Evaluation of an in situ sampling probe for its accuracy in determining particle size distributions from flames

An in situ sampling probe for flames is evaluated for any biases with respect to size distributions in a given size and concentration range. Ultrafine metal particles (Pd, Pt) suspended in argon gas are added to a laminar diffusion flame, burning with argon-diluted methane. Samples of combustion aerosol are extracted from the flame via a small orifice and immediately quenched with nitrogen. The size distribution of the metal particles, recorded at various sampling positions in the flame, is found to be but little affected by the flame and the sampling procedure. The distribution of carbonaceous soot particles from an unseeded flame at a fixed sampling position is not shifted upon seeding but coexists next to the peak from the metallic flame additive. It is concluded that droplets of volatile combustion products, formed in the sampling system through homogeneous nucleation, do not significantly contribute to the soot particle signal. Ar. investigation of the photoelectric yield of the samples implies significant changes of the surface properties of the metal particles during the combustion process. This seeding method offers a new path to study chemical processes in the flame, especially heterogeneous and catalytic reactions of flame gases with particles. (C) 1997 Elsevier Science Ltd. All rights reserved.