Refinement of the Aethalometer Dual Wavelength Technique for Real-time Measurement of Woodsmoke PM: Saturation Compensation and Potential Contributions of Organic PM from Cooking Sources

GEORGE ALLEN (1) Jay Turner (2) Pete Babich (3)

(1) NESCAUM (2) Washington University in St. Louis (3) CT Dept. of Environmental Protection

     Abstract Number: 390
     Last modified: November 9, 2009

Abstract
Woodsmoke-related PM has been identified over the years using various methods, ranging from fine-mode non-soil potassium to levoglucosan and molecular markers. All these techniques are integrated samples, and some are either not specific or not stable over time. The two-wavelength Aethalometer (TM) has been used recently to provide semi-quantitative real-time ambient woodsmoke PM measurements. The difference between the 880 and 370nm channels, UVC minus BC (“Delta-C”), is specific to biomass combustion and in cold winter climates is nearly all driven by woodsmoke PM. The Delta-C response is reasonably stable for homogeneous woodsmoke that has not undergone substantial photochemical processing, but can vary substantially across plumes from individual sources. Results from a recent study in CT show a consistent relationship across six sites across the state between woodsmoke PM as identified using Unmix and the Aethalometer Delta-C signal. The high time resolution of these data show temporal patterns that are consistent with wood heat activity patterns, and provide indications of short-term peak exposure concentrations. Winter woodsmoke has been observed with this technique in substantial amounts in many northern urban areas, and can contribute 20 to 25% of the winter urban PM2.5 during the coldest months of the year even in non-valley cities.

It has recently been shown that sample-spot loading (shadowing) effects can generate substantial and different errors in both Aethalometer channel measurements, but predominantly in the UV-C channel. This results in a substantial negative DC artifact from non-woodsmoke BC sources such as diesel engines, as well as a variable woodsmoke response across different sample-spot loadings. Data processing techniques have recently been developed and implemented in post-processing programs to remove the majority of these saturation artifacts, improving the quality of the Delta-C method for woodsmoke. Techniques for optimizing the hardware and software parameters for saturation compensation are presented. Work characterizing PM emissions from meat cooking and hot oil smoke (frying) show that these aerosols can have enhanced optical absorption at shorter wavelengths, but the cooking signal is weaker compared to woodsmoke.

Nearly all delta-C woodsmoke measurements to date have been made using the AE21 series of Aethalometer. The present model (AE22) has not been carefully evaluated for this use, and there are limited and preliminary indications that some aspects of the current Aethalometer may not provide a similar Delta-C woodsmoke-specific signal. Collocated comparisons across Aethalometer models and at different measurement timebases are presented to investigate this possible change in performance.