Modeling fate and transport of arsenic in a chlorinated distribution system

Abstract

Experimental and modeling studies were conducted to understand the fate and transport properties of arsenic in drinking water distribution systems. Pilot scale experiments were performed in a distribution system simulator by injecting arsenic and measuring both adsorption onto iron pipe material and the oxidation of arsenite by hypochlorite in tap water to form arsenate. A mathematical model describing these processes was developed and simulated using EPANET-MSX, a hydraulic and multi-species water quality software for pipe networks. Model parameters were derived from the pilot-scale experiments. The model was applied to both the distribution system simulator and EPANET example network #3, a real-world model of a drinking water system serving approximately 78,000 customers. The model can be applied to systems-level studies of arsenic fate and transport in drinking water resulting from natural occurrences, accidental spills, or intentional introduction into water.

Introduction

Arsenic (As) is a naturally occurring toxic material that is highly soluble and often found in water sources (U. S. EPA, 2012). The U.S. Environmental Protection Agency reduced maximum contaminant level (MCL) for As from 50 ppb to 10 ppb (μg/L) in 2001 (U. S. EPA, 2015). This MCL reduction and the known toxicity of As have resulted in significant research into technologies to remove As from drinking water (Pierce and Moore, 1980, Pierce and Moore, 1982, Lakshmipathiraj et al., 2006, AWWARF, 2005, Banerjee et al., 2008, Fuller et al., 1993, Zhao et al., 2011, Goldberg and Johnston, 2001, Goldberg, 2002, Dixit and Hering, 2003, O'Reilly et al., 2001, Mamindy-Pajany et al., 2009, Masue et al., 2007, Lafferty and Loeppert, 2005, Wingrich and Wolf, 2002, Bhandari et al., 2012, Payne and Abdel-Fattah, 2005, Jain et al., 1999, Lenoble et al., 2002, Impellitteri and Scheckel, 2006, Waltham and Eick, 2002, Raven et al., 1998, Jeppu et al., 2010, Guan et al., 2008, Habuda-Stanić et al., 2008, Driehaus et al., 1998, Jeppu and Clement, 2012, Catalano et al., 2011, Grossl et al., 1997, van Elteren et al., 2001, Dodd et al., 2006, Manning et al., 1998, Fendorf et al., 1997, Sorlini and Gialdini, 2010, Vasudevan et al., 2006, Jeppu et al., 2012, Philips and Taylor, 1976, Huang et al., 2011a, Huang et al., 2011b, Bissen and Frimmel, 2003, Hug et al., 2001, Dinesh and Pittman, 2007). Due to the high affinity of iron-based materials for As, many of these studies have investigated how much As adsorbs onto different iron-based materials in order to understand their viability for As removal in drinking water treatment (Pierce and Moore, 1980, Pierce and Moore, 1982, Lakshmipathiraj et al., 2006, AWWARF, 2005, Banerjee et al., 2008, Fuller et al., 1993, Zhao et al., 2011, Goldberg and Johnston, 2001, Goldberg, 2002, Dixit and Hering, 2003, O'Reilly et al., 2001, Mamindy-Pajany et al., 2009, Masue et al., 2007, Lafferty and Loeppert, 2005, Wingrich and Wolf, 2002, Bhandari et al., 2012, Payne and Abdel-Fattah, 2005, Jain et al., 1999, Lenoble et al., 2002, Impellitteri and Scheckel, 2006, Waltham and Eick, 2002, Raven et al., 1998, Jeppu et al., 2010, Guan et al., 2008, Habuda-Stanić et al., 2008, Driehaus et al., 1998, Jeppu and Clement, 2012, Catalano et al., 2011, Grossl et al., 1997, Manning et al., 1998, Fendorf et al., 1997, Bissen and Frimmel, 2003, Dinesh and Pittman, 2007).

Iron-based pipe materials can be found in the majority of drinking water distribution systems. Cast iron, ductile iron, and steel have been used in water distribution systems for many years. While some of this infrastructure is slowly being replaced with other materials, such as cement-lined cast iron, many older pipes within a distribution network may still have exposed iron-based surfaces. Over time, these iron-based pipes corrode and can also acquire mineral deposits or biofilms (Mohebbi and Li, 2011, Lytle et al., 2004a, Lytle et al., 2010, Lytle et al., 2014, Gerke et al., 2008). The corroded surface of iron pipes has been found to contain a mixture of different species of iron, such as amorphous ferric hydroxide, goethite, ferrihydrite, siderite, lepidocrocite, and magnetite (Lytle et al., 2004a). If arsenic is present in the water, it can adsorb to these iron-based pipe materials. Segments of pipes taken out of service have been found to contain elevated As levels even though only low As levels were present in the water (Lytle et al., 2004a).

The goal of this study was to develop a predictive model of arsenic fate and transport in drinking water systems. To this end, experimental studies were conducted to measure arsenic adsorption to and desorption from pipe walls and arsenite oxidation in chlorinated water. Cast iron coupons that were allowed to corrode in tap water were used to mimic the inner surface of pipes in older distribution systems, or where defects may have formed in newer cement-lined iron pipes. A mathematical model was developed that incorporates adsorption, desorption, advection, and reaction with chlorine. Model parameters were estimated from data collected during pilot-scale experiments. The model was simulated using EPANET-MSX, a hydraulic and multi-species water quality modeling tool, and applied to models of both a pilot scale pipe loop and a real-world drinking water system. The model is simple enough to apply to full-scale distribution system models to investigate naturally occurring arsenic or accidental contamination with arsenic species.