Raman Spectroscopy for Nitrate Detection in Water: A Review of the Current State of Art

Abstract

The contamination of natural basins by agricultural or industrial activities, and the growing need for potable water due to climate changes accelerate the drive to find versatile, fast, practical, and easy-to-use methods for water analysis. A potentially versatile technique suitable for water analysis is Raman Spectroscopy (RS). Featured by good resolution but low sensitivity, RS detects molecular vibrational modes of an analyte in water. Nitrate is an indicator of chemical and/or biological pollution, it displays Raman active vibrational modes affected by the interaction with other systems in solution, allowing a wide range of applications. Concerning Nitrate analysis in water, a general introduction to the Raman effect and the basic instrumentation were herein discussed. RS is a potential solution to wastewater analysis. This review first reports the theoretical background of the technique and its basic working principles, then, the state-of-the-art scientific contributions related to Nitrate detection are investigated with a particular interest in the instrumental setup and the chemometric techniques employed to improve its sensitivity. In the studies hereby considered, instrumental setup (for example, laser frequency, laser power, acquisition times) and different technical solutions (for example, micro- versus macro-Raman instruments) to increase the technique’s sensitivity on Nitrate detection are described. Concisely, the use of deep-UV lasers, optically active Surface-Enhanced Raman Spectroscopy (SERS) or Fiber-Enhanced Raman spectroscopy (FERS) equipment, coupled with instrumental settings, i.e. acquisition time, variable temperature of acquisition, use of special sampling apparatus (cuvettes or immersion probes), or with ion exchange resins for analyte enrichment, have been reported. Remarkably, examples of large data correction of unwanted fluorescence by mathematical processing or chemical quenching were reported too, suggesting solutions for the Raman analysis of wastewaters. Finally, a short digression on Machine Learning (ML) applied to RS is proposed, showing the promising results reported in other fields. Data-driven methods could be a solution to improve the low sensitivity of the RS for Nitrate detection. Hence, an approach of ML methods for the typical RS spectra processing (spike removal, baseline correction, fluorescence curve elimination, instrumental noise correction) was hereby mentioned, suggesting an improvement in the detection capability of Nitrate ion in water.