Improved specificity for drinking water treatment monitoring

Research Project, 2018 – 2020

Nearly nine billion litres of treated drinking water are produced in Sweden each year, at a cost of around a billion kronor each year. The costs for infrastructure, chemicals and energy needed to produce stable, healthy drinking water have increased in recent years, and are predicted to increase further, due to climate change increasing the amount of biological material that must be removed during treatment. Much of the biological signal in drinking water is due to dissolved organic material, including the decayed remains of plants and other organisms. This dissolved organic material increases the cost of treatment while reducing its effectiveness, by harbouring pathogens, consuming chemicals, fouling filters and providing food to stimulate bacterial regrowth in the distribution network on the way to the consumer.Around the world, the spectral properties of natural organic matter are being used to track the removal of dissolved organic matter from drinking water. Many dissolved organic compounds have the useful property of glowing after irradiation by ultraviolet and visible light, producing a ‘fluorescence fingerprint’ that is characteristic of the dissolved chemical compounds present in the water, and which glows with an intensity proportional to the amount of chemicals present. These fingerprints are collected using fluorescence spectroscopy then broken down mathematically into spectral fingerprints to determine concentrations for individual organic matter fractions. At that point, it is theoretically possible to track the production and removal of specific organic matter fractions at each stage of drinking water treatment.A problem in practice is that the retrieval of specific chemical fingerprints from fluorescence datasets is not working well. Although nearly everyone uses the same algorithm to mathematically decompose fluorescence datasets, almost nobody retrieves identical spectral fingerprints, even when analysing similar kinds of water in similar kinds of ecosystems. Either the model is inappropriate or we aren’t applying it properly. We consider the second explanation to be valid and in this research, we will deliver new tools and methodologies to improve the characterisation and quantification of organic matter fractions detected in water using fluorescence spectroscopy. This will help treatment plant operators to identify small changes in the concentrations of specific DOM fractions during drinking water treatment, facilitating widespread efforts in Sweden and abroad to optimise water treatment processes using fluorescence spectroscopy.