Minh Pham


John E. Tobiason

Publication Date



This research involved full- and pilot-scale studies of treatment of the Aquarion Water Company (AWC) Lantern Hill groundwater source. With elevated levels of both dissolved manganese (~0.19 mg/L), dissolved iron (~1.9 mg/L) and natural organic matter (NOM) (~3 mg/L) the existing treatment plant is having difficulty in achieving required manganese removal while maintaining low concentrations of disinfection byproducts (DBPs) in finished water. At full-scale, dissolved manganese in the raw water is removed through pre-filter oxidation and adsorption on iron precipitates via application of free chlorine and permanganate as well as adsorption of dissolved manganese onto MnOx(s) coated filter media (anthracite and greensand) which is continuously reactivated by free-chlorine oxidation. The addition of pre-filter chlorine to the raw water with high concentration of NOM leads to the formation of elevated levels of regulated DBPs such as trihalomethanes (THMs) and haloacetic acids (HAAs).

To investigate the effect of achieving NOM removal prior to chlorine addition on decreasing DBP formation, a two stage pilot-scale filter system was installed at Lantern Hill. A first-stage filter column (7.5 in diameter) with conventional dual media (anthracite over sand) for NOM and oxidized iron removal is followed by a second stage high-rate coarse media filter for Mn removal. Prior to the first stage filter, permanganate is dosed in the range of 0.5 to 1.25 times the stoichiometric requirement to oxidize most of the reduced iron and a portion of the dissolved manganese to insoluble forms; pH was controlled at 7 to 7.5 by NaOH addition. In addition, synthetic cationic polymer was also applied ahead of the dual media filter to improve particle and NOM removal. Free iii chlorine is dosed to the first stage filter effluent prior to the second stage contactor which is operated at a hydraulic loading rate (HLR) of 10 to 20 gpm/ft2 .

The results show a dramatic decrease in DBP formation and excellent removal of Mn, Fe and NOM. After the first stage filtration, NOM levels decreased from 3 mg/L to 2 mg/L prior to any chlorine addition, dissolved manganese was between 0.03 to 0.2 mg/L while very low concentrations of reduced iron (<0.01 mg/L) were recorded. Post-filter chlorine addition and the second-stage contractor routinely decreased dissolved Mn to levels of 0.01 to 0.02 mg/L except at high HLR and shallower bed depths when very low pre-filter KMnO, dosing caused filter effluent manganese levels to rise 0.15 to 0.2 mg/L. Resulting DBP analyses showed that contractor effluent levels were only 20 to 30% of full-scale levels in the presence of a similar 1 mg/L free chlorine residual and well below regulatory requirements.

An existing model which simulates manganese removal as a function of bed depth was modified and utilized in simulating the experimental data for the second-stage contactor at the LHWTP. The results show that the modified model can capture well the manganese concentration along the second-stage contactor. The model was later used to simulate manganese removal for different designs of the post-contactor to help the Aquarion Water Company to determine the best design for a Lantern Hill Water Treatment Plant upgrade.