The fraction of gasoline direct-injection (GDI) vehicles
comprising the total vehicle pool is projected to increase in the future.
However, thorough knowledge about the influence of GDI engines on important
atmospheric chemistry processes is missing – namely, their contribution to
secondary organic aerosol (SOA) precursor emissions, contribution to SOA formation, and
potential role in biogenic–anthropogenic interactions. The objectives of
this study were to (1) characterize emissions from modern GDI vehicles and
investigate their role in SOA formation chemistry and (2) investigate
biogenic–anthropogenic interactions related to SOA formation from a mixture
of GDI-vehicle emissions and a model biogenic compound,
Annual biogenic volatile organic compound (BVOC) emissions are estimated to be 825–1150 TgC yr
Motor vehicles are an important anthropogenic source of not only VOCs but
also of particulate matter (PM) and of nitrogen oxides (
Gasoline vehicles can be divided into two groups based on fuel injection technologies in their engines, older-technology port fuel injection (PFI), and newer-technology gasoline direct-injection (GDI) vehicles. GDI vehicles have better fuel efficiency compared to PFI vehicles (Zhao et al., 1999). Due to having better fuel efficiency, GDI vehicles are becoming more popular, which necessitates more detailed research about GDI-vehicle emissions to better understand their impact on air quality through secondary air pollutant formation (Davis et al., 2015; Gentner et al., 2017).
Exhaust emissions from modern gasoline vehicles contribute to SOA in the atmosphere (Gentner et al., 2017; Karjalainen et al., 2016; Saliba et al., 2017; Zhao et al., 2018). The extent and details of this contribution remain a subject of ongoing research. In addition, there are gaps in our knowledge about how new technologies specifically, such as GDI engines, influence SOA precursor emissions (Gentner et al., 2017) – crucial information for planning future vehicle emission restrictions. Only a few studies have simultaneously explored SOA formation and SOA precursors emitted from GDI vehicles. The vehicle emissions, including SOA precursors, are dependent on several factors, such as emission standards of the vehicle, fuel injection technology, fuel used, and driving conditions (Platt et al., 2017; Peng et al., 2017; Zhao et al., 2018; Pieber et al., 2018; Du et al., 2018). Vehicles certified with stricter emission standards produce less SOA compared to vehicles certified with less-strict standards due to decreased emissions of nonmethane organic gas (NMOG), including SOA precursors (Gordon et al., 2014; Zhao et al., 2017, 2018). However, concurrent reduction in SOA formation is by a smaller fraction (Gordon et al., 2014; Zhao et al., 2017), highlighting important changes in the NMOG emission profile. Fuel injection technology may affect SOA precursor emissions from gasoline vehicles; e.g., single-ring aromatics (with low SOA mass yields) are important SOA precursors from PFI vehicles, while semi-volatile organic compounds (SVOCs) and intermediate VOCs (IVOCs; with higher SOA mass yields) are major SOA precursors from GDI vehicles (Du et al., 2018). However, other studies have shown no difference in SOA formation when comparing GDI and PFI vehicles in parallel (Saliba et al., 2017; Zhao et al., 2018). Aromatic content of used fuel greatly affects the emissions and SOA formation from gasoline vehicles as well. A recent study demonstrated that as an aromatic content of gasoline fuel increased from 29 % to 37 %, even 6-fold amplification of SOA production was observed (Peng et al., 2017). Moreover, driving conditions significantly affect the emissions and formed amount of SOA. For example cold-start and idling emissions are significantly higher, leading to greater SOA formation compared to hot-start or stable driving conditions with a heated engine (Weilenmann et al., 2009; Schifter et al., 2010; Nordin et al., 2013; May et al., 2014; Saliba et al., 2017).
Due to several factors affecting gasoline-vehicle emissions, more studies under atmospherically relevant conditions with the newest-generation GDI vehicles are required to better understand their effect on atmospheric chemistry and air quality. Previous studies with GDI vehicles where SOA formation and SOA precursor emissions were investigated simultaneously were conducted using different driving cycles representative of specific regions, such as Beijing, California, or Europe (Platt et al., 2017; Du et al., 2018; Pieber et al., 2018; Zhao et al., 2018). However, no studies exist where SOA formation and SOA precursor emissions are characterized when a modern GDI vehicle is driven at a constant load, although this kind of driving occurs constantly on larger roads.
In addition to anthropogenic SOA (ASOA) and
Previous studies have used various approaches to explore the impact of
anthropogenic emission on biogenic SOA formation and its climate-relevant
characteristics. There have been field studies in environments where anthropogenic and biogenic emissions are mixing, chamber studies with
relevant though typically simplified mixtures, and modeling approaches
(Goldstein et al., 2009; Carlton et al., 2010; Shilling et al., 2013; Emanuelsson et al., 2013; Hao et al., 2014). BVOCs can interact with anthropogenic emissions via several mechanisms, which are understood to varying degree and detail (Hoyle et al., 2011). For example, the increase in primary organic aerosol (POA) caused by anthropogenic emissions can enhance the partitioning of BVOC oxidation products to the particle phase if the anthropogenic POA forms a miscible phase with BVOC oxidation products (Pankow, 1994; Odum et al., 1996; Asa-Awuku et al., 2009). In addition,
anthropogenic emissions can increase the amount of acidic seed by contributing to the formation of sulfuric acid (
Several mechanisms can control biogenic–anthropogenic interactions, making it
a challenging subject of research. Moreover, as anthropogenic emissions are
highly complex mixtures of gas- and particle-phase species, they cannot be
thoroughly represented with a few model AVOCs and
This study characterizes GDI-vehicle-derived SOA formation from constant-load driving for the first time. For that we used the GDI vehicle with the
strictest emission standard certification in Europe (Euro 6) that was driven
at a constant load for different periods of time. Therefore, the results of
this study add valuable information about the SOA formation potential of GDI-vehicle exhaust, in particular when driving at constant load. In addition,
our research sheds light on important interactions between anthropogenic and
biogenic emissions, specifically using a gasoline vehicle and
Figure 1 shows the schematic of the experimental setup. In these experiments
a 29 m
Schematic (not to scale) of the experiment setup. The vehicle was
driven at constant load of 80 km h
Table 1 summarizes the instrumentation used in this study. Gas and
particle phases were monitored during the experiments. Gas-phase VOC monitoring was done using an IONICON proton-transfer-reaction time-of-flight
mass spectrometer (PTR-ToF-MS), and acidic gas-phase species were measured
using an Aerodyne acetate-chemical ionization time-of-flight mass spectrometer (acetate-ToF-CIMS). Furthermore, the following trace gases were
also monitored:
Instrumentation used in this study.
To achieve the research objectives described above, four kinds of
photochemistry experiments were conducted (see Table 2 for more details):
pure vehicle exhaust (Pure Vehicle),
Experimental conditions of the experiments conducted in this study.
Before an experiment was started, the chamber was cleaned by evacuating the
chamber and refilling it with humidified air that was produced from a zero air
generator (Model 737-15, Aadco Instruments Inc., USA). The chamber was
flushed with humidified air (Model FC125-240-5MP-02 Perma Pure, LLC., USA)
overnight to get the chamber ready for the experiment of the next day. Also
the relative humidity of the chamber was adjusted to
After the diluted exhaust was fed into the chamber
GDI-vehicle-emitted species prior to any adjustments or photochemistry period.
For some experiments, gasoline-vehicle exhaust was replaced by AS seed
particles (Sigma-Aldrich, 99 %) to determine the effect of GDI-vehicle
exhaust on the photochemistry of
Loss of particles to chamber walls causes significant errors when determining the amount of formed SOA from measurements if these losses are not corrected for. Therefore, the particle wall losses were determined after each experiment by monitoring the decrease in the organic mass fraction of particles due to wall losses after SOA formation inside the chamber had stopped. Particle mass losses to the chamber walls were corrected by calculating the aerosol mass loss rate constant according to Hao et al. (2011).
The GDI-vehicle exhaust was comprised of a complex mixture of species, but
only a subset of all these species, mostly VOCs, was detected by the
PTR-ToF-MS. From the detected VOCs, we identified and quantified the most
important SOA precursors, i.e., VOCs that contain at least one aromatic ring.
These VOCs were considered to be the most important SOA precursors because
they dominated concentration of SOA forming VOCs (
Calculating predicted SOA provided information about the missing fraction of SOA precursors emitted by the GDI vehicle. However, it should be noted that this analysis does not take into account possible non-linear interactions between VOCs. Anthropogenic VOC mixtures and potential non-linear interactions have not been systematically studied, and there is an obvious need for this kind of study, as it would enable the quantification of SOA formation from anthropogenic VOC emissions more accurately.
OH radical concentration and OH exposure were determined from the decay of
butanol-d9 over the length of the experiment. The OH concentration was
determined following the method presented by Barmet et al. (2012). OH exposure was then calculated from OH concentration by multiplying the instantaneous OH concentration by the length of the time averaging interval (TI), which was 1 min in this study. Eventually, OH exposure at time
VOCs were measured using the PTR-ToF-MS (PTR-TOF 8000, IONICON Analytik).
Many of the trace VOCs in the atmosphere, including those emitted by the GDI
vehicle, possess higher proton affinities than water and are consequently
detectable with a PTR-ToF-MS. A detailed description of the PTR-MS has been
given in several previous publications (Hansel et al., 1995; Lindinger et
al., 1998; Jordan et al., 2009; Blake et al., 2009). Therefore, only key
details of the PTR-ToF-MS are described in this paper. The PTR-ToF-MS consists of four main parts: (1) a hollow cathode discharge ion source, where
pure ionization of water vapor generates
The PTR-ToF-MS instrument settings remained unchanged throughout the
measurement campaign. The drift tube voltage and temperature were set to 600 V and 60
The PTR-ToF-MS mass calibration was performed using the protonated water isotope signal (
In the absence of fragmentation, sampled VOCs are measured as their
protonated parent molecules ([RH]
The size-resolved chemical composition and mass concentration of aerosol were measured by SP-AMS (Canagaratna et al., 2007; Onasch et al., 2012). The detailed operational procedure of our SP-AMS was described in previous publications (Hao et al., 2018). In brief, the SP-AMS was operated at 5 min saving cycles, alternatively switching between the electron ionization (EI) mode and SP mode. In the EI mode, only the tungsten vaporizer was used to measure non-refractory chemical species. In the SP mode, both the intracavity laser and tungsten vaporizers were switched on to measure laser-light-absorbing particles such as refractory black carbon (BC).
The AMS data were analyzed using the standard ToF-AMS data analysis toolkits SQUIRREL version 1.57I and PIKA version 1.16I in Igor Pro software (WaveMetrics, Inc.). For the mass concentration calculations, the default relative ionization efficiency (RIE) values 1.4, 1.2, 1.3, and 1.1 were applied for organic, sulfate, chloride, and nitrate, respectively. The RIE for ammonium and BC was 2.7 and 0.09, which was determined from the ionization efficiency calibration by using monodispersed pure ammonium nitrate and black carbon (REGAL 400R black pigment, Cabot Corp.), respectively. The dataset in the EI mode was analyzed for reporting the majority of the results in the paper and also the matrices generated in that mode were used to conduct positive matrix factorization (PMF) simulations. The data in the SP mode were used to report the BC.
The PMF technique was employed to perform further analysis on the
high-resolution mass spectra (Paatero and Tapper, 1994; Ulbrich et al., 2009). The PMF was evaluated with 1 to 7 factors and Fpeak from
An Aerodyne ToF-CIMS with an acetate ionization scheme was used to measure
acidic gas-phase species. The ToF-CIMS is described by Junninen et al. (2010), and the acetate ionization method is discussed extensively in earlier publications (Veres et al., 2008; Bertram et al., 2011; Aljawhary et al., 2013). Briefly, the sample air into the instrument through a 2 L min
Acetate reagent ions were generated by flowing 0.050 L min
The acetate-ToF-CIMS data post-processing was performed using the data analysis package “Tofware” (version 2.5.13) running in the Igor Pro environment.
This study investigated emissions and SOA formation chemistry from GDI
vehicles, including biogenic–anthropogenic interactions of GDI-vehicle
emissions related to SOA formation. In this section, first we characterize
the primary emission from the GDI vehicle and demonstrate the complex nature
of the exhaust. Next, we show that SOA formation from GDI vehicles was
observed in each experiment, but traditional SOA precursors, i.e., identified
aromatic VOCs, could not fully explain observed SOA formation. This suggests
that there were lower-volatility IVOCs and SVOCs in the GDI-vehicle exhaust
that likely contributed to SOA production but were not detected with the
instrumentation used in this study. Finally we demonstrate that the presence
of GDI-vehicle exhaust substantially suppressed
The GDI vehicle used in this study had the strictest emission standard
certification in Europe, Euro 6. Therefore, it only emitted a small amount
of primary particles; a major fraction of particle emissions was
comprised of BC, and only a minority of particle emissions was POA (Fig. 2a and Table 3). In contrast to the low particle emissions, GDI-vehicle exhaust comprised a significant amount of
The average particle-phase mass spectrum of pure vehicle exhaust measured inside the chamber right after the feeding period, before BL lamps were switched on (AMS data;
Figure 3 shows a time series of the gas- and particle-phase compounds during
a typical Pure Vehicle photochemistry experiment. Photochemistry started at
time 0. Different aromatic VOCs reacted with OH radicals at different
reaction rates (Fig. 3a). For example, the decay of toluene is substantially slower compared to the decay of xylene or trimethylbenzene (TMB). SOA precursors reacted with OH radicals to form gas-phase oxidation products, such as MEK,
Temporal evolution of GDI-vehicle-emitted VOCs, gas-phase oxidation products, and organic aerosol
We identified 20 aromatic species as ASOA precursors in the PTR-ToF-MS data,
and with Eq. (1) we calculated the predicted SOA mass generated by the
reactions of these species with OH radicals. Predicted SOA mass based on
these precursors could explain 19 %–42 % of the measured SOA (Fig. 4). In Fig. 4, the ASOA precursors were divided into six different groups based on elemental composition and number of carbon atoms: benzene, C8 aromatics, C9 aromatics, toluene, oxygenated aromatics, and other aromatics. Apparent from Fig. 4 (stacked bars) is the variability between individual
experiments, both in the different groups of SOA precursors and in the amount of SOA formed by their oxidative processing. The variations may simply be related to experiment-to-experiment variability in GDI-vehicle emissions. In any case, however, Fig. 4 demonstrates that we observed much more SOA (circles) than predicted based on the identified SOA precursors, implying that we were missing a substantial fraction of SOA precursors in our data analysis. Error bars of the predicted SOA (stacked bars) in Fig. 4 show the lower- and upper-bound estimates. These were calculated by taking the lowest and the highest reported SOA yield value of each aromatic VOC from the literature, reported under similar conditions compared to our study (see applied yields from Table S2). Hence, this gap between predicted and observed SOA cannot be explained by SOA mass yields applied to calculate predicted SOA values. Previously it has been shown that IVOCs alone create about 50 % of the formed SOA from the gasoline-vehicle exhaust (Zhao et al., 2016). The same observation was previously reported in a more recent publication, where comprehensive organic emission profiles for relatively new gasoline vehicles were constructed based on previous studies (Lu et al., 2018). The conclusion of that study was that IVOCs and SVOCs contribute approximately 50 % to the formed SOA from gasoline-vehicle exhaust even if VOCs clearly dominate the emission profile (Lu et al., 2018). Based on the results of these previous studies, we can assume that also in our study GDI vehicles emitted IVOCs and SVOCs, and these emissions made a substantial contribution to formed SOA. This result is consistent with other studies that have explored modern gasoline-vehicle emissions during driving cycles, which do not represent the constant-load driving that was used in this study (Gordon et al., 2014; Zhao et al., 2016, 2017). These IVOCs and SVOCs have high SOA mass yields. For example, for linear, cyclic, and branched 12-carbon alkanes, SOA mass yields can vary from 11 % up to 160 % under high-
Measured (red circles) and predicted (bars) SOA formed from the photochemistry of GDI-vehicle exhaust. High
The gap between measured and predicted SOA increases with exhaust feeding time; interestingly it does this much more strongly than the increase in predicted SOA (stacked bars). The vehicle heating procedure prior to the exhaust feeding period was the same in each experiment, and the vehicle was driven at a constant load for all experiments, so the longer feeding time corresponds to a longer driving time. A longer driving time decreased the predicted-SOA-to-measured-SOA ratio, indicating that there was a larger fraction of unaccounted SOA precursors in the chamber. As argued above, we attribute the missing SOA precursor to increased IVOCs and SVOCs that were not detected with the PTR-ToF-MS. This is an important and interesting finding, and to our knowledge, this was not reported earlier. Previous studies with relatively new gasoline vehicles have reported that IVOC and SVOC species together would contribute as much to formed SOA as VOC species (Zhao et al., 2016, 2017; Lu et al., 2018). Our results imply that the contribution of IVOCs and SVOCs to formed SOA is driving-time dependent, at least when the modern gasoline vehicle is driven at constant load. This dependence on driving time could be connected to the higher chamber wall loss rate of SVOCs and IVOCs compared to VOCs
Zhao et al. (2016) found that IVOC emissions as a fraction of NMOG emissions are enriched during the hot operation compared to the cold-start operation (Zhao et al., 2016). This may be due to higher combustion efficiency of more volatile fuel components, when the catalyst reaches its optimal temperature. Recently, Pereira et al. (2018) showed that IVOCs are removed less efficiently than VOCs by catalytic converters in diesel vehicles (Pereira et al., 2018). Similarly, we hypothesize that the catalysts in gasoline vehicles (like the vehicle used in this study) remove IVOCs and SVOCs less efficiently than lower-molecular-weight organics. This would essentially enrich the vehicle exhaust with IVOCs and SVOCs, which have oxidation products with higher SOA mass yields. Moreover, the condensation of IVOCs and SVOCs into tubing of the vehicle that is then released by outgassing when the vehicle is driven for longer time could explain the increasing fraction of IVOCs and SVOCs from total emissions as a function of driving time. However, this must be studied more in future. For example, future studies must characterize the emitted SVOCs and IVOCs to see if their chemical composition changes throughout the driving period or if the same compounds are emitted by the vehicle with increasing quantity that would support the hypothesis about the condensation of IVOCs and SVOCs into tubing of the vehicle and into transfer lines (Pagonis et al., 2017; Deming et al., 2019). In our case the feeding time dependence of IVOC and SVOC losses (or outgassing) may be affected by time-dependent changes in tubing and transfer line temperature caused by the hot exhaust. Any induced errors would lead to an underestimation of emission-induced SOA yields, which means that chamber studies in general underestimate SOA yields, and formed SOA mass, from vehicle exhaust due to line losses of these high-yield SOA precursors.
The results of this study offer new useful information from GDI-vehicle emissions during constant-load driving that have not been considered in
previous studies. To our knowledge, this is the first time that the chemistry of modern gasoline-vehicle emissions were explored during constant-load driving and with a preheated engine. This is in contrast to studies using cold-start emissions combined with breaking and acceleration periods – conditions which generally produce higher emissions from the vehicle but do not necessarily represent realistic emissions for most vehicles on the road. This difference in study design makes it impractical to compare the absolute mass of SOA formed between studies. However, predicted SOA offers a more suitable value for comparison purposes, i.e., to evaluate if our results are in agreement with previously published results of GDI and PFI vehicle emission studies under controlled conditions. In those previous studies, emitted VOCs identified as SOA precursors have contributed from less than 20 % up to 100 % to the measured SOA (Platt et al., 2013; Nordin et al., 2013; Gordon et al., 2014; Liu et al., 2015; Peng et al., 2017; Du et al., 2018; Pieber et al., 2018). Our results (19 %–42 %) fall within that reported range even if the driving conditions and emission standard certification of the vehicles were different between the various studies. However, some studies have reported a much smaller “missing source” of SOA precursors. For example, 60 % of observed SOA mass generated from idling-phase emissions could be accounted for by the reactions of C
Predicted SOA is greatly affected by SOA yield values (low- or high-
The largest uncertainty in predicting SOA mass is caused by the missing yield values of some SOA precursors identified from the exhaust. In this study, predicted SOA was calculated using literature SOA yields for all aromatic VOCs using studies with the most similar aerosol mass concentration as observed in this study because SOA production is influenced by the gas-particle partitioning (Odum et al., 1996). Unfortunately, SOA yields for all identified SOA precursors were not available, and in those cases, SOA yields of other compounds with similar structures were used as a proxy. In this study we made every effort to include the full range of possible SOA mass yield values, which provides increased confidence that the ranges reported here actually represent a lower and upper bound for the predicted SOA. In spite of these uncertainties in predicting SOA mass, our calculations provide useful information about the missing fraction of SOA precursors and its relative contribution to SOA formation in different experiments during this study.
In summary, the following reasons can explain the significant differences between predicted and measured SOA. First, as explained above, uncertainties in the calculations to predict SOA may cause some of the difference. However, two more significant issues are that (1) modern GDI vehicles emit SOA-forming IVOCs and SVOCs that have high SOA yields and cannot be detected by the PTR-ToF-MS and (2) the photooxidation of VOCs in a more complex mixture may also result more complex reaction pathways and product distributions and hence an altered SOA yield compared to pure single precursor experiments (Song et al., 2007). To better understand SOA formation from vehicle exhaust, it is crucial that instrumentation that can measure IVOCs and SVOCs be included in future experimental setups. Identifying those compounds will also help us understand why their fractions of total emissions increase with increasing constant-load driving time.
Vehicle-exhaust-derived SOA formation complicates the estimation of
In the first method we applied PMF analysis to AMS data to quantify the
fraction of formed SOA originating from anthropogenic precursors. As shown
in Fig. S1 in the Supplement and explained above in Sect. 2.4.2, PMF analysis conducted for all experiments produced a four-factor solution. Previously it was shown that gasoline-vehicle-emitted IVOCs and POA are correlated (Zhao et al., 2016). Therefore, based on the observed correlation between IVOCs and POA, we estimated the amount of formed ASOA from the linear fit of the HOA factor as a function of the mix_SOA_LVOOA factor shown in Fig. S2. We justify this approach by the strong observed contribution of IVOCs to formed ASOA, and the fraction of ASOA (out of total SOA) should be strongly reflected in the mix_SOA_LVOOA factor. Hence, by using PMF results of the four
Pure Vehicle experiments to create a linear fit, we were able to use the equation of that fit (
In the second method we used a reacted concentration of a single ASOA
precursor with an elemental composition of
In the third method, we first calculated effective SOA yield of the vehicle
exhaust using Eq. (3):
With the effective SOA yield, we estimated the formed ASOA in the Mixed experiments based on the total concentration of reacted ASOA precursors measured with the PTR-ToF-MS. For this method we assumed that the distribution of isomers of each PTR-ToF-MS detected elemental composition in the exhaust did not change between the experiments because different isomers undergo different fragmentation inside the PTR-ToF-MS, which interferes with their quantitation, but they would have different SOA mass yields. Based on previous studies this was a reasonable assumption (Schauer et al., 2002; Gueneron et al., 2015; Schmitz et al., 2000). The average effective SOA yield of Pure Vehicle 1 and Pure Vehicle 2 experiments was used in this calculation as an effective ASOA yield value for Mixed experiments. The average effective SOA yield was applied because the feeding times of the vehicle exhaust in Mixed experiments were between the feeding times of Pure Vehicle 1 and Pure Vehicle 2 experiments (see Table 2).
From all these three methods, independent of each other, we obtained
estimates of the formed ASOA in the Mixed experiments that were relatively
close to each other (Table 4). This agreement gives confidence that the
average ASOA determined from these three methods represents the formed
ASOA in Mixed experiments well. As Table 4 shows, in the Mixed experiments,
Vehicle-derived SOA formed in Mixed experiments estimated with different approaches described in Sect. 3.3. RSD is relative standard deviation.
In this study our second objective was to explore the effect of anthropogenic emissions on biogenic SOA formation. To study this, we used gasoline-vehicle exhaust and
Figure 5 shows that in the presence of gasoline-vehicle exhaust
Figure 6 shows the evolution of
When
Figure S4 demonstrates that in Mixed experiments we observed the formation
of oxidation products that were absent in Pure vehicle experiments and Pure
Our results imply that gas-phase species emitted by gasoline vehicles may play the most important role in the potential second anthropogenic effect by changing the reaction pathways of
It is good to bear in mind that wall losses of SOA-forming vapors may lead to underestimation of SOA yields during chamber experiments (Kokkola et al., 2014; Zhang et al., 2014, 2015; Yeh and Ziemann, 2015; Platt et al., 2017). For this reason the SOA mass yields and the amounts of formed SOA presented in this paper may be lower limits. By studying yields as a function of initial surface area of particles, one can gain insight into the effect of vapor wall losses on SOA mass yield, as shown by Fig. 5; i.e., the observed yields are dependent on the condensation surface area of the seed particles present inside the chamber during the experiment. Obviously, the seed surface area available for the vapors to condense on affects the yield due to competition for vapors between chamber wall surface and particle surface. However, this does not change the conclusion of this work that two distinct mechanisms caused by anthropogenic emissions suppress biogenic SOA formation because all experiments were conducted using the same chamber and similar initial surface area of particle loadings.
This study had two main objectives. First we studied SOA formation chemistry from emissions of modern GDI vehicles during constant-load driving, representing typical freeway-driving conditions. Second, we explored the complex chemical interactions between anthropogenic and biogenic emissions by using GDI-vehicle exhaust and
This study also demonstrated potentially two mechanisms by which anthropogenic emissions suppressed biogenic SOA formation – the well-established
Data are available upon request from Annele Virtanen (annele.virtanen@uef.fi).
The supplement related to this article is available online at:
EK and AV designed the study. EK, LH, AY, AB, AL, PYP, IN, KK, and JJ conducted the experiments. EK, LH, AY, AB, IN, CLF, SS, and AV participated in data analysis and/or interpretation. EK wrote the paper. LH, CLF, SS, and AV edited the paper.
The authors declare that they have no conflict of interest.
This article is part of the special issue “Simulation chambers as tools in atmospheric research (AMT/ACP/GMD inter-journal SI)”. It is not associated with a conference.
Otso Peräkylä and Mikael Ehn from INAR (Institute for Atmospheric and Earth System research) are thanked for the help they provided with acetate-ToF-CIMS operation during the measurement campaign.
This work was financially supported by the European Research Council (ERC Starting grant no. 335478) and the Academy of Finland Center of Excellence program (grant no. 307331). This work has received funding from the European Union’s Horizon 2020 research and innovation program through the EUROCHAMP-2020 Infrastructure Activity (grant no. 730997). The work of Eetu Kari was financially supported by University of Eastern Finland Doctoral Programme in Environmental Physics, Health and Biology.
This paper was edited by Gordon McFiggans and reviewed by two anonymous referees.