A multiple charging correction algorithm for scanning electrical mobility spectrometer data

• Need for multiple-charge correction of SEMS data is examined.
• Errors associated with the existing multiple-charge correction approach is established.
• New approach for multiple-charge correction of SEMS data without impactor in place.
• The performance of the new approach is validated with simulation results and experiments.

Accuracy of particle size distribution measurements from scanning electrical mobility spectrometers (SEMS) is critically dependent on the quality of SEMS data inversion. A critical element of SEMS data inversion is the consideration of the charged particle fraction in the sample flow. In particular, when larger, multiply-charged particles are present in the measured aerosol, significant errors in the calculation of size distributions are possible if their contribution is not correctly accounted for. While ignoring the contribution of multiply charged particles may be acceptable when the particles are mostly in the ultrafine size range, a substantial error is possible when a significant fraction of particles are larger than 100 nm. Accurate calculation of size distributions from SEMS data is possible when an inertial impactor is used to eliminate the contribution of multiply-charged particles larger than a cut-size and an iterative multiple-charge correction (MCC) algorithm is used during the data inversion process. The effectiveness of this approach is, however, strongly dependent on the relationship between the aerodynamic and the electrical mobility diameters of the particles. Here, we demonstrate the limitations of the existing inversion algorithms and propose an alternative MCC algorithm for size distribution calculation from non-ideal SEMS data. In the proposed approach, the zeroth order singly charged particle size distribution is fit using a Gumbel distribution function, and the resultant fit is used to correct for the multiply-charged contribution to the SEMS data. The effectiveness of the proposed approach is tested for a range of particle size distribution scenarios and validated with experimental data.