The Raman spectrum of the MM standard solution only displays peaks attributed to toluene, the solvent matrix. No mercaptan-specific peaks are observed, indicating that a concentration of 2,000 mg/L (ppm) is too low for detection by standard Raman spectroscopy (Figure 2a). However, after analysis with Ag P-SERS, the Raman band at 675 cm⁻¹ associated with S–C stretching becomes detectable even at 100 ppm (Figure 2b) and was observable down to 0.05 ppm (50 ppb; Figure 2c). This result suggests that SERS enables the detection of mercaptan at trace levels significantly below the ASTM limit of 30 ppm [3].

A calibration curve for low-concentration MM was developed and validated against samples measured by different data collection methods (Figure 3). With an R² of 0.975, the model effectively captures the relationship between peak intensity and concentration. The model’s PRESS (predicted residual error sum of squares) value of 0.0632 is high in order to distinguish subtle concentration changes between 0.00 and 0.05 ppm (50 ppb) but is sufficient for differentiating samples with increments of >0.05 ~ 0.10 ppm. The calibration model predicts the concentration of validation sets with good accuracy, achieving an R² of 0.962 and a PRESS value of 0.053. The validation curve was adjusted for bias and slope using Vision software, optimizing sample validation. These results confirm that MIRA XTR, coupled with Ag P-SERS substrates, can be effective for quantitative analysis of low-concentration MM.

A calibration curve plateau above 1 ppm (data not shown) suggests that the adsorption efficiency of MM on the Ag substrate declines at higher concentrations. This, combined with the low detection limit, indicates a high affinity of mercaptans for the Ag P-SERS substrate. Thus, dilution may be required to accurately quantify higher mercaptan concentrations, such as fuels containing 30 ppm mercaptans.