Signal variation in single particle aerosol mass spectrometry

Abstract

Rapid and accurate detection of airborne micro-particles is currently an important problem in national security. One approach to such detection, bioaerosol mass spectrometry (BAMS), is currently under development at Lawrence Livermore National Laboratory. BAMS is a type of single particle aerosol mass spectrometry that rapidly records dual-polarity mass spectra of aerosolized micro-particles. However, the accuracy of the BAMS system is limited by various uncertainties, resulting in shot-to-shot variations in the mass spectra. I found that the variations in mass peak areas in BAMS spectra were significantly larger than those predicted by Poisson statistics based on the mean number of detected ions. Furthermore, these variations were surprisingly consistent as a function of peak area among synthetic, organic, and biological samples. For both positive and negative ions, the standard deviation in a peak's area was approximately proportional to the mean value of that area to the 0.9 power. Using the consistency of this data, I also developed a novel method for quantitatively evaluating the similarity between mass spectra using a chi-square factor. Peak area variations in other single particle aerosol mass spectrometers may be similarly analyzed and used to improve methods for rapid particle identification.