Refined source apportionment of coal combustion sources by using single particle mass spectrometry

Insights into the seasonal characteristics of single particle aerosols in Chengdu based on SPAMS

To investigate the seasonal characteristics in air pollution in Chengdu, a single particle aerosol mass spectrometry was used to continuously observe atmospheric fine particulate matter during one-month periods in summer and winter, respectively. The results showed that, apart from O3, the concentrations of other pollutants (CO, NO2, SO2, PM2.5 and PM10) were significantly higher in winter than in summer. All single particle aerosols were divided into seven categories: biomass burning (BB), coal combustion (CC), Dust, vehicle emission (VE), K mixed with nitrate (K-NO3), K mixed with sulfate and nitrate (K-SN), and K mixed with sulfate (K-SO4) particles. The highest contributions in both seasons were VE particles (24%). The higher contributions of K-SO4 (16%) and K-NO3 (10%) particles occurred in summer and winter, respectively, as a result of their different formation mechanisms. S-containing (K-SO4 and K-SN), VE, and BB particles caused the evolution of pollution in both seasons, and they can be considered as targets for future pollution reduction. The mixing of primary sources particles (VE, Dust, CC, and BB) with secondary components was stronger in winter than in summer. In summer, as pollution worsens, the mixing of primary sources particles with 62 [NO3]- weakened, but the mixing with 97 [HSO4]- increased. However, in winter, the mixing state of particles did not exhibit an obvious evolution rules. The potential source areas in summer were mainly distributed in the southern region of Sichuan, while in winter, besides the southern region, the contribution of the western region cannot be ignored.

Unequal toxic effects of size-segregated single particles emitted from typical industrial plants, vehicles, and road dust

The health risks of particulate matters (PMs) associated with their chemical components and sizes have attracted increasing attention. However, the toxic effect of critical toxic components in size-segregated PMs from specific emission source remains unclear. We present the toxicity of size-segregated elements in PMs via integrating toxic analysis and online single-particle measurements of real-world industrial plants, vehicles, and road dust. The number fractions of elemental carbon (EC)- and Fe-containing particles were 5-11 and 3-12 folds greater than those of other metal-containing particles, respectively. A unimodal distribution with the peak at 0.4 µm was observed for the toxic metals emitted from industrial plants and road dust, while the distribution was relatively flat for vehicles. When integrating the abundance with toxicity of metals, especially Mn, Cu, V, and Fe, the peak for PM toxicity occurred at 0.4 µm for road dust, 0.4-0.7 µm for industrial plants, and 0.8 µm for vehicle-emitted PM. The inhalation risk in the alveolar region increased for these source-emitted PMs due to the efficient deposition of toxic PMs within 0.4-0.8 µm. These results reveal the complex coupling of health risks and size distributions of PMs, and further highlight that the health-oriented control of air pollution should consider PM1.

Mapping the constituent preference of tree species for capturing particulate matter on leaf surfaces using single-particle mass spectrometry and supervised machine learning

Respiratory health is negatively influenced by the dimensions and constituents of particulate matter (PM). Although mass concentration is widely acknowledged to be key to assessing dust retention by urban trees, the role of plant leaves in filtering PM from the urban atmosphere, particularly regarding the particle dimensions and chemical constituents of retained PM on the leaf, remains elusive. Here we combined single-particle aerosol mass spectrometry and a particle resuspension chamber to investigate how urban tree species capture PM constituents. Results indicate that leaves are efficient in capturing relatively larger particles (1.0-2.0 μm). Compositionally, airborne particles were mostly composed of elemental carbon (EC, 20%), organic carbon (OC, 17%), and secondary reaction products (13%). However, leaf surfaces revealed a preference for retaining crustal species, comprising 55% of captured particulates. Notably, specific tree species demonstrated varied affinities for different PM constituents: Osmanthus fragrans Lour. predominantly captured levoglucosan (LEV), indicative of its efficiency against biomass burning particles, whereas Cinnamomum camphora (L.) J.Presl and Sabina chinensis var. kaizuca W.C.Cheng & W.T.Wang were more effective in capturing heavy metals (HMs). XGBoost modelling identified indicator ions, e.g., CN-, NO3-, NO2-, PO3-, with SHAP values surpassing 0.035, suggesting a preferential adsorption of these ions among different tree species. These findings demonstrate that the particulate capture efficiency of urban tree species varies with species-specific leaf properties, particularly in their ability to selectively adsorb particles containing hazardous constituents such as LEV and HMs. This study provides a scientific basis for the strategic selection of tree species in urban forestry initiatives aimed at improving air quality and public health.

Characterizing leaf-deposited particles: Single-particle mass spectral analysis and comparison with naturally fallen particles

The size and composition of particulate matter (PM) are pivotal in determining its adverse health effects. It is important to understand PM's retention by plants to facilitate its atmospheric removal. However, the distinctions between the size and composition of naturally fallen PM (NFPM) and leaf-deposited PM (LDPM) are not well-documented. Here we utilize a single-particle aerosol mass spectrometer, coupled with a PM resuspension chamber, to analyze these differences. We find that LDPM particles are 6.8-97.3 % larger than NFPM. Employing a neural network algorithm based on adaptive resonance theory, we have identified distinct compositional profiles: NFPM predominantly consists of organic carbon (OC; 31.2 %) and potassium-rich components (19.1 %), whereas LDPM are largely composed of crustal species (53.9-60.6 %). Interestingly, coniferous species retain higher OC content (11.5-13.7 %) compared to broad-leaved species (0.5-1.2 %), while the levoglucosan content exhibit an opposite trend. Our study highlights the active role of tree leaves in modifying PM composition beyond mere passive capture, advocating for a strategic approach to species selection in urban greening initiatives to enhance PM mitigation. These insights provide guidance for urban planners and environmentalists in implementing nature-based solutions to improve urban air quality.

Single particle mass spectral signatures from on-road and non-road vehicle exhaust particles and their application in refined source apportionment using deep learning

With advances in vehicle emission control technology, updating source profiles to meet the current requirements of source apportionment has become increasingly crucial. In this study, on-road and non-road vehicle particles were collected, and then the chemical compositions of individual particles were analyzed using single particle aerosol mass spectrometry. The data were grouped using an adaptive resonance theory neural network to identify signatures and establish a mass spectral database of mobile sources. In addition, a deep learning-based model (DeepAerosolClassifier) for classifying aerosol particles was established. The objective of this model was to accomplish source apportionment. During the training process, the model achieved an accuracy of 98.49 % for the validation set and an accuracy of 93.36 % for the testing set. Regarding the model interpretation, ideal spectra were generated using the model, verifying its accurate recognition of the characteristic patterns in the mass spectra. In a practical application, the model performed hourly source apportionment at three specific field monitoring sites. The effectiveness of the model in field measurement was validated by combining traffic flow and spatial information with the model results. Compared with other machine learning methods, our model achieved highly automated source apportionment while eliminating the need for feature selection, and it enables end-to-end operation. Thus, in the future, it can be applied in refined and online source apportionment of particulate matter.

Brake wear-derived particles: Single-particle mass spectral signatures and real-world emissions

Brake wear is an important but unregulated vehicle-related source of atmospheric particulate matter (PM). The single-particle spectral fingerprints of brake wear particles (BWPs) provide essential information for understanding their formation mechanism and atmospheric contributions. Herein, we obtained the single-particle mass spectra of BWPs by combining a brake dynamometer with an online single particle aerosol mass spectrometer and quantified real-world BWP emissions through a tunnel observation in Tianjin, China. The pure BWPs mainly include three distinct types of particles, namely, Ba-containing particles, mineral particles, and carbon-containing particles, accounting for 44.2%, 43.4%, and 10.3% of the total BWP number concentration, respectively. The diversified mass spectra indicate complex BWP formation pathways, such as mechanical, phase transition, and chemical processes. Notably, the mass spectra of Ba-containing particles are unique, which allows them to serve as an excellent indicator for estimating ambient BWP concentrations. By evaluating this indicator, we find that approximately 4.0% of the PM in the tunnel could be attributable to brake wear; the real-world fleet-average emission factor of 0.28 mg km^-1 veh^-1 is consistent with the estimation obtained using the receptor model. The results presented herein can be used to inform assessments of the environmental and health impacts of BWPs to formulate effective emissions control policies.

Heterogeneous changes of chemical compositions, sources and health risks of PM2.5 with the "Clean Heating" policy at urban/suburban/industrial sites

China has enacted the "Clean Heating" (CH) policy in north China. The domain-specific impacts on PM2.5 constituents and sources in small cities are still lacking, which obstruct the further policy optimization. Here, we performed an intensive observation covering the heating period (HP) and pre-heating period (PHP) in winter of 2017 at urban (UR), industrial (IS), and suburban (SUR) sites in one of the "2 + 26" cities. The mean PM2.5 concentrations at UR and IS decreased by 15.2 % and 4.6 %, while increased by 9.8 % at SUR in the HP compared with the PHP, indicating the heterogeneous responses. The lowest contribution percentages of coal combustion (14.6 %) and industrial emissions (17.1 %) to PM2.5 at UR in the HP implied the CH policy played more effective role. The most increase in NO3-/SO42- ratio by 26.8 % and the highest NO3- concentration at UR in the HP were linked mainly with the thermal-NOx emitted from natural gas (NG) burning in view of NOx emission reductions from other sources. The highest concentrations of OC, SO42-, K+, and Cl-, and contribution percentages of biomass burning (20.0 %) and coal combustion (24.8 %) to PM2.5 at SUR in the HP evidenced the enhanced usage of biomass/coal. Coal banning in the HP at IS and UR led to the obvious decreases in OC, SO42-, As, and Sb. Secondary nitrate became the largest PM2.5 source at IS and UR in the HP. Coal banning, emission control on large-size enterprises and ignored control on small-size enterprises efficiently modified the concentrations and health risks of heavy metals. The lowest carcinogenic risks moved from SUR in the PHP to UR in the HP. The policies on de-NOx of NG-burning related enterprises, reduction of biomass/coal usage in suburban area, and strict regulation of small-size enterprises were urgently need to further improve the air quality.

A clear north-to-south spatial gradience of chloride in marine aerosol in Chinese seas under the influence of East Asian Winter Monsoon

Particulate chloride is a major component of sea salt particles and plays a key role in atmospheric chemistry. Anthropogenic pollutants over the northeastern Asia can be transported to the adjacent seas through the northwest monsoon, which profoundly influences the chloride chemistry over the seas. In this study, spatial distribution of particulate chloride and its sources over the Chinese seas were investigated based on shipboard particle samplings especially online Single Particle Aerosol Mass Spectrometer (SPAMS) over Bohai Sea, North Yellow Sea, and South Yellow Sea (SYS) during a cruise in November 2012. A strong north-to-south (N-S) gradience in marine aerosol composition was found. The Cl^-/Na⁺ ratios in PM_2.5 and single particle composition by SPAMS indicated remarkable chloride enrichment in marine aerosol in the north (Bohai Sea), while depletion in southern SYS. The results of size distribution showed that particulate chloride had higher concentration in coarse particles, while the Cl^-/Na⁺ ratio was much higher in submicron particles. In the north (38-40°N), biomass burning, carbonaceous, and Pb-rich type particles had high fractions in all chloride-containing particles identified by SPAMS (on average 66%). Combining chemical composition with back trajectory, it was found that fine-mode chloride enrichment in the north was mainly due to anthropogenic emission especially coal combustion and biomass burning from northern China. However, the high fine-mode chloride depletion in the south (32-34°N) was probably due to acid replacement by sulfate in aged aerosol during atmospheric transport. Our new findings reveal that marine aerosol in Chinese seas would show a clear N-S pattern of more fresh and anthropogenic enriched particles in the north, but more aged aerosol in the south during the East Asia Winter Monsoon, which provides new insights for the quantitative assessment of anthropogenic impact on marine aerosol and future modeling study.

Self-feedback LSTM regression model for real-time particle source apportionment

Atmospheric particulate matter pollution has attracted much wider attention globally. In recent years, the development of atmospheric particle collection techniques has put forwards new demands on the real-time source apportionments techniques. Such demands are summarized, in this paper, as how to set up new restraints in apportionment and how to develop a non-linear regression model to process complicated circumstances, such as the existence of secondary source and similar source. In this study, we firstly analyze the possible and potential restraints in single particle source apportionment, then propose a novel three-step self-feedback long short-term memory (SF-LSTM) network for approximating the source contribution. The proposed deep learning neural network includes three modules, as generation, scoring and refining, and regeneration modules. Benefited from the scoring modules, SF-LSTM implants four loss functions representing four restraints to be followed in the apportionment, meanwhile, the regeneration module calculates the source contribution in a non-linear way. The results show that the model outperforms the conventional regression methods in the overall performance of the four evaluation indicators (residual sum of squares, stability, sparsity, negativity) for the restraints. Additionally, in short time-resolution analyzing, SF-LSTM provides better results under the restraint of stability.

Particle composition, sources and evolution during the COVID-19 lockdown period in Chengdu, southwest China: Insights from single particle aerosol mass spectrometer data

In order to investigate the effects of the Coronavirus Disease 2019 (COVID-19) lockdown on air quality in cities in southwest China, a single particle aerosol mass spectrometer (SPAMS) and other online equipments were used to measure the air pollution in Chengdu, one of the megacities in this area, before and during the lockdown period. It was found that the concentrations of fine particulate matter (PM2.5), nitric oxide (NO), nitrogen dioxide (NO2), sulfur dioxide (SO2) and carbon monoxide (CO) decreased by 38.6%, 77.5%, 47.0%, 35.1% and 14.1%, respectively, while the concentration of ozone (O3) increased by 57.5% from the time before to the time during lockdown. All particles collected during the study period could be divided into eight categories: biomass burning (BB), coal combustion (CC), vehicle emissions (VE), cooking emissions (CE), Dust, K-nitrate (K-NO3), K-sulfate (K-SO4) and K-sulfate-nitrate (K-SN) particles, and their contributions changed significantly after the beginning of lockdown. Compared to before lockdown, the contribution of VE particles experienced the largest reduction (by 14.9%), whereas the contributions of BB and CE particles increased by 7.0% and 7.3%, respectively, during the lockdown period. Regional transmission was critical for pollution formation before lockdown, whereas the pollution that occurred during the lockdown period was caused mainly by locally emitted particles (such as VE, CE and BB particles). Weighted potential source contribution function (WPSCF) analysis further verified and emphasized the difference in the contribution of regional transmission for pollution formation before and during lockdown. In addition, the potential source area and intensity of the particles emitted from different sources or formation mechanisms were quite different.

Investigating the effect of sources and meteorological conditions on wintertime haze formation in Northeast China: A case study in Harbin

Heavy haze pollution has occurred frequently in the past few years in Northeast China during winters, which was distinct from other regions in China because of the particular meteorological conditions. In this study, we analyzed the temporal variation, source appointment, and influencing factors of PM_2.5 from December 1, 2018 to February 28, 2019 in Harbin. The results showed obvious differences between the non-haze and haze periods. The source appointment based on a single-particle aerosol mass spectrometer showed that coal combustion, vehicle emissions, biomass burning, and secondary inorganic aerosols (SIAs) were the major contributors of PM_2.5. It is interesting that from the non-haze to the haze period, contributions of coal combustion and SIAs increased (from 20.2% to 27.3%, and from 17.3% to 18.9%, respectively) while other sources decreased or increased little. It indicated the primary pollutants from heating supply were the most important contributor to haze formation due to the low temperature. Furthermore, from levels I (0 < PM_2.5 ≤ 75 μg m^-3) to III (115 < PM_2.5 ≤ 150 μg m^-3), SIAs increased from 15.3% to 19.4% (increased 4.1%), while coal combustion from 23.7% to 27.1% and increased 3.4%. It implied clearly that SIAs played a comparable role in the early stage of the evolution of haze episode as that of coal combustion. Combining data on prevailing winds and results of potential source contribution function indicated that PM_2.5 during the haze period was primarily influenced by the air masses originating from the southwestern areas via regional transport. A positive correlation was observed between relative humidity (RH) and haze pollution when RH ≥ 60%, indicating that hygroscopic growth may be the principal factor promoting secondary formation. CAPSULE: Coal combustion was the most important source in Harbin due to the low temperature, and secondary aerosols promoted the early stage of the haze evolution.

Chemical characteristics and source apportionment of PM2.5 using PMF modelling coupled with 1-hr resolution online air pollutant dataset for Linfen, China

The chemical species in PM_2.5 and air pollutant concentration data with 1-hr resolution were monitored synchronously between 15 November 2018 and 20 January 2019 in Linfen, China, which were analysed for multiple temporal patterns, and PM_2.5 source apportionment using positive matrix factorisation (PMF) modelling coupled with online chemical species data was conducted to obtain the apportionment results of distinct temporal patterns. The mean concentration of PM_2.5 was 124 μg/m³ during the heating period, and NO₃^- and organic carbon (OC) were the dominant species. The concentrations and percentages of NO₃^-, SO₄^2-, and OC increased notably during the growth periods of haze events, thereby indicating secondary particle formation. Six factors were identified by the PMF model during the heating period, including vehicular emissions (VE: 26.5%), secondary nitrate (SN: 16.5%), coal combustion and industrial emissions (CC&IE: 25.7%), secondary sulfate and secondary organic carbon (SS&SOC: 24.4%), biomass burning (BB: 1.0%), and crustal dust (CD: 5.9%). The primary sources of PM_2.5 on clean days were CD (33.3%), VE (23.1%), and SS&SOC (20.6%), while they were CC&IE (32.9%) and SS&SOC (28.3%) during the haze events. The contributions of secondary sources and CC&IE increased rapidly during the growth periods of haze events, while that of CD increased during the dissipation period. Diurnal variations in the contribution of secondary sources were mainly related to the accumulation and transformation of corresponding gaseous precursors. In comparison, contributions of CC&IE and VE varied as a function of the domestic heating load and peak levels occurred during the morning and evening rush hours. High contributions of major sources (CC&IE and SS&SOC) during haze events originated mainly from the north and south, while high contribution of a major source (CD) on clean days was from the northwest.

Source-oriented characterization of single particles from in-port ship emissions in Guangzhou, China

In this work, we analyzed freshly emitted particles from ship exhaust in the Guangzhou port region before and after the implementation of a clean fuel policy. We used a single particle aerosol mass spectrometer (SPAMS) to measure the changes in the chemical compositions of single particles and evaluate the role of V as a tracer for ship emissions. Particles from high sulfur fuel (S_F) oil (HS) combustion ships consisted of 54.8% elemental carbon-vanadium-sulfate (EC-V-S) and 25.0% vanadium-sulfate (V-S) particles, while particles from low S_F oil (LS) combustion ships were composed of 38.7% organic carbon-sulfate (OC-S) and 28.6% elemental and organic carbon (ECOC) particles. The sulfate-containing particles exhibited a moderate decrease from 95% in HS emissions to 78% in LS emissions, which still suggests the dominant role of sulfate in LS emissions after the implementation of a clean fuel policy. The V-containing particles showed a sharp decrease from 67% in HS emissions to 14% in LS emissions along with the decrease in the relative peak area (RPA) of V, suggesting a remarkable reduction in V in ship exhaust. The count of V-containing particles in urban Guangzhou in June 2017 was generally ten times lower than that in June 2016, which was in accordance with the sharp decrease in V-containing particles in LS emissions rather than in HS emissions. Despite the decrease in V in source-oriented ship emitted particles, the ubiquitous distribution of V in particles from lower S_F combustion ships suggests V is still effective as a tracer of ship emissions in port regions after the implementation of the clean fuel policy. Furthermore, the particles from LS emissions were investigated in comparison to those from gasoline vehicles (GV), diesel vehicles (DV) and coal combustion (CC) sources to better resolve ship-related particles in port regions.