In situ remediation of leaks in potable water supply systems

1.5.1 In situ remediation of concrete in water mains

The Delaware Aqueduct is an extreme exemplar of infrastructure challenges associated with many water mains in the United States. The Aqueduct provides half of New York City’s daily water demand via the world’s longest concrete-lined tunnel (Simpson, Wegner, & Michaels, 2009). After 65 years of service, very small cracks have formed in the concrete liner of this engineering marvel, leaking 35 million gallons of water per day at a cost of $28 million annually for water alone (Kennedy, 2001; NYC, 2008; Simpson et al., 2009). The leaks have also flooded homes, created massive sinkholes, and raised fears of a catastrophic pipeline failure event, which would have devastating economic and public health consequences for the city (Simpson et al., 2009). Conventional repair of leaks via construction of a 2.5-mile bypass tunnel for just one part of this pipeline is scheduled to be completed no sooner than 2021 at a cost of $2.1 billion (NYC, 2011).

The inability to easily access the 91–455 m (i.e., 300–1500 ft) deep pipeline or even shut it off for inspection/repair prompted an evaluation of novel approaches to mitigate leaks, reduce the chance of failure, and extend the lifetime of the pipeline. A bench-scale study of in situ remediation via temporary formation of suspended calcium carbonate by increasing the pH of the water was undertaken with very promising results (Letterman, Chen, Lavrykova, & Snyder, 2008). The change in water chemistry was able to reduce the rate of leaks by 55% in just 2.5 h by clogging simulated cracks (Letterman et al., 2008). This novel approach is currently being pilot-tested as a promising stop-gap measure, to buy time and increase the chances that effective repairs can be implemented on a cost, plan, and schedule suitable to ratepayers and the utility. The approach is justified by the short-term benefit alone, but permanent or semipermanent repair is also a possibility, as was proven by Roman engineers.

1.5.2 In situ remediation of copper pinholes in premise plumbing systems

A case study by Scardina et al. (2008) demonstrated that remediation of copper pinholes is sometimes occurring naturally in premise plumbing without our prior knowledge (Figure 2). Specifically, two adjacent cities used essentially the same source water, but one had very high incidence of copper pinhole leak damages in buildings while the other did not. Thorough investigation revealed that a similar number of leaks were probably occurring in both cities. The difference was that in one city after a few drops of water leaked the copper pinhole would completely heal itself (Figure 2B) before the consumer even became aware of the problem. This seems to be analogous to autogenous (self) repair that sometimes occurs for concrete, but which occurs as a result of metallic corrosion filling the leaks with scale (rust), as opposed to formation of precipitates in the concrete cracks as a byproduct of concrete dissolution reactions. The only evidence of leaks were visually repaired holes roughly every 0.08 m (1/4″) along a 0.9 m (3′) copper pipe at a pressure of 344,738 Pa (50 psi) with no resulting water damages or repair costs. This particular case of in situ remediation has now been documented to persist for more than a decade. In the adjacent city, the leaks grew rapidly after they first formed (Figure 2A), and each leak created up to thousands of dollars of water damage per incident. One city distributed their water at a slightly different pH and used a different type of corrosion inhibitor. It is therefore speculated that differences in water chemistry likely determined whether the leaks eventually caused the failure or in situ remediation, because there were no significant differences in the copper pipes. Preliminary investigation showed that the materials in the sealed holes were mainly copper corrosion precipitates, with some silica, calcium, and magnesium compounds.

Figure 2

Conventional wisdom is that leaks invariably get larger with time (A) (adapted with permission of Scardina & Edwards, 2008), but it has been discovered that, in some real-world situations, copper leaks can be remediated if appropriate water chemistry is present (B) (adapted with permission of Scardina et al., 2008).

Another case study by Lytle and Nadagouda (2010) verified the same observation by studying copper pitting corrosion in buildings of Cincinnati, Ohio. The water dripping rate through certain copper pinhole leaks decreased, due to blue-green corrosion deposits, such as Cu₄(OH)₆SO₄ and Cu₂O, which formed in the pit structure. Aluminum and silica were also identified in the corrosion deposits. No microbiological activity was identified; thus, water chemistry was demonstrated to play an important role to form the deposits, stopping or reducing the leaks. The water had high pH (∼8.8), chlorine dioxide (0.1–0.37 mg/l), low alkalinity (∼50 mg/l as CaCO₃), and significant levels of chloride (∼64 mg/l) and sulfate (∼120 mg/l).

4.1.1 Water velocity

Water velocity within the pipe can influence many aspects of corrosion, including the type of corrosion (i.e., uniform, erosion corrosion, and cavitation) or the chemical and mechanical conditions at the metal-solution interface, and concentration gradients of dissolved and particulate constituents at the pipe wall (Calle et al., 2007; Taxén et al., 2012; Vargas et al., 2010). Crossflow filtration in membrane gives insight into the potential effect of water velocity on the transport of water constituents within a leak. Specifically, the water velocity is expected to influence diffusion, the shear stress between water flow and the wall, and therefore the potential volume of particles potentially deposited (or eroded from within) on the leaks. As an example of how lower water flow can contribute to autogenous repair, stagnant water allowed healing of cement cracks five times faster compared to water velocity of 0.025 m/s in one study, although differences in the test specimens and crack geometry provide only strong suggestive proof (Wagner, 1974).

Water velocity above 3 m/s can cause copper pipe to fail in field studies due to wall penetration from erosion corrosion (Scardina et al., 2008). Conversely, higher water velocities may be beneficial when delivering more oxygen to the holes/cracks, allowing formation of more corrosion precipitates, which can potentially seal leaks. One study showed that more oxide precipitates were formed at higher velocity (1.2 m/s) in mild steel pipes than lower velocity (0.6 m/s) with 9.0±0.1 mg/l DO at 20°C (Rodolfo et al., 1987), and similar trends might occur within scale formed within leaks. One experiment with flushing demonstrated that high water velocity (0.167 m/s) caused detachment of more particles from the copper pipes into the water mainly due to the greater shear stress than the lower water velocity (0.022 m/s) (Vargas et al., 2010).

Overall, it is expected that lower velocities within leaks might enhance the likelihood of particle attachment and physical blocking of plastic pipe holes by waterborne/water-formed particles. While delivering a larger volume of particulate matter to the proximity of holes would be beneficial, higher scouring rates and reduced attachment would be detrimental. It can also be expected that impacts of velocity would be complex and highly dependent on circumstances.

4.1.2 Pressure

Higher pressure in the pipeline would be anticipated to reduce the likelihood of in situ remediation, as it increases the force on deposits that are needed to plug the holes/cracks and the detachment of accumulated particles. Experimental studies showed that four times higher water pressure (97,000 vs. 24,000 Pa; i.e., 14 vs. 3.5 psi) on 2×10^-4 m wide concrete cracks could lead to 50% less healing, as measured by water flow through the cracks (Edvardsen, 1999). Pressure might also be an impediment to in situ remediation if the sealed leaks fail to withstand the imposed pressure, and the pressure load versus strength recovery of the sealed cracks/hole may become one key to long-term success of in situ remediation (Neville, 2002). The capability of sealed cracks or holes to withstand a typical 140,000–420,000 Pa (20–60 psi) pressure in premise plumbing systems, or even higher pressure (420,000–1,400,000 Pa; i.e., 60–200 psi) in water mains (Woodson, 2006), might be limiting factors. One field study cited herein demonstrated that remediation of leaks in copper pipes by forming scales could occur even under typical premise plumbing pressures of 350,000 Pa (50 psi) (Scardina et al., 2008) and the Roman systems also sometimes operated at high pressure (>1,400,000 Pa; i.e., 200 psi; Pollio 15 BC as translated by Morgan, 1960). No experimental or other field data are currently available to demonstrate the role of pressure on the three mechanisms of in situ remediation.

4.1.3 Temperature

Temperature is a master variable due to its role in solubility, kinetics of precipitation, crystallization, diffusion, sedimentation, and virtually all aspects of in situ remediation mechanisms. Temperature can also affect DO solubility, corrosion precipitates’ solubility, redox kinetics, and biological activity in water (McNeill & Edwards, 2000). Neville (2002) found that lower temperature helped heal cracks in concrete and speculated the enhancement was due to thermal contraction of leaks, whereas high temperature increased the crack size and delayed the remediation of leaks. Temperature has a complicated effect on in situ remediation since it would affect the kinetics of the dissolution reaction from metal (tending to increase the hole size) and the corrosion or precipitation reactions (tending to decrease the hole size). The destiny of leaks and possibility of in situ remediation would depend on which trend is dominant.

A short-term (4.5 months) experiment with intact iron pipes showed that lower temperature (5°C) created slightly more corrosion products on the pipe surface and more iron leaching into water compared to higher temperature (20–25°C) (McNeill & Edwards, 2000). The much bigger volume of the corrosion precipitates created in this study might suggest an improved likelihood of in situ remediation. On the contrary, higher temperature would generally accelerate the kinetics of other reactions that enhance growth of leaks for copper including erosion corrosion and release of particulates (Obrecht & Quill, 1960a, b, c, d, e, f). For intact copper pipes, a 6-month study indicated that higher temperature (60°C) would release a higher proportion of particulate copper in soft water (5 mg/l alkalinity as CaCO₃) compared to lower temperature (4°C, 20°C, and 24°C) (Boulay & Edwards, 2001). Hot temperature could, on the contrary, accelerate the aging of copper pipes and help to develop a stable uniform layer of corrosion products on the pipe surface (Lytle & Nadagouda, 2010). As a result, the dominant rate between metal leaching rate and corrosion product formation rate in leaks determines the pathway toward remediation or failure. As was speculated for concrete, any pipe will expand with higher temperature and contract with lower temperature.

4.1.4 Leak size

Larger leaks are expected to be more difficult to seal due to the need to accumulate more adherent precipitates to clog up the leaks, and the greater total force that must be withstood by the area of new deposit in the leaks. Various authors reported that in situ remediation of concrete was not possible in reasonable time periods if the crack width exceeded a certain threshold (Li & Yang, 2007; Neville, 2002; Wang et al., 1997). This threshold ranged from 150 μm for cracked engineered cementitious composite specimens in a study by Li and Yang (2007) to 200 μm in research by Wang et al. (1997). The reported threshold size of a repairable crack width varies in different studies because of variations in water chemistry and other test conditions as well as differences in the concrete test specimens. Considering the novel concept of in situ remediation for metallic and inert pipes, no prior studies have examined the effect of leak size in these materials. It is logical to expect that larger leaks in copper, iron, and plastic pipes would also be more challenging to in situ remediation, similar to observations for larger concrete cracks.

4.2.1 pH

pH is also a master variable in drinking water systems, which controls many chemical reactions in water. Both acidic and basic conditions can be corrosive to metallic pipes, but acidic conditions often tend to favor metallic ions as reaction products (i.e., soluble form), whereas basic conditions tend to form corrosion precipitates (i.e., insoluble form). This trend would be expected to increase the likelihood of success of in situ remediation at higher pH. Indeed, in iron pipes, higher pH (pH 7–9) was generally found to develop a greater volume of corrosion products on pipe surfaces (McNeill & Edwards, 2001), a necessary step in the remediation of leaks. Another study indicated that higher pH (7.5–9.5) also decreased soluble iron release into the water, minimizing the chance of red water (Sarin et al., 2004). At very high pH, however, it is possible that corrosion rates would be substantially reduced, to the point that leak clogging by corrosion precipitates might be impeded.

For copper pipes, some studies identified a difference between the pH of local water near the pinhole (pH 3.5–5.5) and the pH of the bulk water inside the pipes (pH 8.8) (Edwards et al., 1994; Lytle & Nadagouda, 2010; Pourbaix, 1984). This local pH change might also play a role in preventing remediation of copper pinholes, because the lower pH near the leak would accelerate corrosion and creation of soluble ions. For concrete, a 2-year study on reinforced concrete cracks showed that moderate pH (pH 7) lowered the flow rate through cracks, while lower pH (pH 5.2) allowed more flow through cracks (Ramm & Biscoping, 1998). This indicated potential success of in situ remediation at higher pH and a path to failure at lower pH.

Aside from effects on reactive pipe materials, pH may influence physical clogging of leak holes in inert pipes. pH determines the point of zero surface charge (PZC) for waterborne particles and plastic pipes in drinking water systems, which in turn influences the size of suspended particles present in water by coagulation processes. The attractive or repulsive force between particles and crack/leak walls will also be highly dependent on the pH for the point of zero charge (pH-PZC) and therefore affects the chance of particles adhering to block pipeline cracks or holes. Thus, at or near the PZC of waterborne or water-formed particles, the likelihood of larger particles and attachment to holes (Figure 4) will tend to be maximized. The pH-PZC is controlled by the type of particles (Naidu et al., 1997). For example, aluminum hydroxide, calcium carbonate, and silica particles are commonly present in potable water supplies and their pH-PZCs are quite different. The pH-PZC is 8.5 for aluminum hydroxide, 8–10 for calcium carbonate, and 2–4 for silica in relatively pure water (Edzwald, 2011; Kosmulski, 2009; Montgomery, 1985; Moulin & Rogues, 2003; Noh and Schwarz, 1989). Thus, clogging of leaks in inert materials by waterborne particulates is expected to demonstrate different optimal pH dependent on the particles present.

4.2.2 Carbonic species/alkalinity/calcium carbonate

Calcium and carbonate are two important components participating in in situ remediation of concrete materials. Carbonate species participated in self-repair of concrete in experiments, and the material repairing concrete cracks was mainly calcium carbonate (Clear, 1985; Hearn, 1998; Hearn & Morley, 1997; Lauer & Slate, 1956; Neville, 2002; Parks et al., 2010). As a common waterborne particle in natural water sources or by purposeful formation/addition, suspended calcium carbonate was proven capable of physically clogging simulated cracks of aged concrete pipes in a bench-scale study (Letterman et al., 2008).

In copper pipes, one study proved that the soluble copper concentration in bulk water leached from copper pipes increased with the free carbon dioxide (Broo et al., 1998; Rehring & Edwards, 1996). In monitoring for copper to tap water, the highest copper concentration was observed when alkalinity (7.4 mg/l as HCO₃^-) and pH (6.4) were very low (Broo et al., 1998; Dodrill & Edwards, 1995; Rehring & Edwards, 1996), and corrosion rates showed a similar dependence (Broo et al., 1998; Rehring & Edwards, 1996). Overall, the presence of carbon dioxide and alkalinity can be expected to exert a strong effect on the corrosion rates of copper, copper speciation, and the likelihood for in situ remediation.

For cast iron pipes, alkalinity is a controlling factor in multiple steps of the corrosion process and can be expected to influence leak remediation. For example, high alkalinity and higher buffering intensity tend to reduce the corrosion rate (McNeill & Edwards, 2001; Sarver & Edwards, 2012) and the corresponding volume of corrosion precipitates that form. Carbonate could also help remediation by formation of siderite (FeCO₃) scale, which is believed to be more protective in pipes compared to ferric species, such as goethite (FeOOH) and hematite (Fe₂O₃) (McNeill & Edwards, 2001; Mishra et al., 1992). Free carbon dioxide, on the contrary, could complex with iron oxides and cause higher soluble ferrous iron in water (McNeill & Edwards, 2001) and possibly put leaks on a path to failure.

4.2.3 NOM

NOM is an important constituent of drinking water controlling corrosion rates, colloid stability, and solubility (Rehring & Edwards, 1996). NOM interactions are very complex due to its inherently heterogeneous nature, and after decades of research, these reactions are moderately well understood if not perfectly predictable chemically. Thus, leak in situ remediation in the presence of NOM may be complicated due to its complex secondary impacts on water chemistry and microbes.

For iron pipes, NOM (0.2 mg/l Aldrich humic acid) in one study had no effect on iron corrosion rates under both saturated and low oxygen conditions when microbes were absent (Zhang, 2005). However, in another study, NOM increased the formation of iron oxide scale layers and inhibited iron dissolution (Campbell & Turner, 1983), potentially increasing the likelihood of in situ remediation. For copper pipes, NOM was an effective corrosion inhibitor when its concentration was above 0.3 mg/l in aggressive water, as defined by Larson’s index (Edwards et al., 1994; Murray-Ramos, 2006; Rehring & Edwards, 1996). NOM is also known to reduce leaks by stopping copper pitting corrosion (Edwards et al., 1994; Sarver & Edwards, 2012). Korshin et al. (1996) also showed that NOM could markedly enhance nonuniform corrosion tendencies and potentially form more precipitates in a narrow range of concentrations (0.1–0.2 mg/l), therefore enhancing the likelihood of leak in situ remediation.

Calcite (calcium carbonate) precipitation could also be poisoned by NOM (Lin, 2005; Lin & Singer, 2005; Lin et al., 2005) and inhibit self-healing of concrete. NOM can also coat suspended particulates and decrease the zeta potential, thereby stabilizing colloids in smaller size ranges (Stumm & Morgan, 1996), which could tend to prevent clogging of leaks.

4.2.4 Phosphate corrosion inhibitors

Polyphosphate, orthophosphate, zinc orthophosphate, zinc metaphosphate, and bimetallic phosphate are commonly used iron and copper corrosion inhibitors in the United States (Lewandowski et al., 2010; McNeill & Edwards, 2000, 2001). These inhibitors affect both the metal release rate and scale formation rate within existing leaks, thereby influencing in situ remediation. For example, McNeill (2000) showed that the presence of 1 mg/l orthophosphate as P at pH 9.5 decreased iron release from cast iron pipe into water but increased scale buildup on the pipe wall, relative to conditions without phosphate. The greater rate of scale buildup might indicate that phosphates could assist in remediation. Corrosion inhibitors can also slow the copper corrosion rate and reduce the copper release rate by forming a protective insoluble layer [such as Cu₃(PO4)₂] outside the pipe (Edwards & McNeill, 2002; Lewandowski et al., 2010; Vargas et al., 2009). This also indicates a potential for in situ remediation of copper pipe leaks.

4.2.5 Magnesium, silica, and calcium

Silica is naturally present in water and can be added as a corrosion inhibitor. Silicate was reported to decrease the corrosion rate of iron and is expected to become part of the corrosion products via sorption (Davis et al., 2001; McNeill & Edwards, 2001). Rushing et al. (2003) noted that higher levels of silicate (50 mg/l as SiO₂) would release more iron into water, initially causing a red water problem but in the longer term would then be incorporated into a protective scale layer of the pipe, which might stop leaks. Silica has also been identified in the corrosion products of leaks in copper pipes (Lytle & Nadagouda, 2010). In one field study where pits in copper pipes were remediated, the deposits were mainly composed of copper and silica with a small amount of calcium and magnesium (Scardina & Edwards, 2008). Calcium silica hydrate and magnesium silica hydrate were also identified in sealed concrete cracks as facilitators of in situ remediation (Parks et al., 2010; Santhanam et al., 2002, 2003). Magnesium, calcium, and silica can therefore contribute to remediation of leaks in a wide range of circumstances. On the contrary, silicate can coat iron hydroxide particulates and create a highly negative surface charge, maintaining smaller colloids and preventing attachment in leaks (Davis et al., 2001).

4.2.6 Sulfate/chloride

Sulfate has been identified as aggressive to concrete and cause concrete degradation or expansion in concentrated sodium sulfate and magnesium sulfate solutions (Neville, 2004; Santhanam et al., 2002, 2003). This trend might reduce in situ remediation potential of water by enhancing dissolution rates. Sulfate has mixed effects on iron corrosion depending on pH and flow conditions. One study showed that, at 35°C with pH close to 3 in sulfuric acid solution for 6 h, the weight loss of pure iron coupons increased with sodium sulfate concentrations (from 0.1 to 0.5 mol/l; i.e., from 0.1 to 0.5 M) but decreased thereafter (from 0.5 to 1 M sodium sulfate) (Traubenberg & Foley, 1971). That study concluded that sulfate enhanced corrosion rate, but the extrapolation to results at circa neutral pH in potable water is unclear.

Chloride is expected to have a complex influence on in situ remediation. For copper, chloride tends to help seal the leaks in the longer term due to stifling reactions and possible formation of CuCl (Edwards et al., 1994; Mattsson & Fredrikksson, 1968). For iron, one study indicated that chloride would complex with iron in the scale and increase ferrous iron concentration in water that could potentially help remediation by forming more solids (Elzenga et al., 1987; McNeill & Edwards, 2001). However, Van Der Merwe (1988) showed that chloride had no effect on the weight loss of cast iron, reinforcing the possible complex impact of chloride on in situ remediation.

4.2.7 DO

DO is probably the dominant electron acceptor for corrosion in potable water (see cathode reactions in Table 1). This might lead to increased rates and concentrations of corrosion scale relative to situations with low DO. Perhaps more importantly, at low DO for iron, dissolved Fe²⁺ species will predominate relative to very insoluble Fe³⁺ species that tend to favor rust formation (AWWARF, 1996; Lytle & Nadagouda, 2010; Sarin et al., 2004). More precipitates will potentially form compared to metal ion leaching into water with increasing DO concentration, thereby reducing leaks. For example, approximately 18 mg/l less iron was leached into the stagnant water compared to the water without any DO (pH 6.2–6.8) in intact iron pipes in 2 h, and iron scales were observed outside pipe surface with the presence of 9.7 mg/l DO (Sarin et al., 2004). For copper pipes, little or no corrosion occurs in the absence of oxygen, so at very low DO copper pipe might behave as an inert material (Edwards et al., 1994). Overall, DO could be an important parameter in in situ remediation of copper and iron pipes through participation in corrosion reactions but would not be expected to affect remediation of inert concrete and plastic pipes.

4.2.8 Water disinfectant

Free chlorine, monochloramine, and chlorine dioxide are commonly used for drinking water disinfection in the United States. Disinfectants could influence in situ remediation by (1) increasing corrosion rates (metal dissociation rate and scale formation rate), (2) promoting the formation of less soluble species, and (3) controlling or minimizing biofilms in water and on pipe surfaces.

By nature, these disinfectants can accelerate corrosion (e.g., copper and iron pipes) by redox reactions (Frateur et al., 1999; Hallam et al., 2002; LeChevallier et al., 1993; Nguyen, 2005; Sarin et al., 2001; Zhang et al., 2008). Several studies have shown varied effects of chlorine and chloramine on corrosion rate in metallic pipes depending on water chemistry (Nguyen, 2005; Zhang, 2005). Sarver et al. (2011) indicated that high levels of chlorine could encourage copper pitting (nonuniform) corrosion, and several researchers (Lytle & Nadagouda, 2010; Nguyen et al., 2010; Pourbaix, 1974) have demonstrated that the concentration of ions within pits was dramatically different than in bulk water, which could affect the reactions contributing to or hindering self-remediation. For example, very low pH <4 within pits could hinder the likelihood of self-remediation after leaks begin, but very high levels of Cu(I) ions, which are easily oxidized to Cu(II) oxide scale, could enhance the likelihood of remediation. Sarver et al. (2011) found that chloramine did not contribute to pitting corrosion to the extent that chlorine did, but it nonetheless oxidized copper and contributed to scale formation (Nguyen et al., 2010).

These disinfectants can also transform corrosion products and form less soluble species, potentially promoting the chance of in situ remediation. Specifically, both iron and copper pipe scale products were shown to react with chlorine dioxide in a first-order rate (4.2×10^-4 to 1.4×10^-3 s^-1 for iron and 8.7×10^-5 to 1.0×10^-4 s^-1 for copper), which could oxidize the main scale components Fe₃O₄ or Cu₂O into less soluble ferric or cupric solids (Zhang et al., 2008). The ferrous iron, which was commonly found in corrosion products [e.g., FeCO₃ or Fe(OH)₂], could also be oxidized by free chlorine, forming more insoluble Fe (III) oxides, such as Fe₂O₃ or FeOOH (Sarin et al., 2001). Free chlorine was also demonstrated to react with copper corrosion precipitates [e.g., Cu₂O or Cu(OH)₂] to form more aged solids, such as tenorite (CuO) (Nguyen, 2005).

Disinfectants can control or inhibit microbial activity and biofilms both in water and on pipe surfaces and therefore decrease microbially induced corrosion (MIC) and resulting leaks in pipelines (Edwards et al., 2000; Hamilton, 1985; LeChevallier et al., 1993). However, this process is greatly affected by corrosion byproducts, which could provide a habitat for bacteria and decrease the efficiency of disinfection (Volk et al., 2000). In addition, the type of disinfectant would control biofilms, which might have their own propensity to induce leak repair via biofouling. Overall corrosion, disinfectant type, and microbial activity are expected to sometimes work in synergy and sometimes in opposition in relation to in situ remediation.

4.2.9 Turbidity

High turbidity is conventionally considered an indicator of poor water quality due to its indication of poor removal of waterborne microorganisms. However, in Roman pipelines in 15 BC, wood ashes were purposefully added to water to help in situ remediation of leaks by clogging (Pollio 15 BC as translated by Morgan, 1960). Waterborne or water-formed particles have been acknowledged to physically block holes or cracks in concrete materials (Clear, 1985; Edvardsen, 1999; Hearn, 1998; Letterman et al., 2008; Neville, 2002; Ramm & Biscoping, 1998). This indicates that some forms of turbidity could be helpful in in situ remediation.

4.2.10 Ionic strength

Ionic strength can affect the stability of waterborne and water-formed particles by compressing the electrical double layer and increasing the likelihood of attachment of particles to surfaces (Edzwald, 2011). Thus, ionic strength could affect the type and size of particles present in water and the chance of in situ remediation via physical clogging of leak holes.

Ionic strength can also influence corrosion products and corrosion rates (Dunn et al., 2000; Smart & Bockris, 1992). Higher ionic strength generally has a positive effect on the corrosion rate of metals. Dunn et al. (2000) indicated that the corrosion rate for iron at lower ionic strength (0.0028 M Cl^-) was, on average, 74.4% slower (2.1×10^-4 vs. 8.2×10^-4 m/y) than that at higher ionic strength (0.028 M Cl^-). Moreover, corrosion products such as lepidocrocite (γ-FeOOH) and magnetite (Fe₃O₄) formed in low ionic strength (0.028 M Cl^-) solution versus the more soluble akaganeite (β-FeOOH) in high ionic strength (0.028 M Cl^-) solution. In another study by Smart and Bockris (1992), the corrosion rate of iron linearly increased with increasing ionic strength. In a study by Zhang et al. (2002), the corrosion rate of copper increased when ionic strength increased from 0.005 to 0.01 M. The corrosion products [Cu₂O and Cu(OH)₂] were more soluble due to smaller copper ion activity at higher ionic strength. All these studies indicate that ionic strength could change the solubility and type of corrosion products, which could exert considerable influence on leak remediation efforts.