Development of the near infrared spectroscopy model for detecting herbicide concentrations contaminated in water

This research aims to develop the near infrared spectroscopy (NIR) models for detecting herbicide concentrations contaminated in water. Atrazine, the organochlorine herbicide, solution in the concentration range of 0-15 ppm were prepared in distilled water. Near infrared spectra were scanned by using Fourier transform spectrometer at wavenumbers of 12,500-4,000 cm⁻¹ (800-2, 500 nm). Partial least square regression technique was used to establish the NIR models for detecting herbicide concentrations. The developed models showed high prediction potential of herbicide concentrations contaminated in water with the R² of 0.97, RMSEE of 0.899 ppm, bias of -0.0003 ppm and RPD of 5.43. This indicated that these models can apply to analyse contamination of atrazine in the natural water sources. (paper)