Feb 1999

Ensuring the microbiological safety of drinking water is of paramount importance. Conventional drinking water treatment in the UK includes the use of chlorine or another oxidant for final disinfection. In the case of surface waters, pre-treatment such as coagulation, sedimentation and filtration is generally used to prepare the water before final disinfection. The disinfection of drinking water using chemicals such as chlorine has successfully protected public health against waterborne disease. However, chemical disinfection produces potentially toxic by-products whose presence is undesirable. The use of chemical disinfection can also cause problems with taste and odour leading to customer complaints and growth in sales of bottled water and water filters. Some consumers do not like the idea of chemicals being added to drinking water at all. The emergence of water-borne pathogens which are resistant to chemical disinfection, such as Cryptosporidium, has led to a reappraisal of traditional disinfection practice. Water companies and regulators need to consider how they can respond to such complaints and concerns without compromising safety and public health and whilst maintaining scientific credibility.

Other factors may influence the choice of disinfection process. Increasing pressure on water resources means that poorer quality source waters could be used with a range of implications for the use of chlorine or chlorine-based disinfectants. Possibilities include greater by-product formation, inadequate disinfection using conventional processes and other contaminants. This could be exacerbated by the effects of climate change causing flood or drought conditions, either of which could make treatment more difficult. In the longer term, the pressure on the chemical industry to reduce production of chlorine and chlorine-based disinfectants for environmental reasons may force water companies to turn to other disinfectants or to employ non-chemical disinfection processes.

A number of commercially available alternative processes, such as membrane processes, are able to remove bacteria, viruses and protozoa as well as a range of chemical contaminants. These are coming into use but generally only on a small scale. It may be possible to operate these processes with no chemical disinfection or at least to reduce the amount of chemicals used for final disinfection. Alternatives to chemical disinfection, such as UV irradiation, are also being used for disinfection of drinking water. Such ‘non-conventional’ processes and disinfection methods could in principle be used to replace, or at least greatly reduce, the use of chemical disinfection of drinking water.

However, in some northern European countries final chemical disinfection has been abandoned in many areas, reliance being placed on multiple barriers, often including natural barriers and biological treatment stages, to prevent contamination and to remove or inactivate micro-organisms. This approach is not viable in the UK, at present, due to a number of considerations including: less efficient barriers in chalk aquifers; the much greater reliance on surface-derived supplies; lack of biological treatment processes to reduce assimilable organic carbon loadings; and long distribution systems with long retention times of water in the network.

UV irradiation and membrane processes are potentially suitable alternatives to chemical disinfection. UV is capable of inactivating bacteria and viruses, and possibly protozoan parasites. A range of pressure-driven membrane processes – microfiltration, ultra-filtration, nanofiltration and reverse osmosis in order of decreasing pore size – are also capable of disinfection as well as removal of chemical contaminants, depending on pore size. The use of membrane processes would avoid the formation of disinfection by-products and would reduce the concentrations of other undesirable chemicals, giving a net benefit in terms of toxicological issues. The main microbiological concerns with membrane systems are ensuring the integrity of the membrane and monitoring the efficiency of micro-organism removal; with conventional chlorination the residual chlorine concentration is easily monitored and provides reassurance that disinfection has been carried out effectively.

Both UV irradiation and membrane systems are sufficiently developed to be considered as alternative disinfection strategies. A range of hypothetical process streams were considered as potential alternatives to conventional treatment; it would not be sufficient to consider unit processes in isolation because preceding treatments would affect overall process performance, reliability and costs.

In terms of whole-life costs, the alternatives are generally more expensive than conventional treatment. This is particularly so for larger plants because there are economies of scale with conventional plants that are not realised with UV and membrane systems which are essentially modular. Chemical costs for membranes and UV treatment are lower than for conventional processes but their energy consumption is greater. Membrane systems generate liquid waste (containing impurities concentrated during the process) which might require treatment to destroy toxic chemicals or kill micro-organisms; not all available cost data include allowance for waste treatment and disposal.

To compare the relative risks associated with the various processes, the ways in which output quality could deviate from target values were identified and ranked. The likely frequency of failure and consequence of each failure were determined (or estimated) and multiplied together to obtain the risk of each type of failure. These risks were then summed for each process combination to derive an overall system risk rank. The risk rankings did indicate differences between process combinations. However the variations were within the likely error in this analysis. In terms of risk to the consumer, differences between processes are probably less significant than plant design and operation.

An overall assessment of the alternatives and conventional treatment against several criteria (microbiological and chemical safety, cost, reliability etc.) showed that along with comparable disinfection ability, the alternatives can have benefits such as improved chemical safety and customer aesthetics, but also disadvantages such as difficulty in monitoring process performance. The overall assessment is summarised below:

One consequence of not using chemical disinfection would be the lack of a residual disinfectant in distribution. The main benefits of this are that it acts as a preservative and a simple indicator of failure of treatment or post-treatment contamination. To operate without a residual disinfectant other than in newly constructed networks such as new housing estates, would require a considerable amount of mains renovation and cleaning, and careful monitoring of water quality in distribution. As noted above, there would also need to be a means of monitoring the efficiency of disinfection with defined frequency of measurement and action levels.

It is appropriate to consider alternatives to chemical disinfection due to concerns over disinfection by-products, disinfectant-related tastes and odours and the emergence of some chlorine-resistant protozoan parasites. Changes in the quantity and quality of water sources and environmental pressure to reduce the production and use of chlorine-based chemicals provide additional reasons to consider alternative methods of water treatment.

The alternatives to chemical disinfection are the use of more physical and chemical barriers to contamination, UV irradiation, and membrane processes. In the UK only UV and membranes would be feasible alternatives without fundamental changes to methods of water abstraction, treatment and distribution. UV irradiation is effective against bacteria and viruses, and may be effective in inactivating parasitic protozoa. Membrane processes could be used to replace conventional treatment and provide adequate disinfection. Membrane-based processes may remove or leach undesirable substances as well as producing waste streams. A reduction in the concentrations of chemical disinfectants and their by-products, as well as other contaminants, could provide a net benefit in terms of toxicological issues.

Membrane techniques tend to be more expensive, on a whole-life basis, than conventional treatment, especially for large plants as there is no economy of scale with membranes as there is with conventional plant. Costs increase with the degree of particle size removal i.e. microfiltration < ultrafiltration < nanofiltration < reverse osmosis. The use of UV irradiation can add substantially to both capital and operating costs.

Risk analysis indicated differences between the process streams considered. However, all the differences in risks were within the accuracy of the analysis; actual design and operation of disinfection processes are likely to have a greater impact on risk of failure than the disinfection method used.

The main consequence of operating without a disinfectant residual would be lack of a preservative effect and loss of a useful indicator for failure of disinfection or serious post-treatment contamination. Before operating without a residual, considerable mains renovation and cleaning would be required. Intensive water quality monitoring would be needed until it was established that water quality in distribution was acceptable.

The main implications for water treatment and process monitoring of changing to alternatives to chemical disinfection would be the need to ensure the efficacy of disinfection. Whilst there is no reason to impose tighter standards for non-conventional treatment methods, there is no simple surrogate, like chlorine residual measurement, to assess the integrity of a membrane-based process.

Taking these concerns into account, it is considered that if alternative processes are to be used for primary disinfection, a low dose of chemical disinfectant should still be added to provide a residual within distribution. This would enable effective disinfection and disinfection by-product precursor removal, resulting in low production of by-products whilst still ensuring microbiological safety, together with the benefits in terms of monitoring and preservative effect of having a residual in distribution. The applicability of alternative processes would need to be considered on a case-by-case basis, taking into account source water characteristics, existing treatment, the nature of the distribution system and the need to ensure protection of public health. For example, with a new surface water treatment plant feeding a new distribution system it would be appropriate to consider membrane and/or UV treatment alone whereas if the plant was to supply an existing network then membrane or UV treatment should be followed by residual chlorination. On the other hand, if an existing treatment plant is performing adequately a change to non-chemical disinfection could not be justified on technical grounds.

Copies of this report may be available as an Acrobat pdf download under the 'Find Completed Research' heading on the DWI website.