Performance Comparison of TSI's New Wide-Range and Established Long Differential Mobility Analyzer
Justin S. Koczak, ANDREA J. TIWARI, Torsten Tritscher, Sebastian H. Schmitt, Timothy Wright, Markus Petters,
TSI Incorporated Abstract Number: 107
Working Group: Instrumentation and Methods
AbstractThe TSI Long Differential Mobility Analyzer (Long DMA, Model 3081/3081A) has become an established reference instrument in aerosol sizing (Kinney et al 1991). They have been in use for more than 40 years and have been employed in a variety of applications, including atmospheric chemistry (Cocker et al 2001) and ambient aerosol monitoring (Birmili et al 2016).
The TSI Long DMA is considered a general-purpose DMA for many different applications. Some ambient air quality researchers, particularly in the European community, however, have instead used the DMA design originally developed by the University of Vienna (Winklmayer et al 1991). Notably, TROPOS in Leipzig introduced a DMA based on the Vienna design. The “TROPOS DMA” has since become established in the ambient monitoring literature, and is a now a widely-respected instrument within the European ambient monitoring community (e.g. ACTRIS and GUAN). In response to growing interest in ambient monitoring research, TSI has recently introduced the Wide-Range DMA (W-DMA, Model 3083), which is based on the TROPOS design, to expand the applications of the Scanning Mobility Particle Sizer™ Spectrometer (SMPS, Model 3938).
The Long DMA and the TROPOS-designed DMA have an overlapping size range. While transfer function statistics for the TROPOS DMA and Long DMA have been investigated in prior work (Birmili et al 1997), an in-depth performance comparison between the new W-DMA and the Long DMA has not yet been presented.
This work examines the size accuracy of the W-DMA for monodisperse particles, comparing theoretical computations to experimental work. Modeling of DMA transfer was performed using an open-source model (Petters 2021). The model predicts the output of a hypothetical tandem DMA setup, where DMA 1 selects quasi-monodisperse particles and DMA 2 scans over a range of mobility. The model assumes the diffusion-broadened theoretical DMA transfer function computed based on the dimension and flow rates in each DMA. The normalized concentration corresponds to the ratio of particle concentration measured by a condensation particle counter after DMA 2 divided by the particle concentration after DMA 1. Example model calculations for a sheath flow rate of 5 L min-1 and over a range of particle sizes show that for particles 50 nm and larger, the transfer functions are nearly identical between the two DMA models. We anticipate presenting observational data that shows the degree to which either DMA model conform to this theoretical expectation.