Atmospheric aerosols modulate the radiative forcing of the Earth and adversely affect human health. Nitrogen (N) is a significant element that constitutes ambient aerosols and exists in the forms of inorganic nitrogen (IN) and organic nitrogen (ON). While IN aerosols such as nitrate and ammonium have been widely measured, the abundances and sources of ON aerosols are much less evaluated due to the lack of a suitable method for determining ON quantity over a long term. Recently, we have developed an aerosol IN&ON analyzer that uses programmed thermo-evolution and chemiluminescent detection coupled with multivariate curve resolution data treatment to achieve sensitive and simultaneous quantification of aerosol IN and ON without any pretreatment. The method breakthrough enables quantitative and accurate surveying of aerosol ON in a large number of samples with ease. In this work, we applied the new method and quantified ON in 609 aerosol filter samples collected from 12 sites of varying urban influence in China. It is the first time that the relative quantities of aerosol ON and IN and their typical varying ranges in urban and suburban atmospheres are established. This work helped to address the following questions: (1) how abundant of aerosol ON and its contribution to total nitrogen (TN) in the atmospheres of urban influences? (2) what proportion of nitrogenous organic aerosols (OA_N) account for in ambient organic aerosols (OA)? (3) what are the major sources of ON aerosols? The major findings of this research are as follows:

1. Revelation of the abundance of aerosol ON and its contribution to aerosol TN in southern and northern China.

Average aerosol ON concentrations ranged from 0.4 to 1.4 μg N m^-3 across 12 sampling sites in China, including two suburban sites in the North China Plain (NCP), three suburban/rural sites, and seven urban sites in the Pearl River Delta (PRD). The percentages of ON accounting for TN were mostly 10-35% at the sites, indicating ON represents a significant fraction of aerosol N and therefore atmospheric N deposition budget. We segregate the 12 sites into three groups – Suburban NCP, Suburban/rural PRD, and urban PRD – based on geographic locations and the degree of urban influences. Both IN and ON showed significantly higher levels in northern China than southern China (p<0.0001), while the ON-in-TN percentages were comparable in the two regions (p = 0.08) (Fig. 1). Between the urban and suburban/rural site groups in the PRD region, the average of ON was slightly higher in urban areas (p<0.001) while the IN abundances were comparable (p = 0.62). This makes the median of ON-to-TN ratio higher in the urban environment (22%) than in suburban/rural environment (17%) (p<0.001) (Fig. 1).

Fig. 1. Distribution histograms of aerosol IN and ON concentrations and ON-to-TN ratio in three groups of sampling sites in China. Statistical data (i.e., averages, medians, and standard deviations) are presented in each plot. The 12 sampling sites are segregated into three groups, namely urban PRD, suburban/rural PRD, and suburban NCP.

2. Estimation of OA_N contribution to OA

We take advantage of the jointly measured aerosol organic carbon (OC) and ON to estimate the contribution of OA_N to OA. Mathematically, the fraction of OA_N (i.e., the sum mass of all nitrogen-containing organic molecules) in OA can be expressed to be the products of three ratios, as shown in Eq. (1A) and a simplified version in Eq. (1B):

OA_N/OA=(OA_N/ON) × (ON/OC) × (OC/OA)         (1A)

OA_N/OA=A× B×C                 (1B)

Where A is the OA_N-to-ON ratio, B is the ON-to-OC ratio, and C is the reciprocal of OA-to-OC ratio, all in the unit of g/g. More specifically, A is the factor converting ON mass to mass of N-containing organic molecules, i.e, OA_N mass. At a molecular level, this information is exactly known for individual molecules of known molecular formula (e.g., 2.14 for urea, etc). Estimation of A for ambient OA can be made by assuming average molecular formulas for constituent groups of OA_N by surveying the literature. B is the N-to-C mass ratio of the bulk OA. This ratio has been measured for more than 600 ambient PM_2.5 aerosol samples collected at 12 sites in China as mentioned above. The reciprocal of C, better known as the OA-to-OC ratio in the literature, converts OC mass to OA mass.

Fig. 2. Estimating the contribution of nitrogenous molecules to organic aerosol (i.e., OA_N/OA) using three ratio quantities (OA_N/ON, ON/OC, and OA/OC). (A) The distribution of OA_N/ON is obtained by synthesizing available studies on OA_N constitutes and conducting Monte Carlo simulations. (B) The distribution of ON/OC is from measurements of 600+ aerosol samples, represented by the grey bars in the plot. (C)The OA/OC values are constrained in the range of 1.3-2.5 and assumed to follow a log-normal distribution. (D) The resulting OA_N/OA ratio is determined through Monte Carlo simulations of the constrained distributions of OA_N/ON, ON/OC, and OA/OC.

We have estimated the population of possible OA_N/OA values using Monte Carlo (MC) simulation. The population distribution of B (i.e., ON/OC) is well described by the 609 samples collected from 12 sites. It exhibits a lognormal distribution (Fig. 2B), with a population mean of 0.11, a standard deviation of 0.06, and a median of 0.10. The reciprocal of C (i.e., OA/OC) is constrained in the range of 1.3-2.5, based on the numerous past studies of organic aerosols and assumed to follow a log-normal distribution (Fig. 2C). We constrain its distribution domain of A by taking a bottom-up approach, i.e., the bulk OA_N/ON is calculated by summing up individual constituent groups of nitrogenous compounds, such as urea, amino acids, etc, for which the converting factors of ON-to-organic compound mass are known from their chemical formulas. Due to the complex mixture nature of OA_N, MC simulations were carried out to produce the distribution in Fig. 2A. The mean of A is estimated to be 7.16, with a 95% confidence interval of [6.45, 7.78]. Based on the MC simulation of the constrained distributions of A, B, and C, the resultant OA_N/OA ratio is estimated to have a mean of 42.2%, a median of 38.6%, and a 95% confidence interval of [14.1%, 87.3%] (Fig. 2D). We further conducted comprehensive sensitivity tests for the estimation. Finally, we estimate that, in the environments of urban influence, the mean of OA_N/OA is ~42%, likely in the range of 37–50%, while the 95% confidence interval of OA_N/OA would be [12%, 94%].

3. Source apportionment of ON by PMF receptor model

So far, we know little about the major sources of aerosol ON. In this study, we have gained for the first-time quantitative information on the sources of aerosol ON by positive matrix factorization (PMF) analysis of a set of samples collected at urban TW, suburban BJ, and rural NS sites (Fig. 3). About 70% of aerosol ON was derived from primary emissions, including biomass burning, primary biological particles, vehicular emissions, cooking, ship emissions, soil dust, and sea salt at the sites. Among them, biomass burning was a major contributor and accounted for 21-24% of aerosol ON masses (Fig. 3). Seasonally, the contribution of biomass burning was the highest during the winter and lowest in the summer. In comparison, the proportion of ON derived from primary biological particle emissions exhibits distinctly different seasonality, with the highest occurring in the summer at all the three sites. Secondary formation accounted for a larger proportion of ON during spring at TW and NS sites of southern China, while it was more prominent during summer at BJ site of northern China. These results provided quantitative understanding of aerosol ON sources, which bridged the knowledge gap regarding the geochemical processes of the organic part of N in aerosols.

Fig. 3. Annual and seasonal average contributions of major source groups to aerosol ON at three sites of varying urban influence in China. The three sites include an urban site in Hong Kong (TW), a rural site in south China (NS), and a suburban site in north China (BJ). The size of pie is proportional to aerosol ON mass concentration (μg N m^-3). Individual contributions are shown for three major source groups – biomass burning, secondary formation, and biological particle emission. A lumped source contribution is shown for other primary sources, which include industrial emission, fossil fuel combustions (coal combustion, vehicle emission, ship emission), soil dust, and sea salt.