Abstract View
Influence of the Peaks Method Assumption on Variation in Volatility Tandem Differential Mobility Analyzer Measurements
CHRISTOPHER OXFORD, Brent Williams, Washington University in St. Louis
Abstract Number: 367
Working Group: Instrumentation and Methods
Abstract
There is a significant amount of variation in Volatility Tandem Differential Mobility Analyzer (V-TDMA) aerosol measurements. As an example of such discrepancies, four authors have published V-TDMA volatility measurements for adipic acid (Tao 1989; Bilde 2003; Selah 2003; Salo 2010). The percent difference between the largest and smallest volatility measurement is: 129 % (surface energy), 142% (vapor pressure at 298 K), and 45% (enthalpy). Bilde et al. (2015) observed that the differences between the studies were often larger than the errors within each study. Thus, exploring the differences between the experiments could yield the cause of the variation.
The oven used in each V-TDMA was different, and each experiment used different experimental settings (aerosol-to-sheath ratio, etc.) to obtain the raw data and used different computational methods to solve for the volatility variables. In contrast, every study used the peaks method assumption to reduce the distributions to single values. In this assumption, the initial diameter is assumed to be the diameter at the peak of the size distribution entering the oven; the final diameter is assumed to be the diameter at the peak of the size distribution exiting the oven; and the residence time is assumed to be the residence time at the peak of the residence time distribution. Theoretically, differences in the experimental settings and oven designs should interact with the peaks method assumptions, which provide inputs to the different calculation methods.
In this work, we use a model to evaluate two different calculation methods (solving for vapor pressure, while assuming a surface energy value, versus simultaneously solving for both variables) used by V-TDMAs. We show that different V-TDMA settings and oven designs generate errors through the peaks method assumptions, and these errors can be amplified by the choice of calculation method.