The statistical convergence of aerosol deposition measurements

The spatial distribution of deposited aerosol particles shows substantial statistical variation. This variation is especially important for sampling techniques having a high spatial resolution, such as laser-based methods, or for experiments involving a small sample size (few identical locations measured) or a short duration. This paper develops a criterion to ensure adequate statistical convergence in deposition measurements. The time evolution of statistical variations is modeled using standard Poisson statistics; this model provides a bound on measurement uncertainties. The model is compared to experimental measurements of the spatial variation of deposition. Numerical simulations are used to examine the effects of aerosol polydispersity on statistical convergence. Both the experiments and the numerics show excellent agreement with the model. List of symbols Roman letters A area of one measurement nN sample mean number of particles deposited N number of measurements in a sample n number particles deposited in area A during time T n sample standard deviation of number of particles deposited v sample standard deviation of voltage from LIF deposition measurements T time duration of deposition process vN sample mean voltage from LIF measurements Greek letters j particle deposition rate per unit area k n true mean number of particles deposited in an area A during a time T p n true standard deviation of number of particles deposited in an area A during a time T Received: 17 July 1996/Accepted: 17 October 1996 J. C. Moran, J. H. Lienhard V Warren M. Rohsenow Laboratory Department of Mechanical Engineering Room 3-162 Massachusetts Institute of Technology Cambridge, MA 02139-4307, USA Correspondence to: J. H. Lienhard V