Water suppliers are facing the challenge of achieving higher levels of pathogen removal while having to lower residual concentrations of disinfection byproducts. A key factor for meeting these goals is the removal of naturally-occurring organic matter (NOM), which can be obtained by the new magnetised ion exchange (MIEX) process. This process was developed in Australia specifically for the removal of NOM, using a micro-size, macroporous, strong base anion exchange resin. It is applied in a completely mixed contact basin, as opposed to conventional resins that are used in fixed bed contactors.
The structural and chemical characteristics of the resin provide a high degree of resistance to organic fouling, thereby allowing its effective use through multiple regeneration cycles. The effectiveness of NOM removal is derived from its highly negative charge and related affinity to anion exchange sites.
Anion exchange offers an alternative to processes such as granular activated carbon (GAC) adsorption and nanofiltration (NF) for NOM removal. The resin has been reported to remove both small and large DOC molecular weight fractions, allowing it to work well in concert with the medium and large molecular weight removal that typically occurs with coagulation.
Using anion exchange in conjunction with coagulation results in removal of various broad spectrum DOC components, up to 80% or higher in some cases. One significant aspect of this process is the removal of other anions, such as arsenate, arsenite, phosphate, sulphide and bromide. The latter is significant to control the formation of brominated disinfection by-products (DBPs). A number of the brominated DBPs are of a significant health concern, therefore the ability to further limit formation of brominated DBPs represents an additional public health benefit of the anion exchange approach as compared with most of the other NOM removal strategies.WHAT IS RESIN?
Resin is a macroporous, polyacrylic structure, with a mean particle diameter of 150-180µm and quartenary amine functional groups, serves as the charge sites. While the internal specific surface area is comparable to that of conventional macroporous resins, the smaller resin size provides a greater external surface area, benefitting the DOC exchange kinetics and improving resistance to fouling. In addition to these features, the resin incorporates a magnetic component in its polymeric structure, which acts as a weighting agent and makes individual beads behave like small magnets capable of forming large, heavy agglomerates that facilitate recovery and reuse of the resin through multiple regeneration cycles. An additional property that is significant to water treatment application is a high degree of physical stability that allows recycle pumping with minimal damage to the resin. Process is when the magnetised anion exchange resin is applied continuously to the raw water in a mixing tank where the ion exchange occurs, see Figure 1. A small amount of resin (5-20mL of resin per litre of water) is used to exchange organics from water during a 5-30 min detention time in a completely mixed flow reactor.
The resin is recovered in a high-rate settling basin where the 'magnetically' enhanced flocculation agglomerates individual resin beads for better separation. Settled resin is pumped back to the contactor as a concentrated slurry. A small amount (5-10%) of recycled resin is continuously removed for regeneration with sodium chloride and replaced with regenerated resin.
This approach to ion exchange has several advantages over conventional fixed-bed designs. One significant advantage is flexibility to regenerate small batches of resin rather than an entire bed. The use of mixing tanks for contacting eliminates the need for pre-treatment for particle removal, thereby allowing removal of naturally occurring organic matter before the addition of coagulants. This can result in large reductions in the coagulant dose, particularly where organic constituents rather than particulate constituents determine the coagulant demand. Use of magnetised anion exchange beads also allows recovery in the settling step without the need for a coagulant or filtration, both of which would complicate recovery. The new process was evaluated, under sponsorship of AwwaRF, at several locations in the US to allow an assessment of the application of the process under different water quality and treatment conditions.
An important aspect of this testing was to focus on simulation of operation under conditions that replicate steady state operation in a flow-through mode rather than rely on testing of virgin or freshly regenerated resin.Test locations
Pilot plant operations were conducted at four test locations:
- Maytum WTW in New Port Richey, Florida. The Tampa Bay water system presently treats ground water with aeration for sulphide removal and chlorination for disinfection, oxidation of sulphide and reduction of colour. Tests targeted the removal of total organic carbon (TOC) and control of DBPs to meet future regulations, as well as sulphide removal,
- Hanahan WTW in Charleston, South Carolina, treats surface water that contains high levels of TOC and bromide that affects disinfection by-product formation. In addition, the effects on downstream coagulation was an important consideration as NOM can have a significant effect on coagulant demand,
- Fort Thomas WTW in Kentucky provides treatment of Ohio River water with conventional coagulation, flocculation, sedimentation and filtration. The river water has moderate alkalinity and TOC concentrations and turbidities can periodically be high. Testing was performed during a period of high rainfall and raw water turbidity, with an average of 78NTU. The treatment goals at this location are primarily the removal of DBP precursors, enhancement of the coagulation process and bromide removal,
- Alfred Merritt Smith WTW in Las Vegas, Nevada, is serving Las Vegas. Lake Mead is characterised by low turbidity and organic content that can be treated with relatively low doses of ferric chloride as the coagulant in conjunction with a polymer for conditioning prior to direct filtration. The plant is also in being upgraded with pre-ozonation.
The direct filtration limits the opportunities for enhanced coagulation through increased ferric chloride dose. In addition, high alkalinity restricts the pH adjustment for enhancing organics removal by coagulation. Traces of other anions include bromide, arsenate, arsenite and perchlorate.
Each location has different conditions under which anionic resins might be applied, thereby providing a broad basis for evaluation. Testing was performed over a period of four weeks using a 2.5m3/h (10gpm) skid-mounted pilot plant, designed to simulate full-scale operation of the (MIEX) process. Organics removal - a significant variation was observed in the DOC removal:
- at Maytum and Hanahan, generally in excess of 60% was achieved,
- at Fort Thomas ranges from 40-60%,
- at the Alfred Merritt Smith ranges from 23-44%.
The lower removal at the latter two locations may be a result of the relatively high alkalinity and total dissolved solids of the raw water as contrasted with the other two locations.
Bicarbonate and carbonate anions comprising the alkalinity compete with DOC for the anion exchange sites of the resin along with sulphate and other components of total dissolved solids. Other factors that contribute to observed differences may include the nature and quantity of DOC, leading to significant site-specific considerations that will need to be assessed to determine process performance.
As far as DOC dose-response testing, the results for Maytum exhibited much less sensitivity to changes in resin concentration and detention time than did the results for the Hanahan, which exhibited a much greater benefit from increases in resin concentration and contact time. At Hanahan, both MIEX and enhanced coagulation were individually capable of achieving high levels of DOC removal, from 50-60% but still greater removals up to 74% were achieved when the two processes were combined. Removal of sulphide was examined at Maytum, with raw water sulphide concentrations that can exceed 1.0mg/l. Pilot testing confirmed removal capability in a continuous flow system and sulphide removal was effective throughout the pilot plant testing to below odour thresholds.
Removal of bromide was observed under some conditions, depending on the source of salt (sodium chloride) that was applied for make-up of the regenerant brine. Much of the salt that is available for this type of application is a sea salt that contains an appreciable quantity of bromide.
Application of such a salt appeared to result in limited bromide removal. Use of a lower bromide salt from an alternative source appeared to provide conditions that allowed for effective removal of bromide at the Hanahan. However, results at Alfred Merritt Smith with a much concentration of higher total dissolved solids (TDS) as competing anions showed less effectiveness with a low bromide salt.
Removal of arsenic and perchlorate were assessed at Alfred Merritt Smith. Potential for removal was anticipated as both occur in an anionic form. Some degree of removal of both contaminants was observed, thereby indicating a potential for this type of application. Again, competition by the high TDS content of this water may have impacted resin effectiveness.Effect on coagulation
Applied as a pre-treatment step prior to the addition of coagulants, the MIEX process can have an effect on downstream coagulation, as a result of the removal of highly negative charge content in NOM, which otherwise impart a coagulant demand. The results in Hanahan show significant changes in the effect of alum dose and pH when the raw water is pre-treated, allowing a settled water turbidity of less than 1NTU over a wide range of dose and pH conditions.
These findings have significance in terms of capability to reduce alum dose and production of waste solids and for improved stability of flocculation and sedimentation sequences, with a possibility for operating at higher loading rates. In addition, improved filter efficiency may also prove to be a consequence of improved coagulation.
Figure 2 shows the positive experience of the first large-scale plant in Wanneroo, Western Australia (56Ml/d) which, together with the good results in the US, pilots have now led to tests of the process in northern UK, as well as evaluation of the technology in Wales, Finland and the Netherlands