Air pollution and traffic: Searching for the missing emissions

Dr Jacqueline F. Hamilton
Wolfson Atmospheric Chemistry Laboratories, University of York
jacqui.hamilton@york.ac.uk
ECG Bulletin July 2017

View and download slides

Since the U.S. Environment Protection Agency issued a Violation Notice of the Clean Air Act to the car manufacturer Volkswagen in 2015, nitrogen dioxide (NO₂) emissions from diesel engines have received considerable public attention (1). Diesel cars also emit other pollutants, including volatile organic compounds (VOCs). Recent studies suggest that models underestimate the contribution to VOCs from diesel cars.

Over the last 10 years, diesel fuel usage has increased dramatically worldwide. Exxon Mobil predicts that diesel fuel will be the number one transport fuel globally by 2040 (2). In the UK, the fraction of newly registered vehicles using diesel fuel has risen from less than 10% in 1991 to almost 50% in 2015. Recent measurements from London suggest that emission inventories underestimate the emission flux of NO₂ by 30 to 40%, mainly as a result of under-representation of road traffic emissions (3).

Although many diesel cars pass emission testing during stringent but predictable laboratory tests, emissions of NO₂ can be many times higher under normal driving conditions (4). Diesel vehicles use a range of different emission control technologies to oxidise or remove harmful pollutants such as carbon monoxide, particulate matter, nitrogen oxides (NOx) and volatile organic compounds (VOCs). The latter two are controlled together as NOx + VOC, and new vehicles must meet the EURO VI emission standard of 170 mg/km. Mounting evidence suggests that diesel vehicle emissions of NOx are too high under normal driving conditions, but what about the other pollutants? And should we care?

VOCs play a key role in urban atmospheric chemistry. Their oxidation can lead to the formation of additional toxic pollutants, including ozone and particulate matter. The UK’s national atmospheric emission inventory includes diesel VOC emissions, but they are predicted to be small compared to those from gasoline and natural gas burning. Measurements of VOCs in the atmosphere can be difficult, because the concentrations are very low (parts per trillion) and a wide range of volatilities and polarities are present. The diesel fraction is particularly challenging to measure as a result of increasing isomeric complexity as carbon number increases.
 
Gasoline vapours are usually analysed with gas chromatography, but the higher carbon number in diesel leads to poor resolution and an unresolved complex mixture. Comprehensive two-dimensional gas chromatography (GC×GC) is a high-resolution technique that couples two GC columns of different selectivity. A modulator concentrates discrete bundles of eluent from the first column and injects them onto the second column, where a very fast separation takes place (<10 seconds). This approach gives higher peak capacity and sensitivity and provides structured chromatograms.
 
The University of York has developed a field portable GC×GC instrument to study the role of diesel VOC emissions. The instrument was deployed during the Clean Air For London (Clearflo) project in North Kensington in 2012 (5). This collaborative project involved multiple institutions from across the UK and was funded by the Natural Environment Research Council. Hourly VOC measurements were made using the GC×GC to study VOCs with 6 to 13 carbons, and a dual channel GC system was used to study VOCs with 1 to 7 carbons (6). In the example GC×GC chromatogram shown in Figure 1, each spot represents an individual compound. As the carbon number increases along the x-axis, the complexity of the sample also increases; by the end of the analysis, there are too many isomers to allow each individual spot to be isolated. The structured nature of the GC×GC contour space allows us to group similar compounds (such as all aliphatic compounds with 12 carbons) and quantify their contribution to the atmosphere. The results showed that although diesel-related VOCs only represented 20 to 30% of the total hydrocarbon mixing ratio, they comprised more than 50% of the atmospheric hydrocarbon mass and are a dominant local source of secondary organic aerosols. Thus, emissions from diesel vehicles can dominate gas-phase reactive carbon in cities with high proportions of diesel vehicles and, in London, were predicted to contribute up to 50% of the ozone-production potential.

Figure 1. Example GC×GC-FID chromatogram from London during winter. The x- and y-axes show retentions on the first and second columns, respectively, with the intensity of the compound shown by the coloured contours. The number of carbon atoms in molecules increases along the x-axis. The areas labelled 1 to 8 represent aliphatic groups from C6 to C13. The other labels are (9) benzene, (10) toluene, (11) C2 substituted monoaromatics, (12) C3 substituted monoaromatics, (13) C4 substituted monoaromatics, (14) naphthalene, and (15) C10 monoterpenes.

Comparing these urban air measurements with the UK emission inventory shows a substantial under reporting of diesel-related hydrocarbons—an underestimation of a factor 4 for C9 species, rising to a factor of over 70 for C12 during winter. Laboratory and ambient measurements give different results for the importance of diesel versus gasoline emissions in the formation of secondary organic aerosols, a type of particulate matter (7). To determine the impact of the diesel VOCs observed in London, this additional source needs to be added to emission inventories used in air quality modeling. Ots et al. used the EMEP4-UK model to study the impact of diesel VOCs on secondary organic aerosol production in the UK (8). They added the diesel emissions as pentadecane and used the same emission profile as in London for every European country. Addition of diesel VOCs improved the model-measurement agreement for SOA, with diesel emissions accounting, on average, for 30% of the annual SOA produced in the model.
 
Further work is needed to understand the factors that control VOC emissions from diesel vehicles. The COM-PART (Combustion Particles in the Atmosphere, Properties, Transformation, Fate and Impact) project aims to improve understanding of diesel exhaust emissions and what happens after they are released. This NERC funded project, involving the Universities Manchester, York and Birmingham, will study the emissions of a diesel engine under different real world driving conditions. The exhaust from a 1.9 L VW diesel engine is injected into an atmospheric simulation chamber (Manchester Aerosol Chamber). The chamber has a series of lights, which can be turned on to simulate different sunlight conditions. Initial results show that the temperature of both the engine and the emissions controls are key factors in determining not only the amount of VOC but also the composition of emissions.

These studies highlight that emissions of VOCs from diesel engines can significantly impact other toxic pollutants in London and other cities across Europe where diesel use is high. As diesel use has risen, the impact on VOC emissions has been largely unnoticed due to the lack of meaningful measurement infrastructure. The control of NO₂ from modern diesel vehicles entails significant policy challenges for many developed cities. There may also be a similar, but currently unrecognized, policy challenge to controlling reactive carbon emissions and their contributions to secondary pollutants.

References

1. https://www.epa.gov/vw/documents-related-volkswagen-violations-model-years-2009-2016.
   
2. Exxon Mobil, Outlook for Energy: A view to 2040 (2016).
   
3. A. R. Vaughan et al., Faraday Discussions 189, 455 (2015).
   
4. D. C. Carslaw , G. Rhys-Tyler, Atmospheric Environment 81, 339 (2013).
   
5. S. I. Bohnenstengel et al., Bulletin of the American Meteorological Society 96, 779 (2015).
   
6. R. E. Dunmore et al., Atmospheric Chemistry and Physics 15, 9983 (2015).
   
7. D. R. Gentner et al., Environmental Science and Technology 51, 1074 (2017).
   
8. R. Ots et al., Atmospheric Chemistry and Physics 16, 6453 (2016).
   

Banner image: Rush hour traffic on a city road. Credit: Repina Valeriya/Shutterstock