AAAR 34th Annual Conference
October 12 - October 16, 2015
Hyatt Regency
Minneapolis, Minnesota, USA
Abstract View
Development of Multivariate Calibration Approach for Measurement of Aerosol Elemental Concentration Using Microplasma Spectroscopy
LINA ZHENG, Pramod Kulkarni, M. Eileen Birch, Dionysios Dionysiou, Centers for Disease Control and Prevention, NIOSH
Abstract Number: 538 Working Group: Instrumentation and Methods
Abstract A method for conducting multivariate calibration for elemental measurement of aerosols using microplasma spectroscopy is presented. Conventionally, the univariate approach is widely used for calibration of spectroscopic methods to determine elemental concentration. However, bias from matrix effects can be significant depending on the characteristics of the plasma and the nature of matrix containing the analyte. The measured concentration of target element can depend on the presence of other elements and their abundance in the aerosol particle. This can sometimes lead to large uncertainties in the measured concentration of elements using an univariate approach. The objective of this study was to investigate effectiveness of multivariate calibration approaches, such as partial least square regression (PLS) and principal components regression (PCR), to account for matrix effects and thereby minimize measurement uncertainties. A large training set was developed consisting of 25 orthogonal aerosol samples with 9 factors (elements: Cr, Mn, Fe, Ni, Cu, Zn, Cd, Pb, Ti) and 5 levels (elemental concentrations). Elemental concentrations in the aerosolized samples were measured with Aerosol Spark Emission Spectrometer (ASES) at a time resolution of 1 minute. Simultaneous filter samples were collected at every 30 minutes for determination of elemental concentration using inductively coupled plasma mass spectrometry (ICP-MS). Calibration curves were developed using PLS and PCR algorithms with the mass spectrometry measurements as the reference values. The performance of the multivariate calibration model was then tested for near real-time prediction of concentration of Cr, Mn, Fe, and Ni in the aerosol emitted from welding processes in an actual workplace. We will present the comparison with simultaneous ICP-MS measurements, and discuss implications for uncertainty of measured elemental concentration using ASES.