Chemical Fingerprinting of Biomass Burning Organic Aerosols from Sugar Cane Combustion: Complementary Findings from Field and Laboratory Studies

Abstract

Agricultural fires are a major source of biomass-burning organic aerosols (BBOAs) with impacts on health, the environment, and climate. In this study, globally relevant BBOA emissions from the combustion of sugar cane in both field and laboratory experiments were analyzed using comprehensive two-dimensional gas chromatography time-of-flight mass spectrometry. The derived chemical fingerprints of fresh emissions were evaluated using targeted and nontargeted evaluation approaches. The open-field sugar cane burning experiments revealed the high chemical complexity of combustion emissions, including compounds derived from the pyrolysis of (hemi)cellulose, lignin, and further biomass, such as pyridine and oxime derivatives, methoxyphenols, and methoxybenzenes, as well as triterpenoids. In comparison, laboratory experiments could only partially model the complexity of real combustion events. Our results showed high variability between the conducted field and laboratory experiments, which we, among others, discuss in terms of differences in combustion conditions, fuel composition, and atmospheric processing. We conclude that both field and laboratory studies have their merits and should be applied complementarily. While field studies under real-world conditions are essential to assess the general impact on air quality, climate, and environment, laboratory studies are better suited to investigate specific emissions of different biomass types under controlled conditions.

Figure 1

Quantification of gas phase components, namely, n-alkanes, (methylated) PAHs, and BTX from GCB gas phase tubes by TD–GC–MS for (A) field experiments and (B) laboratory experiments. Concentrations (ng m^–3) were adjusted for blank and background levels and normalized to the corresponding sampling volume. Error bars represent the standard deviations (laboratory: n = 9; field: n = 5). The break in the y-axis at 1 × 10³ ng m^–3 (only in plot A) indicates a nonlinear scale to better visualize low concentration levels in the bar chart.

Figure 2

Semiquantification of the targeted reference particle phase compounds (see Table S12) from the (a) sugar cane burning field campaign in South Africa and the (b) laboratory combustion experiments. Each bar corresponds to the average concentration of a specific compound derived from TD–GC × GC–TOFMS measurements of the filter samples. Concentrations (ng m^–3) were derived from a 4-point calibration using the internal standard compound fluorene-D10 and normalized to the corresponding sampling volume (see Figure S4). A tentative assignment of compounds to their respective molecular compositions was done with a NIST mass spectral library search (match quality ≥70%) and retention indices. Error bars represent the standard deviation values (laboratory: n = 9; field: n = 5).

Figure 3

Bar charts illustrating the relative abundance of different compound classes for the 100 compounds with the highest signal intensities from TD–GC × GC–TOFMS measurements. The classified most abundant peaks contribute to 53.8 and 55.7% of the total TIC signal in the laboratory and field measurements, respectively. The tentative assignment of compounds to their respective molecular compositions was based on a NIST mass spectral library search (match quality ≥70%) and retention indices. (a) Categorization into eight chemical compound classes (n-alkanes, PAHs, furan derivatives, methoxyphenols, monosaccharide derivatives, phytosterols, triterpenoids, and unidentified compounds) based on the likewise classification of the previously discussed targeted reference compounds (see Figure 2). The bordered red bar illustrates the remaining compounds that could not be classified into these groups (other). (b) Subclassification of these “other” compounds into aromatic (blue), noncyclic (yellow), heterocyclic and heteroaromatic (red), and cyclic (green) compounds. The respective pattern illustrates a subclassification of the chemical functionalities.

Figure 4

Volcano plot depicting the differential abundance of chemical compounds found by TD–GC × GC–TOFMS measurements of background and stationary filters from the field campaign, excluding unknowns, standard substances, and substances found in the background filter (sampled before fire initiation). The statistical significance (p-values from the Welch two-sample t-test) and fold changes were calculated using TIC areas, which were corrected for the respective sampling volume and normalized using fluorene-D10 as the internal standard (one-point normalization). The horizontal lines indicate the significance thresholds of p-values of 0.05 (black, dashed) and 0.01 (black). Compounds which surpass the p-value threshold of 0.01 (red rectangular markers) and exhibit increased abundance in comparison to the background filter measurements, as indicated by the vertical lines showing different fold change ranges, are discussed. The compound names for peaks 1–49 are provided in the Supporting Information (Table S13).