Particle size inversion of flowing aerosol using Bayesian inference in dynamic light scattering.

Area of Science:

• Aerosol science
• Optical physics
• Statistical modeling

Background:

• Dynamic light scattering (DLS) is crucial for aerosol characterization.
• High flow velocities complicate accurate particle size distribution (PSD) determination using classical inversion methods.
• Ill-conditioned inversion equations in DLS hinder precise measurements with increasing flow velocity.

Purpose of the Study:

• To develop a robust method for accurate PSD retrieval in flowing aerosols.
• To overcome limitations of classical inversion methods at elevated flow velocities.
• To enable simultaneous online retrieval of PSD and flow velocity.

Main Methods:

• Established a probability model using Bayesian inference.
• Derived a posterior probability density function (PDF) for PSD parameters.
• Employed Markov chain Monte Carlo (MCMC) algorithm for parameter sampling and averaging.

Main Results:

• The Bayesian method accurately determines PSD, avoiding artifacts like peak shifts or broadening seen in regularization inversion.
• Measurement errors in PSD remain stable irrespective of increasing flow velocity.
• Successfully retrieved flow velocity information from the intensity autocorrelation function.

Conclusions:

• Bayesian inference offers a superior approach for DLS analysis of flowing aerosols.
• The developed method enhances accuracy and reliability in dynamic aerosol characterization.
• Enables real-time monitoring of both particle size distribution and flow velocity.