Effect of iron corrosion on the fate of dosed copper to inhibit nitrification in chloraminated water distribution system

Abstract

Nitrification has been acknowledged as one of the major barriers towards efficient chloramination in water supply distribution systems. Many water utilities employing monochloramine as the final disinfectant have been encountering unwanted microbiologically assisted choramine decay and find it difficult to maintain desired chloramine residual at distribution system extremities. A novel method of using cupric sulphate (< 0.4 mg-Cu(II)/L) to inhibit ammonia oxidizing bacteria was recently granted a US patent (7465401). Efficient inhibition was achieved in bench scale work and a pilot reservoir in the field. However, unexpected dissolved Cu(II) loss occurred when copper salt was dosed into one pipe section of the Goldfield & Agricultural Water Supply System (G&AWSS) in Western Australia. It prevented dissolved copper from reaching extremities to protect chloramine from microbiologically assisted decay.Our previous research and evaluation of the pipe environment suggested that severe dissolved copper loss could be related to iron pipe corrosion due to aging of the cementlined steel pipe, extensive temperature fluctuation, chloramination and nitrification. A large amount of copper and iron found in sediments after the pipes’ flushing provided further evidence. Although scale formation could be a complicated process that depends on a variety of physical and chemical conditions and the composition of corrosion scale can be distinct in each particular system, based on the literature review, ferric hydroxide flocs are acknowledged as one of the major corrosion products. Consistent severe copper loss over the three-year trial indicated that iron pipe corrosion is continuously occurring in the distribution system and thus supplying fresh iron salts. Ferrous ions could be released from new crevices during the initial stage of corrosion and oxidized to ferric ions. Ferric ions are further converted to ferric hydroxide flocs under drinking ii water pH and oxidation conditions. Therefore, ferrous and ferric ions as well as ferric hydroxide were chosen as the major corrosion products in this research.Bearing the goal of improving the inhibition strategy, this research investigated aqueous copper speciation in bulk waters, quantified dissolved Cu(II) removal by the iron corrosion products at trace concentrations (< 2 mg-Fe/L) and modelled dissolved Cu(II) loss subject to iron pipe corrosion. Mundaring raw water (MRW), which is the source water of G&AWSS, was employed as the main water source in this study. In addition, the nitrified water (NW) which was sourced from our laboratory reactors and the water containing humic substance (HAW) were used to investigate the effects of natural organic matter (NOM) of different characters on the fate of dissolved copper. Batch experiments were undertaken to measure Cu(II) solubility under various aqueous conditions. MINEQL+® (chemical equilibrium modelling system) was used to analyse Cu(II) speciation and cross-examine aqueous Cu(II) concentrations measured in the laboratory experiments. Aqueous ferrous ions, ferric ions and ferric hydroxide flocs, which are believed to be representatives of iron corrosion products, were added at low concentrations to remove dissolved Cu(II) in various bulk water samples. Their ability to remove dissolved Cu(II) was assessed individually, then the theory that a two-stage corrosion process removes Cu(II) was developed.The impact of NOM character on Cu(II)-NOM chelation in bulk water and their subsequent removal by ferric salts were elucidated by means of apparent molecular weight distribution and differential absorption spectra analysis. Finally, the dynamic process of dissolved Cu(II) removal by ferric salts was investigated. Combining both equilibrium and dynamic studies of dissolved Cu(II) removal by ferric salts, a model was established to predict dissolved Cu(II) loss in a corroded iron pipe distribution system. Aquasim® was used for estimating parameters and simulating corrosion patterns and thus predicting Cu(II) loss in the field.Cu(II)-NOM complexes were found to be the dominant forms in MRW, NW and HAW. Cu(II) solubility varied slightly in these bulk waters due to different NOM characters. Generally, intermolecular dicarboxylate chelation was considered to be the dominating chelation type between Cu(II) and organic compounds in MRW and NW, while salicylate chelation was prevalent in HAW due to more salicylate type binding sites available in humic substances. The removal of dissolved Cu(II) was assumed to occur via a two-stage process during corrosion: Stage I-coagulation and aggregation by released ferrous/ferric ions; Stage II-adsorption by iron hydroxide flocs formed afterwards. Both ferrous/ferric ions and ferric hydroxide flocs showed considerable capacity to remove dissolved Cu(II). Addition of 2 mg-Fe/L ferric ions was sufficient to remove the majority of dissolved Cu(II) in MRW and NW. The dissolved Cu(II) removal by Fe(OH)3 flocs in MRW and NW could be explained as multilayer adsorption obeying a Freundlich isotherm. In addition, the adsorption process could be interfered with by the presence of heterogeneous Cu(II)-containing particles (CuO and Cu(OH)2), which rendered less dissolved Cu(II) removal. Cu(II)-NOM in HAW demonstrated a relatively high resistance to removal by ferric salts.From these observations, Cu(II) was thought to preferentially complex with small organic molecules until saturation is reached. Slightly higher Cu(II) solubility and less dissolved Cu(II) removal observed in NW indicated that a proportion of small soluble organic substances was probably produced and chelated with Cu(II) during nitrification. In humic acid water (HAW), Cu(II) bound with small MW organic matter could be shielded by a relatively high proportion of large MW organic matter. The dynamic process of dissolved Cu(II) removal by Fe(OH)3 flocs can be described by Pseudo second order decay. From the comparison of dissolved Cu(II) loss between the modelling results and field data, the loss of dissolved Cu(II) could be due to removal by iron corrosion products and modelled by a reasonable assumption of the iron corrosion situation in the distribution system. At the end, Cu(II)-based inhibition and chloramination strategies are recommended.