Cationic polymer and clay or metal oxide combinations for natural organic matter removal

Abstract

The effect of adding suspended matter in the form of clay or metal oxide when a cationic polymer was employed as the primary coagulant was found to be beneficial. The solids provide both an adsorbent for natural organic matter (NOM) and a nucleating species for precipitating the NOM-polymer complex. Metal oxides in conjunction with a cationic polymer were more promising than clay, with effectiveness in the order Fe₂O₃>Fe₃O₄>Al₂O₃>MnO₂. Magnesium oxide at a much lower dose was nearly as effective as ferric oxide, but of course raised the pH level significantly. A simpler and more convenient way of having reactive solids present was to add alum to form flocs; for one of the waters studied the alum dose could be reduced by 67% by adding 1 mg/L of polymer, to give equal or better performance than alum alone at the optimum dose.

Introduction

Natural organic matter (NOM) can be removed from drinking water supplies by primary coagulation with cationic polyelectrolytes, followed by settling and filtration. The aim is to convert soluble organic matter into an insoluble form. The use of polymers as primary coagulants avoids the production of extra solids such as metal hydroxides formed by hydrolysis when alum or metal salts are used as the coagulant. NOM reduction was only partial when NOM solutions free of suspended solids were treated with long-chain cationic polymers of high charge content (Bolto et al (1998), Bolto et al (1999)), whereas in direct filtration with a floating filter, where additional finely divided solids in the form of kaolinite were present as well, much better removals were obtained (Kancharla et al., 1997). This behaviour parallels that reported for treatment of a pond water with the polyelectrolyte chitosan (Murcott and Harleman, 1993). Here chitosan alone performed poorly in terms of turbidity and colour removal, but a chitosan/bentonite mix gave a synergistic improvement in colour removal; bentonite itself was not effective. Likewise, tertiary treatment of sewage effluent with cationic polymers was successful if the turbidity was above 1 NTU, being raised if necessary by the addition of clay (Isais and Al Rammah, 1999). The combination was more useful than metal salt coagulants, which produced a voluminous sludge that resulted in shorter filter runs.

The aim in adding clay is to convert the NOM to an insoluble form; the role of the polymer is to coagulate the organic-laden particles. On addition of kaolinite to humic acid solutions, the organic material is mostly adsorbed onto the clay surface; kaolinite has been found to have a much stronger affinity for humic acid than montmorillonite (Parazak et al., 1988). The interaction is believed to involve polyvalent metals at the clay surface, and since kaolinite has a much higher aluminium content than montmorillonite it adsorbs humic acid more strongly. Also, montmorillonite has a high negative charge, largely invariant with pH, while that of kaolinite is much lower at about one tenth the level and dependent on pH. Hence, there will be a strong electrostatic repulsion between montmorillonite and NOM.

The adsorption of NOM by clays is generally accepted as a ligand exchange between the carboxylate groups on the organic molecules and the OH groups bound to a metal on the surface of the solid (Kaiser and Zech, 1998):

$$\text{>}\text{M}\text{–}\text{OH}\text{+}\text{RCOO}{}^{\mathrm{-}}\text{→>}\text{M}\text{–}\text{OOCR}\text{+}\text{OH}{}^{\mathrm{-}}$$

There are other analogous interpretations such as that involving H-bonding (Schulthess and Huang, 1991), but as the organic acids are more than half ionised under normal water treatment conditions the above model is more pertinent. The coating of charged organics on a clay has a strong impact on the amount of coagulant required (Gibbs, 1983). The charge on the particles is controlled by the adsorbed layer of NOM; because of the polyelectrolyte character of some of the organics, there will be an enhancement of the negative charge on the particles because other carboxylate groups on the same organic molecule are not bound to the metal. The charge on the particles is also influenced by salinity and the amount of divalent cations in the water. The surface potential is an important parameter influencing colloid stability and adsorption behaviour, and in natural systems is invariably negative, irrespective of the nature of the primary particle (Beckett and Le, 1990). It decreases in magnitude as the ionic strength of the water increases because of charge shielding (Beckett, 1990).

A similar rationale holds for metal oxides, which are stronger adsorbers of NOM than clays as there are more metal ions accessible to interact with the NOM via its chelating carboxylic acid/phenolic ligands. Iron oxide in particular has received considerable attention (Kolarik, 1983; Van Velsen et al., 1991; Korshin et al., 1997). On adding a highly charged long-chain cationic polymer as well, any unadsorbed humic acid reacts with the polymer to form insoluble hydrophobic precipitates before flocculation of the solids commences. Essentially all of the added polymer ends up as compact insoluble aggregates with the NOM (Parazak et al., 1988). These authors comment that in some cases the enhanced polymer performance could not be totally ascribed to bridging or electrostatic patch flocculation, and proposed that there was hydrophobic interaction between polymer molecules adsorbed onto adjacent particles.

The purpose of the current study is to explore the removal of NOM by primary coagulation with cationic polyelectrolytes, assisted by the presence of suspended matter in the form of clays, silica or metal oxides.