Interactive comment on “ Contributions of nitrated aromatic compounds to the light absorption of water-soluble and particulate brown carbon in different atmospheric environments in Germany and China ” by Monique Teich

I am also missing a direct comparison of babs, the light absorption coefficient of particles, and Abs, and as well the calculated MAE (why was MAE not calculated for the Aethalometer data, based on babs and total PM mass/total PM organic mass, if available?) At least a comparison plot of the relative temporal evolution of these parameters for the Waldstein (summer) and TROPOS (winter) campaigns is highly interesting from both a scientific and methodological point of view and should be added to the paper (could also be in the SI).


Introduction
Organic components of atmospheric aerosols are usually treated as solely light scattering in global climate models over the near-ultraviolet and visible range (UV/Vis) of the solar spectrum and therefore only make a negative contribution to radiative forcing (Feng et al., 2013).However, the existence of light-absorbing organic carbon (OC) has become more and more evident in the past decade and can be a regionally important phenomenon (see reviews of Andreae and Gelencser, 2006, and, more recently, Laskin et al., 2015, and references therein and Ulevicius et al., 2010).In contrast to black carbon (BC), which absorbs light efficiently over the whole visible and UV region, light absorbtion by OC exhibits a distinct wavelength dependence.The light absorption sharply increases with decreasing wavelength, making it an efficient absorber in the UV/Vis range.Due to its characteristic yellowish to brownish color, light-absorbing OC is also often addressed as brown carbon (BrC).
The light absorption by BrC over the whole solar spectrum are found to be relatively weak compared to BC (J.Liu et al., 2013).Nevertheless, at near-UV/Vis wavelengths (300-500 nm) BrC has a non-negligible effect on radiative forcing and the regional and global climate (e.g., Bahadur et al., 2012;Feng et al., 2013;Jo et al., 2016;Park et al., 2010).For instance, modeling studies showed that the radiative forcing of BrC relative to BC is up to 25 % (Feng et al., 2013).Furthermore, BrC light absorption in the UV range of the spectrum may alter the concentrations of atmospheric oxidants due to reduced photolysis rates (Jacobson, 1999).
The light absorption of ambient particles is generally quantified by the determination of the particulate light absorption coefficient b abs , (in Mm −1 ), which can be normalized by the sample mass to give the mass absorption efficiency (MAE; in m 2 g −1 ).MAE can then further be used to estimate the radiative forcing of particulate BrC, which makes it an important parameter for modeling studies.These models start with modeled emitted mass, which is converted into concentrations using a dispersion model.The absorption of aerosols in the atmosphere is then determined using the appropriate MAE (Feng et al., 2013).The wavelength dependence of light absorption of a particle sample is described by the absorption Ångström exponent (AAE) based on the power-law dependence of b abs : BC has a weak wavelength dependence and often an AAE of 1.0 is assumed, which is the Mie theoretical value, though measured values range between 0.8 and 1.1 (Gyawali et al., 2009).Values greater than 1.1 indicate a stronger wavelength dependence and give evidence for the presence of absorbers at lower wavelengths, such as BrC and mineral dust (Wang et al., 2013).
Several methods exist to quantify the light absorption properties of BrC.Filter-based and multi-wavelength measurement techniques are commonly used to assess the contribution of BrC to the total aerosol light absorption (Moosmüller et al., 2009).One of these instruments is the seven-wavelength aethalometer, which operates from 370 to 940 nm.Due to the main light absorption of BrC at short wavelengths, the 370 nm channel is of particular interest.To estimate the contributions of BrC and BC to aerosol light absorption, a simple approach based on wavelength pairs and assumed AAEs to distinguish between the light absorption of BC and BrC has often been used (Lack and Langridge, 2013;Sandradewi et al., 2008).
Another method is to extract aerosol particles into solvents, mostly water or methanol (e.g., Bosch et al., 2014;Chen and Bond, 2010;Cheng et al., 2011;J. Liu et al., 2013).The advantage of this approach is that BC is insoluble in these solvents and therefore removed by filtration.The extracts are then further analyzed by UV/Vis spectrophotometry.Commonly, light absorption at 365 nm is used to characterize water-soluble BrC (e.g., Bosch et al., 2014;Cheng et al., 2011;Hecobian et al., 2010).This wavelength is in a range where the light absorption of inorganics does not interfere, whereas the light absorption of organics shows a sufficiently high intensity (Hecobian et al., 2010).Observed MAE values (related to the water-soluble OC, WSOC) for water-soluble BrC typically range between 0.41 and 1.80 m 2 g −1 (Cheng et al., 2016;Hecobian et al., 2010), where higher values are usually found in winter.
Until now, field measurements of the light absorption properties of water-soluble BrC have been performed mostly in the US and Asia, and measurements in Europe are scarce.Duarte et al. (2005) investigated the optical properties of WSOC from aerosol samples collected at a rural site in Portugal.They found a seasonal variation with a higher absorptivity in autumn than in summer.Some studies have reported the optical properties of humic-like substances (HULIS), which are a part of WSOC (e.g., Baduel et al., 2010;Utry et al., 2013).
The molecular composition of BrC still remains largely unknown due to its complex nature.Several attempts have been made to characterize BrC at a molecular level (Lin et al., 2015a, b).Among the identified compound groups are large macromolecules, like HULIS (Hoffer et al., 2006), aldol condensation products (Noziere andEsteve, 2005, 2007) and nitrogen-containing compounds formed by the reaction of atmospheric aldehydes with ammonia or amines, e.g., imidazoles (Galloway et al., 2009;Lin et al., 2015b;Yu et al., 2011).Nitrated aromatic compounds (NACs), which are the focus of the present study, comprise another group of major contributors to BrC (Jacobson, 1999;Mohr et al., 2013).Nitrophenols (NPs) can be either emitted directly into the atmosphere, e.g., by traffic exhaust (Nojima et al., 1983;Tremp et al., 1993) and wood burning (Hoffmann et al., 2007), or secondarily formed by the nitration of precursor compounds like phenol either in the gas phase or liquid phase (Bolzacchini et al., 2001;Vione et al., 2002).NPs have been quantified at many different locations, especially in Europe (Cecinato et al., 2005;Iinuma et al., 2010;Kahnt et al., 2013;Zhang et al., 2010), but there is surprisingly little information on NP concentrations in Asia (Chow et al., 2015).Nitrated salicylic acids (NSAs) have also been recently detected in atmospheric aerosol particles (Kitanovski et al., 2012;van Pinxteren and Herrmann, 2007).A strong correlation with nitrate was found by Kitanovski et al. (2012) in samples from Ljubljana, Slovenia, suggesting secondary formation from precursor compounds such as salicylic acid, which has been found in biomass-burning aerosols as a lignin degradation product (Iinuma et al., 2007).
Studying BrC at a molecular level is considered important, since even trace levels of a compound could have a significant impact on the light absorption properties of particles (Kampf et al., 2012).However, little is known about the contributions of specific light-absorbing compounds to the light absorption of either ambient aerosols or aqueous extracts from ambient particles.Mohr et al. (2013) estimated the relative contribution of NACs to particulate BrC light absorption to be about 4 % at 370 nm at a measurement site in the United Kingdom with high influence of BB aerosols.Zhang et al. (2013) calculated a contribution of NACs to aqueous extract light absorption of 4 % from the Los Angeles Basin (USA).
The present study aims to expand the understanding of BrC and water-soluble BrC by investigating its spatial and temporal variation in different, very diverse environments (urban, rural, biogenic, high BB influence).Furthermore, the contribution of individual light-absorbing organic compounds to the BrC light absorption is included.

Measurement campaigns
Measurements took place at five sampling sites (three German and two Chinese), in six different atmospheric conditions.An overview over the sites, sampling periods, geographical coordinates and the type of atmospheric environment is given in Table 1.Average temperatures and wind speeds are given as campaign averages.
The Leibniz Institute for Tropospheric Research (TRO-POS (winter) campaign, 2014) is located in Leipzig in eastern Germany and can be regarded as a moderately polluted urban background site impacted by a mixture of various sources (van Pinxteren et al., 2016).During the measurement period this site was strongly influenced by BB aerosols, as indicated by high levoglucosan concentrations.The first half of the campaign (24 to 31 January) was characterized by low 12 h mean temperatures ≤ 0 • C. Afterwards, the temperature gradually increased towards the end of the campaign (up to 10 • C as 12 h mean).The periods of lower temperatures often coincided with increased emissions from residential heating.The average wind speed was 2.2 m s −1 .
The measurement site Waldstein in the Bavarian Fichtel Mountains (Waldstein (summer) campaign, 2014; Plewka et al., 2006) is located about 180 km southwest of Leipzig in a low mountain range.The site is surrounded by forest, where spruce is the dominant species.Measurements took place on a tower at a height of 21 m (about 780 m above sea level).The measurement period was characterized by low wind speeds (< 1.3 m s −1 ) and predominantly sunny weather (average temperature of 19 • C).Because of the surrounding forest, the influence of freshly emitted biogenic organics is expected to be high.
Measurements at Melpitz, a rural background site, located about 50 km northeast of Leipzig, were carried out in winter (Melpitz (winter) campaign, 2014) and summer (Melpitz (summer) campaign, 2014).The Melpitz (winter) campaign was characterized by an average temperature of −3 • C and an average wind speed of 2.8 m s −1 .The Melpitz (summer) campaign was characterized by an average temperature of 22 • C and a mean wind speed of 2.0 m s −1 .The dates in winter and summer correspond to the measurement periods of the TROPOS (winter) and Waldstein (summer) campaigns, respectively.Due to the proximity of TROPOS and Melpitz, the two sites are influenced by similar regional air masses.
The Chinese measurement sites Xianghe (Xianghe (summer) campaign, 2013) and Wangdu (Wangdu (summer) campaign, 2014) are both located in the Hebei Province in the NACs and UV/Vis spectra were determined from the aqueous extract by extracting 11-28 pieces of the filter (1.54 cm 2 each) into 10 mL of ultrapure water.NPs were analyzed according to the method described in Teich et al. (2014) based on hollow fiber liquid-phase micro extraction for analyte enrichment and capillary electrophoresis electrospray ionization mass spectrometry (CE-ESI-MS, Agilent 3D CE instrument, Bruker Esquire 3000 + ion trap mass spectrometer).Samples from the Waldstein (summer) campaign were not analyzed for NPs due to the limited availability of filter material.
NSAs were enriched by evaporating an alkalinized aliquot of the aqueous filter extract (1.8 mL plus 200 µL 10 mM NaOH) to dryness in a vacuum concentrator (miVac, Genevac Ltd., UK) and redissolving the residue into 40 µL ultrapure water.Subsequent analysis was carried out by CE-MS as described in van Pinxteren et al. (2012).Standard addition was carried out for quantifying the NSA compounds by co-injecting a standard solution plug at different concentration levels into the CE capillary after the sample plug.
OC and elemental carbon (EC) were determined from the filter by a thermal-optical method using the Sunset Laboratory Dual-Optical Carbonaceous Analyzer (Sunset Laboratory Inc., Tigard, OR, USA) following the EUSAAR 2 temperature protocol and applying a charring correction using light transmission (Cavalli et al., 2010).
The aqueous extract for determining WSOC was prepared by extracting 21.5 cm 2 of the filter samples from the Xianghe (summer), Wangdu (summer), Melpitz (winter) and Melpitz (summer) campaign and 38.5 cm 2 for samples from the campaigns Waldstein (summer) and TROPOS (winter) into 25 and 30 mL of ultrapure water, respectively, followed by 20 min of ultrasonication.After filtration, the extract was injected into a TOC-V CPH analyzer (Shimadzu, Japan) operating in the NPOC (nonpurgeable organic carbon) mode.More details on the method can be found in van Pinxteren et al. (2009).
The air mass origin was estimated by 96 h back trajectories calculated using the HYSPLIT model (Stein et al., 2015).

Instrumentation for aerosol light absorption measurements
Particulate light absorption was measured by a sevenwavelength aethalometer, model AE33 (Aerosol d.o.o., Slovenia), during the TROPOS (winter) and Waldstein (summer) campaigns.A detailed description of the instrument is given in Drinovec et al. (2015).Briefly, aerosol particles are collected on a filter tape.Light attenuation is measured continuously through this aerosol laden filter (time resolution of 1 s).When a fixed attenuation threshold is reached, the tape advances to a new filter spot.Filter-based measurements feature nonlinear loading effects, caused by the increasing deposition of the sample in the filter loading during the measurement, and filter matrix light scattering effects (Weingartner et al., 2003).The aethalometer AE33 measures the loading effect by using a dual-spot approach, where attenuation measurements are carried out simultaneously on two differently loaded spots.Multiple scattering effects were compensated by normalizing the particulate light absorption coefficient b abs against data from a multi-angle absorption photometer (MAAP; Petzold and Schonlinner, 2004) which was also located at the measurement sites.The determined normalization factor (for light scattering effects) C was 1.69 for the TROPOS (winter) campaign and 2.06 for the Waldstein (summer) campaign.More details are given in the Supplement Sect.S1.The relative contribution of BrC and BC to b abs,370 was determined following the method used by Kirchstetter and Thatcher (2012).For the calculations, it was assumed that BC was the only absorbing species at 940 nm, that OC and BC were externally mixed and that the AAE of BC was 1.0.These assumptions are consistent with recently published studies (Martinsson et al., 2015;Mohr et al., 2013).Using Eq. ( 1), b abs,370,BC was determined by extrapolating from 940 nm to 370 nm, and the contribution of BrC was then calculated using Eq. ( 2 (2) The AAE was fitted over all seven wavelengths and is given as the average of the wavelength dependencies with the measurement time resolution.

UV/Vis spectrophotometry
Light absorption of aqueous solutions were measured with a Lambda 900 UV/Vis spectrophotometer (Perkin Elmer) using quartz cells (Secomam, France).The aqueous filter extracts were analyzed at pH≈2 (acidified with H 2 SO 4 , in the following also indicated with the subscript "A") and at pH≈10 (addition of NaOH, in the following also indicated with the subscript "B").Hence, the target compounds were either in their neutral or deprotonated state to obtain the lower and upper limit for the contribution of NACs to the BrC light absorption In principle, it could be possible that introducing acids or bases into the system induces unforeseen chemical reactions influencing the total light absorption of the aqueous extract.However, to minimize the risk of potential chemical modification the solutions where kept in the dark and analyzed as soon as possible after preparation by UV-Vis spectrophotometry.The influence of the pH on MAE and the aqueous light absorption coefficient (Abs) is discussed in Sect.3.1.1.Spectra were recorded from 300 to 800 nm.The interpretation of the aqueous extract light absorption follows the method described by Hecobian et al. (2010).In general, according to the Beer-Lambert law the absorbance (A λ ) of a solution is defined as where I 0 and I are the intensity of the incident and transmitted light, respectively, l the absorbing path length, c the concentration of absorbing species in solution and ε λ is the wavelength-dependent molar extinction coefficient.The resulting data were then converted to Abs λ : where A 800 is the reference wavelength to account for any baseline drift, V l is the volume of water used for extraction (10 mL), V a is the volume of air passed through the filter and ln(10) was used to convert the common logarithm to a natural logarithm.The path length l was either 2 or 5 cm.Abs λ is the light absorption coefficient of the aqueous extract solutions and is not to be mistaken for b abs , the light absorption coefficient of ambient particles.This nomenclature was chosen in accordance with other studies (J.Liu et al., 2013Liu et al., , 2015)).
Using the mass concentration of WSOC (in µg m −3 ), MAE was calculated by . (5) The molar extinction coefficient at 370 nm, ε 370 , was determined for each target compound by preparing a dilution series of standard compounds in water.Subsequently, a three point curve was recorded twice.According to the Beer-Lambert law, the slope of the regression in an absorbanceconcentration plot is proportional to ε.The molar extinction coefficients for each compound under acidic and alkaline conditions are presented in Table S2.
In other studies dealing with aqueous extracts, Abs values were given for 365 nm (e.g., Bosch et al., 2014;Cheng et al., 2011;Hecobian et al., 2010).We chose 370 nm to characterize the aqueous extract light absorption properties to match the 370 nm channel of the aethalometer.

Calculation of the contribution of nitrated aromatic compounds to light absorption of aqueous extracts and particulate brown carbon
The contribution of the target compounds to the aqueous extract light absorption was computed by first calculating the absorbance of the single compound using Eq. ( 3) and then determining the percentage on the total aqueous solution light absorption.
To assess the contribution of the NACs to particulate BrC light absorption at 370 nm, the molar extinction coefficient of each compound was converted to the liquid-phase molecular absorption cross section b liq (Jacobson, 1999): = 1000 ln(10) where N A is the Avogadro constant and ln( 10) is used to convert the common logarithm of the molar extinction coefficient (base 10) to the natural logarithm.The MAE comp of each compound was then calculated by dividing b liq by the mass of one molecule.The subscript "comp" denotes the specific compound.If the MAE comp and the mass concentration in air of the compound are known, b abs,comp of the single compound in ambient particles can be calculated by the following equation: where [comp] is the mass concentration of the compound in the aerosol.By knowing b abs,comp and b abs,BrC , the relative contribution of the individual compound on the particulate BrC light absorption can be calculated.) and the AAE.The results of the optical properties are summarized in Table 2.

Influence of acidic and alkaline conditions on the mass absorption efficiency and the aqueous light absorption coefficient
Values for MAE and Abs have usually been directly derived from filtered aqueous filter extract without considering the pH of the solutions (e.g., Bosch et al., 2014;Cheng et al., 2011;Hecobian et al., 2010).However, depending on the nature of the absorbing species, pH alters the absorptivity.Alif  (1987,1990,1991) showed for NPs that their anionic form is a much stronger absorber than their neutral form.Moreover, the absorption maximum shifts towards longer wavelengths under alkaline conditions.The dissociation in a solution depends on the pK a value of the compound and at different pH levels different mixtures of neutral/ionized compounds exist.Generally, ambient particles are estimated to be more acidic (pH < 4, Scheinhardt et al., 2013), making the findings under acidic conditions more relevant for ambient conditions.However, Hinrichs et al. (2016) reported that NPs adsorbed to a particle surface exhibit a significant red shift similar to the aqueous light absorption spectrum of their deprotonated form.Hence, under certain conditions the deprotonated forms can also be important.
Comparing alkaline and acidic conditions, strong correlations have been found for Abs, MAE and the relative contribution of the individual compounds to Abs over all campaigns (R 2 ≥ 0.93, see Fig. S3).On average, Abs 370 increased by a factor of 1.4 and MAE by a factor of 1.6 under alkaline conditions compared to the acidic conditions.As indicated by the good correlation, this factor is quite consistent and independent of the measurement site.
The molar extinction coefficient at 370 nm increases with pH for the individual compounds (see Table S2).A higher pH leads to deprotonation of hydroxyl and carboxyl groups, which may lead to a shift of the absorption maximum towards longer wavelengths.The light absorption of the aqueous extracts is a sum of many different light-absorbing species.Therefore, the relative contribution to the aqueous extract light absorption for individual compounds is affected differently.For example, the relative contribution of 3NSA and 2,6DM4NP to Abs 370 decreases at pH 10.No difference in the contribution to Abs 370 was observed for 5NSA.The highest increase in the contribution to Abs 370 was found for 2,4DNP with a factor of 6.8.The observed differences in the contribution to Abs 370 are related to both the molar extinction coefficient, which alters with the pH, and the composition of the solution.A decrease in the relative contribution to the light absorption at higher pH is most likely due to greater contributions of other, now overlapping light-absorbing compounds at this specific wavelength.
The increase of Abs 370 with higher pH is relatively modest (a factor of 1.4) compared to the increase of the molar extinction coefficient of individual NACs (an average factor of 4.5).This might indicate that the absorption properties of the majority of light-absorbing compounds are barely affected by the pH.These findings are only valid for the investigated wavelength of 370 nm and are likely to differ over the whole spectral range.Due to the close relationship between MAE and Abs under alkaline and acidic conditions, only the values for the acidic conditions are discussed in the following section (Sect.3.1.2),while both of them are always displayed in the figures and tables.

Mass absorption efficiency and aqueous light absorption coefficient of water-soluble brown carbon
The temporal evolution of Abs 370 and MAE 370 for each measurement site is shown in Fig. 1.The highest mean MAE 370A values were found for the campaigns Melpitz (winter, 0.86 ± 0.33 m 2 g −1 ) and TRO-POS (winter, 0.84 ± 0.20 m 2 g −1 ) followed by Wangdu (summer, 0.55 ± 0.15 m 2 g −1 ) and Xianghe (summer, 0.38 ± 0.06 m 2 g −1 ).The lowest values were observed under summer conditions in Germany, where the values for the Melpitz (summer) campaign slightly exceeded the MAE 370A observed at the Waldstein (summer) campaign (0.22 ± 0.03 m 2 g −1 over 0.17 ± 0.03 m 2 g −1  2015) observed 0.7 m 2 g −1 (all values determined at 365 nm).The findings for the TROPOS (winter) and Melpitz (winter) campaigns suggest regional behavior and are comparable to observed MAEs from urban sites in the US (Hecobian et al. (2010), 0.47-0.87m 2 g −1 , annual mean, determined at 365 nm).Low MAE values, similar to the observed MAEs at the Melpitz (summer) and Waldstein (summer) campaigns, have been found at urban and rural sites in the US (0.21 and 0.12 m 2 g −1 , respectively, determined at 365 nm) when the concentration of levoglucosan was low (< 50 ng m −3 ).Although the average Abs 370A is similar for the campaigns of Wangdu (summer), TROPOS (winter) and Melpitz (winter) (4.64, 3.92 and 4.13 Mm −1 , respectively), the observed mean MAE 370A for the Wangdu (summer) campaign is lower, indicating that despite higher concentrations of OC and WSOC during the Wangdu (summer) campaign (see Table S3), fewer light-absorbing compounds were present.
A correlation between levoglucosan and Abs 370 was found during the TROPOS (winter) campaign (R 2 = 0.8) and during the BB episode of the Wangdu (summer) campaign (R 2 = 0.9), when BB aerosol was abundant (see Fig. 2).This correlation suggests that BB aerosol is a major contributor to the aqueous extract light absorption.The connection between BB and light-absorbing BrC was also observed in a number of other studies (e.g., Desyaterik et al., 2013;Hoffer et al., 2006;Lack et al., 2013;Mohr et al., 2013).During non-BB episodes of the Wangdu (summer) campaign and the Xianghe (summer) campaign, only a weak or no correlation at all was observed between levoglucosan and Abs 370 (see Fig. 2).This indicates the presence of sources for lightabsorbing compounds other than BB aerosols.A decay of levoglucosan more rapid than the decay of NACs might also be a cause for this observation (Hoffmann et al., 2010).
Regarding seasonal differences, the aqueous extract light absorption was much lower in summer compared to winter samples.This is a general trend, observed at many measurement sites (e.g., Du et al., 2014;Hecobian et al., 2010;Kim et al., 2016).Kim et al. (2016) suggested that this is either because the amount of emitted BrC is much higher during winter or that summer BrC is less water soluble.By comparing aqueous extract light absorption to ambient particle light absorption measurements, this issue will be further explored and discussed in the following section.

Particulate light absorption coefficient and absorption Ångström exponent of particulate brown carbon
The temporal evolution of the measured particulate light absorption coefficient b abs at 370 nm as well as the determined contributions of particulate BrC and BC to the light absorption at 370 nm are shown in Fig. 3. Normalizing the determined b abs for particulate BrC by the according OC content gives the mass absorption efficiency for BrC in the particle (MAE 370,BrC,particle ).The temporal variation of MAE 370,BrC,particle is displayed in Fig. S4.The AAE (as campaign average) obtained by fitting over all aethalometer wavelengths was 1.5 ± 0.1 and 1.2 ± 0.3 for the campaigns TROPOS (winter) and Waldstein (summer), respectively.An AAE close to 1 indicates that BC is the dominant species (Kirchstetter et al., 2004).The increased AAE values indicate that BrC plays a larger role for the TROPOS (winter) campaign than for the Waldstein (summer) campaign, where its contribution is minor.
During the BB influenced TROPOS (winter) campaign, the mean total b abs,370 was 54 Mm −1 .Particulate BrC contributed about 40 % to the overall light absorption at 370 nm and an average MAE 370,BrC,particle of 1.95 m 2 g −1 was calculated.The influence of BB on the light absorption of ambient particles has also been observed in other studies.For instance, Kirchstetter and Thatcher (2012) estimated that the contribution of BrC to the light absorption of wood smoke is 49 % below 400 nm from samples collected in wintertime California (USA).Mohr et al. (2013) found a BrC contribution of 46 % to b abs,370 at a UK site that was highly influenced by BB.Our findings on the contribution of BrC to b abs, 370 during the TROPOS (winter) campaign are consistent with these earlier measurements.The light absorption was observed to be higher for easterly air masses than those from the west, a trend also seen in the NAC concentrations of the TROPOS (winter) campaign (see Sect. 3.2).Easterly winds were often accompanied by lower temperatures or thermal inversion layers or both.Therefore, higher concentrations of light-absorbing compounds (and higher particulate light absorption) might be a result of both increased emissions (locally and in the source region) and lower dispersion of absorbing pollutants due to the thermal inversion layer.
In contrast, the light absorption of aerosol particles was much lower during the Waldstein (summer) campaign with a mean total b abs,370 of 6 Mm −1 , a MAE 370,BrC,particle of 0.21 m 2 g −1 and a BrC contribution to the particulate light absorption of about 15 % at 370 nm.The measurements at Waldstein (which is surrounded by a mixed forest) were carried out in a sunny summer period when high biogenic emissions would strongly influence the results.Several laboratory studies have shown that the formation of lightabsorbing compounds from biogenic precursors is negligible (e.g., P. F. Liu et al., 2013;Nakayama et al., 2010;Song et al., 2013).However, a few studies have suggested that BrC might be formed in the presence of high ammonia concentrations (Flores et al., 2014;Updyke et al., 2012) or from isoprene epoxydiols (IEPOX; Lin et al., 2014).Findings from field measurements in the Amazon Basin confirmed that the light absorption by biogenic aerosol is much lower than the light absorption of BB aerosols or BC (Rizzo et al., 2011).Compared to the TROPOS (winter) campaign, b abs for BrC is very small for the Waldstein (summer) campaign, with an average of 0.94 Mm −1 compared to 21.8 Mm −1 .Our findings corroborate those of Rizzo et al. (2011) and laboratory studies stating that biogenic aerosols alone produce only small amounts of light-absorbing compounds.
The relative contribution of light-absorbing WSOC to particulate light absorption can be estimated by applying a conversion factor.Based on Mie theory calculations, J. Liu et al. (2013) suggested a factor of 2 to convert the aqueous extract light absorption coefficient (Abs) into the particulate light absorption coefficient (b abs ).The temporal variation of the fraction of the converted Abs 370 to the particulate BrC light absorption is given in Fig. S4.For our data, a good agreement between the Waldstein (summer) average Abs 370A (0.51 Mm −1 × 2) and the Waldstein (summer) average b abs,370,BrC (0.94 Mm −1 ) was achieved, suggesting that the observed particulate light absorption is mainly associated with water-soluble compounds during this campaign.For the TROPOS (winter) campaign, doubling the average Abs 370A gives a value of 7.8 Mm −1 , which is still much lower than the average b abs,370,BrC (21.8 Mm −1 ).Thus, a large fraction of non-water-soluble OC likely exists as well, which is highly light absorbing at 370 nm.

Concentrations of nitrated aromatic compounds under different atmospheric conditions
This section presents an overview and a comparison of the NAC concentrations for all campaigns.Table 3 summarizes the measured concentrations in comparison with other studies at similar locations.The results of the temporal variation of the target compounds for each campaign are displayed in Fig. 1.
A general concentration trend of 4NP > 3M4NP > 2M4NP > 2,6DM4NP > 2,4DNP > 3,4DNP and 3NSA > 5NSA was observed for all campaigns of the present study and is in good agreement with findings from most other studies.Regarding the sum of NPs, the mean concentrations were highest during the TROPOS (winter) campaign (14.0 ng m −3 ) followed by the Melpitz (winter) campaign (11.1 ng m −3 ).These concentrations are a factor of 3 to 10 higher than those observed during the campaigns Wangdu (summer) and Xianghe (summer) (4.4 and 1.4 ng m −3 , respectively).During the Melpitz (summer) campaign, the NP concentration was 100 times lower (average of 0.1 ng m −3 ) than during the Melpitz (winter) campaign.For the sum of NSAs, the highest mean concentrations were found at the Wangdu (summer) campaign (4.8 ng m −3 ) followed by the campaigns Xianghe (summer) and TROPOS (winter) (average of 2.2 ng m −3 , each).Average concentrations for NSAs of ≤ 1 ng m −3 were observed at the campaigns Melpitz (winter, 1.0 ng m −3 ), Waldstein (summer, 0.4 ng m −3 ) and Melpitz (summer, 0.2 ng m −3 ).
A clear seasonality can be seen in the data from the German sites.The observed concentrations of the different species are generally lower in summer than in winter because the sources of NACs are stronger during winter and generally connected to polluted air, as discussed in the Sect. 1.
The campaigns Melpitz (winter) and TROPOS (winter) were carried out in the same period and the two sites are directly comparable due to their geographic proximity.Concentrations of the NACs are in the same range at both sites and show a common pattern.Two periods with concentration maxima were found.The first period occurred around 25 January and the second peak was around 30 January 2014 (see Fig. 1a and b).Both periods are characterized by prevailing easterly winds and shorter mean lengths of 96 h back trajectories.In contrast, air masses during other periods originated from the southwest and had longer trajectories and increased residence times over the Atlantic Ocean.Higher concentrations of PM and many of its constituents with easterly winds are often observed at Melpitz (Spindler et al., 2010).The long residence time above the continent results in increased concentrations of pollutants in those air masses.Furthermore, as described in Sect.3.1.3,the temperature was generally lower when the air mass originated from the east, which likely led to higher emissions from residential heating both locally and in the air mass source regions.
The campaigns Melpitz (summer) and Waldstein (summer) were also conducted in the same time period.The concentrations of NSAs are comparable at both sites (0.01-0.51 ng m −3 ) and are very low compared to the winter measurements.NPs were not analyzed from the Waldstein (summer) campaign as explained above.
Day and night samples were collected during the campaigns TROPOS (winter), Waldstein (summer), Xianghe (summer) and Wangdu (summer).Nighttime concentrations of the sum of NACs were found to be slightly higher than during the day (not statistically significant at 95 % confidence level).The lower daytime concentrations are probably caused by the higher boundary layer heights during the day, i.e., dilution effects, and/or lower emission/formations rates during daytime or photolysis processes.
The sum of all target compounds at the TROPOS (winter) campaign was only weakly correlated with levoglucosan (R 2 = 0.47, Fig. S5), indicating that BB was not the dominant source of NACs during the campaign.As mentioned above, the formation or decay processes of levoglucosan and NACs might be very different, which resulted in no correlation of those species.It is also known that the sources of NACs are very diverse (see Sect. 1).
At the Wangdu (summer) campaign, a BB episode was observed between 11 June and 19 June (see Fig. 1f), indicated by a peak in measured levoglucosan concentrations.During  a Estimated from graph.b Sum of corresponding m/z ratios, including all isomers.c Concentrations are given as the average of three years.n.d.: not detected that period, increased concentrations of NACs were found.Although the NAC concentration generally increased with increasing levoglucosan concentration, no clear connection between NACs and BB aerosols was found.For the Xianghe (summer) campaign, levoglucosan concentrations (average of 109 ng m −3 ) were lower than for the Wangdu (summer) or TROPOS (winter) campaign, suggesting that BB played a minor role at this site during the time of measurements.No correlation was found between the sum of NACs and levoglucosan.This implies that BB was neither the major source of NACs in the Wangdu (summer) nor in the Xianghe (summer) campaign.It was found that NACs can be a product of the photochemical processing of anthropogenic volatile organic compounds (Jaoui et al., 2008), which might be a possible source for NACs at the Chinese sites besides BB.
NACs seem to be ubiquitous and are found at each site in measurable amounts although they are highly variable at the different measurement sites.The sum of NAC concentrations at the different campaigns followed the order TRO-POS (winter) > Melpitz (winter) > Wangdu (summer) > Xianghe (summer) > Waldstein (summer) > Melpitz (summer).This trend might be considered surprising, since the average PM 10 and OC values were higher at the Chinese sites than at the German sites (see Table S3).However, the seasonality and influence of BB aerosol seem to play a large role in the NAC concentration resulting in lower concentrations at the Chinese summer sites than at the German winter sites.

Contribution of nitrated aromatic compounds to water-soluble and particulate brown carbon
In this section, the contribution of NACs to the aqueous extract light absorption and then the contribution of NACs to the particulate BrC light absorption are discussed.The relative contribution of NACs to Abs 370 is shown in Fig. 4 together with the contribution of NACs to the WSOC concentration.Table 2 summarizes the contribution of NSAs and NPs to aqueous extract light absorption (Abs) and particulate BrC light absorption (b abs ).The focus is on the relative contribution to the aqueous extract light absorption at the wavelength of 370 nm to match the 370 nm channel of the aethalometer.To give additional information, the relative contribution of NACs to Abs (acidic conditions) over a spectral range of 300 to 500 nm is presented in Fig. 5 as their campaign averages.
The relative NAC contribution to Abs 370 varied greatly across the different sites, ranging from 0.02 to 4.41 % for acidic conditions and from 0.02 to 9.86 % for alkaline conditions.This result likely reflects the dependence of NAC source strength on location and seasonality.The maximum contribution of NACs to Abs 370 was found for the TRO-POS (winter) campaign.The contribution of NACs to Abs 370 was low for the campaigns Waldstein (summer) and Melpitz (summer).For the Waldstein (summer) campaign, a mean contribution of 0.15 % (acidic) or 0.13 % (alkaline) was ob- served.Average values of 0.13 % (acidic) and 0.24 % (alkaline) were obtained from the Melpitz (summer) campaign.This result is not surprising due to the low influence of BB aerosols or traffic and thus low NAC concentrations.A recent study by Nguyen et al. (2013) suggested the formation of BrC from ketoaldehydes derived from biogenic monoterpenes in the presence of ammonium ions.Thus, this reaction may play a role in regions with higher influence of biogenic emissions and might be one explanation for the observed absorption at the Melpitz (summer) and Waldstein (summer) campaigns.The formed species were suggested to consist of conjugated aldol condensates, secondary imines and nitrogen containing heterocycles, The results for the Xianghe (summer) and Wangdu (summer) campaigns are about half the values of the TROPOS (winter) or Melpitz (winter) campaigns.This shows that NACs in the present study had a much higher impact on the light absorption properties during the German winter campaigns than during the Chinese summer campaigns and indicates that other compounds play a larger role in particle light absorption at the Chinese sites.More studies are needed to resolve the molecular identity and abundance of these light-absorbing compounds.
The relative contribution of NACs to Abs is wavelength dependent.Under acidic conditions, the contribution of NACs to the aqueous extract light absorption increases towards lower wavelength, reaching a maximum that was found to generally lie in the range of 330-350 nm (see Fig. 5).Moreover, it can be seen that maximum values differ for individual compounds.Under alkaline conditions, this maximum shifts towards the range of 400-420 nm (see Fig. S6).As stated above, the results for acidic conditions are likely to be more atmospherically relevant.Therefore, a higher influence of NACs towards shorter wavelengths suggests that they can be more important in terms of their influence on atmospheric photochemistry, e.g., O 3 photolysis (Jacobson, 1999).
The relative mass contribution of NACs to the WSOC concentration at the different campaigns follows the same pattern as the contribution of NACs to the aqueous extract light absorption and also the same pattern as the NAC concentrations (see Sect. 3.2 and Fig. 4).The highest mass contributions of NACs were found during the TROPOS (winter) campaign followed by the Melpitz (winter) campaign, and lowest values were observed for the campaigns Waldstein (summer) and Melpitz (summer).The average mass contribution of NACs to WSOC is less than 0.2 % at each site.In contrast, NACs have a mean contribution to aqueous extract light absorption of 1 %, which is a factor of 5 larger than their mass contribution to WSOC.Our results corroborate the findings of other studies, which show that even small amounts of light-absorbing compounds can significantly impact the light absorption properties of particles.As a result, a detailed investigation of the molecular composition of BrC is important (Kampf et al., 2012;Laskin et al., 2015).
The correlation between NAC concentrations and Abs 370 for each campaign is displayed in Fig. 6.The correlation is good for the TROPOS (winter) and Melpitz (winter) campaign (R 2 = 0.7, each) and weaker for the Wangdu (summer) campaign (R 2 = 0.6).No correlation was found for the campaigns Xianghe (summer), Melpitz (summer) and Waldstein (summer).In general, the presence of high NAC concentrations also indicates higher values for Abs 370 than at sites with lower NAC concentrations.However, due to the diverse sources and sinks (e.g., photobleaching, Zhao et al., 2015) of NACs and BrC, NACs cannot be considered as the only tracers for BrC for very different atmospheric conditions.
The average contribution of NACs to particulate BrC light absorption was found to be 0.10 % for the Waldstein (summer) campaign and 0.25 % for the TROPOS (winter) campaign for the protonated forms and 0.13 and 1.18 %, respectively, for their deprotonated forms.The contribution during the Waldstein (summer) campaign is similar for the aqueous extract and aerosol light absorption, which is consistent with the assumption discussed above that the observed light absorption is mostly derived from WSOC and no further waterinsoluble BrC was present at this site.
To the best of the author's knowledge, up to now only two other field studies have determined the contribution of individual compounds to BrC light absorption.Zhang et al. (2013) estimated the contribution of eight NACs to aqueous extract light absorption in Los Angeles to be about 4 % at 365 nm, while Mohr et al. (2013) estimated a NAC contribution of 4 % to particulate BrC light absorption at 370 nm at a BB influenced site in the UK.The findings from Zhang et al. (2013) are in the same range as our findings for the TROPOS (winter) campaign (for Abs 370 ), whereas the estimated contribution of NACs to particulate BrC light absorption by Mohr et al. (2013) is a factor of 16 larger than our observed contributions.One possible reason for this discrepancy is that the mean b abs,BrC , measured in the present study is about twice as high as observed by Mohr et al. (2013) (10 Mm −1 ) while the concentrations of NACs are in the same range in both studies.Moreover, Mohr et al. (2013) considered a different suite of NACs than the present study.Another important point is that the calculation by Mohr et al. (2013) used values based on the absorption maxima of the individual species, which usually lie below 370 nm (for their protonated form).In contrast, in the present study the molar extinction coefficient of each compound was directly determined at 370 nm.When calculations were made based on the absorption maxima of protonated NACs, an average contribution to the particulate BrC light absorption of 1 % was found for the present study, i.e., a factor of 4 higher than the calculation at 370 nm.

Conclusions
In the present study, the contributions of eight NACs to the light absorption of aqueous particle extracts and particulate BrC were determined in Germany and China at five measurement sites under six different atmospheric conditions.
The effect of pH on the aqueous extracts light absorption was investigated.Absorption of extracts at 370 nm, Abs 370 , increased by a factor of 1.4 and MAE by a factor of 1.6 when the pH was increased from 2 to 10.A general seasonal trend of MAE 370 and Abs 370 being higher in winter than in summer was observed, which is in agreement with other studies.The MAE values during winter times impacted by BB in Germany exceeded those of the Chinese background stations during summer.Very low MAE and b abs,370 were observed at the forest site Waldstein in summer, indicating that freshly emitted biogenic aerosols are only weakly absorbing.In contrast, a strong relationship was found between the light absorption properties and BB aerosol concentrations, corroborating findings from other studies.An average contribution of particulate BrC to b abs,370 of 46 % was observed during the TROPOS (winter) campaign and 15 % during the Waldstein (summer) campaign.The AAE of particulate BrC was found to be 1.5 and 1.2 for the campaigns TROPOS (winter) and Waldstein (summer), respectively.The relative contribution of NACs to the aqueous extract light absorption was highly variable depending on the measurement site, ranging from 0.02 to 4.41 % under acidic conditions and 0.02 to 9.86 % under alkaline conditions.This indicates that the emission strength of light-absorbing compounds and the composition of BrC differed between the sites.In addition, the formation and decay processes might be very different in the respective environments.The mean contribution of NACs to the particulate BrC light absorption was 0.10 % (acidic conditions) and 0.13 % (alkaline conditions) during the Waldstein (summer) campaign and 0.23 % (acidic conditions) and 1.15 % (alkaline conditions) during the TRO-POS (winter) campaign.A correlation between the NAC concentration and Abs 370 was observed during the campaigns TROPOS (winter) and Melpitz (winter), at other sites the correlation was weak or nonexistent.
The mass contribution of NACs to WSOC was 5 times lower than their contribution to the aqueous extract light absorption.This corroborates conclusions of other studies that even small amounts of light-absorbing compounds can have a disproportionately high impact on the aerosol light absorption properties (Kampf et al., 2012).
Field studies on the molecular composition of BrC are still scarce, especially for sites that have no or little influence of BB aerosols.Therefore, more efforts are needed to assess BrC on a molecular level, since a deeper knowledge of the present light-absorbing compounds can improve the understanding and prediction of BrC aging processes and its radiative forcing (Laskin et al., 2015).A further investigation of BrC compounds may also result in finding tracer compounds for specific BrC sources.

Data availability
The dataset is available upon request to the corresponding author.The aqueous light absorption coefficient is given at 370 nm for acidic conditions (indicated by the subscript "A", black squares) and for alkaline conditions (indicated by the subscript "B", red dots).
The Supplement related to this article is available online at doi:10.5194/acp-17-1653-2017-supplement.

Figure 1 .Figure 2 .
Figure1.Temporal variation of the concentrations of nitrated aromatic compounds and levoglucosan (Levo) at each measurement site (a-f) and their optical properties.The aqueous light absorption coefficient (Abs) and mass absorption efficiency (MAE) are given at 370 nm for acidic conditions (indicated by the subscript "A") and for alkaline conditions (indicated by the subscript "B").

Figure 3 .
Figure 3. Temporal evolution of the particulate light absorption coefficient b abs at 370 nm derived from the aethalometer measurements during (a) the TROPOS (winter) campaign and (b) the Waldstein (summer) campaign.The fractions of black carbon (dark grey area) and brown carbon (light grey area) to the total measured b abs were calculated as described in Eq. (2).

Figure 4 .
Figure 4. Top and center: contribution of the sum of NACs to the aqueous extract light absorption at each site under acidic (a) and alkaline (b) conditions.Bottom: mass contributions of the sum of NACs to the WSOC (c).

Figure 5 .
Figure 5. Relative contribution of NACs to Abs λ over the spectral range of 300 to 500 nm for each measurement campaign (a-f) under acidic conditions.The data are presented as campaign averages.Due to instrumental issues, data for lower wavelengths are not always available.

Table 1 .
Overview over sampling sites, used terminology, sampling period, sampling time given in local time (LT) and according atmospheric conditions during the sampling period.

Table 2 .
Optical properties and calculated contribution of NACs to the light absorption of aqueous extracts and particulate BrC.The values are given as minimum-maximum (mean ± standard deviation).Acidic conditions are indicated by the letter "A" and alkaline conditions are indicated by the letter "B".Highest and lowest mean values for each category are marked in bold.

Table 3 .
Measured concentrations of nitrophenols and nitrated salicylic acids for each campaign in comparison to other studies.The concentrations are given as minimum-maximum (mean ± standard deviation).