NOM removal in drinking water treatment using dead-end ceramic microfiltration: Assessment of coagulation/flocculation pretreatment

Abstract

In Nordic countries, surface water is a common source for potable water production. Such waters are often characterised by high Natural Organic Matter (NOM) content, resulting in high colour, very low turbidity, low alkalinity and low hardness due to natural conditions. Treatment of such waters basically comprises the removal of NOM and colour, corrosion control and disinfection. Although the largest part of NOM is not harmful, some fractions can cause colour, taste and odour problems or can even be toxic. Complexation of heavy metals and organic micro-pollutants, mobilizing them and making them available in the water phase, and increasing the amount of necessary disinfectants as well as being recognized as precursors for disinfection by-product formation which can be carcinogenic, are issues to be considered. In addition, there is a growing concern on the increase of NOM in natural water sources. Recent studies have shown that the concentration of NOM in surface water has increased in the last decades and may increase further during the coming decades, caused by progressing climate change issues and/or changed precipitation patterns. Thus, the removal of NOM is one of the major concerns and makes advanced drinking water treatment necessary.

In the past, conventional treatment processes such as coagulation/rapid sand filtration and filtration by nanofiltration (NF) membranes have been used successfully in Norway. However, a number of disadvantages have been identified in these processes. While the quality off effluent from rapid filters is high if coagulation is optimised, it may be severely compromised if coagulation fails, or operation is unstable. NF membranes have been subject to high irreversible fouling. High operating pressures at low fluxes, permeability loss over time and high chemical demands for cleaning have been observed frequently. By contrast, low pressure membrane filtration has emerged as an alternative for direct surface water treatment in recent years, with number of installations increasing rapidly around the world. Disadvantages as described above can be avoided to a large extent with this technology. The usage of low pressure membrane processes leads to high, stable effluent quality compared to conventional technologies, especially with regards to hygienic aspects since water borne parasites and most of bacteria can be retained. By applying ceramic microfiltration (MF) a more energy efficient operation is possible, the higher investment costs for ceramic membranes can be offset by the ability to operate with significantly higher fluxes. However, such membranes demand coagulation pre-treatment in order to achieve sufficient NOM and colour removal, fulfil the requirements as a hygienic barrier as well as for the control fouling by NOM.

This study showed that coagulation, combined with subsequent ceramic MF, is a successful concept for the direct treatment of Nordic waters. In spite of high NOM content in the raw water (DOC 6.8 mg C/L, colour 55 mg Pt/L), stable operation was demonstrated at high membrane fluxes of up to 250 L/(m2 h), achieving irreversible membrane fouling below 1 mbar/h, a DOC removal of 70% and colour removal of around 90%, at a coagulant dosage of 0.65 mg Al per mg DOC (using PACL at pH 6, with 60s of inline flocculation). While NOM removal depended only on coagulant dosage and coagulation pH, membrane fouling was also influenced by flocculation type, time and G-value. The study also found that the optimization of coagulation pretreatment is crucial. If, for example, an insufficient amount of coagulant is dosed, membrane fouling increases drastically, residual metal concentration is high and NOM removal is minimal. Inline coagulation with a static mixer, followed by pipe flocculation, showed promising results compared to conventional tank coagulation, where irreversible fouling rates observed were in the same range. However, reversible fouling was significantly lower after tank coagulation and flocculation, in spite of floc breakage in the membrane feed pump. Indications are thus that, either the properties of the broken flocs are significantly different from freshly created ones, or that flocs regrow in the membrane capillaries.

Additional findings of this study show that if the coagulation pre-treatment is optimised, MS2 bacteriophages can be efficiently removed from surface water characterized by high NOM content. Removal rates of 6 log units up to complete virus retention were achieved, at pH values ranging from 5.0 to 5.5. However, at such conditions the residual metal concentration exceeded legal limits. In order to avoid this, it is advised that the coagulation pH should be increased to 6 or higher. At these pH-values efficient virus removal can still be achieved, however, higher coagulant dosages are required. By optimizing coagulation/flocculation pre-treatment, an efficient removal of NOM can be achieved while complying with hygienic barrier requirements, resulting in virus removal of 4 log units and higher. Coagulation/flocculation pre-treatment coupled with ceramic MF filtration is thus a viable and flexible treatment scheme for the production of high quality potable water from surface waters having high NOM concentrations.