Ambient characteristics of elemental carbon and black carbon in the Pearl River Delta and the application of elemental carbon as a combustion aerosol tracer
by Wu Cheng
Ph.D. Environmental Science
xviii, 195 pages : illustrations (chiefly color), color maps ; 30 cm
Black carbon (BC) or elemental carbon (EC) aerosols is a class of refractory and light absorbing fine particulate matter originating from incomplete combustion activities. BC is not only an air pollutant that has an adverse health impact on human beings, but is also a vital climate forcer on both the global and regional scale due to its light absorption ability. Better understanding of BC is needed before effective mitigation measures can be formulated in the Pearl River Delta (PRD) region of China. Existing studies are either ad-hoc studies limited by spatial coverage or have temporal limitation. This thesis work aims to improve the understanding of BC in terms of analytical uncertainties, temporal and spatial distribution, and optical properties. For this study, multiple analytical te...[ Read more ]
Black carbon (BC) or elemental carbon (EC) aerosols is a class of refractory and light absorbing fine particulate matter originating from incomplete combustion activities. BC is not only an air pollutant that has an adverse health impact on human beings, but is also a vital climate forcer on both the global and regional scale due to its light absorption ability. Better understanding of BC is needed before effective mitigation measures can be formulated in the Pearl River Delta (PRD) region of China. Existing studies are either ad-hoc studies limited by spatial coverage or have temporal limitation. This thesis work aims to improve the understanding of BC in terms of analytical uncertainties, temporal and spatial distribution, and optical properties. For this study, multiple analytical techniques were employed for characterizing BC and the main findings are summarized below:
(1) Evaluations of the two most widely used carbon analyzer models and analysis protocols are performed by analyzing EC and organic carbon (OC) from PRD samples. The two instruments agree well but discrepancy is found between the two protocols, which is much larger than previous reports on US samples. Temperature ramping parameters and laser correction methods are identified to be the dominant factors leading to the protocol discrepancy. Degree of charring is found to positively correlate with the discrepancy. These results suggest that OC and EC determined by different protocols are not comparable and extra care should be taken for data interpretation from different protocols.
(2) Observation of light absorption coefficients (σabs) and BC concentrations by aethalometers were conducted at five sites in the PRD region in two periods: 16 May-20 June 2008 during the rainy season and 12 December 2008- 8 January 2009 during the dry season. Little difference in BC level was found at a rural site in the PRD between the two sampling periods (2.88 and 2.62 μg m-3). At the PRD urban sites, the daily average concentration of BC ranged from 1.56 to 37.9 μg m-3, higher in the dry season sampling period (12.31 μg m-3) and lower in the rainy season sampling period (6.17 μg m-3). At the urban/suburban PRD sites, BC was observed to have a diurnal pattern of higher concentrations at night and at the urban sites a small peak in the early morning rush hour, implying the significance of vehicular emission.
(3) Due to the refractory nature and its exclusive combustion source origin, EC tracer method has been extensively used for distinguishing primary and secondary OC but seldom assessed. Key to the EC tracer method is to identify an appropriate OC/EC ratio that represents primary combustion emission sources (i.e., (OC/EC)pri) at the observation site. A numerical study is performed to evaluate the accuracy of secondary organic carbon (SOC) estimation by various OC/EC)pri calculation schemes, including minimum OC/EC method, percentile (10%) OC/EC method and minimum R squared (MRS) method. Different scenarios are considered, including single primary source, two independent primary sources, and two correlated primary sources scenario. The results show among the three (OC/EC)pri methods, the MRS method consistently produces the most accurate SOC estimates. Factors affecting the accuracy of SOC estimates are also identified through the numerical experiments.
(4) Application of EC tracer method is conducted on one year hourly OC and EC data (1 February 2012 to 31 January 2013) of a suburban site Nancun (NC) in the city of Guangzhou, China. Results show that an interference of EC co-emitted SOC precursor can cause (OC/EC)pri overestimation in the afternoon, and diurnal minimum MRS (OC/EC)pri is adopted to minimized this artifact. In addition, seasonal variations of (OC/EC)pri are observed (e.g. elevated (OC/EC)pri in the biomass burning season), therefore (OC/EC)pri is calculated for individual months to account for such variations. In winter, SOC exhibits higher concentrations, but in summer higher formation activity is observed as evidenced by an elevated SOC/OC ratio. SOC by resolved carbon fractions are also investigated. SOC contributed 38% of low temperature OC and 78% of high temperature OC, implying that most of the high temperature OC (probably more oxygenated) are secondarily formed. Results from 49 individual rain events show that carbonaceous particle mass was reduced 50% on average due to wet removal. The removal rate is uniform for POC, SOC and EC, suggesting that the mixing state of carbonaceous particles is probably predominated by internal mixing at the sampling site. Higher EC/CO ratio was observed at night time, suggesting the high EC at night is associated with shallow mixing depth as well as higher intensity of primary emissions from diesel trucks, which is consistent with the traffic regulations in the city of Guangzhou.
(5) Optical properties of BC are explored by both numerical study and field measurement. Core-shell Mie simulation is performed to investigate the possible range of the light absorption enhancement factor (Eabs). Considering the typical size distribution of BC particles in the PRD, the upper limit of Eabs is estimated at ~2. The annual average mass absorption efficiency (MAE) at NC is 19.75 m2g-1. MAEpri is proposed to represent the MAE of soot particles without coating. Linear regression and Minimum R Squared (MRS) method are employed for MAEpri estimation. The annual average estimated MAEpri by linear regression and MRS is 13.16 and 12.7 m2g-1 respectively, implying an annual average enhancement factor of 1.50~1.55. Seasonal variation of MAE exhibit much higher enhancement in summer. Enhancement apportionment shows that in the rainy season the enhancement is mainly associated with SO42- and NO3-, while in the dry season SOC, SO42- and NO3- are all responsible for the enhancement. BC concentrations from an aethalometer and EC from a RT-ECOC analyzer are compared and the causes of discrepancy are discussed. Retrieval of the light absorption coefficient (σabs) from the RT-ECOC analyzer is demonstrated, and the results are shown to be consistent with the aethalometer.