Multi-laser-power calibration for quantitative determination of N and P in fertilizers by Raman spectroscopy

Abstract

A new method called multi-laser-power calibration (MLPC) is described for quantitative analysis by Raman spectroscopy. MLPC is based on the premise that, under controlled operating conditions, the intensity of Raman spectroscopy signals for a given analyte is directly proportional to the laser power applied to generate that signal. Different from traditional methods, in which calibration is performed by varying analyte concentrations while keeping instrumental parameters fixed, MLPC uses a single calibration standard and different laser applied powers to generate several calibration points. Samples and standard solution are measured under identical conditions, and multiple scattering intensities (I) are generated by varying the Raman's laser applied power (P). The MLPC curve is built with I values for the calibration standard (I^std) on the x axis and the respective values for the sample (I^sample) on the y axis. The analyte concentration in the sample (C^sample) is then calculated from the calibration curve slope (m). As a proof-of-concept, the MLPC method was applied for the determination of N and P species in fertilizer and raw material samples using a 30 s exposition time and laser applied powers in the 35–319 mW range. Calibration curves with correlation coefficients greater than 0.9986 were consistently obtained. The method's accuracy was evaluated by analyzing three certified reference materials of fertilizer. MLPC results were in agreement with reference values at a 95% confidence level (t-test). Ten commercially available fertilizer and raw material samples were also analyzed, and the results were compared with values obtained by both Raman spectroscopy using external standard calibration (EC-Raman), and official methods based on Kjeldahl and spectrophotometry. Limits of quantification (dry mass, % w/w) for phosphate, phosphite, urea, and nitrate were calculated as 0.03, 0.09, 0.04, 0.05, respectively. Relative standard deviations (RSDs) were 10%, 7%, 11%, and 3% (9%, 4%, 9%, and 2% for EC), respectively, for these analytes. Total N and total P determination by Kjeldahl and spectrophotometry had RSDs of 4% and 10%, respectively. The MLPC method using 5000 mg L⁻¹ single calibration standards was successfully used to speciate N and P in fertilizer analyses.