INTRODUCTION: Turmeric has been widely used in curry powders as the main spice. Conventional chemical analysis such as high-performance liquid chromatography (HPLC) may take several hours to extract curcuminoids and prepare samples in many turmeric processing industries. OBJECTIVE: This study was conducted to evaluate curcuminoids in turmeric powder using near-infrared reflectance spectroscopy (NIRS). METHODS: All spectral acquisition ranged from 1100 to 2500 nm and a chemometrics analysis using partial least-squares (PLS) regression was performed to quantify the contents of individual curcuminoids. The HPLC was carried out (n = 129) to develop a PLS model based on the reference values. RESULTS: High correlation coefficient (R(2) > 0.93) and low standard error of cross-validation (SECV < 0.20 g/100 g) and standard error of prediction (SEP < 0.13 g/100 g) values were obtained for precision and accuracy. In addition, the ratio of prediction to deviation (RPD > 2.65) values was also calculated. CONCLUSION: Our results indicate that NIRS could be utilised as a control procedure or as an alternative rapid and effective quantification method.
INTRODUCTION:Turmeric has been widely used in curry powders as the main spice. Conventional chemical analysis such as high-performance liquid chromatography (HPLC) may take several hours to extract curcuminoids and prepare samples in many turmeric processing industries. OBJECTIVE: This study was conducted to evaluate curcuminoids in turmeric powder using near-infrared reflectance spectroscopy (NIRS). METHODS: All spectral acquisition ranged from 1100 to 2500 nm and a chemometrics analysis using partial least-squares (PLS) regression was performed to quantify the contents of individual curcuminoids. The HPLC was carried out (n = 129) to develop a PLS model based on the reference values. RESULTS: High correlation coefficient (R(2) > 0.93) and low standard error of cross-validation (SECV < 0.20 g/100 g) and standard error of prediction (SEP < 0.13 g/100 g) values were obtained for precision and accuracy. In addition, the ratio of prediction to deviation (RPD > 2.65) values was also calculated. CONCLUSION: Our results indicate that NIRS could be utilised as a control procedure or as an alternative rapid and effective quantification method.