Assessment of Fe(II), Fe(III) and NaMnO4 dosing for As removal <1 µg/L during aeration-filtration at WTP Prinsenbosch

More Info


In 2013, Brabant Water (BW) adopted a new goal of <1 μg/l for As level in the drinking water, resulting in the need to enhance As removal at seven water treatment plants (WPT’s), including WTP Prinsenbosch (As = 2.6 μg/l). Earlier experiences of BW at WTP Dorst had shown that dosing of NaMnO4 to an aeration-filtration system efficiently reduces As concentrations. However, the so-called AOCF method had undesired operational side-effects, including decreased filter run time, breakthrough of particles and poor thickening of backwash sludge. The aim of this thesis was to investigate 3 process alternatives (respectively dosing of NaMnO4, Fe(III) and Fe(II)) to enhance the As removal to < 1 μg/l at WTP Prinsenbosch, specifically including the impact on operational aspects like the filtration process and the backwash water production and sludge properties. The assessment is based on the results of (full-scale and) pilot-plant research at WTP Prinsenbosch.

It was found that Fe(III) was slightly more effective than Fe(II) and NaMnO4 for the removal of As, but at doses of 1.3 mg/l NaMnO4, 0.7 mg/l Fe(III) or 1.2 mg /l Fe(II) a drinking water quality of 0.7 μg As/l could be obtained with all 3 chemicals. In rapid sand filters, As(III) is oxidized biologically leading to subsequent adsorption of As(V) onto Fe(III)oxyhydroxides and Mn(IV) oxides. It was found that the adsorption capacity of the filter precipitates was relatively high, because to obtain a concentration of 0.7 μg As/l, the acceptable adsorption load was between 1.4-1.7 μg As/ mg (Fe+Mn).

Although As removal was similar for dosing either 1.3 mg/l NaMnO4, 0.7 mg/l Fe(III) or 1.2 mg /l Fe(II), the impact on the filtration process and the backwash water production and sludge properties proved to be quite different. Dosing of NaMnO4 and Fe(III) led to shorter filter run times, increased vulnerability to breakthrough, higher backwash water production and reduced thickening of the sludge. Contrarily, dosing of Fe(II) and the reference case (without dosing) led to longer run times, no breakthrough, lower backwash water production and superior thickening of the sludge. The differences in the impact on the filtration process and sludge properties between Fe(II) and Fe(III) and NaMnO4 are most likely explained by a different density of the Fe(III)oxyhydroxides and Mn(IV) oxides formed in the filter. The different density may be related to the formation process (biotic or abiotic oxidation of Fe and Mn, homogeneous or heterogeneous precipitation of Fe(III)oxyhydroxides and Mn(IV) oxides) and the crystal structure and density (which may be influenced by Eh and pH).

The conclusion of the research for WTP Prinsenbosch is that Fe(II) is the preferred chemical in view of the superior operational aspects. It is also cheaper than NaMnO4 (both chemical costs and sludge disposal costs) and easier to handle.