PM2.5 levels, particularly pronounced for Fe-rich FPs. The number proportions of Fe-rich FPs associated with toxic metals were 65–76% on WHO days, 56–78% on GB days, 50–73% on haze days, and decreased to 24–26% on dust days during winter. In spring, these proportions were 68–75% on WHO days and decreased to 14–45% on dust days (Figure c). When considering the mass fraction of toxic metals in Al-, Si-, and Fe-rich FPs across seasons, only Fe-rich FPs demonstrated statistically significant differences, exhibiting 1.5 times higher values in winter compared to spring (p < 0.01; Figure S15). Moreover, at varying PM2.5 levels, the mass fractions of toxic metals in Al-, Si-, and Fe-rich FPs followed a similar trend, peaking on WHO days, decreasing with rising PM2.5 levels, and reaching their lowest values on dust days (Figure S16). A larger proportion of mmFPs, particularly Fe-rich FPs, were associated with toxic metals in PM2.5 collected on clean days compared to polluted days, suggesting a potentially higher toxic risk at lower PM2.5 levels, at the single particle level.

3.4. Unequal Toxic Potencies of PM2.5 during Different Periods

In this study, BEAS-2B cells were utilized to assess the lung cytotoxic potency of PM2.5 collected during different periods. EC1.5 (effective concentration causing a 1.5-fold increase in oxidative stress) and IC20 (inhibitory concentration reducing cell viability by 20%) values were employed to evaluate the toxic potencies of PM2.5. Lower EC1.5 and IC20 values indicate a higher cytotoxic potential per unit mass concentration. Seasonal analysis revealed that EC1.5 and IC20 values for PM2.5 collected in the winter were 1.6 and 1.4 times lower than those in spring, suggesting that winter PM2.5 exhibits greater lung cellular toxicity.

Oxidative potential of PM2.5 collected on winter WHO days was the highest, which was 8.1 times greater than the lowest one on dust days in spring. Specifically, in winter, the EC1.5 values for PM2.5 were 185 ± 10 μg/mL on WHO days, 295 ± 53 μg/mL on GB days, 300 ± 49 μg/mL on haze days, and 1130 ± 160 μg/mL on dust days. In the spring, significant differences in EC1.5 values were observed across PM2.5 levels, with values of 281 ± 23 μg/mL on WHO days, 576 ± 63 μg/mL on GB days, 785 ± 129 μg/mL on haze days, and 1505 ± 166 μg/mL on dust days (Figure a). These findings indicate that the oxidative stress potential of PM2.5 tends to increase as PM2.5 pollution levels decrease.

4.

4

Open in a new tab

Comparison of EC1.5 (a) and IC20 (b) of PM2.5 from various pollution periods in both seasons, and results of variable importance analysis of EC1.5 (c) and IC20 (d) using random forests model (*, significant; ns: nonsignificant; % Inc.MSE: the increment of mean square error). Metal-rich FPs and metal-smFPs represent their number proportions in the metal-containing FPs (e.g., Fe-rich FPs represent the number proportions of Fe-rich FPs within Fe-containing FPs). TMs associated with metal-rich FPs represent the number proportions of metal-rich FPs associated with toxic metals in their respective mmFPs. Metal-dissolved, OC and EC represent their concentrations (μg/g) in PM2.5, respectively.

For IC20, the lowest value was also found on winter WHO days, which was 6.3 times lower than that on spring dust days. In detail, PM2.5 collected during winter exhibited lower IC20 values on WHO days (88 ± 11 μg/mL) compared to GB days (116 ± 10 μg/mL), with both 3–4 times lower than those on polluted days (350 ± 42 μg/mL on haze days and 358 ± 39 μg/mL on dust days). Similarly, in spring, IC20 values on clean days were 4–5 times lower than those on polluted days, with values of 115 ± 20 μg/mL on WHO days, 129 ± 25 μg/mL on GB days, 511 ± 55 μg/mL on haze days, and 553 ± 127 μg/mL on dust days (Figure b). The consistently lower IC20 values for PM2.5 collected on clean days, compared to polluted periods, indicate a higher cytotoxic effect. These pronounced differences in toxic potencies across different periods likely stem from variations in the sources and chemical compositions of PM2.5. ,,

3.5. Key Toxic Components of PM2.5

To identify the key components of PM2.5 influencing lung cytotoxic potency (as indicated by EC1.5 and IC20 values), correlation analysis and variable importance analysis were conducted. A total of 112 columns of data were analyzed, with a focus on the high-resolution elemental signatures of MCFPs. The analysis also incorporated concentrations of EC, OC, and dissolved metals in PM2.5 (Figures S17 and S18). Notably, Fe-rich FPs, toxic metal-rich FPs (e.g., Mn- and Pb-rich FPs), smFPs (e.g., Mn-, Pb-, and Ba-smFPs), toxic metals associated with Fe-rich FPs, dissolved metals (e.g., Cr, Fe, and Sn), and OC exhibited significantly negative correlations with both EC1.5 and IC20 (p < 0.05; Figures S19 and S20). Variable importance analysis further highlighted that Fe-rich FPs and toxic metals associated with Fe-rich FPs ranked as the top 2 primary factors influencing PM2.5-induced oxidative stress, underscoring the critical role of Fe-rich FPs in modulating oxidative stress across different periods (Figure c). These Fe-rich FPs may release iron ions intracellularly, contributing to oxidative stress through catalysis or direct participation in the Fenton reaction. − In addition, Ba-smFPs, OC, dissolved Cr, and Zn-, Mn-, and Cu-smFPs may also play a significant role in regulating PM2.5-derived oxidative stress. As shown in Figure d, IC20 values were predominantly influenced by toxic metals associated with Fe-rich FPs, suggesting that Fe-rich FPs act as key carriers for toxic metals, potentially driving lung cell apoptosis upon PM2.5 exposure. Fe2+/Fe3+ in FPs drives Fenton or Fenton-like reactions, while co-associated redox-active metals (e.g., Mn2+, Cr3+) may further amplify ROS generation via electron transfer, creating a synergistic oxidative stress. Furthermore, Mn- and Pb-dominant FPs, along with dissolved Cr, significantly contributed to the PM2.5-induced cytotoxicity. OC also played a crucial role in the toxic potency of PM2.5, although its importance ranking was relatively low. OC has a complex composition, with some components potentially playing a more significant role in the cytotoxicity. For example, as evidenced by previous studies, PAHs in PM2.5 substantially influence the short-term toxicity of PM2.5 on the human respiratory tract. ,

3.6. Source Apportionment of MCFPs across Different Periods

High-resolution elemental signatures of individual MCFPs from natural sources (represented by desert dust) and anthropogenic sources (including coal combustion, diesel vehicle emissions, gasoline and electric vehicle emissions, and biomass combustion) were collected to conduct atmospheric MCFP source apportionment. A random forest classifier was employed to develop the traceability model, which was validated using both a test data set and mechanically mixed samples. During training, recognition thresholds for source-discriminative elemental combinations were iteratively optimized through automated tuning to maximize the classification accuracy. MCFP types not observed in the training data set were classified as “uncertain”. The confusion matrix for the test data set demonstrated that the model achieved an accuracy of nearly 70% in its predictions (Figure S21a). Additionally, 40 mechanically mixed samples, simulating actual PM2.5 compositions from the five sources, were prepared for validation. The model-predicted source contributions of MCFPs showed strong linear correlations with the expected contributions (R 2 = 0.65, p < 0.0001; Figure S21b), indicating the model’s capability to provide reliable preliminary insights into MCFP source contributions.

Using this validated model, source apportionment analysis was conducted on MCFPs in PM2.5 samples collected during the observation period. As illustrated in Figure , significant variations were observed in the contributions of anthropogenic and natural sources to MCFPs across different periods. Overall, anthropogenic sourcesincluding coal combustion, traffic emissions, and biomass combustioncontributed more substantially to MCFPs in PM2.5 during the winter than in the spring (p < 0.01; Figure S22a), particularly during haze days in both seasons (Figure c,g). In winter, anthropogenic sources accounted for an average of 70% of MCFPs in PM2.5, with the highest contribution of 78% (range: 66–87%) on WHO days, followed by 76% (61–82%) on haze days, 71% (65–76%) on GB days, and 52% (50–57%) on dust days (Figure a–d). On clean days, MCFPs were predominantly derived from anthropogenic sources, such as coal combustion, diesel vehicle emissions, and gasoline/electric vehicle emissions. Winter haze events, often occurring under conditions of high RH and low WS, are exacerbated by increased local emissions, , leading to elevated contributions from anthropogenic sources, particularly traffic emissions and coal combustion. Conversely, the reduced anthropogenic contributions during dust periods are largely due to the influx of external air masses, which disperse local particulate matter while introducing substantial dust particles, − as corroborated by backward trajectory analysis (Figure S23d).

5.

5

Open in a new tab

Contributions of MCFPs from natural and anthropogenic sources on WHO (a), GB (b), haze (c), and dust days (d) in winter (a–d) and spring (e–h). Anthropogenic sources include coal combustion, diesel vehicle emissions, gasoline/electric vehicle emissions, and biomass combustion.

In spring, the contributions of anthropogenic sources to MCFPs in PM2.5 exhibited a declining trend with increasing pollution levels, averaging 72% (range: 67–76%) on WHO days, 56% (44–61%) on GB days, 53% (47–58%) on haze days, and decreasing to 38% (34–43%) on dust days (Figure e–h). Notably, the contribution of naturally occurring MCFPs from sand dust increased significantly on dust days, averaging 60% (Figure h). This shift is attributed to the transport of dust-laden air masses from frequent dust storms in northern China during spring (Figure S23h).

Source apportionment was further conducted for Fe-rich FPs, which were identified as one of the most significant toxic components in PM2.5 in this study. Based on the results of the random forest-based machine learning model for Fe-rich FPs, contributions from anthropogenic sourcesparticularly traffic emissions and coal combustionwere generally higher in the winter than in the spring (Figure S22b). These contributions were relatively high on clean days compared to polluted days. On average, anthropogenic activities accounted for over 80% of Fe-rich FPs during all periods in winter and more than 71% in spring, even on dust days (Figure S24). Among these anthropogenic sources, traffic emissions and coal combustion contributed largely to the Fe-rich FPs, especially during winter WHO days (48% for traffic and 30% for coal combustion, respectively). Given that traffic emissions dominate Fe-rich FP production on clean days (Figure S24), targeted interventions (e.g., nonexhaust PM control from brakes/tires) should be prioritized alongside traditional combustion sources.

4. Environmental and Health Implications

Recognized as a major environmental health hazard, PM2.5-associated health risks have been extensively documented in epidemiological and toxicological studies. − In this study, the sampling site was strategically located in a heavily urbanized zone of Shanghai, a premier Chinese megacity, where anthropogenic activities exert exceptionally strong influences. This representative location provides an ideal setting for investigating urban particulate pollution.

This study, using advanced spICP-TOF-MS analysis, elucidated the elemental signatures of individual MCFPs across varying PM2.5 concentrations during winter and spring seasons, providing crucial insights into the particulate components responsible for PM2.5 toxicity. Although MCFP number concentrations were higher in spring than in winter and increased with PM2.5 pollution levels, our findings reveal that PM2.5 collected during clean days demonstrated greater pulmonary cytotoxic potencies compared to polluted periods. The counterintuitive observation can be attributed to the distinct elemental composition at the single-particle level. Among all characterized particulate properties, including OC, EC, dissolved metals, and single-particle metal composition, Fe-rich particles emerged as a critically important contributor to ROS generation, while simultaneously serving as effective carriers for other toxic metals, thereby exacerbating cellular damage. While spICP-TOF-MS enabled high-throughput elemental characterization, its inherent resolution limits particle size determination. Future investigations should incorporate complementary analytical approaches (e.g., TEM-EDX) to better resolve size-morphology-toxicity relationships. Despite these constraints, our findings, for the first time, elucidate that even when PM2.5 mass concentrations comply with WHO guidelines, chronic exposure to low concentrations of MCFPsparticularly Fe-rich particlesmay still pose significant health risks due to their prolonged biological persistence in critical organs. While further studies across diverse urban settings are needed to confirm these findings, this study highlights the importance of component-specific PM2.5 monitoring, particularly for Fe-rich particles in heavily industrialized areas. Given the evidence linking Fe-rich FPs to various diseases, ,, further research into the health risks posed by different airborne Fe-rich FPs is warranted.

Machine learning-driven source apportionment analysis revealed distinct anthropogenic contributions to MCFPs, with higher proportions observed during clean periods compared to polluted days. Notably, Fe-rich particles demonstrated substantial anthropogenic origins, particularly from traffic-related emissions and coal combustion processes. It is notable that the source apportionment in this study only focused on widely recognized anthropogenic emission sources (e.g., coal combustion and vehicle emissions). Future studies incorporating additional sources, such as steel manufacturing and shipping, are urgently needed. Nevertheless, the data clearly demonstrate that PM2.5 toxicity is primarily determined by emission sources rather than concentration levels alone. , This paradigm shift necessitates a strategic reorientation of air quality policies from mass-based controls to source-specific mitigation strategies. Furthermore, the implementation of advanced analytical techniques for high-resolution characterization of source-specific PM composition, particularly at the single-particle level, coupled with systematic toxicological assessment, should be prioritized to develop effective prevention measures against PM-induced health risk.

Supplementary Material

eh5c00157_si_001.pdf (2.6MB, pdf)

Acknowledgments

This study was funded by the National Key R&D Program of China (2023YFC3708301) and the National Natural Science Foundation of China (No.42125102). Additional funding for this work was provided by the Fundamental Research Funds for the Central Universities, and East China Normal University (ECNU) Multifunctional Platform for Innovation (004).

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/envhealth.5c00157.

  • Specific sampling information on PM2.5 and source samples; Supplement information on spICP-TOF-MS conditions; Cytotoxicity test experiment; Supplement information on machine learning model; PM2.5 concentrations and related parameters during the observation; Specific data of other MCFPs; Toxic metal mass fractions in Al-, Si-, and Fe-rich FPs; OC, EC, and dissolved metal concentrations; Correlation analysis between cytotoxicity and high-resolution elemental signatures of MCFPs, OC, EC, and dissolved metal concentrations; Source apportionment results of Fe-rich FPs; Particle size distributions of Al-, Si-, and Fe-rich FPs (PDF)

The authors declare no competing financial interest.

References

  1. Li X., Jin L., Kan H.. Air pollution: a global problem needs local fixes. Nature. 2019;570(7762):437–439. doi: 10.1038/d41586-019-01960-7. [DOI] [PubMed] [Google Scholar]
  2. Cohen A. J., Brauer M., Burnett R., Anderson H. R., Frostad J., Estep K., Balakrishnan K., Brunekreef B., Dandona L., Dandona R., Feigin V., Freedman G., Hubbell B., Jobling A., Kan H., Knibbs L., Liu Y., Martin R., Morawska L., Pope C. A.. et al. Estimates and 25-year trends of the global burden of disease attributable to ambient air pollution: an analysis of data from the Global Burden of Diseases Study 2015. Lancet. 2017;389(10082):1907–1918. doi: 10.1016/S0140-6736(17)30505-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Hill W., Lim E. L., Weeden C. E., Lee C., Augustine M., Chen K., Kuan F.-C., Marongiu F., Evans E. J., Moore D. A., Rodrigues F. S., Pich O., Bakker B., Cha H., Myers R., van Maldegem F., Boumelha J., Veeriah S., Rowan A., Naceur-Lombardelli C.. et al. Lung adenocarcinoma promotion by air pollutants. Nature. 2023;616(7955):159–167. doi: 10.1038/s41586-023-05874-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Burnett R., Chen H., Szyszkowicz M., Fann N., Hubbell B., Pope C. A., Apte J. S., Brauer M., Cohen A., Weichenthal S., Coggins J., Di Q., Brunekreef B., Frostad J., Lim S. S., Kan H., Walker K. D., Thurston G. D., Hayes R. B., Lim C. C.. et al. Global estimates of mortality associated with long-term exposure to outdoor fine particulate matter. Proc. Natl. Acad. Sci. U. S. A. 2018;115(38):9592–9597. doi: 10.1073/pnas.1803222115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Liu J., Song R., Li X., Liu L., Wei N., Yuan J., Yi W., Pan R., Cheng J., Zhang X., Su H.. Effects of PM2.5 and Its Components on Disease Severity in Patients with Schizophrenia and the Mediating Role of Thyroid Hormones. Environ. Health. 2024;2(5):290–300. doi: 10.1021/envhealth.3c00194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. World Health Organization . Billions of people still breathe unhealthy air: new WHO data. 2022, https://www.who.int/news/item/04-04-2022-billions-of-people-still-breathe-unhealthy-air-new-who-data.
  7. Weichenthal S., Pinault L., Christidis T., Burnett R. T., Brook J. R., Chu Y., Crouse D. L., Erickson A. C., Hystad P., Li C., Martin R. V., Meng J., Pappin A. J., Tjepkema M., van Donkelaar A., Weagle C. L., Brauer M.. How low can you go? Air pollution affects mortality at very low levels. Sci. Adv. 2022;8(39):eabo3381. doi: 10.1126/sciadv.abo3381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Hu W., Wang Y., Wang T., Ji Q., Jia Q., Meng T., Ma S., Zhang Z., Li Y., Chen R., Dai Y., Luan Y., Sun Z., Leng S., Duan H., Zheng Y.. Ambient particulate matter compositions and increased oxidative stress: Exposure-response analysis among high-level exposed population. Environ. Int. 2021;147:106341. doi: 10.1016/j.envint.2020.106341. [DOI] [PubMed] [Google Scholar]
  9. Weichenthal S., Lavigne E., Traub A., Umbrio D., You H., Pollitt K., Shin T., Kulka R., Stieb Dave M., Korsiak J., Jessiman B., Brook Jeff R., Hatzopoulou M., Evans G., Burnett Richard T.. Association of Sulfur, Transition Metals, and the Oxidative Potential of Outdoor PM2.5 with Acute Cardiovascular Events: A Case-Crossover Study of Canadian Adults. Environ. Health Perspect. 2021;129(10):107005. doi: 10.1289/EHP9449. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chaudhary E., George F., Saji A., Dey S., Ghosh S., Thomas T., Kurpad A. V., Sharma S., Singh N., Agarwal S., Mehta U.. Cumulative effect of PM2.5 components is larger than the effect of PM2.5 mass on child health in India. Nat. Commun. 2023;14(1):6955. doi: 10.1038/s41467-023-42709-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Shen Y., Yu G., Liu C., Wang W., Kan H., Zhang J., Cai J.. Prenatal Exposure to PM2.5 and Its Specific Components and Risk of Hypertensive Disorders in Pregnancy: A Nationwide Cohort Study in China. Environ. Sci. Technol. 2022;56(16):11473–11481. doi: 10.1021/acs.est.2c01103. [DOI] [PubMed] [Google Scholar]
  12. Adams P. J., Seinfeld J. H., Koch D., Mickley L., Jacob D.. General circulation model assessment of direct radiative forcing by the sulfate-nitrate-ammonium-water inorganic aerosol system. J. Geophys. Res. Atmos. 2001;106(D1):1097–1111. doi: 10.1029/2000JD900512. [DOI] [Google Scholar]
  13. Wang Y., Salana S., Yu H., Puthussery J. V., Verma V.. On the Relative Contribution of Iron and Organic Compounds, and Their Interaction in Cellular Oxidative Potential of Ambient PM2.5 . Environ. Sci. Technol. Lett. 2022;9(8):680–686. doi: 10.1021/acs.estlett.2c00316. [DOI] [Google Scholar]
  14. Liu S., Tian H., Bai X., Zhu C., Wu B., Luo L., Hao Y., Liu W., Lin S., Zhao S., Wang K., Liu K., Gao J., Zhang Q., Zhang K., Kan H., Liu Y., Hao J.. Significant but Spatiotemporal-Heterogeneous Health Risks Caused by Airborne Exposure to Multiple Toxic Trace Elements in China. Environ. Sci. Technol. 2021;55(19):12818–12830. doi: 10.1021/acs.est.1c01775. [DOI] [PubMed] [Google Scholar]
  15. Kelly F. J., Fussell J. C.. Size, source and chemical composition as determinants of toxicity attributable to ambient particulate matter. Atmos. Environ. 2012;60:504–526. doi: 10.1016/j.atmosenv.2012.06.039. [DOI] [Google Scholar]
  16. Wu D., Zheng H., Li Q., Wang S., Zhao B., Jin L., Lyu R., Li S., Liu Y., Chen X., Zhang F., Wu Q., Liu T., Jiang J., Wang L., Li X., Chen J., Hao J.. Achieving health-oriented air pollution control requires integrating unequal toxicities of industrial particles. Nat. Commun. 2023;14(1):6491. doi: 10.1038/s41467-023-42089-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Wang W., Lin Y., Yang H., Ling W., Liu L., Zhang W., Lu D., Liu Q., Jiang G.. Internal Exposure and Distribution of Airborne Fine Particles in the Human Body: Methodology, Current Understandings, and Research Needs. Environ. Sci. Technol. 2022;56(11):6857–6869. doi: 10.1021/acs.est.1c07051. [DOI] [PubMed] [Google Scholar]
  18. Hammond J., Maher B. A., Ahmed I. A. M., Allsop D.. Variation in the concentration and regional distribution of magnetic nanoparticles in human brains, with and without Alzheimer’s disease, from the UK. Sci. Rep. 2021;11(1):9363. doi: 10.1038/s41598-021-88725-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Calderón-Garcidueñas L., González-Maciel A., Reynoso-Robles R., Hammond J., Kulesza R., Lachmann I., Torres-Jardón R., Mukherjee P. S., Maher B. A.. Quadruple abnormal protein aggregates in brainstem pathology and exogenous metal-rich magnetic nanoparticles (and engineered Ti-rich nanorods). The substantia nigrae is a very early target in young urbanites and the gastrointestinal tract a key brainstem portal. Environ. Res. 2020;191:110139. doi: 10.1016/j.envres.2020.110139. [DOI] [PubMed] [Google Scholar]
  20. Wei W., Yang B., Zhu X., Liu X., Song E., Song Y.. Silica Nanoparticle Exposure Caused Brain Lesion and Underlying Toxicological Mechanism: Route-Dependent Bio-Corona Formation and GSK3β Phosphorylation Status. Environ. Health. 2024;2(2):76–84. doi: 10.1021/envhealth.3c00119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Calderón-Garcidueñas L., González-Maciel A., Mukherjee P. S., Reynoso-Robles R., Pérez-Guillé B., Gayosso-Chávez C., Torres-Jardón R., Cross J. V., Ahmed I. A. M., Karloukovski V. V., Maher B. A.. Combustion- and friction-derived magnetic air pollution nanoparticles in human hearts. Environ. Res. 2019;176:108567. doi: 10.1016/j.envres.2019.108567. [DOI] [PubMed] [Google Scholar]
  22. Maher B. A., González-Maciel A., Reynoso-Robles R., Torres-Jardón R., Calderón-Garcidueñas L.. Iron-rich air pollution nanoparticles: An unrecognised environmental risk factor for myocardial mitochondrial dysfunction and cardiac oxidative stress. Environ. Res. 2020;188:109816. doi: 10.1016/j.envres.2020.109816. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Maher B. A., Ahmed I. A. M., Karloukovski V., MacLaren D. A., Foulds P. G., Allsop D., Mann D. M. A., Torres-Jardon R., Calderon-Garciduenas L.. Magnetite pollution nanoparticles in the human brain. Proc. Natl. Acad. Sci. U. S. A. 2016;113(39):10797–10801. doi: 10.1073/pnas.1605941113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. He X., Zhao Q., Chai X., Song Y., Li X., Lu X., Li S., Chen X., Yuan Y., Cai Z., Qi Z.. Contribution and Effects of PM2.5-Bound Lead to the Cardiovascular Risk of Workers in a Non-Ferrous Metal Smelting Area Considering Chemical Speciation and Bioavailability. Environ. Sci. Technol. 2023;57(4):1743–1754. doi: 10.1021/acs.est.2c07476. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Qi Y., Wei S., Xin T., Huang C., Pu Y., Ma J., Zhang C., Liu Y., Lynch I., Liu S.. Passage of exogeneous fine particles from the lung into the brain in humans and animals. Proc. Natl. Acad. Sci. U. S. A. 2022;119(26):e2117083119. doi: 10.1073/pnas.2117083119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Hochella M. F., Mogk D. W., Ranville J., Allen I. C., Luther G. W., Marr L. C., McGrail B. P., Murayama M., Qafoku N. P., Rosso K. M., Sahai N., Schroeder P. A., Vikesland P., Westerhoff P., Yang Y.. Natural, incidental, and engineered nanomaterials and their impacts on the Earth system. Science. 2019;363(6434):eaau8299. doi: 10.1126/science.aau8299. [DOI] [PubMed] [Google Scholar]
  27. Bland G. D., Battifarano M., Liu Q., Yang X., Lu D., Jiang G., Lowry G. V.. Single-Particle Metal Fingerprint Analysis and Machine Learning Pipeline for Source Apportionment of Metal-Containing Fine Particles in Air. Environ. Sci. Technol. Lett. 2023;10(11):1023–1029. doi: 10.1021/acs.estlett.2c00835. [DOI] [Google Scholar]
  28. Oroumiyeh F., Jerrett M., Del Rosario I., Lipsitt J., Liu J., Paulson S. E., Ritz B., Schauer J. J., Shafer M. M., Shen J., Weichenthal S., Banerjee S., Zhu Y.. Elemental composition of fine and coarse particles across the greater Los Angeles area: Spatial variation and contributing sources. Environ. Pollut. 2022;292:118356. doi: 10.1016/j.envpol.2021.118356. [DOI] [PubMed] [Google Scholar]
  29. Liu S., Wu T., Wang Q., Zhang Y., Tian J., Ran W., Cao J.. High time-resolution source apportionment and health risk assessment for PM2.5-bound elements at an industrial city in northwest China. Sci. Total Environ. 2023;870:161907. doi: 10.1016/j.scitotenv.2023.161907. [DOI] [PubMed] [Google Scholar]
  30. Wu J., Tou F., Guo X., Liu C., Sun Y., Xu M., Liu M., Yang Y.. Vast emission of Fe- and Ti-containing nanoparticles from representative coal-fired power plants in China and environmental implications. Sci. Total Environ. 2022;838:156070. doi: 10.1016/j.scitotenv.2022.156070. [DOI] [PubMed] [Google Scholar]
  31. Xu M., Niu Z., Liu C., Yan J., Peng B., Yang Y.. Oxidative Potential of Metal-Containing Nanoparticles in Coal Fly Ash Generated from Coal-Fired Power Plants in China. Environ. Health. 2023;1(3):180–190. doi: 10.1021/envhealth.3c00040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mehrabi K., Günther D., Gundlach-Graham A.. Single-particle ICP-TOFMS with online microdroplet calibration for the simultaneous quantification of diverse nanoparticles in complex matrices. Environ. Sci-Nano. 2019;6(11):3349–3358. doi: 10.1039/C9EN00620F. [DOI] [Google Scholar]
  33. Hendriks L., Gundlach-Graham A., Hattendorf B., Günther D.. Characterization of a new ICP-TOFMS instrument with continuous and discrete introduction of solutions. J. Anal. At. Spectrom. 2017;32(3):548–561. doi: 10.1039/C6JA00400H. [DOI] [Google Scholar]
  34. Praetorius A., Gundlach-Graham A., Goldberg E., Fabienke W., Navratilova J., Gondikas A., Kaegi R., Günther D., Hofmann T., von der Kammer F.. Single-particle multi-element fingerprinting (spMEF) using inductively-coupled plasma time-of-flight mass spectrometry (ICP-TOFMS) to identify engineered nanoparticles against the elevated natural background in soils. Environ. Sci-Nano. 2017;4(2):307–314. doi: 10.1039/C6EN00455E. [DOI] [Google Scholar]
  35. Shi Z., Xu M., Wu L., Peng B., Yang X., Zhang Y., Li S., Niu Z., Zhao H., Ma X., Yang Y.. Size-Dependent Elemental Composition in Individual Magnetite Nanoparticles Generated from Coal-Fired Power Plant Regulating Their Pulmonary Cytotoxicity. Environ. Sci. Technol. 2024;58(44):19774–19784. doi: 10.1021/acs.est.4c05570. [DOI] [PubMed] [Google Scholar]
  36. Zhang Q., Chen L., Zhao H., Qin J., Zhang L., Yang H., Liu L., Fu S., Maher B. A., Liu Q., Jiang G.. Deposition of Air Pollution-Derived Magnetic Nanoparticles in Human Kidney Revealed by High-Resolution Microstructural Characterization. Environ. Sci. Technol. 2025;59(13):6745–6756. doi: 10.1021/acs.est.4c13858. [DOI] [PubMed] [Google Scholar]
  37. Xu M., Niu Z., Shi Z., Zhang Y., Meng M., Yang X., Wang M., Ma X., Zhao H., Yang Y.. High-Resolution Characterization of Coal Combustion-Derived Metal-Containing Nanoparticles and Their Health-Related Implications. Environ. Sci. Technol. Lett. 2024;11(6):611–618. doi: 10.1021/acs.estlett.4c00292. [DOI] [Google Scholar]
  38. Tou F., Niu Z., Fu J., Wu J., Liu M., Yang Y.. Simple Method for the Extraction and Determination of Ti-, Zn-, Ag-, and Au-Containing Nanoparticles in Sediments Using Single-Particle Inductively Coupled Plasma Mass Spectrometry. Environ. Sci. Technol. 2021;55(15):10354–10364. doi: 10.1021/acs.est.1c00983. [DOI] [PubMed] [Google Scholar]
  39. Shamsollahi H. R., Yunesian M., Kharrazi S., Jahanbin B., Nazmara S., Rafieian S., Dehghani M. H.. Characterization of persistent materials of deposited PM2.5 in the human lung. Chemosphere. 2022;301:134774. doi: 10.1016/j.chemosphere.2022.134774. [DOI] [PubMed] [Google Scholar]
  40. Jin L., Xie J., Wong C. K. C., Chan S. K. Y., Abbaszade G., Schnelle-Kreis J., Zimmermann R., Li J., Zhang G., Fu P., Li X.. Contributions of City-Specific Fine Particulate Matter (PM2.5) to Differential In Vitro Oxidative Stress and Toxicity Implications between Beijing and Guangzhou of China. Environ. Sci. Technol. 2019;53(5):2881–2891. doi: 10.1021/acs.est.9b00449. [DOI] [PubMed] [Google Scholar]
  41. Raudoniute J., Stasiulaitiene I., Kulvinskiene I., Bagdonas E., Garbaras A., Krugly E., Martuzevicius D., Bironaite D., Aldonyte R.. Pro-inflammatory effects of extracted urban fine particulate matter on human bronchial epithelial cells BEAS-2B. Environ. Sci. Pollut. Res. 2018;25(32):32277–32291. doi: 10.1007/s11356-018-3167-8. [DOI] [PubMed] [Google Scholar]
  42. Chen X., Wu D., Tan Y., Song X., Chen J., Li Q.. Absence of a Causal Link between Elemental Carbon Exposure and Short-Term Respiratory Toxicity in Human-Derived Organoids and Cellular Models. Environ. Sci. Technol. 2025;59(1):668–678. doi: 10.1021/acs.est.4c11256. [DOI] [PubMed] [Google Scholar]
  43. Fan H., Zhao C., Yang Y., Yang X.. Spatio-Temporal Variations of the PM2.5/PM10 Ratios and Its Application to Air Pollution Type Classification in China. Front. Environ. Sci. 2021;9:692440. doi: 10.3389/fenvs.2021.692440. [DOI] [Google Scholar]
  44. Lu D., Li H., Tian M., Wang G., Qin X., Zhao N., Huo J., Yang F., Lin Y., Chen J., Fu Q., Duan Y., Dong X., Deng C., Abdullaev S. F., Huang K.. Secondary aerosol formation during a special dust transport event: impacts from unusually enhanced ozone and dust backflows over the ocean. Atmos. Chem. Phys. 2023;23(21):13853–13868. doi: 10.5194/acp-23-13853-2023. [DOI] [Google Scholar]
  45. Zhang R., Jing J., Tao J., Hsu S. C., Wang G., Cao J., Lee C. S. L., Zhu L., Chen Z., Zhao Y., Shen Z.. Chemical characterization and source apportionment of PM2.5 in Beijing: seasonal perspective. Atmos. Chem. Phys. 2013;13(14):7053–7074. doi: 10.5194/acp-13-7053-2013. [DOI] [Google Scholar]
  46. Cwiertny, D. M. ; Baltrusaitis, J. ; Hunter, G. J. ; Laskin, A. ; Scherer, M. M. ; Grassian, V. H. . Characterization and acid-mobilization study of iron-containing mineral dust source materials. J. Geophys. Res. Atmos. 2008, 113, D5 DOI: 10.1029/2007JD009332. [DOI] [Google Scholar]
  47. Michalik J. M., Wilczyńska-Michalik W., Gondek L̷., Tokarz W., Żukrowski J., Gajewska M., Michalik M.. Magnetic fraction of the atmospheric dust in Kraków–physicochemical characteristics and possible environmental impact. Atmos. Chem. Phys. 2023;23(2):1449–1464. doi: 10.5194/acp-23-1449-2023. [DOI] [Google Scholar]
  48. Ito A., Adebiyi A. A., Huang Y., Kok J. F.. Less atmospheric radiative heating by dust due to the synergy of coarser size and aspherical shape. Atmos. Chem. Phys. 2021;21(22):16869–16891. doi: 10.5194/acp-21-16869-2021. [DOI] [Google Scholar]
  49. Zhang Q., Lu D., Wang D., Yang X., Zuo P., Yang H., Fu Q., Liu Q., Jiang G.. Separation and Tracing of Anthropogenic Magnetite Nanoparticles in the Urban Atmosphere. Environ. Sci. Technol. 2020;54(15):9274–9284. doi: 10.1021/acs.est.0c01841. [DOI] [PubMed] [Google Scholar]
  50. Bai X., Tian H., Zhu C., Luo L., Hao Y., Liu S., Guo Z., Lv Y., Chen D., Chu B., Wang S., Hao J.. Present Knowledge and Future Perspectives of Atmospheric Emission Inventories of Toxic Trace Elements: A Critical Review. Environ. Sci. Technol. 2023;57(4):1551–1567. doi: 10.1021/acs.est.2c07147. [DOI] [PubMed] [Google Scholar]
  51. Chen X., Wu D., Zheng L., Chen Y., Cheng A., Li Q.. Variable Valence State of Trace Elements Regulating Toxic Potencies of Inorganic Particulate Matter. Environ. Sci. Technol. Lett. 2024;11(3):223–229. doi: 10.1021/acs.estlett.4c00019. [DOI] [Google Scholar]
  52. Borgie M., Ledoux F., Verdin A., Cazier F., Greige H., Shirali P., Courcot D., Dagher Z.. Genotoxic and epigenotoxic effects of fine particulate matter from rural and urban sites in Lebanon on human bronchial epithelial cells. Environ. Res. 2015;136:352–362. doi: 10.1016/j.envres.2014.10.010. [DOI] [PubMed] [Google Scholar]
  53. Valko M., Morris H., Cronin T. D. M.. Metals, Toxicity and Oxidative Stress. Curr. Med. Chem. 2005;12(10):1161–1208. doi: 10.2174/0929867053764635. [DOI] [PubMed] [Google Scholar]
  54. Eteshola E. O. U., Haupt D. A., Koos S. I., Siemer L. A., Morris D. L.. The role of metal ion binding in the antioxidant mechanisms of reduced and oxidized glutathione in metal-mediated oxidative DNA damage. Metallomics. 2020;12(1):79–91. doi: 10.1039/c9mt00231f. [DOI] [PubMed] [Google Scholar]
  55. Torti S. V., Torti F. M.. Winning the war with iron. Nat. Nanotechnol. 2019;14(6):499–500. doi: 10.1038/s41565-019-0419-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Wu D., Zheng H. T., Li Q., Jin L., Lyu R., Ding X., Huo Y. Q., Zhao B., Jiang J. K., Chen J. M., Li X. D., Wang S. X.. Toxic potency-adjusted control of air pollution for solid fuel combustion. Nat. Energy. 2022;7(2):194–202. doi: 10.1038/s41560-021-00951-1. [DOI] [Google Scholar]
  57. Zhang S., Xing J., Sarwar G., Ge Y., He H., Duan F., Zhao Y., He K., Zhu L., Chu B.. Parameterization of heterogeneous reaction of SO2 to sulfate on dust with coexistence of NH3 and NO2 under different humidity conditions. Atmos. Environ. 2019;208:133–140. doi: 10.1016/j.atmosenv.2019.04.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Anastasopolos A. T., Hopke P. K., Sofowote U. M., Zhang J. J. Y., Johnson M.. Local and regional sources of urban ambient PM2.5 exposures in Calgary, Canada. Atmos. Environ. 2022;290:119383. doi: 10.1016/j.atmosenv.2022.119383. [DOI] [Google Scholar]
  59. Huang, K. ; Zhuang, G. ; Li, J. ; Wang, Q. ; Sun, Y. ; Lin, Y. ; Fu, J. S. . Mixing of Asian dust with pollution aerosol and the transformation of aerosol components during the dust storm over China in spring 2007. J. Geophys. Res.: Atmos. 2010, 115, D7 DOI: 10.1029/2009JD013145. [DOI] [Google Scholar]
  60. Liu J., Ding J., Rexiding M., Li X., Zhang J., Ran S., Bao Q., Ge X.. Characteristics of dust aerosols and identification of dust sources in Xinjiang, China. Atmos. Environ. 2021;262:118651. doi: 10.1016/j.atmosenv.2021.118651. [DOI] [Google Scholar]
  61. Wang N., Zheng P., Wang R., Wei B., An Z., Li M., Xie J., Wang Z., Wang H., He M.. Homogeneous and heterogeneous atmospheric ozonolysis of acrylonitrile on the mineral dust aerosols surface. J. Environ. Chem. Eng. 2021;9(6):106654. doi: 10.1016/j.jece.2021.106654. [DOI] [Google Scholar]
  62. Chen Y., Ye X., Yao Y., Lv Z., Fu Z., Huang C., Wang R., Chen J.. Characteristics and sources of PM2.5-bound elements in Shanghai during autumn and winter of 2019: Insight into the development of pollution episodes. Sci. Total Environ. 2023;881:163432. doi: 10.1016/j.scitotenv.2023.163432. [DOI] [PubMed] [Google Scholar]
  63. Tran K. B., Lang J. J., Compton K., Xu R., Acheson A. R., Henrikson H. J., Kocarnik J. M., Penberthy L., Aali A., Abbas Q., Abbasi B., Abbasi-Kangevari M., Abbasi-Kangevari Z., Abbastabar H., Abdelmasseh M., Abd-Elsalam S., Abdelwahab A. A., Abdoli G., Abdulkadir H. A., Abedi A.. et al. The global burden of cancer attributable to risk factors, 2010–19: a systematic analysis for the Global Burden of Disease Study 2019. Lancet. 2022;400(10352):563–591. doi: 10.1016/S0140-6736(22)01438-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Shi L., Zhu Q., Wang Y., Hao H., Zhang H., Schwartz J., Amini H., van Donkelaar A., Martin R. V., Steenland K., Sarnat J. A., Caudle W. M., Ma T., Li H., Chang H. H., Liu J. Z., Wingo T., Mao X., Russell A. G., Weber R. J.. et al. Incident dementia and long-term exposure to constituents of fine particle air pollution: A national cohort study in the United States. Proc. Natl. Acad. Sci. U. S. A. 2023;120(1):e2211282119. doi: 10.1073/pnas.2211282119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Huang W., Xu H., Wu J., Ren M., Ke Y., Qiao J.. Toward cleaner air and better health: Current state, challenges, and priorities. Science. 2024;385(6707):386–390. doi: 10.1126/science.adp7832. [DOI] [PubMed] [Google Scholar]
  66. Liang F., Xiao Q., Huang K., Yang X., Liu F., Li J., Lu X., Liu Y., Gu D.. The 17-y spatiotemporal trend of PM2.5 and its mortality burden in China. Proc. Natl. Acad. Sci. U. S. A. 2020;117(41):25601–25608. doi: 10.1073/pnas.1919641117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Sun H., Chen X., Huang W., Wei J., Yang X., Shan A., Zhang L., Zhang H., He J., Pan C., Li J., Wu J., Wang T., Chen J., Guo Y., Tong S., Dong G., Tang N.-J.. Association between Long-Term Exposure to PM2.5 Inorganic Chemical Compositions and Cardiopulmonary Mortality: A 22-Year Cohort Study in Northern China. Environ. Health. 2024;2(8):530–540. doi: 10.1021/envhealth.4c00020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Yu H. Y., Xu L. T., Cui T. L., Wang Y., Wang B. Q., Zhang Z., Su R. J., Zhang J. X., Zhang R., Wei Y. H., Li D. C., Jin X. T., Chen W., Zheng Y. X.. The Foam Cell Formation Associated With Imbalanced Cholesterol Homeostasis Due to Airborne Magnetite Nanoparticles Exposure. Toxicol. Sci. 2022;189(2):287–300. doi: 10.1093/toxsci/kfac079. [DOI] [PubMed] [Google Scholar]
  69. Du P., Du H., Zhang W., Lu K., Zhang C., Ban J., Wang Y., Liu T., Hu J., Li T.. Unequal Health Risks and Attributable Mortality Burden of Source-Specific PM2.5 in China. Environ. Sci. Technol. 2024;58(25):10897–10909. doi: 10.1021/acs.est.3c08789. [DOI] [PubMed] [Google Scholar]
  70. Park M., Joo H. S., Lee K., Jang M., Kim S. D., Kim I., Borlaza L. J. S., Lim H., Shin H., Chung K. H., Choi Y.-H., Park S. G., Bae M.-S., Lee J., Song H., Park K.. Differential toxicities of fine particulate matters from various sources. Sci. Rep. 2018;8(1):17007. doi: 10.1038/s41598-018-35398-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

eh5c00157_si_001.pdf (2.6MB, pdf)

Articles from Environment & Health are provided here courtesy of American Chemical Society

RESOURCES