Biomass burning is a global source of climate- and health-affecting emissions. The impacts of biomass burning emissions (BBE) are tied to their complex and variable chemical makeup. For instance, the nitrogen content of BBE influences their capacity to absorb light, and therefore affect the Earth's radiative budget. Factors such as temperature, biomass type, or air flow rate during the combustion all modify the composition of BBE, making accurate characterization challenging. Herein, for the first time, principal component analysis (PCA) was applied to emissions gathered during laboratory-based combustion of wood and cow dung biomass in a tube furnace. A thermal desorption two dimensional time-of-flight gas chromatography mass spectrometry (TD-GC × GC-ToF-MS) setup was employed to separate and identify chemical species. By combining these techniques with a feature selection algorithm, we determined that low temperature and air flow rate lead to greater feature separation on PCA scores plots. Of the 729 variables used to construct the plots, 61 were identified as significant. These species - including sugars such as d-Allose and melezitose, as well as tracers such as levoglucosan and guaiacol - significantly differentiated emissions from wood versus cow dung biomass, especially at lower temperatures. In particular, combustion of either fuel at 0.2 slpm and 500 °C, lead to 20 times the variability in levoglucosan peak area over more efficient furnace parameters. Chemical species evolved only from dung burning contained on average 0.595 nitrogen atoms versus 0.515 for wood, indicating that a higher nitrogen content of the base fuel may not necessarily translate into emission of unique nitrogen containing species, potentially causing the underestimation of dung burning impacts. Overall, TD-GC × GC-ToF-MS coupled to PCA reliably separated emissions from wood and dung biomass while simultaneously identifying significant chemical features, displaying the suitability of this combination of techniques towards characterizing complex BBE matrices in the future.
Read full abstract