Microelectrode Investigation on the Corrosion Initiation at Lead–Brass Galvanic Interfaces in Chlorinated Drinking Water

Abstract

In this study, the effects of pH, dissolved inorganic carbon (DIC), and flow on changes in surface chemistry (pH, dissolved oxygen, and free chlorine) of lead-brass joints at initial stages of corrosion were investigated using microelectrodes. Surface measurements showed that the water chemistry at the metal surfaces was highly heterogeneous. At pH 7 and during water stagnation, local pH difference between anodic (leaded-solder) and cathodic (brass) regions differed by as much as 7.5 pH units. High DIC water under the water flowing condition showed minimal pH changes on the surface, whereas in low DIC water, a pH range of 7.6−5.4 (ΔpH 2.2) was observed over the surface. Free chlorine consumption near the lead-brass surface was greater under stagnation, regardless of bulk pH. It was also found that flow can move the low pH plume that originated at the anode. Overall, this study provides direct evidence for highly localized galvanic corrosion in a chlorinated drinking water environment.

Graphical Abstract

INTRODUCTION

Excessive corrosion of drinking water distribution system (DWDS) materials can cause pipe damage, water leaks, and degrade water quality, which can create structural integrity issues, economic losses, and public health concerns.¹ Galvanic corrosion refers to an electrochemical process whereby two dissimilar metals are in electrical contact in an electrolyte and one corrodes preferentially to the other. Galvanic corrosion in a DWDS can be an important source of lead (Pb) in drinking water. The general concepts and theories of galvanic corrosion processes with lead have been widely studied with the following chemical reactions:^2–4

• reaction in the anode

  $$\text{Pb}\to {\text{Pb}}^{2+}+2{\text{e}}^{-}$$ (1)
• reaction in the cathode

  $${\text{OCl}}^{-}+{\text{H}}_2\text{O}+2{\text{e}}^{-}\to {\text{Cl}}^{-}+2{\text{OH}}^{-}$$ (2)

  $${\text{O}}_2+2{\text{H}}_2\text{O}+4{\text{e}}^{-}\to 4{\text{OH}}^{-}$$ (3)

In the reaction in the anode, Pb can be oxidized to lead ions (Pb²⁺) (eq 1). In the reaction in the cathode (eqs 2 and 3), oxygen and chlorine species can both serve as the oxidizing reagents in the galvanic corrosion. These reactions can be accelerated in a very short time under the stagnation process. Many water quality factors, including pH,^4,5 alkalinity,^5,6 dissolved oxygen (DO),⁶ chlorine residual,^7,8 chloride,^9,10 sulfate,¹⁰ corrosion control chemical agents (i.e., phosphate and silicate),^5,11 chloride/sulfate mass ratio (CSMR),^4,12 natural organic matter (NOM),¹³ and type of disinfectant,⁸ can impact the extent of corrosion.

Despite the importance of galvanic corrosion, our understanding of the process could be improved. Although many related studies have been conducted, experimental approaches have largely been based on macroscale measurements and observations.^4,12,14 Given the local nature of galvanic interactions, building from past efforts to understand water-metal interfacial reactions could provide valuable new insights.^1,4,15–20 For example, Church et al. successfully measured direct pH change and a monochloramine concentration decrease at the surface of a real pitting copper sample from a house plumbing system.²⁰ Most recently, Ma et al. demonstrated the ability of constructing in situ two-dimensional (2D) maps at the surface of corroding galvanically connected metals using a microelectrode technique.¹

The objective of this study was to use microprofiles and 2D maps generated from microelectrode measurements to evaluate the effects of water flow, pH, and dissolved inorganic carbon (DIC) concentration on the pH, free chlorine, and DO concentration changes at the surface of freshly prepared galvanically coupled brass-leaded-solder coupons. The novelty of this study was to provide direct evidence of cathodic-anodic reactions at the water-metal interface (100 μm above the surface) by constructing 2D maps of pH and free chlorine concentrations in various water conditions during the initiation state of galvanic corrosion processes.

EXPERIMENTAL SECTION

Materials and Reagents.

Metals used in this study were CDA 443 brass (UNS C44300, admiralty brass,$\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$16$}\right.$ in. thick × 0.47 in. wide × 5.51 in. long, with a density of 8.52 g cm⁻³, Metal Samples Co., Munford, AL, U.S.A.), which contains 70–73% copper, 25–28% zinc, 0.07% lead, and other metal impurities, and 50:50 Pb/Sn solder (38110, Forney). The original brass slide was cut into several small pieces (1.6 mm thick × 12 mm wide × 10 mm long) and cleaned using a combination of two ASTM International coupon wash procedures: ASTM G31–72 and ASTM D2688–83.^21,22 Two brass coupons were joined by a 1.6 mm thick × 12 mm wide × 2.0 mm long band of Pb/Sn (38110, Forney) solder, and a cold epoxy specimen mounting technique was applied to provide a smooth even surface and polished with 800 grit sandpaper afterward (Figure S1 of the Supporting Information). A matrix of five experimental conditions were evaluated to examine the impact of flow, pH, and DIC in prepared chlorinated drinking water on galvanic corrosion (Table 1). Reagent-grade chemicals (NaCl, Na₂SO₄, and NaHCO₃, Fisher Scientific, Fair Lawn, NJ, U.S.A.) were used to prepare water to desired chloride (100 mg of Cl⁻ L⁻¹), sulfate (100 mg of SO₄²⁻ L⁻¹), and DIC (10 or 50 mg of C L⁻¹) concentrations. The prepared water was then air-saturated by bubbling air through the water before pH adjustment and free chlorine addition. Free chlorine stock solution (8000 mg of Cl₂ L⁻¹) was then prepared from 6% sodium hypochlorite (NaClO, Fisher Scientific, Fair Lawn, NJ, U.S.A.) and spiked to the prepared experimental water (4 L) to achieve the desired free chlorine concentration (2 mg Cl₂ L⁻¹). HCl and NaOH (Fisher Scientific, Fair Lawn, NJ, U.S.A.) were used to adjust pH.

Table 1.

Experimental Conditions for Microprofiling and 2D Mapping of Galvanic Joint Coupons in Synthetic Drinking Water^a

experimental group	flow condition	pH	free chlorine (mg of Cl2 L−1)	DIC (mg of C L−1)
Ex 1, control	Ex 1F, flow	7	2	10
	Ex 1S, stagnation	7	2	10
Ex 2, DIC	Ex 2F, flow	7	2	50
	Ex 2S, stagnation	7	2	50
Ex 3, pH	Ex 3F, flow	9	2	10
	Ex 3S, stagnation	9	2	10
Ex 4, pH + DIC	Ex 4F, flow	9	2	50
	Ex 4S, stagnation	9	2	50
Ex 5, non-continuous stagnationb	Ex 5F, stagnation	7	2	10

^aAll conditions include 100 mg of Cl⁻ L⁻¹ and 100 mg of SO₄²⁻ L⁻¹ to provide background ionic strength with a chloride/sulfate mass ratio (CSMR) of 1:1.

^bNo previous flow condition.

Experimental Conditions.

In each experiment, except Ex 5S (Table 1), two consecutive conditions were applied for each prepared water environment to investigate the effect of flow pattern: flow and then stagnation (e.g., Ex 1F and Ex 1S in Table 1). First, a freshly prepared galvanic joint coupon was pre-conditioned in the prepared chlorinated water in an acrylic glass flow cell (1.6 mm depth × 12 mm wide × 22 mm long, 150 mL of water volume) (Figure S2 of the Supporting Information) under the flow condition (2 mL min⁻¹) using a peristaltic pump (Cole-Palmer, Vernon Hills, IL, U.S.A.) for 1 h before the profile measurements. The flow rate of 2 mL min⁻¹ was selected considering electrochemical sensor utilization without noise and high spatial resolution.^1,19,23 The influent part of the flow cell has three inlet holes to distribute the flow evenly. After one-dimensional (1D) microprofiling and 2D mapping under flow, the pump was turned off and the consecutive stagnant condition was followed. The coupons were then pre-conditioned under stagnation for 2 h before microelectrode investigation. To eliminate the effect of the previous flow condition on the coupon surface chemistry, non-continuous stagnation was also conducted as a control experiment for comparison (Ex 5S). In the non-continuous stagnation, the newly prepared coupon was immersed initially in the same stagnant water environment (Ex 1S), and then profiles were measured after 2 h of stagnation pretreatment. The measured results were compared to the data from Ex 1S for reference purposes.

Microelectrode Measurement: 1D Microprofiling and 2D Mapping.

Multiple 1D concentration microprofiles were measured using pH (pH 10, 10 μm tip diameter, UNISENSE A/S, Denmark), free chlorine (10 μm tip diameter, self-made), and DO (O₂ 50, 50 μm tip diameter, UNISENSE A/S, Denmark) microelectrodes.^23,24 The microelectrodes used in this study have been successfully applied to many applications (e.g., biofilm,^25,26 corroded material,^1,19 emulsion,²⁷ photocatalytic metal surface,²⁸ and plants²⁹). For the microprofiling experiment (Table 1), first each fresh coupon was placed in a transparent flow cell under each water condition for 1 h. During the pre-conditioning period, a well-calibrated microelectrode was connected along with a dip-type Ag/AgCl reference electrode (MI-401, Microelectrodes, Inc., Bedford, NH, U.S.A.) and the sensor tip was positioned above the target metal surface by carefully controlling a three-dimensional (3D) manipulator (UNISENSE A/S, Denmark) under a microscope (World Precision Instruments, Sarasota, FL, U.S.A.). Then, the microelectrode tip was positioned at 2000 μm above the target coupon surface, and the 1D microprofile was measured at 50 μm intervals until 50 μm above the surface. There was a 15 s waiting time before the stabilization for the pH and DO microelectrode and 5 s for free chlorine microelectrode between each single-point measurement. Every microprofile was measured in duplicate to ensure that profiles were representative and reproducible. The average time for completing a 1D microprofile was about 10 min. All microprofile measurements were conducted in a Faraday cage (81-334-04, Technical Manufacturing Co., Peabody, MA, U.S.A.) to minimize electrical interference.³⁰ Representative calibration curves (pre- and post-) for pH and free chlorine microelectrodes were shown in Figure S3 of the Supporting Information.

The 2D pH and free chlorine maps were constructed by moving the microelectrode tip across a grid (x and y axis) at a distance of 100 μm above the metal coupon surface (Figures S2 and S4 of the Supporting Information). For each experiment, 8400 μm (x) × 20 000 μm (y) of surface area was mapped at intervals of 1200 μm (x) × 1500 μm (y) with a total of 112 data points. Approximately 20 s was taken for recording and averaging data for each point. The average time for one 2D map construction was 2.5 h. The details on the procedures of 2D map construction can be found elsewhere.¹⁹ All microelectrodes were pre- and post-calibrated. Free chlorine concentrations were validated using a colorimetric test kit (HACH-8021) using a DR 5000 spectrophotometer (HACH Co.). Data measured using microelectrodes were further processed using Python script (Python 2.7, Python software). Surface characterization of galvanic joints was also conducted for surface morphology and element analysis.

Coupon Surface Characterization.

Surface characterization was conducted using scanning electron microscopy (SEM), backscattered electrons (BSE), and energy-dispersive spectroscopy (EDS) (ZEISS ULTRA-55, Germany) to investigate element distribution and coupon surface morphology under a given water condition. In this study, a galvanically connected leaded-solder and brass coupon was exposed under one flow condition (Ex 1F), where lead leaching is expected as a result of aggressive galvanic reaction¹ for 2 h, and the surface was characterized to investigate the flow effect and identify lead leaching from the galvanic joint. SEM images and BSE images were used for investigating surface morphology and element distribution across the galvanic coupon, and EDS images provided information on elements on multiple locations.

RESULTS AND DISCUSSION

2D Mapping of Surface Chemistry Changes at pH 7 under Flow and Stagnation.

The pH, DIC, and free chlorine are important parameters that can affect the corrosion rate depending upon the type of plumbing materials.³¹ In this study, the corrosion of galvanically connected leaded solder and brass in chlorinated water (2 mg of Cl₂ L⁻¹) at pH 7 and 9 and DIC (10 and 50 mg of C L⁻¹) were evaluated. The 2D maps clearly illustrate the impact of DIC on surface pH and galvanic corrosion. In low DIC (10 mg of C L⁻¹) water at pH 7, pH differed greatly across surfaces under both flow and stagnant conditions, whereas at high DIC (50 mg of C L⁻¹) at pH 7, large pH differences were only observed under the stagnant condition (Figure 1a). In low DIC water at pH 7, pH ranged between 9.3 and 3.8 (ΔpH of 5.5) under the flowing condition, while the pH difference during stagnation was broader, ranging between 10.3 and 2.8 (ΔpH of 7.5) across the galvanic joint coupon surface. As a result of a galvanic current, the brass as a cathode can accelerate corrosion of the lead anode, releasing Pb²⁺.^1,3 The production of free lead (Pb²⁺), which is a Lewis acid, can cause a local pH drop upon removal of OH⁻ ions and draw chloride ions (Cl⁻) to the lead anode surface via the formation of soluble complexes or insoluble precipitations that contain OH⁻ and Cl⁻.^4,6 In addition, the local low pH at the leaded surface would coincide with a region of high lead solubility and a source of lead release to the bulk water.¹

Figure 1.

2D mapping of (a) pH and (b) free chlorine concentration on the galvanic coupon surface under different flow and DIC concentrations (pH 7, free chlorine of 2 mg of Cl₂ L⁻¹, 100 mg of Cl⁻ L⁻¹, and 100 mg of SO₄²⁻ L⁻¹). Dashed lines indicate leaded-solder area boundaries.

The low pH plume in low DIC water appeared to originate from the leaded solder and partially migrate with water flow downstream of the brass surface. This observation was confirmed by repeating the surface 2D pH mapping without a preceding flow period. While the preceding flow resulted in a pH decrease on the downstream brass surface (Figure 2a), without the preceding flow condition, the pH decrease was limited primarily to the leaded-solder region of the coupon because there was no flow during the galvanic reaction, and thus, the changes of pH with distance from the leaded-solder region were asymmetric (Figure 2b). This was probably due to the diffusion and transport of the lead ions, which were released from the galvanic reaction, depending upon the flow condition. The comparison between different flow conditions clearly demonstrated that flow (2 mL min⁻¹ in this study) would greatly influence the Pb²⁺ migration to the downstream and consequent deposition on the brass surface, resulting in a local pH decrease, even on the cathodic area (e.g., brass surface).

Figure 2.

2D pH mapping of (a) consecutive stagnation after initial flow condition versus (b) stagnation without initial flow under pH 7, free chlorine of 2 mg of Cl₂ L⁻¹, and DIC of 10 mg of C L⁻¹. Dashed lines indicate leaded-solder area boundaries.

Specifically, the pH change at the anodic region (lead) can be attributed to the Lewis acidity of lead and tin, and the increase of pH at the cathode (brass) may be attributed to the reduction of free chlorine species and possibly oxygen. In high DIC water at pH 7, the pH across the surfaces was greatly different depending upon flow. With flow, the pH across the surfaces was relatively uniform and was similar to the bulk water pH of 7 (Figure 1a), except for a small region over the anodic leaded solder that dropped to 5.4. Under stagnation, however, large pH changes were observed primarily over the leaded-solder portion of the coupon, while pH increases were observed on both brass coupon sections. Unlike in low DIC water, no low pH plume in the downstream was observed in high DIC water at pH 7. The pH ranged between 10.0 (inflow brass end) and a low of 3.1 (leaded solder) (ΔpH of 6.9).

The 2D map of the free chlorine concentration shows the distribution of free chlorine concentrations at the metal surface (Figures 1b and 3b). At pH 7, the complex pattern of the free chlorine concentration in higher DIC (50 mg of C L⁻¹) with flow indicates that the surface cathodic reaction by free chlorine (OCl⁻ + H₂O + 2e⁻ → Cl⁻ + 2OH⁻) was heterogeneous, while the reaction in lower DIC (10 mg/L) and under stagnation was less heterogeneous with free chlorine depletion at the surface, indicating active reaction between metal and free chlorine probably as a result of low buffering intensity. It was clearly visualized that a well-buffered system (e.g., DIC of 50 mg of C L⁻¹ and flow) at pH 7.0 showed relatively low free chlorine consumption at the metal surface compared to lower DIC and/or stagnant conditions.

Figure 3.

2D mapping of (a) pH and (b) free chlorine concentration on the galvanic coupon surface under different flow and DIC concentrations (pH 9, free chlorine of 2 mg of Cl₂ L⁻¹, 100 mg of Cl⁻ L⁻¹, and 100 mg of SO₄²⁻ L⁻¹). Dashed lines indicate leaded-solder area boundaries.

1D Microprofiles and Flow Effect at pH 7.

Figure S5 of the Supporting Information showed representative microprofiles of pH, free chlorine, and DO concentration under stagnation in a selected water environment (Ex 1S, pH 7, 2 mg of Cl₂ L⁻¹, 10 mg of C L⁻¹, 100 mg of Cl⁻ L⁻¹, and 100 mg of SO₄²⁻ L⁻¹). The galvanically connected coupon was conditioned in the test water for over 5 h (3 h of the flow condition and then 2 h of stagnation). Microprofiles showed that pH decreased or increased depending upon the lead surface location, with approximately 500 μm of diffusion boundary layer (DBL) under stagnation. The measured pH microprofiles (Figure S5a of the Supporting Information) showed that, during the initiation of galvanic corrosion, pH levels were significantly different (9.3 versus 2.7) in two different locations, even in a small lead joint area (0.84 × 0.3 mm) (Figure S6 of the Supporting Information), where the area of high pH reflects the cathodic region and the low pH area reflects the anodic region, where elevated levels of soluble Pb²⁺ ions from leaded solder locally build up.¹ The localized pH differences at the lead joint surface required construction of 2D surface maps to examine the pH dynamics across the galvanic joint. The free chlorine concentration decreased from 1.0 to 0.37 mg of Cl₂ L⁻¹ with 600 μm of DBL in a selected location of the brass surface (Figure S5b of the Supporting Information). However, a free chlorine concentration microprofile in a different location showed a different trend (e.g., free chlorine depletion at the brass surface), indicating that the free chlorine concentration, like pH, was heterogeneous across the metal surfaces. Therefore, 2D maps of pH as well as free chlorine concentrations were constructed to investigate the changes of pH and free chlorine concentrations across the brass-lead joint surface under different water conditions.

Interestingly, DO concentration microprofiles from bulk water to the coupon surface (Figure S5c of the Supporting Information) showed minimal oxygen consumption (less than 0.5 mg of O₂ L⁻¹) across the brass-lead joint surface in this study. The similar trends of minimal oxygen consumption on metal surface were also found under other experimental conditions (Table 1) within the time frame of this study (i.e., 5 h). Minimal oxygen consumption at the metal surface would be expected, where free chlorine residuals were present, which are much stronger oxidants than oxygen, in the system.

2D Mapping of Surface Chemistry Changes at pH 9 under Flow and Stagnation.

In low DIC and high pH 9 water, the plume of local pH decrease at the anodic leaded-solder region was evident but much smaller and the drifting effect was not apparent when compared to the pH 7 cases (Ex 3 and Ex 4) (Figure 3) suggesting that lower pH water is more aggressive to galvanic attack.⁴ The galvanic reaction was still clearly defined by the pH difference across the coupon surfaces between 10.1 and 4.7 under the flowing condition (ΔpH of 5.4), while under stagnation, the pH ranged between 9.8 and 6.3 (ΔpH of 3.5). In high DIC water at pH 9 under the flowing condition, there were minimal pH changes across the galvanic joint, as observed at pH 7. Under stagnation, however, the pH ranged between 9.6 and 5.9 (ΔpH of 3.7) in the high DIC water, which was similar to the pH distribution in the low DIC water at pH 9 (ΔpH of 3.5). The identical pH trends (evenly distributed regardless of galvanic joints) in high DIC were observed for both pH 7 and 9 under the flowing condition, and it was found that each surface pH distribution relatively followed each bulk pH (Figures 1a and 3a).

DIC is necessary to form lead carbonates (cerussite and hydrocerussite) and is directly associated with the buffer intensity of water.³¹ Buffer intensity reflects the ability of water to resist a change in pH and is pH- and DIC-dependent. For example, in 10 mg of C L⁻¹ DIC water at pH 7 and 9, the buffer intensity (β) is 2.88 × 10⁻⁴ and 1.09 × 10⁻⁴ mol L⁻¹ pH unit⁻¹, respectively (Figure S7 of the Supporting Information). An increase in DIC to 50 mg of C L⁻¹ at pH 7 can increase β by a factor of 5 to 1.44 × 10⁻³ mol L⁻¹ pH unit⁻¹. The increase in β could explain the absence of pH change (pH 7 and 9) on the metal surfaces under the flowing condition (Figures 1a and 2a). Thus, the continuous supply of sufficient carbonate to maintain buffering capacity may inhibit galvanic reaction on the galvanic joints or associated pH changes during flow. However, under water stagnation, regardless of bulk pH and DIC, the pH was stratified across the galvanic joints with low pH in the anodic leaded-solder region and high pH in the cathodic brass region (Figures 1a and 2a). The pH values across the coupon surfaces under stagnation are very similar, despite bulk pH, suggesting that galvanic-corrosion-associated reactions played a dominant role on local water chemistry. Considering experimental time over 5 h (3 h of the flowing condition and then 2 h of stagnation), it seems that, once the system is under stagnation, even high initial input of DIC as carbonate was used up for buffering the system and lost the buffering capacity in a short time (i.e., 2 h), resulting in the similar 2D pH map with low DIC. It can be concluded that, under stagnation, the diffusion of ions to and from the surface is limited and the local pH changes driven by galvanic interactions likely overtake buffering at both selected pH environments.

At pH 9, although the heterogeneity of the free chlorine reaction was decreased, more free chlorine reaction was found at the surface compared to pH 7 (Figures 1b and 3b). The 2D maps of free chlorine concentrations reflected that over 90% of the free chlorine concentration was consumed in some locations over the galvanic coupon surface. Under the flow condition, the free chlorine concentration at the surface was 0.1–1.65 mg of Cl₂ L⁻¹ for DIC of 10 mg of C L⁻¹ and 0.2–1.1 mg of Cl₂ L⁻¹ for DIC of 50 mg of C L⁻¹, respectively. Under stagnation, the free chlorine concentration at the surface was 0–1.2 mg of Cl₂ L⁻¹ for both DIC conditions (Figure 3b). Generally, free chlorine depletion was greater under stagnation for both pH 7 and 9. It seems that the consumption of free chlorine was relatively more prominent in the cathodic region (high pH area) in the presence of the anodic region (low pH) and/or in a lower buffered system (e.g., DIC of 10 mg of C L⁻¹ and stagnation). Therefore, under different DIC conditions in pH 9, all cathodic areas (brass surface) showed higher free chlorine consumptions under both flow and stagnation, reflecting free chlorine high oxidation ability with reactive material, and its reaction rate can be slowed with increased alkalinity.

Cross-Sectional Image of the pH Distribution of Galvanic Corrosion.

Figure 4 shows a cross-sectional 2D map of pH microprofiles under stagnation (Ex 4S, pH 9, 2 mg of Cl₂ L⁻¹, and DIC of 50 mg C L⁻¹). For the cross-sectional map, a total of 11 1D pH microprofiles were measured at 2000 μm intervals across the galvanic joint. The DBL of pH defined from the microprofiles was approximately 600 μm. The cross-sectional pH map clearly distinguishes between the anodic leaded sold (low pH) and cathodic brass (high pH). The map also illustrates the zone of pH influence above the metal surface under stagnation compared to the bulk water pH. Lastly, the profiles and other aspects of this work clearly illustrate the valuable insights that only in situ microelectrode measurements can provide. In addition to rapid high-resolution pH measurements made in microenvironments, there is no disruption of DBL, especially under stagnation, which is essential for such microscopic studies. A few studies made similar microelectrode pH measurements for understanding corrosion^4,15,16 and demonstrated the same conclusion of pH change near the galvanic surface and the concept of electrochemical reversal, in which the metal that is normally the anode becomes the cathode under certain flow and chemistry conditions. However, while their microelectrodes provide a certain level of pH changes in a bulk solution, it was not capable of providing such detailed information on local corrosion and chemistry changes near the metal surface as a result of their relatively large electrode tip size (e.g., 0.8 mm, which is 80 times larger than the microelectrode in this study) (Figure 4), which are not able to discriminate DBL and create the high-resolution 2D maps shown in this study. This study presented a novel approach for 2D in situ mapping of the microscale water chemistry near a galvanic joint to provide a more detailed understanding of the extent of galvanic corrosion and related water quality changes in the microenvironment.

Figure 4.

Cross-sectional 2D map of pH microprofiles (Ex 4S, pH 9, 2 mg of Cl₂ L⁻¹, and DIC of 50 mg of C L⁻¹), with a total of 480 data points with high resolution of 20 μm (twice the size of a tip diameter). During the preceding flow condition (Ex 4F), lead may transfer as ions and precipitate on the downstream brass surface, resulting in low pH. The diameters of different pH probes were compared for reference purposes. Dashed lines indicate leaded-solder area boundaries.

Surface Investigation of the pH Distribution and Deposition of Corrosion.

Deposition of corrosion byproducts has been well -investigated in previous studies, showing a direct relationship between elevated lead levels in water and flow patterns.^16,32–34 However, the flow effect on localization of the pH distribution and the associated corrosion byproduct deposit is not completely understood. To correlate the surface having low pH with the deposit of corrosion byproducts during the initiation of galvanic corrosion, a separate test with a freshly prepared brass-lead joint coupon was conducted by immersing the coupon in the same solution of Ex 1F (pH 7, DIC of 10 mg of C L⁻¹, and 2 mg of Cl₂ L⁻¹) at a flow rate of 2 mL/min for 2 h. Similar to the previous Ex 1F, partial expansion of the white particles from the surface of the lead joint to the outflow end of brass after 2 h of flow was found by visual observation (Figure 5 and Figure S8 of the Supporting Information). The area where the white particles deposit was similar to the pH drift found in the previous Ex 1F (Figure 1a). SEM images indicated that deposits were formed on the outflow end of brass in multiple locations (locations 1 and 2). BSE images with higher contrast showed elements with different atomic numbers found on the brass surface. It was found that the white particles represent a high concentration of lead-contained particles, along with copper, zinc, and oxygen located in the selected area (Figure 5e). This finding implies that high levels of Pb and Pb precipitating near the galvanic joints are likely caused by the carryover by the flow and deposited on the outflow end of the brass surface, causing the significant pH change at the downstream brass surface with a microgalvanic corrosion reaction and further oxidation of Pb²⁺ oxides.^1,35 The repetition of the flow and stagnation cycle would promote the lead precipitate migration further downstream and affect local pH at the surface of the initial metal, where the lead precipitates have been sitting. The deposition of corrosion byproducts observed in this study was also found in long-term pilot-scale experiment operation, resulting in an elevated Pb concentration, even after partial replacement of the Pb pipe.⁶

Figure 5.

Surface characterization of a galvanized coupon in location 1 (after 2 h of flow under pH 7, free chlorine of 2 mg of Cl₂ L⁻¹, DIC of 10 mg of C L⁻¹, 100 mg of Cl⁻ L⁻¹, and 100 mg of SO₄²⁻ L⁻¹): (a) observation of the corroded coupon and location for surface characterization, (b) SEM image of the selected location with magnification (22×), (c) BSE image of the location with magnification (22×), (d) magnification (100×) of the BSE image of the selected location in Figure S8c of the Supporting Information, and (e) element distribution using EDS.

SUMMARY AND CONCLUSION

In this study, 2D maps of pH and free chlorine on the brass-lead galvanic joint surface demonstrated direct evidence of surface pH variation and disinfectant consumption during the early stage of galvanic corrosion. The 2D surface maps along with 1D concentration microprofiles also showed that galvanic corrosion could be greatly influenced by the anodic and cathodic surface area, water flow, water quality, and metals. This information can greatly improve our understanding by connecting with other existing findings in galvanic corrosion. For example, recently, DeSantis et al. identified antlerite, a copper sulfate mineral, from visual and mineralogical analysis of lead pipe joints.³⁶ Given that the bulk water pH is 7.7 and antlerite is only formed at low pH (pH < 3.0–4.0),³⁶ the pH maps constructed in this study can provide direct evidence of the reflected aggressive water at the metal surface. In combination with mineralogical investigation and theoretical solubility modeling, microscopic investigation using microelectrodes can advance our current knowledge regarding lead release and mitigation in a DWDS by providing direct measurements of pH, free chlorine, DO, and specific ions of interest (e.g., Pb²⁺).

Results from these experiments may have the following important implications for a better understanding of lead leaching in chlorinated drinking water systems during early stages: (1) Galvanic reactions in chlorinated low DIC (10 mg of C L⁻¹) water resulted in dramatic local pH changes and heterogeneous pH distribution on the brass–lead metal surface under both flow and stagnant conditions, although greater under stagnation. (2) Galvanic reactions in chlorinated high DIC (50 mg of C L⁻¹) water resulted in dramatic local pH changes and heterogeneous pH distribution on the brass-leaded-solder metal surfaces under only the stagnant condition. The increased buffer capacity resulting from the higher DIC was likely responsible for the absence of pH change under flowing conditions. (3) The measured pH and free chlorine concentration microprofiles and 2D maps provided direct evidence of the water flow and DIC effect on galvanic corrosion, confirming what others have previously speculated or indicated by alternate “macro” methods. (4) Flow on the galvanic coupon can result in the lead particle deposit on the outflow end of brass, creating microgalvanic cells and forming the localized electrochemical reversal on the brass surface.