Abstract:

Confocal Raman micro-spectrometry combined with surface enhanced Raman scattering (SERS) was used for
collecting surface enhanced Raman spectra of apple samples containing the pesticide malathion. Stepwise multiple linear
regression (SMLR) method was used to build two mathematical models for quantitative analysis based on golden and silver
sols, respectively. The results showed that within the linear range of 0.01–0.5 mg/kg, the correlation coefficient of the gold
sol model was 0.999 6, with root mean square error of calibration (RMSEC) of 0.001 86 mg/kg, and the maximum difference
between the actual and calculated values was less than 0.003 mg/kg. The correlation coefficient of the silver sol model was 1,
with RMSEC of 0.000 78 mg/kg, and the maximum differences between the actual and calculated value did not exceed 0.001 2 mg/kg.
The prediction standard deviations for the golden and silver sol models were 0.213 mg/kg and 0.146 mg/kg, respectively.
These results indicate that the effect of silver sol in quantitative analysis of malathion is better than that of gold sol.

Key words: surface enhanced Raman scattering (SERS), stepwise multiple linear regression (SMLR), golden(Au) sol, silver (Ag) sol, malathion