NUMERICAL-ANALYSIS OF ESR-SPECTRA FROM ATMOSPHERIC SAMPLES

Improvements of the matrix isolation/electron spin resonance technique for the measurement of NO2, NO3, and RO2 radicals in the atmosphere are described. The use of D2O instead of H2O as the matrix yields a better spectral resolution and, as a consequence, larger a signal-to-noise ratio as well as better identification of the different species. Reference spectra of the different radicals in H2O and D2O matrices are compared. While a large degree of correlation exists amongst the spectra of the different (unsubstituted and substituted) alkylperoxy radicals, the spectra of HO2, CH3C(O)O2, and NO3 show significant differences that allow their distinction in atmospheric samples. A numerical procedure for the analysis of the composite ESR spectra obtained from atmospheric samples was developed. Subtraction of the dominant NO2 signal is performed by matching a reference NO2 spectrum with respect to amplitude, spectral position, and line width to the sample spectrum. The manipulations are performed with the virtually noise-free reference spectrum and are based on physical information. The residual spectrum is then analyzed for RO2 (and/or NO3) by simultaneously fitting up to six different reference spectra. The method was applied to laboratory samples as well as to atmospheric samples in order to demonstrate the ability of retrieving small amounts of HO2 in the presence of large amounts of NO2 and other peroxy radicals. The new algorithm allowed, for the first time, the identification of the HO2 and CH3C(O)O2 radical in tropospheric air at concentrations ranging up to 40 ppt. © 1990 Kluwer Academic Publishers.