Carbonaceous Particulate Matter Emitted from a Pellet-Fired Biomass Boiler

Biomass pellets are a source of renewable energy; although, the air pollution and exposure risks posed by the emissions from burning pellets in biomass boilers (BBs) are uncertain. The present study examines the organic species in fine particle matter (PM) emissions from an BB firing switchgrass (SwG) and hardwood (HW) biomass pellets using different test cycles. The organic and elemental carbon (OC and EC) content and select semivolatile organic compounds (SVOCs) in filter-collected PM were identified and quantified using thermal-optical analysis and gas chromatography–mass spectrometry (GC–MS), respectively. Fine PM emissions from the BB ranged from 0.4 g/kg to 2.91 g/kg of pellets burned of which 40% ± 17% w/w was carbon. The sum of GC–MS quantified SVOCs in the PM emissions varied from 0.13 to 0.41 g/g OC. Relatively high levels of oxygenated compounds were observed in the PM emissions, and the most predominant individual SVOC constituent was levoglucosan (12.5–320 mg/g OC). The effect of boiler test cycle on emissions was generally greater than the effect due to pellet fuel type. Organic matter emissions increased at lower loads, owing to less than optimal combustion performance. Compared with other types of residential wood combustion studies, pellet burning in the current BB lowered PM emissions by nearly an order of magnitude. PM emitted from burning pellets in boilers tested across multiple studies also contains comparatively less carbon; however, the toxic polycyclic aromatic hydrocarbons (PAH) in the PM tested across these pellet-burning studies varied substantially, and produced 2–10 times more benzo[k]fluoranthene, dibenz[a,h]anthracene and indeno[1,2,3-c,d]pyrene on average. These results suggest that further toxicological evaluation of biomass pellet burning emissions is required to properly understand the risks posed.


Figure S1
Error! Bookmark not defined.

Figure S2
Error! Bookmark not defined.

Figure S3
Error! Bookmark not defined.

Figure S4
Error! Bookmark not defined.

Figure S5
Error! Bookmark not defined.

Table S1
Error! Bookmark not defined.

Table S2
Error! Bookmark not defined.

Emission factor calculations 11
Dilution Factor 11 Figure 1. Study-wide distributions, quantiles, and summary statistics for OC-EC concentrations in the PHBB emissions. In the outlier box plot, the confidence diamond contains the mean and the 95% confidence interval about the mean. The red bracket outside the box is referred to as the 'shortest half' or the densest 50% of the observations.
Mean Conc (µg/g OC)  All values are presented as averages and represent a wide variety of power outputs. The solid line at 5.2 g/kg is the study-wide mean; the median was 2.5 g/kg. .

Emission factor calculations
For each test run, emission factors for the target pollutants were calculated in terms of mass of fuel burned. The mass emission (Mx) for a set burn time t for each target compound is calculated in the dilution tunnel as follows: Where: Cx,t = the concentration (mass/volume) of the target compound x in the dilution duct, Cx,a = the ambient concentration (mass/volume) of the target compound x, and V = the volumetric flow rate (volume/time) in the dilution tunnel at time t. Most concentrations were determined on a test average basis. The ambient concentration for total PM utilized an ambient air sample extracted beside the boiler during 100% load operation firing switchgrass pellets. Remaining concentrations were corrected with concentrations determined from dilution duct sampling with no firing (i.e. cold) of the boiler.
Volumetric flow rate was determined by multiplying the average of the dilution duct velocity measured before and after each test by the cross-sectional area of the dilution duct at the point measured. The 10-inch duct had a cross-sectional area of 0.0506 m 2 (0.545 ft 2 ).
Emission factors were calculated and reported per mass of fuel burned. The emission factor per mass of fuel burned (EFm,x) is calculated as: ,

Dilution Factor
Measurements made at the stack or after secondary dilution were corrected to dilution duct concentrations using a dilution factor. Due to failure in the CEM measurements, estimates were made for CO2 concentrations in the dilution duct for use in dilution factor calculations. The average CO2 in the dilution duct was estimated based on the mass of fuel burned in each test (Mf), the carbon concentration in the fuel from the ultimate analysis, and the volumetric flow rate determined for the test. The calculation estimates the volume of CO2 emitted divided by the volume of flow in the dilution duct corrected for ambient CO2 concentration: Where: CCO2,t = the CO2 concentration in the duct, CCO2,a = the ambient concentration of CO2, %C = the carbon concentration in the fuel (weight percent), MWCarbon = the molecular weight of carbon, and SV = the specific volume of an ideal gas at 20 °C and 1 atmosphere. Concentrations measured at the stack or on secondary dilution were corrected to dilution duct concentrations for emission calculations.
The dilution factor at the stack were based on the estimated dilution duct CO2 concentration and the average stack CO2 concentration: The dilution used for secondary dilution was performed with nitrogen containing minimal CO2: Where: SDCO2,t = Secondary dilution CO2 concentration during sample interval Because the nature of operations and the short-term nature of the measurements taken on secondary dilution, each measurement was corrected using time specific dilution factors. Due to the quality of duct CO2 measurement, time specific dilution duct CO2 concentrations were estimated using the average dilution duct CO2 measurements for the specific sampling time, average dilution duct CO2 concentrations for total test, and the average fuel based estimate of dilution duct CO2 concentrations calculated as above: