Yaghobian F, et al. Lab on a Chip, 2011, 11(17), 2955-2960.
This work presented an on-chip method for biomarker quantification based on isotope dilution surface-enhanced Raman scattering (IDSERS). The entire process was implemented on a few square millimeters of SERS-active substrate. The experimental results showed that SERS spectroscopy can be used to generate and validate PLS models and predict creatinine concentrations of unknown serum samples. To illustrate the effectiveness of the proposed method for analyzing a variety of potential biomarkers, urea was examined as a secondary example.
Internal cross-validation results obtained for a PLS model generated from 20 urea/13C,15N2-urea mixture spectra:
· Initial experiments revealed that no SERS spectra were detectable when using colloidal nanoparticle suspensions, suggesting that urea does not spontaneously adhere to metal surfaces. In these instances, it is necessary to position the analyte molecules close to the nanostructures through drop coating deposition, which can only be accomplished on flat nanostructured substrates.
· The SERS spectra of natural urea show a notable peak at 1002 cm-1, while the corresponding vibration for 13C, 15N2-urea occurs at 980 cm-1. A gradual transition between the two isotopologues is evident across the spectral series. Due to the overall signal intensity being lower than that of creatinine under the same experimental conditions, the number of reference mixtures for the PLS model was increased. Additionally, despite a larger wavenumber shift of 22 cm-1 for the isotopically enriched urea, a comparable model prediction uncertainty was achieved, reflected in an RMSECV of 3.4%.
Hansen E S S, et al. Magnetic Resonance in Medicine, 2016, 76(6), 1895-1899.