The Significant Contribution of Polycyclic Aromatic Nitrogen Heterocycles to Light Absorption in the Winter North China Plain

Abstract

By quantifying the absorption of black carbon (BC), brown carbon (BrC) and the lensing effect, we found that BrC dominates the total absorption at 450 nm, and the largest absorption contribution proportion of BrC could reach 78.3% during heavy pollution. The average absorption enhancement (E_abs) at 530 nm was only 1.38, indicating that BC is not coated well here. The average value of the absorption Ångstrom exponent (AAE) between 450 nm and 530 nm was 5.3, suggesting a high concentration of BrC in Wangdu. CHN+ was the greatest contributor to the light absorption of molecules detected in MSOC with a proportion of 12.2–22.4%, in which the polycyclic aromatic nitrogen heterocycles (PANHs) were the dominant compounds. The C₆H₅NO₃ and its homologous series accounted for 3.0–11.3%, and the C₁₅H₉N and its homologous series, including one C₁₆H₁₁N and three C₁₇H₁₃N compounds, accounted for 5.1–12.3%. The absorption of these PANHs is comparable to that of nitro–aromatics, which should attract more attention to the impact of climate radiative forcing.

1. Introduction

Aerosols have not only become a major issue in the field of atmospheric science, such as the formation of atmospheric pollutants, human health and the impact on the ecological environment [1,2,3], but also because they change the reflectance of clouds by scattering and absorbing solar radiation, which has a great impact on the atmospheric radiation balance [4,5,6], and thus people have carried out more extensive research on its optical properties. Model and observational studies revealed that uncertainty of the radiative forcing of aerosols’ direct and indirect effects is the greatest uncertainty in climate prediction [7]. The impact of black carbon (BC) aerosol emissions on radiative forcing has been well established [7], but the absorption of brown carbon (BrC) remains poorly understood. If BC is coated by non-absorbing materials, then these coatings can enhance the magnitude of absorption via the well-known lensing effect [8,9,10]. The uncertainties of absorption caused by BrC and the lensing effect are the main limitation for estimating aerosol radiative forcing [11].

BC particles generated by the incomplete combustions of biomass, biofuels and fossil fuels may be coated by other aerosol species during aging [12]. Model studies indicate that the coatings on BC can double absorption via the lensing effect [13,14]. Although Mie calculations and lab studies support the model studies [15,16], observations of this effect in ambient air have yielded a wide range of results [17,18,19]. Cappa et al. [17] found that the absorption enhancement of BC due to an absorptive coating was 1.06, which is much lower than that from the core-shell mode. Zhang et al. [20,21] indicated that the absorption enhancement of mixing BC is decided by the BC’s position inside the coating and may not increase if the BC is near the coating boundary. Previous studies in Chinese cities showed that the E_abs values range from 2 to 3 in polluted environments [22,23,24,25]. Although some modified models can better simulate the results of some field observations [26], the absorption enhancement of BC remains one of the key limitations in assessing BC’s radiation effects due to its complicated mixing state.

BrC refers to complex mixtures of organic compounds that absorb light in the near-ultraviolet and visible light regions [27]. It is generally believed that biomass burning is a major source of BrC [28,29,30]. Fossil fuel-related sources of BrC also exist [31], and the concentrations of BrC depend on the fuel and burning conditions [32]. Atmospheric oxidation processes are the main driving reaction for the transformation of atmospheric chromophore [33]. Various approaches used to estimate the absorption contribution of BrC were reported, and the percentage of absorption contribution varies greatly [19,34,35,36,37]. Ultrahigh-resolution Fourier transform ion cyclotron resonance mass spectrometry and three-dimensional excitation emission matrix fluorescence spectroscopy were used to identify BrC chromophores [38,39,40]. The absorption and molecular characterization of BrC have mainly been studied in biomass burning-influenced areas or based on lab studies [11,41,42,43]. Several classes and subclasses of BrC chromophores have been identified [44,45,46,47,48], including nitro-aromatics, aromatic compounds with hydroxyl and carboxyl, polycyclic aromatic hydrocarbon (PAH) derivatives, nitrogen-containing heterocyclic compounds and sulfur-containing compounds. Nitro-aromatics have always been regarded as the most abundant BrC chromophores in OAs in many regions [49,50]. Due to the complicated composition of BrC, knowledge of the molecular properties of BrC and the relationship to its light absorption properties is still limited.

The primary goal of this study was to understand the relative importance of BC, BrC and the lensing effect by quantifying the percentage of real-time absorption contribution due to BC, BrC and the lensing effect. In addition, by combining high-performance liquid chromatography (HPLC) with a diode array detector (DAD) and quadrupole time-of-flight mass spectrometry (Q-ToF-MS), the key BrC chromophores were identified, and their relationship to its light absorption properties was discussed.

2. Materials and Methods

2.1. Field Site and Inlet System

Field observation was conducted in a field station at the Wangdu site (38.67° N, 115.25° E), which is located in a rural area surrounded by vast farms and scattered villages. The Wangdu site is ~90 km north of Shijiazhuang (the capital of Hebei province) and ~160 km south of Beijing. The observation position is shown in Figure S1. The sampling was carried out from 22 January 2019 to 3 February 2019. The inlet was installed on the roof of the field station 3 m above the ground. There are no main pollution sources nearby.

2.2. Real-Time Measurement Equipment

The meteorological parameters, including the temperature, relative humidity, wind speed, wind direction and air pressure, were obtained by an automatic weather station (Vaisala, HADMETTM Vantaa, Finland). A scanning mobility particle sizer (SMPS) consisting of a differential mobility analyzer (DMA, Model 3085, TSI, Auburn, IL, USA) and a condensation particle counter (CPC, Model 3776, TSI, Auburn, IL, USA) was used to measure the aerosol size distributions every 10 min. A time-of-flight aerosol chemical speciation monitor (Tof-ACSM, Aerodyne Inc., Billerica, MA, USA) was used to measure the mass concentration of water-soluble ions (Cl⁻, NO₃⁻, SO₄²⁻, NH₄⁺ and K⁺). Details about the ACSM can be found elsewhere [51]. The calculation of the (NH₄)₂SO₄ and NH₄NO₃ concentrations refers to the method of Gysel et al. [52]. The organic carbon (OC) and elemental carbon (EC) concentrations were measured by a semi-continuous carbon analyzer (Sunset Laboratory, Forest Grove, OR, USA). The particulate organic matter (POM) was estimated by multiplying the OC content by a factor of 1.6 [53].

Two cavity-attenuated phase shift spectroscopies (CAPS, Shoreline Science Research Inc., Tokyo, Japan) with light-emitting diode (LED) light sources were used to measure the aerosol extinction and scattering coefficients at wavelengths of 450 and 530 nm. Aerosol light scattering was measured at wavelengths of 450, 550 and 700 nm with an integrating nephelometer (Model 3563, TSI, Auburn, IL, USA). Separation of the volatile and refractory components of particles was achieved by a laboratory-made thermodenuder (TD) set at 250 °C. Detailed descriptions of the set-up and TD are given in Supporting Information and Figure S2.

2.3. Quantification of the Absorption Contribution of BC, BrC and the Lensing Effect

Aerosol absorption enhancement (E_abs) is defined as the ratio of the absorption coefficient of ambient aerosols at a specific wavelength (b_{abs-λ-unheated}) and the absorption coefficient at the same wavelength after the aerosols have been heated in a thermal denuder to remove volatilities (b_{abs-λ-heated}), as shown in Equation (1):

$$E_{abs\_\lambda }^{}=\frac{b_{abs,\lambda ,unheated}}{b_{abs,\lambda ,heated}}$$

(1)

Several approaches used to estimate the contribution to the total absorption from brown carbon and lensing were discussed by Pokhrel et al. [54]. Here, we assumed that the thermal denuder could remove the majority of non-BC components, and the E_abs for the lensing effect was constant at 450–530 nm. Since the TD cannot remove all non-BC components of a particle, these assumptions may lead to errors in E_abs, which was discussed by Lack et al. [55]. The aerosol light absorption due to BrC is calculated using the following equations:

$$b_{abs,\lambda ,BC}=b_{abs,\lambda ,heated}$$

(2)

$$b_{abs,450,Lens}=b_{abs,450,BC}*(E_{abs\_530}^{}-1)$$

(3)

$$b_{abs,450,BrC}=b_{abs,450,\mathrm{u}\mathrm{n}\mathrm{h}\mathrm{e}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}}-b_{abs,450,BC}-b_{abs,450,Lens}$$

(4)

where b_abs,λ,BC is the absorption coefficient of BC at a specific wavelength, b_abs,λ,Lens is the absorption coefficient of the lensing effect at a specific wavelength and b_abs,λ,BrC is the absorption coefficient of BrC at a specific wavelength.

2.4. Aerosol Sample Collection and BrC Analysis

A high-volume cyclone sampler (TH150-A, Wuhan Tianhong INST Group, Wuhan, China) was used for aerosol sample collection at a rate of 1.13 m³ min⁻¹, and all the PM_2.5 samples from 15 January 2019 to 2 March 2019 were collected on quartz fiber filters (180 mm × 230 mm), which were baked at 600 °C for 6 h before sampling. Each aerosol sample was collected for about 24 h. After sampling, the samples were stored in aluminum foil and refrigerated at −20 °C to avoid aging by light until they were analyzed. For each sample, a part of the quartz-fiber filter (diameter 47 mm) was extracted with 15 mL of methanol under ultrasound in a nicely chilled bath for 30 min. The extracts were filtered through a Teflon^® filter (0.22 µm) and reduced by volume to approximately 0.2 mL under a gentle stream of nitrogen. For the analysis, 0.2 mL of the reconstituted extract was diluted by adding 1 mL of methanol. The extracts of the field blanks were prepared in the same way.

The HPLC/DAD/Q-ToF-MS (Agilent 1290 Infinity II HPLC coupled with an Agilent 6546 QTOF−MS, Agilent Technologies) system was employed to examine the molecular formula and related UV absorption of molecules detected in MSOC in the PM_2.5 samples. Both the positive electrospray ionization (ESI+) and negative electrospray ionization (ESI−) modes were used in the QTOF−MS. The detailed procedures for data processing can be found in our previous study [35].

3. Results and Discussion

3.1. General Overview

Figure 1 shows the temporal variations in particle size distribution, concentration of PM_2.5 and concentrations of the major chemical compositions of the particles. Both the concentration of PM_2.5 and the chemical compositions of the particles presented obvious diurnal variations, which were higher at night and lower during the day, mainly reflecting the decrease in boundary layer height and the increase in pollution emission at night. Figure 1a shows the temporal variations in size distribution, in which the concentration of particles also shows an obvious diurnal variation, and the size distribution shows a single peak mode, mainly concentrated in the 46.14~201.69 nm range. The average particle size was 102.56 nm on the lightly polluted days, and it reached 269.32 nm on the highly polluted days. Figure 1b,c presents the absorption coefficients (b_abs) of heated and unheated aerosols collected by the CAPs at 450 nm and 530 nm, respectively. The average values of b_abs for unheated and heated aerosols were 198.23 ± 119.46 Mm⁻¹ and 108.08 ± 70.21 Mm⁻¹ at 450 nm and 94.58 ± 66.60 Mm⁻¹ and 67.80 ± 48.12 Mm⁻¹ at 530 nm. The single-scattering albedo (ω) varied widely (from 0.34 to 0.86), with an average value of 0.60 ± 0.08 at 530 nm, which is considerably low, suggesting a higher proportion of light-absorbing compounds. Despite the heavy pollution in Wangdu, the average E_abs at 530 nm was only 1.38, which is lower than the values measured in most Chinese cities, such as in Xi’an, Beijing [25,56] and Nanjing [22], indicating that BC is not coated well here. The average value of absorption Ångstrom exponent (AAE) between 450 nm and 530 nm was 5.3, suggesting a high concentration of BrC in Wangdu. Figure 1e shows the temporal variations of the mass concentrations of PM_2.5, OC, EC, Cl⁻, NO₃⁻, SO₄²⁻, NH₄⁺ and K⁺. Among them, the mass concentration of OC was the highest, with an average value of 55.13 ± 42.89 μg m⁻³. It was 9.71 ± 7.62 μg m⁻³ for EC. The mass concentration of NO₃⁻ was the highest among inorganic ions, with an average value of 14.45 ± 9.82 μg m⁻³, while the values of SO₄²⁻ and NH₄⁺ were 6.68 ± 4.36 μg m⁻³ and 10.26 ± 6.77μg m⁻³, respectively. This indicates that nitrate dominates the secondary inorganic aerosols in Wangdu.

3.2. Absorption Attribution of BrC, BC and the Lensing Effect

As shown in Figure 2, we quantified the contributions of BC, BrC, and the lensing effect to the total absorption at 450 nm, and the average absorption contribution percentages of BrC, BC and the lensing effect at 450 nm were 46.9 ± 15.7%, 39.0 ± 11.8% and 14.4 ± 6.7%, respectively. The histograms of the fractional absorption for BC, BrC and the lensing effect at 450 nm are presented in Figure 2b, and they show that the absorption of BC did not dominate the total absorption at 450 nm, while the absorption contribution of BrC exceeded that of BC during heavy pollution. The largest absorption contribution proportion of BrC could reach 78.3%. Since the average value of AAE between 450 nm and 530 nm was 5.3, the absorption at 300–400 nm would be 2–4 times that at 450 nm [27], implying that the radiative forcing induced by BrC at 300–450 nm was significantly greater than that caused by BC in Wangdu.

3.3. Diurnal Variation of the Optical Properties

The diurnal variations of the absorption coefficients of BC, BrC, and the lensing effect at 450 nm, AAE (450/530), O₃ and E_abs at 530 nm and the mass percentages of NH₄NO₃, (NH₄)₂SO₄, BC and POM in PM_2.5 are shown in Figure 3. Figure 3a shows that the proportion of BrC’s absorption contribution began to increase at 10:00 a.m. in the daytime and reached its maximum value (57.2%) at 6:00 p.m., while the BC and the lensing effect both showed the opposite downward trend, suggesting that some BrC generated by photooxidation during the daytime led to the increase in BrC’s absorption contribution. The AAE also showed an obvious diurnal variation, which was similar to the pattern of BrC’s absorption contribution, confirming the formation of BrC during the day. The proportion of nitrate increased sharply in the daytime, indicating the photochemical oxidation reaction was active in the daytime. Although the proportion of organic matter in PM_2.5 decreased due to the increase in (NH₄)₂SO₄ and NH₄NO₃, the increase in the proportion of BrC absorption and the increase in AAE in the daytime indicate that there were some UV-absorbing compounds formed in the daytime. The previous research reported that oxidation of phenolic compounds contributes to the majority of secondary organic aerosols (SOAs), in which particulate nitrophenolic compounds were estimated to explain 29 ± 15% of the average BrC light absorption at 405 nm despite accounting for just 4 ± 2% of the average organic aerosol (OA) mass [49]. It was also found that nitrophenols were the main light-absorbing compounds here, suggesting that the increase in AAE in the daytime may be caused by the formation of nitrophenols. The diurnal cycle of E_abs showed a dual-peak pattern, with one peak at 1:00 p.m. in the daytime and the other at 3:00 a.m. at night. The first peak could be attributed to the increasing thickness of the BC coating, which resulted from the formation of SOAs by photooxidation since the variation in E_abs was consistent with the variation in the concentration of O₃ during the daytime, and some studies have shown that O₃ will undergo photochemical aging with BrC to form SOA [57,58,59,60].The second peak at night may be related to the direct emission of a large amount of OAs at night, as shown in Figure 1e, which may have come from burning coal for heating in the surrounding rural areas at night. These OAs can then be adhered to BC’s surface to form a coating, enhancing E_abs.

3.4. Quantifying the Contributions of the Identified BrC in the Ambient Samples to Light Absorption

As shown in Figure 4, 52 molecular formulas, comprising 35–55% absorption of BrC dissolved in methanol, were determined. Among them, none of the identified formulas appeared in both the ESI− and ESI+ modes, and CHN+ was the greatest contributor to the total BrC absorption, with a proportion of 12.2–22.4%, followed by CHON− (9.6–19.0%), CHO− (3.9–5.9%), CHO+ (3.0–4.2%) and CHON+ (1.7–5.1%). The CHO in the determined BrC compounds was likely O-heterocyclic PAHs (O-PAHs), some of which have been found in previous research, including C₁₃H₈O (9-fluorenone) [61,62], C₁₆H₁₀O (benzo [b]naphtho [2,3-d]furan) [61,62], C₁₇H₁₀O (benzo [de]anthracen-7-one) [61], C₈H₆O₄ (phthalic acid) [63], C₉H₈O₄ and C₁₄H₁₀O₅. In addition, a number of CHON+ compounds containing one N atom and one O atom in their formulas with Xc ≥ 2.5, such as C₅H₅NO, C₉H₇NO and C₁₀H₉NO, also had absorption contributions.

The CHON− of the determined BrC compounds is generally considered to be nitroaromatics. The main light-absorbing contributors of them were C₆H₅NO₃, C₇H₇NO₃, C₁₀H₇NO₃ and C₈H₉NO₅, which were assigned to the nitrophenol [11], methyl nitrophenol [61], nitronaphthol [35,61] and nitrosyringol [47,61], respectively. Among them, the C₆H₅NO₃ and its homologous series (two C₇H₇NO₃ and six C₈H₉NO₃) accounted for 3.0–11.3% of the light absorption of molecules detected in MSOC. Nitro-aromatics have always been regarded as the most abundant BrC chromophores in OAs in many regions [49,50]. However, we found that the absorption contribution of CHN+ was comparable to CHON− in Wangdu, which was seriously affected by the emission of CC in winter [64,65,66]. Several reduced nitrogen compounds, such as C₁₂H₁₀N₂, C₁₆H₁₁N, C₁₇H₁₃N, C₁₃H₁₁N and C₁₅H₉N, were observed to have absorption in the UV spectrum [67]. Their high Xc values (Xc ≥ 2.71) suggest that they are PANHs with 3–5 aromatic rings fused together [68], and most of these PANHs contain an N atom in their formulas. The total light absorption of these PANHs comprised approximately 12.2–22.4% of the total measured absorption of molecules detected in MSOC at 365 nm. PANHs are toxicologically relevant organic compounds which are significantly different from the well-studied polycyclic aromatic hydrocarbons (PAHs) [69,70]. Benjamin et al. reported high concentrations of the total PANHs in Xi’an and suggested that other Chinese cities possibly have higher concentrations of PANHs because of the more pronounced use of coal [71]. A wide range of N-heterocyclic aromatics have been detected in soil and sediment samples contaminated with coal tar and creosote [68]. We found that C₁₅H₉N and its homologous series, including one C₁₆H₁₁N and three C₁₇H₁₃N, accounted for 5.1–12.3% of the light absorption of molecules detected in MSOC, which is comparable to C₆H₅NO₃ and its homologous series. Azapyrene and azafluoranthene are assumed structures for C₁₅H₉N, according to the literature data [68]. The research on the light absorption contribution of PANHs in OAs is relatively limited, and PANHs have never been reported to have such a large amount of light absorption in the atmosphere. This may be related to the large amount of CC emissions in Wangdu in winter, since CC has been reported to produce PANHs [67,72,73,74]. Molecular-level investigations of the ketolimononaldehyde (KLA, C₉H₁₄O₃) browning reaction found that N-heterocycles may contribute to the light-absorbing properties [75]. Bones et al. suggested that protonated Schiff bases may play a role in visible light absorption [76]. The N atoms in PANHs have non-bonding electron lone pairs, which are responsible for the n-π* transitions, resulting in the UV−vis absorption spectra of PAH experiencing a red shift and increasing climatic radiative forcing, which should attract more attention.

4. Summary and Conclusions

Investigations were conducted on the light absorption of BC, BrC and the lensing effect, as well as the chemical compositions of BrC in Wangdu. Quantifying the contribution of identified BrC to light absorption can help people better understand the sources of atmospheric radiation and more accurately evaluate atmospheric radiation forcing in model research. In addition, the absorption coefficients of 450 nm and 520 nm aerosols in the local area were higher than that in the background area of China. Combined with the high consumption of coal for domestic use in the surrounding rural areas, the organic aerosol pollution is relatively serious. Therefore, the local area should pay more attention to coal combustion.

BrC dominated the total absorption at 450 nm, and the largest absorption contribution proportion of BrC could reach 78.3% during heavy pollution. The E_abs at 530 nm was only 1.38, indicating that BC was not coated well here. The CHN+ was the highest contributor to the light absorption of molecules detected in MSOC with a proportion of 12.2–22.4%, in which the PANHs were the dominant compounds. The C₁₅H₉N and its homologous series, including one C₁₆H₁₁N and three C₁₇H₁₃N, accounted for 5.1–12.3%. The absorption of these PANHs is comparable to that of nitro-aromatics, which should attract more attention to the impact of climate radiative forcing.