Final Report Summary - AQUASENSE (Development of Novel Sensors for Contaminant Detection in Water using Near Infrared Light and Aquaphotomics) Project objectives The overall objective of the proposed research was to investigate the potential use of aquaphotomics combined with near-infrared spectroscopy (NIRS) and chemical imaging (NIR-CI) for detection of contamination in water. Knowledge gained on the theory and application of aquaphotomics acquired during the outgoing stage was transferred to the European Union (EU) during the incoming stage. As such, this project investigated non-destructive monitoring of contamination in water using NIRS and NIR-CI with the following specific aims: (1) to develop new knowledge in the theory of aquaphotomics through application of perturbation spectroscopy to water systems; (2) identification of water matrix coordinates (WaMaCs) and water absorption patterns (WAPs) for water systems subjected to a wide range of physical and chemical perturbations; (3) using (1) and (2) develop algorithms for the detection of contaminants in water using NIRS and NIR-CI; (4) evaluate performance of NIRS and NIR-CI compared with standard methods for water contaminant detection. Description of the work performed since the beginning of the project - WP1: Training in aquaphotomics theory (1 May 2010 - 31 May 2011) (i) Literature review: applications of vibrational spectroscopy to water quality monitoring (ii) Review of techniques applied to obtaining water matrix coordinates (iii) Review of assignment of water matrix coordinates (iv) Analysis of existing laboratory data - WP2: Selection of contaminants and experimental design (1 May 2010 - 10 October 2011) Contaminants from the main groups of organic and inorganic contaminants were chosen based on the following criteria: safety, availability, comparative results in the literature. The experiments were randomised designed to include regular control measurements to characterise any additional variability in the experiments due to laboratory conditions and instrumental drift. - WP3: Baseline assessment and variance modeling (1 August 2010 - 31 December 2010) Baseline spectral measurement of pure water in controlled environmental conditions was carried out to measure spectral features and identify sources of variation in the pure water. - WP4: Perturbation spectroscopy of contaminated water (1 November 2010 - 30 June 2012) Contaminated samples were prepared by dissolving selected contaminants in pure water at a range of different concentration levels and spectra were obtained. WP5: Algorithm development (01 November 2011 - 30 September 2012) Algorithms were developed for contaminant identification, using the data obtained in WP4. - WP6: Validation study using NIR and NIR-CI (1 May 2012 - 1 October 2013) Spectral acquisition for selected range of contaminants was repeated (both in Dublin and in Rome) to verify the findings in the outgoing phase. The capability of NIR-CI for contaminant detection was investigated. A validation study was carried out to test the performance of the developed algorithms, leading to robust limit of detection (LOD) estimates. - WP7: Comparison with traditional methods (1 October 2013 - 30 April 2013) The researcher was trained in high-performance liquid chromatography (HPLC). Samples obtained from rivers were analysed using NIR-CI and standard water quality testing methods. Description of the main results achieved so far - WP1: Training stage completed, with fellow being trained in Aquaphotomics by Prof. Tsenkova in Japan. - WP2: Selection of contaminants and experimental design completed. - WP3: Baseline assessment completed. It was found that the main interfering factors, ambient temperature and humidity, could confound the elucidation of more interesting spectral responses. These sources of variability could be characterised by obtaining control spectra during the period of each experiment. The observed variation could then be used as an input to extended multiplicative signal correction, a method developed to orthogonalise a spectral response of interest to that of the interferences. - WP4: NIRS experiments for contaminants selected in WP2 have been carried out. WaMaCs and WAPs have been characterised. - WP5: Variable selection algorithms have been developed for contaminant identification using the data obtained in WP4. The 1300 - 1600 nm range was found to be the optimal wavelength range for prediction of contamination. - WP6: Spectral acquisition for selected range of contaminants was repeated (both in Ireland and in Italy) to verify the findings in the outgoing phase (in Japan). WAPs obtained in each test location exhibited similar profiles. Limits of detection for each location / contaminant are shown. The attained LOD ranged from 162 - 2414 ppm. - WP7: NIRS and NIR-CI developed methods were not capable of detecting contamination in water at levels similar to those detectable by traditional techniques (i.e. ppb-ppm range) due to higher inherent LOD, confirming findings from WP6. Expected final results, their potential impact and use The research has allowed characterisation of the detection limit for various contaminants in pure water matrices, showing that a LOD within the range 162 - 2414 ppm was achievable with NIRS and aquaphotomics in the most optimistic conditions (i.e. single salt solutions with no matrix interference effects). This range represents the lowest LOD achievable with the proposed method, indicating that the proposed method is not suitable for measuring contaminants in drinking or river water at relevant concentration ranges (ppb-ppm). However, it may be useful for screening of wastewater.
