Colloidal gold clusters formation and chemometrics for direct SERS determination of bioanalytes in complex media

In this work, we report the sensitive and selective sensing of the purine bases adenine and guanine in urine matrix by using surface-enhanced Raman spectroscopy (SERS) and a colloidal SERS substrate. To identify suitable conditions for quantitative analysis, the pH dependence of spectra of adenine, guanine, urine simulant and their mixtures was studied on gold nanoparticles suspension. Interestingly, although the urine matrix promotes the analytes signal suppression and overlapping bands, it can also cause an improvement in repeatability of the SERS measurements. This effect was associated to the relatively controlled formation of small-sized gold clusters and it was investigated both experimentally and theoretically. Furthermore, a correlation constrained multivariate curve resolution-alternating least squares (MCR-ALS) method was developed to resolve overlapping SERS bands and to quantify physiologically relevant (micromolar) concentrations of the bioanalytes. The performance of the proposed MCR-ALS approach (assessed in terms of figures of merit) was similar to that obtained by using partial least squares regression, but with the additional advantage of retrieving valuable spectral information. Therefore, this method can be used for improving selectivity of colloidal clusters in qualitative and quantitative SERS analysis of complex media, avoiding the need for tedious nanoparticle-surface modification or preliminary chromatographic separation.

Keywords

Correlation constraint, Mie theory, Multivariate curve resolution, Partial least squares, Surface-enhanced Raman spectroscopy, Theoretical calculations