NOM: that is another matter

The water supply industry is governed by increasingly stringent water quality guidelines and legislation, ensuring customers receive the highest quality drinking water at all times. One area receiving considerable interest is disinfection by-products (DBPs) produced as a result of using chlorine. These by-products can take the form of trihalomethanes (THMs), haloacetic acids (HAAs) and a host of other halogenated DBPs, a number of which have been shown to cause cancer in laboratory animals and are produced by the reaction of chlorine with organic molecules. Recent legislation has reduced the THM standard in the US from 100-80µg/l. A similar standard of 100µg/l is in force in the UK and has recently tightened from a three-month average of 100µg/l to a limit of 100µg/l. As any organic molecule in water can act as a precursor for disinfection by-products the tightening of standards has generated a further challenge for those utilities treating waters high in colour and natural organic matter (NOM).

NOM is a complex mixture of organic material, such as humic and fulvic acids, hydrophilic acids, proteins, lipids, amino acids and hydrocarbons, present in all raw water sources. The variety of the components in NOM changes from water to water (influenced seasonally, by climate effects and drought) which leads to variation in the reactivity with chemicals such as chlorine or ozone. Since the drought of the mid-1990s there has been a significant year-on-year increase in the colour and organics levels entering upland WTWs, see Figure 1. The colour levels in the raw water entering the WTW have increased from an average of 40°Hazen to in excess of 160°Hazen leading to a 100% increase in THMs (Banks et al, 2002). In addition to promoting the formation of DBPs the presence of NOM has been shown to:

• enhance transportation and distribution of organic micro pollutants,

• provide an undesirable biological growth substrate in distribution systems,

• cause odour and taste problems,

• lower the efficiency of treatment processes.

Fate of NOM during conventional water treatment

The removal of NOM and hence reduction in THM precursors is therefore a major goal in the treatment of any water source. The most commonly used processes for the reduction in organic carbon are coagulation/flocculation, carbon adsorption, biofiltration and membrane filtration. To improve the performance of NOM removal it is key to firstly identify its character, as the type of organics will affect not only the choice of treatment process but also the performance of the selected process. One method for assessing character is to fractionate the NOM into hydrophobic and hydrophilic material, as NOM found in all water consists of both hydrophobic and hydrophilic components where the largest fraction is generally hydrophobic acids, which makes up approximately 60% of the dissolved organic carbon. Figure 2 shows how the quantity of each fraction changes through a WTW and shows the two hydrophobic fractions (humic acid fraction – HAF, fulvic acid fraction – FAF) are significantly reduced through treatment but the hydrophilic non-acid (HPI-NA) fraction is relatively untouched by treatment.

Working with nature

For once nature seems to favour the operators as typically the hydrophilic fraction is less reactive than the hydrophobic fraction with chlorine in raw water samples, although recent work at Cranfield has showed the reactivity of the hydrophobic and hydrophilic fractions with chlorine varied significantly over the year and from source to source. It was found that in the summer, the hydrophobic acid fraction formed 38mg THM/mg C compared to 37mg THM/mg C for the hydrophilic acid fraction (HPI-A), whilst in autumn the hydrophobic and hydrophilic acid fractions both formed 171mg THM/mg C.

Whilst there are obvious issues regarding the reaction of hydrophilic organics with chlorine there are a number of other questions relating to what organics are present in this fraction. The hydrophilic organic fraction is a heterogeneous mixture of organic molecules characterised by high solubility in water and examples of chemical groups identified in this fraction include alcohols, ketones, esters, alkylhalides and aromatics. Assessment of the hydrophilic fraction has shown it to be made up of molecules of <3K Daltons in size. Whilst few individual compounds have been identified any organic compound with a high water solubility has the potential to be in this fraction including pharmaceutical products, pesticide products used on arable crops in the UK and emerging pollutants such as methyl tertiary-butyl ether (MTBE), vinyl chloride and haloacetic acids. Recently, both US and German researchers have shown how these hydrophilic organics are poorly removed during wastewater treatment and can enter surface and ground waters from wastewater discharges and then on in to potable supply.

Emerging Processes

The conventional treatment process for waters containing NOM is coagulation/flocculation using inorganic coagulants such as ferric chloride or aluminium sulphate. These processes remove NOM by adsorption onto flocs and can typically achieve 50-80% removal of DOC (Parsons et al., 2002). Increasing coagulant dose will increase removal but the process is limited when trying to reach DOC levels less than 1mg/l as it is poor in removing hydrophilic organics. In addition DOC removal by increasing coagulant dose will also lead to an accompanying increase in the amount of plant residuals (sludge) generated and associated increases in costs and operators’ time. The sludge produced during the coagulation of NOM is also difficult to dewater because of the increased metal ion and organic content.

Two new processes, however, have shown considerable potential in increasing DOC removal. The first process is MIEX, a commercially available processes developed by Orica in Australia. The name MIEX comes from magnetic ion exchange, because the ion exchange resin particles contain a magnetised component within their structure which allows the particles to act as weak individual magnets. The very small particle size of around 180µm provides a high surface area allowing rapid adsorption kinetics. In a settler these magnetic particles agglomerate into rapidly settling resin flocs. Two full scale plants have been constructed in Australia, the largest being a ground water site in Wanneroo. Cranfield has recently evaluated the technology on two UK upland waters and both have shown significant improvement in the removal of DOC and control of THMs over inorganic coagulants.

The second process showing considerable potential is Fenton’s reagent. Fenton’s reagent is an advanced oxidation process in which hydroxyl radicals are produced during the decomposition of hydrogen peroxide in the presence of ferrous salts and these strong oxidising species mineralise organic molecules to CO2 and water.

Recent tests on Fenton’s reagent showed they have a great potential for treating water high in NOM with increased removal of colour, DOC and THMs shown over coagulation. Figure 3 shows the effect pH has on the performance of Fenton’s reagent and shows the levels of THMs left in the treated water, reaching levels of less than 10mg THM/l.

Conclusions

It has become increasingly important to understand how water treatment processes can remove not only increasing amounts of DOC but to also to understand how their performance is effected by the character of organic material or NOM found in raw waters. It is clear that whilst conventional treatment options such as coagulation and adsorption are able to reduce levels of hydrophobic organics they are unable to remove hydrophilic polar molecules. There are a number of new process options under investigation at Cranfield such as MIEX and Fenton’s reagent which have shown significant potential in reducing both DOC and THM levels and are showing potential in targeting the low molecular weight hydrophilic compounds currently untouched by conventional treatment