Source apportionment of organic gas and particulate matter pollutants using concentration and flux measurements in Delhi

One of the world’s worst cities for air pollution is Delhi, India. Past studies have shown air pollution in Delhi has a significant burden on the population’s respiratory and cardiovascular health. It is therefore important to analyse the sources of these pollutants in order to improve air quality mitigation strategies. This thesis presents time-resolved source apportionments of the organic fraction of particulate matter (PM) and of the gaseous volatile organic compounds (VOCs) that were measured in Delhi. The results of this analysis are used to identify the sources of these air pollutants and determine their atmospheric implications. Online aerosol mass spectrometry (AMS) measurements taken in Old Delhi during pre-monsoon, monsoon and post-monsoon seasons revealed concentrations of submicron particulate matter (PM1) reaching ~750 μg m-3 The largest contributing inorganic species to PM1 in the post-monsoon is nitrate (8%) but this changes to sulphate during the monsoon (24%) and premonsoon (24%). PM1 is dominated by the organics throughout the three measurement periods (54% - 68%). Source apportionment on the organics fraction using Positive Matrix Factorisation (PMF) revealed that PM1 traffic emissions are the largest contributor of primary organic aerosol (POA) in the pre-monsoon (12%) and monsoon (16%) periods. Two traffic factors were resolved: a hydrocarbon-like organic aerosol (HOA) and another HOA factor rich in nitrogen (NHOA). The N within NHOA is previously undocumented within AMS measurements and is found to be linked to nitrile species. The two traffic factors are found to have separated due to fuel-type where NHOA is related to heavy goods vehicles (HGVs) and diesel emissions while HOA is related to petrol and compressed natural gas emissions. The highest PM1 concentrations are observed in the post-monsoon period during which the highest form of POA is from a burning-related factor. These factors alone contribute 35% to the total post-monsoon increase when concentrations increase by 188%. A high correlation between one burning factor, semi-volatility biomass burning organic aerosol (SVBBOA), and Earth observation fire counts in surrounding states demonstrates its link to crop residue burning. Another burning factor, solid-fuel OA (SFOA), is found to have a high composition of polyaromatic hydrocarbons (PAHs) and novel AMS-measured marker species for polychlorinated dibenzodioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs). SFOA is therefore linked to local open-waste burning. There is also a 522% increase in chloride concentrations from pre-monsoon to post-monsoon. High correlations with SVBBOA and SFOA, suggest the post-monsoon increase in chloride is due to crop residue and open-waste burning. Future air quality mitigation strategies should concentrate on traffic emissions in order to cause a reduction PM1 across the whole year. To reduce peak PM1 concentrations during the post-monsoon, requires targeting of burning-related activities in future air quality policy. The monitoring and measurements of volatile organic compounds (VOCs) in Delhi is often overlooked. While the effects of VOCs on people’s heath is still being debated, the outcome of their interaction with the atmosphere can play a pivotal role in PM formation, OH reactivity and photochemical ozone creation potentials (POCP). For the first time, a novel approach to source apportionment was developed which allows for emission and deposition to be calculated for positive matrix factorisation (PMF) VOC source factors. Spatial flux patterns for PMF factors are presented, which give in-depth detail of the localised fluxes. A total of nine factors were resolved, of which traffic emissions are the largest (70% of total flux) and are the most significant source of OH reactivity flux and photochemical ozone creation potential (POCP). The two traffic factors separate based on the HGV restrictions causing the composition of one factor (TRAF1) to have a strong connection with diesel emissions and the other (TRAF2) with petrol and compressed natural gas. The second most significant source of emissions is an evaporative VOC (EVOC) source (10% of total flux) which has a high composition of marker species related to asphalt emissions. These are found to be naphthalene and tetralin-based structures, along with 1,3-butadiene and styrene which are constituent species of the adhesives used in asphalt. The peak in emissions of EVOC occurs during midday and follows surface temperature and solar radiation. Due to their similar meteorological controls, most of the biogenic VOCs reside within the EVOC factor. It is estimated that biogenic emissions contribute 25% to total monoterpene concentrations and 18% to the localised monoterpene fluxes. However, the traffic factors have the highest composition of monoterpenes and, as a result, they dominate the monoterpene contributions to concentrations (60%) and localised fluxes (78%). Two burning-related factors were separated, one associated with solid-fuel combustion VOCs (SFVOC) and another associated with pyrotechnical activity VOCs (PVOC). The two factors are shown to share the largest percentage of the concentration-weighted OH reactivity. This is caused by a high content of furan-based species. The first observation of a strong urban deposition flux for PMF factors is also seen for SFVOC and an oxidised VOC (OVOC1) factor. Spatial deposition flux patterns suggest vegetation could act as a VOC sink in Delhi, however, the magnitude of which could show it to be small. Over recent years, multiple PM mitigation strategies have been implemented in Delhi, but, how effective these changes have been to reduce PM is not clear. It is however possible to chart the changes in PM by analysing large archives of PM filter samples that have been collected over multiple years. An automated offline-AMS method was developed which has enabled high-throughput analysis, allowing PM concentrations to be charted over multiple years. This novel offline-AMS method uses an organic solvent mix of acetone and water which can extract high quantities of OA (95.4 ± 8.3 %). PM10 filter samples collected in Delhi for the years of 2011, 2015 and 2018, were analysed. PMF analysis was performed on the organic fraction and nine factors were resolved. These factors can be grouped into four source categories: cooking, traffic, coal-combustion and burning-related (solid fuel or open burning) OA. Burning-related OA is the highest contributor when total OA concentrations are also at their highest, during the winter and post-monsoon. As a result, burning-related OA likely contributes significantly to acute PM-related health affects in Delhi. Annual averages show a decline in burning related OA concentrations from 2015 to 2018 (47%). This could be linked to the 2015 ban on open-waste burning; however, compositional analysis of OA factors suggests municipal waste burning is still present in 2018. Further mitigation strategies are therefore required to reduce burning-related OA further. The shutdown of the two coal power stations in Delhi, along with initiatives to reduce the popularity of coal-use in businesses, residential homes, and industry, has led to a significant decrease (87%) in coal-combustion OA. Total OA concentrations were also predicted to be reduced by 17%, therefore, proving these measures brought effective reductions in PM10. The Bharat stage emissions standards for vehicles likely supressed PM10 OA concentrations as the annual averages of traffic OA factors do not reflect the coinciding rapid increase in population and registered vehicles. However, restrictions on HGVs entering during the day shows possible links to increases in PM10 concentrations during the winter and post-monsoon months. This is likely due to the large influx of diesel-engine HGVs entering the city when the boundary layer is particularly low during the early mornings and evenings