Rapid prediction of caffeine in tea based on surface-enhanced Raman spectroscopy coupled multivariate calibration

Abstract

This study was focused on the quantitation of caffeine in black tea by surface-enhanced Raman spectroscopy coupled with gold nanoparticles. Caffeine has its own importance in tea due to its significant role against cardiovascular diseases and many other benefits. Caffeine was predicted for the first time as low cost and rapid by surface-enhanced Raman spectroscopy (SERS) coupled chemometrics in black tea. Gold nanoparticles (AuNPs) were synthesized successfully with high enhancement factors as SERS substrate used for SERS detection coupled partial least squares (PLS) algorithms. Caffeine exhibited several SERS characteristic peaks after adsorption on AuNPs owing to electromagnetic enhancement while excited by laser excitation. Quantification of caffeine in black tea was predicted using four build models, PLS, synergy interval-PLS (Si-PLS), genetic algorithm-PLS (GA-PLS), and Si-GA-PLS on preprocessed spectral data by standard normal variate (SNV). The better results were noted by using Si-GA-PLS while latent variables, (LVs) was 5, the correlation coefficient of calibration (RC) = 0.9705 where root mean square error of cross validation (RMSECV) = 0.114% and correlation coefficient of prediction (RP) = 0.9233 where root mean square error of prediction (RMSEP) = 0.165% and residual predicted deviation (RPD) was noted 2.43 and relative standard deviation (RSD) for precision was recorded as ≤3.42%. Based on the predicted results it is obvious that the purposed AuNPs nanosensor coupled Si-GA-PLS model could be successfully employed for caffeine prediction in tea with high sensitivity and rapidity.