DEVELOPMENT OF A STABLE SPECTRAL SIGN FOR REAGENT-FREE MONITORING OF NITRATES IN COMPLEX MATRICES

Abstract

Direct ultraviolet (UV) spectroscopy is a promising method for reagent-free express monitoring of nitrate ions (NO3 ).. This study details a multi-wave optical system based on UV light-emitting diodes (265, 308, 365, 405 nm). The initial analysis of the "Base Model," calibrated on distilled water, demonstrates excellent performance with a high coefficient of determination (R= 0,9994). However, this model proves to be completely inadequate for analyzing real samples with complex matrices ("dirty" water), where its error increases to unacceptable 300 mg/l. This complete failure of the simple model under real conditions motivates the further search for robust features.

As an intermediate step, a differential feature is analyzed   D А = А 310 -А365, , which successfully compensates for the additive shift caused by turbidity. This success becomes the basis for the development of this work's main innovation—the local curvature feature based on a spectral triad  (LC205 -310-365 ). This feature, calculated using a formula that estimates the deviation of the central peak from the baseline, demonstrates high selectivity in mixed matrices, simultaneously compensating for both organic chromophores (CDOM) and scatterers (turbidity). The local curvature feature is a robust, computationally efficient, and physically interpretable solution for reagent-free monitoring of nitrates in real conditions, serving as a logical continuation of previous research that employed machine learning methods. Keywords: optics, nitrates, in liquid, ultraviolet spectroscopy, multichannel analysis, differential method, spectral curvature, control, water quality, water, automatic monitoring, photometry, humic substances, turbidity, optical measurements, non-destructive testing.