Guohua Zhang 1, Long Peng1,2, Xiufeng Lian1,2, Qinhao Lin1, Xinhui Bi1, Duohong Chen3, Mei Li4, Lei Li4, Xinming Wang1, Guoying Sheng1

 

State Key Laboratory of Organic Geochemistry and Guangdong Key Laboratory of Environmental Resources Utilization and Protection, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou 510640, China
University of Chinese Academy of Sciences, Beijing 100039, China
State Environmental Protection Key Laboratory of Regional Air Quality Monitoring, Guangdong Environmental Monitoring Center, Guangzhou 510308, China
Institute of Mass Spectrometer and Atmospheric Environment, Jinan University, Guangzhou 510632, China


 

Received: December 28, 2017
Revised: May 31, 2018
Accepted: June 7, 2018

Download Citation: ||https://doi.org/10.4209/aaqr.2017.12.0566  


Cite this article:

Zhang, G., Peng, L., Lian, X., Lin, Q., Bi, X., Chen, D., Li, M., Li, L., Wang, X. and Sheng, G. (2019). An Improved Absorption Ångström Exponent (AAE)-Based Method for Evaluating the Contribution of Light Absorption from Brown Carbon with a High-Time Resolution. Aerosol Air Qual. Res. 19: 15-24. https://doi.org/10.4209/aaqr.2017.12.0566


HIGHLIGHTS

  • Light absorption coefficient (σ) of ambient/heated aerosols were measured.
  • The AAE-based method was improved to evaluate the σ of brown carbon (σabs,BrC).
  • The estimated σabs,BrC is > 20% larger than that estimated by assuming AAE = 1.
  • AAE around 0.7 for “pure” BC is suggested to evaluate σabs,BrC in the PRD region.
 

ABSTRACT


While brown carbon (BrC) might play a substantially important role in radiative forcing, an estimation of its light absorption contribution with high-time resolution is still challenging. In this study, a multi-wavelength (370–950 nm) Aethalometer was applied to obtain the wavelength dependent light absorption coefficient (σabs) of aerosols both before and after being heated to 250°C. An improved absorption Ångström exponent (AAE)-based method was developed to evaluate the contribution of BrC to light absorption at a wavelength of 370 nm (σabs,BrC/σabs,370nm). The σabs,BC at 370 nm was determined from the field measured AAE values for the wavelengths from 880 to 950 nm with a one-hour resolution. The simultaneous measurements of heated aerosols help confirm the negligible influence of BrC on the σabs values across the range of 880–950 nm. Meanwhile, σabs,BrC/σabs,370nm was also estimated with previously reported methods by assuming that the AAE was equal to 1 (Method I) as well as a new approach based on the light absorption enhancement (Method II). While the estimated σabs,BrC/σabs,370nm based on our developed method and Method I is highly correlated (r2 = 0.78), the difference could be as large as > 20% on average. The obtained mean σabs,BrC/σabs,370nm was negative with Method II, indicating the net production of BrC when the aerosols were heated. The difference between the values for σabs,BrC/σabs,370nm obtained by our developed method and by Method II was ~40% on average and much higher (> 50%) during the noon hour, when secondary organic aerosols and sulfate were abundant. We propose that it is more suitable to use an AAE around 0.7 for “pure” BC to evaluate the contribution of BrC to light absorption in the PRD region. The developed method thus helps improve our understanding of the light absorption and climate forcing of BrC.


Keywords: Brown carbon; Black carbon; Light absorption; Absorption Ångström exponent; Aethalometer; PRD region.


INTRODUCTION


Most climate models regard black carbon (BC) as the sole light-absorbing particulate material. However, some organic species (i.e., brown carbon/BrC) could also absorb solar radiation, and thus play an important role in radiative forcing (Liu et al., 2014; Liu et al., 2015; Shamjad et al., 2015). Recently, high contribution of BrC to light absorption is reported worldwide from satellite and ground observations, and modeling studies (Bahadur et al., 2012; Feng et al., 2013; Laskin et al., 2015, and references therein). Overwhelming presence of BrC above the Yellow Sea shows the necessity for understanding more than just bulk BC and organic carbon fractions of Asian aerosols (Alexander et al., 2008; Zhu et al., 2013). However, the variable absorptivity of light-absorbing carbonaceous aerosols is poorly known, making it challenging to account for their roles in radiative forcing (Andreae and Gelencser, 2006; Laskin et al., 2015). Therefore, an accurate estimation of BrC light absorption is essential both for constraining large uncertainties in current aerosol forcing estimates and for control policy development (Bergstrom et al., 2007; Zhang et al., 2013b; Saleh et al., 2014).

The BrC light absorption is strongly dependent on the wavelength (Favez et al., 2009; Moosmüller et al., 2011), parameterized by absorption Ångström exponent (AAE). AAE has commonly been used to estimate the contribution of BrC to light absorption (Corr et al., 2012; Kirchstetter and Thatcher, 2012). Previous studies indicate that the AAE of “pure” BC is around 1 in the climate relevant wavelength region of the solar spectrum, and the presence of BrC will increase the AAE due to its enhanced absorption at UV/near-UV wavelengths (Kirchstetter et al., 2004; Yang et al., 2009; Rathod et al., 2017). So, the measured wavelength dependence of AAE can be used to provide a quantitative estimate of BrC (Sandradewi et al., 2008; Harrison et al., 2012), and also for further distinction of sources (Ajtai et al., 2011; Pintér et al., 2018). However, theoretical calculations and field studies have shown that the likely range of AAE for internally mixed BC varies in a range of approximate 1–1.7, depending on the size and optical properties of the particle core and non-absorbing coating and the wavelength pairs used to determine AAE (Gyawali et al., 2009; Lack and Cappa, 2010; Lin et al., 2016). Using a three wavelength photo-acoustic soot spectrometer (PASS-3) and aerosol mass spectrometers (AMS), Yuan et al. (2016) explored the relationship between the measured AAE and the relative abundance of organic aerosol to BC and derived more realistic AAE values for “pure” BC aerosol, which is significantly lower than 1. Contributions to total light absorption at 370 nm by BrC were also estimated by comparing the wavelength dependence of the absorption coefficients of aerosols before and after heating (Guo et al., 2014). This is based on the assumption that the change of AAE for aerosols after heating is due to the evaporation of BrC, but the potential error of which has not been fully assessed.

In the present work, the optical, physical, and chemical properties of atmospheric aerosols were measured before and after heating in the Pearl River Delta (PRD) region, China. In this region, organics are the second largest contributor for light extinction (Zhang et al., 2013a; Chen et al., 2016), and the average light absorption contributions of BrC at 405 nm were quantified to be ~10%. This work presents spectral measurements of light absorption coefficient (σabs) in a broad spectral range (370–950 nm) by an Aethalometer, highlighting some characteristics of light-absorbing aerosols. An improved method was developed, based on wavelength segregated AAE, in order to estimate the contribution of BrC to light absorption at 370 nm (σabs,BrCabs,370nm). Meanwhile, the BrC contribution was also compared to those estimated based on the previously reported methods.


METHODS


 
Ambient Observation

Ambient measurements were performed at Guangdong Atmospheric Supersite (22.73N, 112.93E), a suburban site (Zhang et al., 2014) in the PRD region from July 8 to August 2, 2014. Over the sampling period, ambient temperature and relative humidity generally varied between 24–37°C and 40–93%, with an average of 29°C and 73%, respectively. The concentration of PM2.5 varied in a wide range of 0.78–166 µg m–3, with mean value of 29.7 µg m–3. Wind speed ranged between 0.3–5.1 m s–1 and the prevailing direction was from the southwest. Wind speed was lower (~1 m s–1) during the early morning hours and was relatively higher (~2 m s–1) in the afternoon.

To compare the properties of aerosols before and after heating, an automatic-controlled actuated ball valve that switches between “bypass” and “heating” on a 5-min cycle was used. The thermodenuder (TD, Dekati Ltd.) consists of a 95-cm long stainless steel tube, where the particles can be heated with controlled temperature, i.e., 250°C. The residence time of particles in the TD section is estimated to be 1.2 s, which should be enough for the effective vaporization of semi-volatile species (Villani et al., 2007). Volatile or semi-volatile species, such as most organic compounds and ammonium nitrate, are supposed to evaporate during the heating (Bi et al., 2015). Subsequently, these species were adsorbed in active charcoal in the adsorber section. The bypass consists of a silica tube of similar length for heating condition. In order to eliminate the influence of resident aerosols in the tubes, only the measurements during the latter 3-min of the 5-min cycle were included in the following data analysis. Particle loss could be typically represented as a function of flow rate and particle size (Stevanovic et al., 2015). The test made by the manufacturer shows that particles with a physical diameter dp ≤ 70 nm have a size dependent particle loss, increasing from ~20% to ~40% when dp decreases from 70 nm to 10 nm; particle loss at dp > 70 nm is estimated to be approximately 20% at the total flow rate of 10 L min–1, mainly attributed to the diffusion of particles. Regarding that the dp of refractory BC was reported to be around 200 nm in the PRD region (Huang et al., 2011), the particle loss might lead to a 20% negative bias for σabs measured through the heated line. The BC around 200 nm was most probably attributed to be from traffic emission, with their diurnal variation similar to that of NOx (Qin et al., 2017), as shown in Fig. S1.

The σabs was determined in 1-min resolution by an Aethalometer (AE-33, Magee Scientific) with seven wavelengths (at 370, 470, 520, 590, 660, 880 and 950 nm, respectively). A single particle aerosol mass spectrometer (SPAMS, Hexin Analytical Instrument Co., Ltd.) was deployed to measure the size and mass spectra of individual particles (Li et al., 2011). Aethalometer and SPAMS shared an aerosol inlet (dried total suspended aerosols) and were placed downstream of the TD. Additionally, the concentrations of water soluble ions associated with PM2.5 were obtained from a Monitor for AeRosols and GAses in ambient air (MARGA) (Rumsey et al., 2014). Hourly variations of organic carbon (OC)/elemental carbon (EC) in PM2.5 were provided by a Sunset Semi-Continuous Aerosol Analyzer. Concentration of O3 was measured using an ozone analyzer (Model 49i, Thermo-Fisher Scientific). For comparison, all the measured parameters are presented as hourly averages in this paper. A detailed description of the instruments and the data analysis can be found in Supporting Information (SI).

 
Light Absorption Measurements

Generally, filter-based optical measurement instruments (e.g., Aethalometer) must take into account the “filter loading effect” to accurately determine the σabs (Arnott et al., 2005; Ajtai et al., 2011). The new version Aethalometer (AE-33) used in the present study has been improved with the filter loading correction part, based on a two parallel spot measurement of optical absorption, through providing a real-time output of the “loading compensation” parameter to compensate for the “loading effect,” in detail described in the SI. The details of operation principle, data deduction, error budget of AE-33, inherent uncertainties in its technique and the corrections are extensively available in the literature (Drinovec et al., 2015). The BC mass concentration is calculated from the change in optical attenuation at 880 nm in the selected time interval using the mass absorption cross section (MAC) of 7.77 m2 g–1, which is close to that reported in the PRD region (Cheng et al., 2011). The noise level of AE-33 at the time-base of 1 minute is < 0.016 Mm–1 for the σabs (http://www.mag escientific.com/). The contribution of dust to σabs can be considered to be negligible in the PRD region (Yuan et al., 2016), due to limited mass fraction (Cheng et al., 2008; Zhang et al., 2013a), and lower mass absorption efficiency values (Favez et al., 2009).


RESULTS AND DISCUSSION


 
Previous Methods to Estimate the Light Absorption from BrC


Method I: Assuming AAE = 1 for “Pure” BC

The wavelength (λ) dependence of σabs is usually approximated by a power-law expression (e.g., Bergstrom et al., 2007), defined as:

σabs = × λ–AAE                                                    (R1)

where K is a constant. In practice for multi-wavelength measurements, AAE can be retrieved by fitting the σabs to the λ (370–950 nm), or to the different wavelength ranges (Bergstrom et al., 2007). Thus, a wavelength segregated AAE value can be obtained through the logarithmic fit of R1, or AAE = ln(σabs1abs2)/ln(λ12) for two wavelengths.

The AAE has often been used as a simplified method for attributing short visible wavelength absorption to BrC (Lack and Cappa, 2010). Light absorption at longer wavelength (950 nm) is assumed to be due only to BC, and is extrapolated to the shorter wavelength using an assumed AAE, most commonly 1, which is illustrated in Fig. 1. It estimates the σabs,BrC at 370 nm by subtracting the measured σabs to the extrapolated σabs,BC at 370 nm (“Method I”):

σabs,BrC = σabs,370nm – σabs,950nm × (370/950)–AAE             (R2)


Fig. 1. An illustration of the methods to estimate the contribution of light absorption from BrC, based on the assumption of AAE = 1 for Method I, and AAE = AAE880–950 for our developed method.Fig. 1. 
An illustration of the methods to estimate the contribution of light absorption from BrC, based on the assumption of AAE = 1 for Method I, and AAE = AAE880950 for our developed method.


Method II: 
Based on Light Absorption Enhancement

The σabs,BrCabs,370nm was also estimated based on the enhancements of σabs (Eabs), defined as the ratio of σabs for ambient and heated aerosols (Eabs = σabs,ambientabs,TD), at both 370 and 950 nm in previous studies (Cappa et al., 2012; Guo et al., 2014).

σabs,BrCabs,370nm = (Eabs,370nm – Eabs,950nm)/Eabs,370nm       (R3) 

where σabs,BrC refers to the effective absorption coefficient at 370 nm for the BrC evaporating under the TD temperature (i.e., 250°C), which is slightly different from the definition in R2. When there is no evaporative BrC, Eabs,950nm should equal Eabs,370nm, since evaporation of BrC would result in larger Eabs,370nm than Eabs,950nm (Cappa et al., 2012). It is worthy to note that this estimation (R3) actually represents the net result of evaporation and production of BrC in measured aerosol fraction. 


An improved AAE
-Based Method Developed by this Study

As far as the wavelength segregated AAE is concerned (Sandradewi et al., 2008; Mogo et al., 2012; Utry et al., 2014), the AAE determined over the wavelength of 370–520 nm and 880–950 nm and their diurnal variations were compared in Fig. 2. AAE370-520 were 1.36 ± 0.25 and 1.41 ± 0.22, and AAE880-950 were 0.70 ± 0.07 and 0.68 ± 0.08 for ambient and heated particles, respectively. AAE880-950 were quite similar, indicating there was a thin coating and negligible lensing effect, which could be supported by the chemical composition of the ambient and heated BC-containing particles detected by the SPAMS (Fig. S2). It can be seen that AAE880-950 varied in a narrow range (0.63–0.78, 10th–90th of the observed data), and significant difference (p < 0.001) was found between AAE370-520 and AAE880-950 for both ambient and heated particles, respectively. It reflects that σabs becomes more sensitive to spectral variations at lower wavelengths than at higher wavelengths, which is attributed to the presence of BrC (Mogo et al., 2012). While AAE is sensitive to these microphysical properties (Moosmüller et al., 2009; Lack and Cappa, 2010), the influence of particle size on the AAE for heated aerosols should be limited in the present study, since the difference of hourly average AAE880-950for aerosols before and after heating was in a small range (2% ± 2%). Furthermore, there is no significant correlation (p > 0.05) between AAE880-950 and the mass fraction ratio of available coating materials (PM2.5-BC) to BC, which further indicates limited influence of coating materials on the AAE880-950. In contrast, the difference between AAE370-520 for aerosols before and after heating was relatively higher (–5% ± 11%), which is more evidently shown in their diurnal variation (Fig. 2(b)). Correlation coefficients (r2) between AAE370-520 and AAE590-950, AAE660-950, and AAE880-950 are 0.56, 0.48 and 0.17, respectively, implying limited influence of BrC on the AAE880-950. Overall, these results supported that AAE880-950 may well represent the AAE of BC dominant aerosols in the present study. “Pure” BC AAE (i.e., AAE880–950) of 0.70 ± 0.07 inferred in this study is significantly lower than the widely adopted value (AAE = 1) (Bond and Bergstrom, 2006). While it might not be typical to have AAE of the “pure” BC lower than 1, the retrieved value in this study is quite similar to that (AAE between the wavelength at 532 and 781 nm) estimated for the “pure” BC dominated from traffic emission in the same region (Yuan et al., 2016). Additionally, lower AAE values (< 1) have also been reported for the “pure” BC in traffic and biomass burning emissions (Kirchstetter et al., 2004), and at urban site in Beijing (Wu et al., 2015). Theoretical calculation shows that AAE decreases with the increasing of “pure” BC sizes (Wu et al., 2015). Regarding the reported volume equivalent diameter peak of “pure” BC to be around 200 nm in this region (Huang et al., 2011), the corresponding theoretical AAE would most probably vary between 0.4 and 1 (Wu et al., 2015). In addition, the fractal like structure of combustion generated BC could also be one possible explanation of the smaller value of AAE measured in this study (Moosmüller et al., 2009; Lack and Cappa, 2010). Therefore, based on the assumption that BC is the dominant absorbing species over the near-IR wavelength (Shrestha et al., 2014), it is proposed to apply AAE = AAE880-950, instead of AAE = 1, in R2 to estimate σabs,BrC. Additionally, the input of hourly variation of AAE880-950 further accounts for the factors (e.g., sources and atmospheric processing) that have influence on AAE880-950 of BC dominant aerosols.


Fig. 2. (a) Comparison of AAE at the wavelength of 370–520, 880–950, and 370–950 nm, and diurnal variations of AAE at the wavelength of (b) 370–520 nm and, (c) 880–950 nm, for aerosols before and after heated, respectively.Fig. 2. 
(a) Comparison of AAE at the wavelength of 370–520, 880–950, and 370–950 nm, and diurnal variations of AAE at the wavelength of (b) 370–520 nm and, (c) 880–950 nm, for aerosols before and after heated, respectively.

The uncertainty associated with this method was estimated through a sensitivity test. The upper and lower bound of hourly average σabs (σabs ± std (σabs), std (σabs) refers to the standard deviation of σabs) were inputted in R1 and R2 to obtain an estimate of the upper and lower bound of σabs,BrC. Std (σabs) at 1-min time base is assumed to be three times the noise level (0.016 Mm–1) of AE33, and thus hourly based std (σabs), averaged over 18-min measurements of ambient aerosols could be calculated as: 3 × × 0.016 Mm1 = 0.2 Mm1. The upper and lower bounds of σabs,BrC was calculated to be 28.8 and 27.1 Mm–1 on average, respectively, which is within 3% difference compared to the mean value of σabs,BrC(28.0 Mm–1).

Based on the improved AAE-based method, the temporal profiles of the estimated σabs,BrC in Guangdong Atmospheric Supersite as well as theconcentration of potassium, the correlation between σabs,BrC and BC concentration, and the frequency histogram of the σabs,BrCabs,370nm are shown in Fig. 3. The estimated σabs,BrCabs,370nm distributed within a range of 30%–50%, with an average value at 37.4% ± 7.1%. Significant correlation (r2 = 0.72) between the estimated σabs,BrC and BC demonstrates their similar sources (Andreae and Gelencser, 2006; Moosmüller et al., 2009; Saleh et al., 2013; Guo et al., 2014). The correlation between the estimated σabs,BrC and the concentration of potassium, which is commonly used as a marker for biomass burning activities, was also observed to be significant (r2 = 0.38, p < 0.001), indicative of potential contribution of biomass burning to the observed BrC (Lack et al., 2013; Saleh et al., 2014; Hsiao et al., 2016; Rathod et al., 2017) in the PRD region. Potassium originated from sea salt and dust (Pio et al., 2007) was estimated to be < 5% of the observed potassium in our study. Moderate correlation could be partly explained by that emission of BrC from biomass burning is linked to both burning condition and the emitted organic tracers (Lack et al., 2013; Saleh et al., 2014), and also other sources.


Fig. 3. (a) Temporal profile of σabs,BrC and the concentration of potassium, (b) correlation analysis of σabs,BrC and the concentration of BC, and (c) the histogram summarizing of the σabs,BrC/σabs,370nm over this study.Fig. 3. 
(a) Temporal profile of σabs,BrC and the concentration of potassium, (b) correlation analysis of σabs,BrC and the concentration of BC, and (c) the histogram summarizing of the σabs,BrCabs,370nm over this study.


Comparisons of th
e Method Developed in this Study with the Previous Methods


Comparison 
with Method I

It is necessary to note that applying AAE = 1 to anchor absorption for BC and BrC may introduce potential errors, since the AAE for ambient particles (measured between a short and long visible wavelength) has often been observed to be variable, dependent on their physical and chemical properties (Mogo et al., 2012; Lack and Langridge, 2013). Fig. 1 shows the average AAE (1.36 ± 0.25 and 0.70 ± 0.07) of the ambient particles estimated from the linear fit (R1) of σabs over the wavelength of 370–520 nm and 880–950 nm, respectively.

Based on AAE = 1 assumption, σabs,BrCabs,370nm was calculated to be 16.8% ± 11%. While the estimated σabs,BrCabs,370nm based on AAE = AAE880-950 (the developed method) and AAE = 1 is highly correlated (r2 = 0.78, p < 0.001, Fig. 4), the difference can be as large as 20% on average. There were some negative values obtained when assuming AAE = 1, because the observed AAE370-950 was lower than 1 during some periods. Therefore, an arbitrary attribution of AAE to be 1 may substantially underestimate the σabs,BrCabs,370nm in this study.

 Fig. 4. Comparison of σabs,BrC/σabs,370nm estimated by our developed method and Methods I based on AAE = 1. Fig. 4. Comparison of σabs,BrCabs,370nm estimated by our developed method and Methods I based on AAE = 1.


Comparison 
with Method II

Based on Method II, the average estimated σabs,BrCabs,370nm over the study was –2.4% ± 11.5%, which indicates that production of BrC dominated over evaporation, resulted in a net production of BrC when ambient aerosols were heated. Similar estimation in previous studies also resulted in the negative values for σabs,BrCabs,370nm (Guo et al., 2014; McMeeking et al., 2014; Nakayama et al., 2014), and it was attributed to the production of BrC when ambient aerosols were heated (Nakayama et al., 2014). A comparison of the σabs,BrCabs,370nm estimated by Method II and the developed method in Fig. 5 indicates that the production of BrC during heating periods can be substantially important, which may account for > 40% difference between the estimated σabs,BrCabs,370nm, and thus should be emphasized when estimating σabs,BrCabs,370nm with Method II. The uncertain evaporation state of BrC during heating might be partly explained by the residue chemical information (Fig. S2). The increased difference during afternoon hours (~50%) (Fig. 5) may further indicate the influence of photochemical activity on the production of BrC during the heating periods. This could be evidenced by negative correlations (p < 0.001) between the estimated σabs,BrCabs,370nm by Method II and the concentrations of secondary organic carbon (SOC, see SI) and O3 (Fig. 6). This connection is probably explained by the conversion of SOC to BrC when aerosols were heated, and the conversion should be improved when there are high levels of SOC and O3. Such fraction of BrC may be attributed to low volatility oxygenated organic compounds, which has also been detected in heated (300°C) particles (Poulain et al., 2014). Therefore, the estimated σabs,BrCabs,370nm based on Method II should suffer substantial error when there is high abundance of SOC.


Fig. 5. Comparisons of (a) diurnal variations and (b) hourly averaged σabs,BrC/σabs,370nm estimated by Methods II and our developed method.Fig. 5. 
Comparisons of (a) diurnal variations and (b) hourly averaged σabs,BrCabs,370nm estimated by Methods II and our developed method.

From the co-located SPAMS, it was observed that detected number for internally mixed OC and sulfate (OC-sulfate) particles, which is frequently observed in the atmosphere of PRD region (Zhang et al., 2015), substantially increased during noon hours when aerosols were heated (Fig. S3). More efficient detection of these particles in this case is probably linked to increased absorption of ablation laser by heated OC-sulfate particles (Hatch et al., 2014). Without strong absorbing EC cores, the increased absorption is therefore more likely attributed to the formation of absorptive species (i.e., BrC) in heated OC-sulfate particles. Together with significant correlation between the estimated σabs,BrCabs,370nm and SOC and sulfate (Fig. 6), it is suspected that the formed BrC during heating measurements might be linked to organosulfate, which was ionized to fragments representative of OC and sulfate in the internally mixed OC and sulfate particles (Hatch et al., 2011). Light absorption at shorter wavelength had been observed for organosulfate (Song et al., 2013). Although secondary organic aerosol (SOA) produced by photo-chemical ageing may contribute to BrC (Lambe et al., 2013; Saleh et al., 2013), our results indicate primary emission plays a dominant role. As discussed above, significant correlations between the estimated σabs,BrC and BC, and potassium were observed. Furthermore, there is no correlation (r2 < 0.01) between the estimated σabs,BrC and SOC (Fig. S4). As noted by Lambe et al. (2013), the extent of absorption by SOA is linked with the magnitude dependent on both precursor type and oxidation level.


Fig. 6. Correlation plots between the σabs,BrC/σabs,370nm estimated by Method II and the concentrations of O3, SOC, OC and sulfate.Fig. 6.
 Correlation plots between the σabs,BrCabs,370nm estimated by Method II and the concentrations of O3, SOC, OC and sulfate.


CONCLUSIONS


This paper provides an improved AAE-based method of quantitatively estimating the light absorption by BrC in real-time based on multi-wavelength light absorption measurements. The analysis shows that the wavelength dependence of BC light absorption can be better approximatedfor higher (880–950 nm) wavelengths. Based on simultaneous measurements of σabs for both ambient and heated aerosols, AAE880–950 was found to show a negligible response to the presence of BrC; thus, the application of AAE = AAE880–950 (0.70 ± 0.07) for “pure” BC, instead of AAE = 1, should be more accurate for estimating the light absorption by BrC in the PRD region. However, the estimated AAE value of approximately 0.7 in this work cannot represent all kinds of situations, in particular, when the BrC absorption over longer wavelengths is non-negligible. BrC was estimated to contribute 37.4% to the light absorption at 370 nm by the developed method, and it was largely attributed to primary emissions. The contribution is > 20% larger on average than that estimated by assuming the AAE to be equal to 1. The results also suggest that the estimated contribution of BrC to the total light absorption at 370 nm based on the light absorption enhancement method might suffer from potential errors during high loadings of SOA, which might serve as evidence for the production of BrC when aerosols are heated—an aspect neglected in previous studies.

 
ACKNOWLEDGEMENTS


This work was supported by the National Nature Science Foundation of China (No. 41775124), the State Key Laboratory of Organic Geochemistry (SKLOGA201603A), and Guangdong Provincial Key Laboratory of Environmental Protection and Resources Utilization (2014B030301060). The authors would also like to thank all participants for the sampling at the Guangdong Atmospheric Supersite.


Aerosol Air Qual. Res. 19 :15 -24 . https://doi.org/10.4209/aaqr.2017.12.0566  


Don't forget to share this article 

 

Subscribe to our Newsletter 

Aerosol and Air Quality Research has published over 2,000 peer-reviewed articles. Enter your email address to receive latest updates and research articles to your inbox every second week.

Latest coronavirus research from Aerosol and Air Quality Research

2018 Impact Factor: 2.735

5-Year Impact Factor: 2.827


SCImago Journal & Country Rank

Aerosol and Air Quality Research (AAQR) is an independently-run non-profit journal, promotes submissions of high-quality research, and strives to be one of the leading aerosol and air quality open-access journals in the world.