A literature review of bench top and pilot lead corrosion assessment studies

Abstract

Bench top and pilot lead corrosion studies are gaining more interest, considering revisions and upcoming improvements to the Lead and Copper Rule. This literature review identified studies ranging from simpler month(s)-long bench top dump-and-fill stagnant water tests (coupon tests/standing pipe tests) to more complicated year(s)-long intermittent flow pilot studies (recirculating pipe loops/once through pipe rigs). With increasing complexity in design and operation, studies more closely approximated real plumbing conditions (e.g., by incorporating harvested lead pipes and intermittent flow regimes) at increased cost, footprint, and duration. Comparison of bench top and pilot designs (in terms of lead test piece age/dimensions/configuration/replicates, study duration, sample collection, and other factors) can assist drinking water utilities, consultants, academics, and others to select a design that matches their needs and constraints. No matter the choice, surrogate systems cannot replace actual system water testing and are best complemented by other corrosion assessment tools.

1 |. INTRODUCTION

The drinking water industry relies on lead corrosion assessments to meet various objectives. Such assessments can be performed for fundamental or applied laboratory research by academic groups or for industry certification of plumbing products on behalf of plumbing manufacturers. They can also be used for voluntary operational improvements or for mandatory compliance with policy requirements by drinking water utilities and their consultants. The Lead and Copper Rule Revisions and the upcoming Lead and Copper Rule Improvements (eCFR, 2021) are expected to attract more interest in lead corrosion assessment tools in the United States.

The intent of the lead corrosion assessment affects the choice and utilization of specific tool(s), as demonstrated by the wide range of approaches that will be covered in subsequent sections. It may also affect the terminology used for these tools, since academic research might use looser terms than those defined in policy documents. Recognizing that other classifications and terminologies may exist, corrosion assessment tools can be classified into two general categories from an applied research angle herein: simulated studies and real-world studies (Figure 1, Table 1).

FIGURE 1.

Corrosion assessment tools as categorized in this work from an applied research angle. Other classifications and terminologies exist. This literature review focuses on bench top and pilot studies (highlighted in green). Monitoring of background water quality parameters is also important for any corrosion assessment, although it is not explicitly listed as a tool.

TABLE 1.

Qualitative relative comparison of corrosion assessment tools (low, medium, high).

Corrosion assessment tool		Cost	Time/effort			Overall time/effort	Results
			Set up	Implementation	Result analysis
SIMULATED STUDIES
Bench top study	Finished drinking water Synthetic drinking water	Low	Medium	Low Medium	Low	Low Medium	Relative comparisons of lead leaching among test conditions
Pilot study	Typically finished drinking water	High	High	Medium	Low-high	High	Relative comparisons of lead leaching among test conditions
Modeling	Drinking water parameters as inputs	Low	Medium-high	Low	Low	Medium	Theoretical solubility (i.e., predicted lead release) based on input and modeling assumptions
REAL WORLD STUDIES
Diagnostic drinking water sampling for metals	Tap water	Medium	Medium high	Low-high	Low-high	High	Lead release at home or other building taps
Pipe scale analysis	Excavated pipes from water system	Medium-high	High	Medium	Medium	High	Characterization of scale from excavated pipes

Note: This literature review focuses on bench top and pilot studies (emphasized with underline). Preliminary evaluation with feedback from a large US drinking water utility. Qualitative descriptors are rough and assume in-house capabilities. If consulting/contractor/external lab analysis is needed, then the cost, time, and other parameters may greatly vary; Qualitative ranges depend on number of samples, number of analyses, etc.; Overall relative time/effort roughly sums up the individual time/effort components.

Simulated studies may be entirely theoretical and/or might involve bench or pilot work. For example, computer modeling might be used as part of a theoretical desktop study along with the theoretical comparison of a water system to similar water systems with available corrosion control performance information. Bench top studies typically use new metal pieces or coupons submerged in stagnant water within glass jars (coupon tests) or short pipe segments filled with stagnant water (standing pipe tests) to determine how specific water treatments affect lead leaching. Pilot studies also test specific water treatments but consist of more complicated pipe loops or pipe rigs made of various plumbing materials with water intermittently pumped through them to simulate real-world flow conditions more closely.

Real-world studies may include diagnostic drinking water sampling and pipe scale analysis. Diagnostic drinking water sampling can rely on various sampling approaches to assess water lead levels at actual consumer taps (e.g., residences, schools, daycares, or other buildings), including profile sampling. Pipe scale analysis may rely on various analytical techniques (e.g., optical microscopy, x-ray diffraction, and scanning electron microscopy) to assess the corrosion of excavated lead pipes or other plumbing materials, based on the actual accumulation of corrosion deposits on their interior surface while they were in service. Pipe scale analysis has also been used in pilot and bench top studies to investigate pre-existing scale from harvested pipes and scale formed during the experiments (Aghasadeghi et al., 2021; Bae et al., 2019; Masters et al., 2022), but those specific pipe scale analyses reflect simulated experiments with excavated materials (and not real-world conditions).

Monitoring of background water quality parameters (e.g., pH, disinfectant residuals, alkalinity, corrosion inhibitor residuals, chloride, sulfate, iron, manganese, calcium, magnesium, and aluminum) is also important in planning any corrosion assessment study, although it is not explicitly listed as a tool herein. The information obtained from background water quality monitoring can inform desktop studies, serve as a starting point for water chemistry in bench top and pilot studies, and assist in the interpretation of diagnostic drinking water sampling and pipe scale analysis data.

Every corrosion assessment tool has pros and cons in terms of relative cost, time/effort, and results (Table 1). An evaluation before beginning a corrosion assessment study can determine what tool(s) will work best for the specific situation. If resources allow, a combination of tools can be used to gain a comprehensive understanding of metal release within a system (Harmon et al., 2022; Roth et al., 2021). While all tools are important and can each provide complementary information, this review focuses on bench top and pilot studies. Information on bench top studies has not yet been synthesized in the peer-reviewed literature, whereas information on pilot rigs has not been updated since 2003 (Eisnor & Gagnon, 2003). Other recent work explained corrosion assessment tools that will not be detailed here, including pipe scale analysis (Harmon et al., 2022) and diagnostic drinking water sampling for lead (Lytle et al., 2021; Triantafyllidou et al., 2021).

For this literature review, emphasis was placed on the evolution of bench top and pilot studies that were available in journal articles, conference proceedings, and reports. Historical studies (pre-1991) were briefly reviewed prior to summarizing studies conducted after the 1991 Lead and Copper Rule (LCR) promulgation and until 2022, with emphasis on their key design considerations. Industry experiences have not always been published in peer-reviewed journals, conference proceedings, or reports. Therefore, it was not possible to include every study performed on lead corrosion.

2 |. GENERAL DESIGN CONSIDERATIONS FOR BENCH TOP AND PILOT STUDIES

When designing a bench top or pilot study, there are several parameters to be considered, including plumbing type/configuration, water chemistry, flow/stagnation regimes, sampling protocol, lead quantification approach, and study duration (Figure 2). These parameters are introduced here and will be expanded under specific study examples in later sections. As with all studies, compromises must be made to balance the availability of resources (materials, personnel, time, and cost). Before designing a study, it is important to understand the conditions within the water system(s) represented in the study and to determine the overall intent of the study (i.e., research, operational improvements, regulatory compliance).

FIGURE 2.

General design considerations for lead corrosion bench top and pilot studies, which are the focus of this literature review.

2.1 |. Test piece materials

There are numerous types of plumbing materials that can be used for both bench top and pilot studies to simulate lead sources of local concern. Options for test pieces include lead pipes, copper pipes with lead solder, leaded brass, and galvanized iron pipes. Depending on the type of study (bench top vs. pilot) sheets of metal, segments of actual pipe, solder, and/or actual plumbing fixtures (faucets, valves, joints, etc.) have been used. Another important factor to consider is the age of the plumbing material. Aged pipes and fixtures harvested from an actual drinking water system are more representative. However, they can also cause more variability in the study results, due to the complex nature of pre-existing scale and/or due to scale destabilization from the harvesting process (Eisnor & Gagnon, 2003; Wysock et al., 1995). EPA research experience suggested this to be especially true for excavated lead pipes dominated by Pb(IV) minerals on the pipe scale (CDM Smith, 2018). While new plumbing materials may reduce variability in results, they are not representative of an aged plumbing system and may not produce realistic results (Eisnor & Gagnon, 2003; Roth et al., 2021). Numerous studies have highlighted the importance of having at least triplicate tests for each water condition to give a representative picture of the inherent variability of lead release from plumbing materials (Nguyen et al., 2010; Parks et al., 2014; Masters et al., 2016, 2022).

2.2 |. Duration

In both cases (new or aged materials), a pre-testing phase, a conditioning phase, and/or a stabilization phase may be performed prior to the actual testing phase. This is typically done to attempt to reduce variability between and within replicate test pieces over time. While a conditioning and/or stabilization phase was more typically encountered, a pre-testing phase was observed in only three studies.

The pre-testing phase allows for the collection of preliminary lead leaching data from materials to determine which replicate test pieces to ultimately select for the study (Bradley, 2018; Parks et al., 2014; Tang et al., 2018), with more information on specific studies detailed later. Pre-testing evaluates more test pieces than would actually be used during the study. That way, pieces with extremely high or extremely low levels of lead leaching could be excluded from testing, in an effort to reduce variability (Bradley, 2018; Parks et al., 2014). In one research study, however, high- and low-leaching pieces were instead deliberately selected, to test extreme lead leaching conditions (Tang et al., 2018). The pre-testing phase used the same water as the conditioning phase in one study (Tang et al., 2018) and different water chemistry that was considered to be corrosive (i.e., a synthetic water recipe based on National Sanitation Foundation [NSF] Standard 61 Section 9 protocol) (NSF, 2007) in two other studies (Bradley, 2018; Parks et al., 2014).

The conditioning phase exposes all test pieces that will be used for the study to the same control water condition. This typically reflects the current water chemistry used by a water system under the same protocols that will be used during the actual study (typically dump-and-fill for bench top studies and intermittent flow for pilot studies). The stabilization phase may follow the conditioning phase to expose plumbing materials to the same water conditions that will be used during the testing phase of the study. It accounts for a transitional period of time when there might be a potential increase and variability in lead leaching due to metal exposure in a new water chemistry (CDPHE, 2019).

The testing phase uses the water conditions of interest and is where the actual comparative analysis of data will occur. Data collected during pre-testing, conditioning, and stabilization is typically not used for the comparative analysis that follows the actual testing phase, as will be detailed in subsequent sections. While some studies did not specifically mention a stabilization phase, they either did not include the first few data points of the testing phase in their analysis (CDPHE, 2019) or noted the drop in lead leaching after an initial period of time that indicated the levels stabilized (Bae et al., 2019). It is important to note that while the data from the stabilization phase might be too variable to draw conclusions on corrosion control treatment effects, they can still be beneficial to understand how the system may initially react to the water chemistry change (e.g., if pipes scales will break off and how long it will take to establish stable conditions).

2.3 |. Flow/stagnation regime

Pilot studies can be designed to either have water flow through the system once then go to waste (pipe rigs) or have water flow in a recirculating loop that keeps supplying the pilot (pipe loops). While the once-through pipe rigs are a more realistic representation of actual water systems, they require significantly more water, which may not be feasible depending on the number of water conditions being tested (Eisnor & Gagnon, 2003). Other important factors to consider in pilot studies are the intermittent flow and stagnation periods. Ideally, the flow regimes should be as realistic as possible, mimicking real world situations. Just as with flow rates, stagnation times should also represent real-world situations. Bench top studies typically do not have any water flow for prolonged stagnation periods between each dump-and-fill.

2.4 |. Water chemistry conditions

For once-through pipe rigs that use a significant amount of water, it is generally more common to set up the pipe rigs at a water treatment facility where there is an unlimited water supply and staff who can readily oversee the test system operation. Bench top studies and recirculating pipe loop studies do not use as much water (Bae et al., 2019; Nguyen et al., 2010). They can therefore either use finished drinking water shipped from a treatment plant (Nguyen et al., 2010; Tang et al., 2006, 2018) or synthetic water made in the lab by adding chemicals to deionized water to simulate finished drinking water from a treatment plant (Bae et al., 2019, 2020; Edwards & Dudi, 2004; Masters & Edwards, 2015; Roth et al., 2018; Triantafyllidou & Edwards, 2011). Water from a treatment plant will be more representative of the real world. However, it may also reflect larger variability and seasonal changes, which can affect the reproducibility of water chemistry within the testing period. Synthetic water is more easily reproducible but is not completely representative of real-world variations.

2.5 |. Sampling protocol and lead quantification approach

Older studies assessed lead corrosion by using electrochemical techniques (e.g., current density and resistance change), measuring oxygen depletion, and/or calculating the weight loss of the plumbing material after study completion (ASTM, 2004; Byars & Gallop, 1975; Reiber, 1995; Reiber et al., 1988, 1996; Reiber & Benjamin, 1990; Singley & Lee, 1984). Such setups were sophisticated, including coupon sleeves (Reiber et al., 1988; Ryder, 1978), corrosion cells (Reiber, 1995), and closed loop recirculation systems for polarization measurements (Reiber & Benjamin, 1990; Ryder, 1978; Silverman, 1995), and were able to provide some information about lead corrosion. But they fell out of favor because electrochemical measurements of corrosion rate do not directly translate to actual metal release (Eisnor & Gagnon, 2003).

In most modern studies, water samples are collected periodically throughout the study for analysis (i.e., inductively coupled plasma mass spectrometry or inductively coupled plasma atomic emission spectroscopy) to directly determine the lead concentration in addition to other water quality parameters. The concentration of lead is generally plotted over time to show trends of lead leaching throughout the study duration. The frequency of sample collection, stagnation time prior to sampling, and duration of the study are additional parameters that can affect the results.

Analyzing the results is the last step to decision-making, with Wysock et al. (1995) discussing statistical tests that can be used to analyze data from corrosion studies. The reader is directed to Wysock et al. (1995) and Kimeyer et al. (1996) for such statistical considerations, including the determination of an adequate sample size and frequency for proper statistical analysis, the importance of the stabilization phase for lead data stabilization, evaluating data outliers, and parametric (e.g., Pearson correlation coefficient, t-test) versus non-parametric (e.g., Spearman rank correlation coefficient, Wilcox test) statistical tests. Modern studies, such as Masters et al. (2022) have adapted the statistical methods described in Wysock et al. (1995) for their data analysis. Masters et al. (2022) found that the lead data tended to follow a log-normal distribution. Therefore, they conducted a t-test and an analysis of variance after log-transformation of their data to compare the means of two groups and multiple comparisons of means, respectively. In addition, they used a stepwise regression method for evaluating water quality factors that influenced lead release, Spearman’s rank correlation coefficient for assessing the relationship between lead concentration and orthophosphate dose, and Levene’s test for homogeneity of variance for comparing variability in lead release between studies (i.e., coupon study, pipe rack study, and water distribution sampling) (Masters et al., 2022).

The following sections review bench top studies followed by pilot studies (i.e., in order of increasing study complexity), which are the focus of this review. For both categories, studies conducted pre-1991 are termed historical studies herein and are summarily presented, whereas post-1991 studies generally contain more information. Bench top studies are classified into two general sub-categories herein, namely coupon tests (Table 2) and standing pipe tests (Table 3). Pilot studies are also classified into two general sub-categories, namely recirculating pipe loops (Table 4) and once through pipe rigs (Table 5). Tables 2–5 are not meant to be exhaustive. For example, they exclude follow-up studies using adapted or similar setups (often by the same research group), for brevity purposes. Design considerations will be discussed in more detail for all these categories, using specific study examples. While this review focused on lead corrosion assessment, some aspects may apply to studies investigating the corrosion of other metals within a system (e.g., copper, iron).

TABLE 2.

Overview of coupon bench top studies reviewed.

Studya	Intent	Water chemistry	Plumbing material	Stagnation regime	Sampling	Durationb	Replicates
Lytle et al. (1992)	Compare the % of lead (and other metals) in brass alloys to metal leaching	Water from tap w/pH adjustments (30 mL per Teflon sample cell)	New metal sheet coupons of brass, lead, copper, zinc, and 60:40 Pb/Sn solder (1″ x 2″ × 1/8″)	1 day	1× per day for several weeks then 2× per week	150 days (21.4 weeks)	Not specified
Lin et al. (1997)	Compare effects of free and combined chlorine on lead leaching	Finished water from treatment plant w/adjustments (disinfectant, pH, inhibitor) (1 L per plastic beaker)	New metal sheet coupons of lead, copper (coated w/lead solder), and brass (0.5″ × 4″ × 1/32″)	Mixed once per day for 30 min. No mention of water change except for inhibitor experiments	1× per week	5 weeks	Not specified except for brass coupons (duplicates)
Nguyen et al. (2010)	Examine the impacts of chloride-to-sulfate mass ratio (CSMR) on lead leaching	Finished water from a water utility (40 mL per jar)	New brass rod coupons (0.25″ dia, 0.38″ h)	2–3 days	3× per week (weekly composites of all 3 water changes)	11 weeks	Triplicates
	Examine the impact of treatment changes on lead leaching	Finished water from 10 utilities w/adjustments (inhibitor dosing, pH, treatment processes that affect CSMR) (25–100 mL per jar)	New brass (1/4″ dia), and copper (1/2″ dia) pipe coupons (1″); 50:50 Pb/Sn solder (0.125″ dia, 0.854″ h)	3–4 days	2× per week (weekly composites of both water changes)	18–35 weeks	Duplicates/ triplicates
Bradley (2018)	Examine the effects of inhibitors on high CSMR water from two utilities	Raw water from two utilities treated in lab to simulate plant treatment w/adjustments (inhibitor, chloride) (50–100 mL)	Copper pipe coupons (3/8″ dia) with 50:50 lead: tin solder (1/2″ h)	Jars mixed at 25 rpm; water change every 2– 3 days	3× per week (weekly composites of all 3 water changes	14–16 weeks	5 replicates
Roth et al. (2018)	Examine the impact of blended surface and ground water on lead leaching	Synthetic water representative of 2 water utilities (500 mL per jar)	New lead coupons (3″ × 1/2″ × 1/16″)	3–4 days	2× per week	Phase 1: 7 weeks Phase 2: 5 weeks	Phase 1: duplicates Phase 2: duplicates split into two conditions (one replicate per condition)
Tang et al. (2018)	Examine the effect of water conditions on metal leaching	Finished water (surface and ground water sources) (125 mL per jar)	Harvested galvanized iron pipe coupons (0.5–1″ dia; 0.79″ h)	Weeks 0–5.5: 2–3 days Weeks 5.5–9: 7 days Weeks 9–13.5: 2–3 days	Weeks 0–5.5: 3× per week Weeks 5.5–9: 1× per week Weeks 9–13.5: 3× per week (weekly composites water changes)	Pre-testing: 4 weeks Study: 13.5 weeks	Pre-testing period used to determine “good” and “bad” coupons. Triplicate “good” and “bad” coupons for each water condition
Cornwell and Wagner (2019)	Determine effective corrosion control treatment (CCT) on lead leaching	Finished water from utilities w/adjustments (pH, inhibitor) (250–500 mL per jar)	New metal sheet coupons of lead, copper, mild steel, and brass(1/2″ × 3″)	3–4 days	1–2 × per week	~ 1 year	Duplicates
CDPHE (2019), Arnold et al. (2021)	Compare effects of CCT on lead leaching Examine the effect of changing coagulants at a surface water treatment plant (WTP)	Finished water from 2 utilities w/adjustments (pH, inhibitor) Finished water and raw water (with simulated treatment in lab) provided by WTP (125 mL per jar)	New leaded brass and copper pipe coupons (1/2″ dia, 1″ h), copper pipe (1/2″ dia, 1″ h), with 50:50 lead/tin solder (0.125″ dia, 1″ h)	2–3 days	3× per week (weekly composites of all 3 water changes)	6 weeks	Triplicates for leaded brass and copper pipe coupon tests 5 tests for copper pipes with 50:50 lead/tin solder
Masters et al. (2022)	Examine the effect of pH and orthophosphate on lead leaching	Water from treatment plant, after filtration but before secondary disinfection, with adjustments (pH, inhibitor) (240 mL per jar)	New lead coupons (2″ × 2″ × 0.063″) w/ 0.188″ dia hole.	Jars mixed at 50 rpm; water change 2–3 days	3× per week (weekly composites of all 3 water changes)	Conditioning: 6 weeks Study: 7 weeks	Triplicates

Note: This list is not exhaustive and follow-up studies using similar set-ups were excluded for brevity.

Clarifications: Finished drinking water: water directly from the drinking water treatment plant that has not come in contact with any plumbing within the distribution system; Synthetic water: water made in the lab by adding chemicals to deionized water to simulate finished drinking water; Tap water: water from the tap in the lab (this differs from finished water which comes directly from the treatment plant and does not come in contact with any plumbing material within the distribution system).

Conversions: 1″ = 25.4 mm

Abbreviations: Dia, pipe diameter (inner/outer diameter not specified in paper); h, height of pipe coupon; w/, with.

^aThis table includes studies (historical to 2022) with enough level of detail to fill each column.

^bDuration accounts for pre-testing and conditioning (if mentioned) as well as actual testing phases.

TABLE 3.

Overview of pipe bench top studies reviewed.

Studya	Intent	Water chemistry	Plumbing material	Stagnation regime	Sampling	Durationb	Replicates
Edwards and Dudi (2004)	Investigate the effects of oxidants on lead leaching	Synthetic water representative of a utility (13 mL per pipe)	New lead pipes (0.6″ ID, 2.5″)	Before sampling: 16 h Weekdays: 24 h Weekends: 60 h	2–4 × per month	4 weeks	Duplicate/triplicate
		Synthetic water representative of a utility	8 types of new brass hose bibs	Before sampling: 16 h Weekdays: 24 h Weekends: 60 h	1–2 × per month	8.5 weeks	One of each hose bib type per water condition
Triantafyllidou et al. (2012)	Examine the effects of new brass plumbing devices on lead leaching	Synthetic water representative of a utility	New brass ball valves w/ and w/out soldering fluxes	2–3 days	3× per week (weekly composites of all 3 water changes)	w/out flux: 19 weeks w/ flux: 15 weeks	w/ out flux: triplicate w/ flux: single sample
Nguyen et al. (2010)	Examine the effects of chloride-to- sulfate mass ratio (CSMR) on lead leaching	Synthetic water representative of a high CSMR water	Simulated pipe joint corrosion macro-cell consisting of new: copper pipe (0.5″ ID, 12.2″ long); clear Tygon tubing; 50:50 lead: tin, pure tin, or pure lead solder inserted through center of a copper pipe (0.5″ ID, 2.5″ long)	2–3 days	3× per week (weekly composites of all 3 water changes)	15 weeks	Triplicate
Triantafyllidou and Edwards (2011)	Examine the effects of partial lead service line replacements (PLSLRs) on lead leaching	Synthetic water representative of a utility (Low and high CSMR)	New copper pipes (0.75″ ID) electrically connected to new lead pipes (0.75″ ID)	2–3 days	3× per week (weekly composites of all 3 water changes)	Conditioning: 0–1 year Study: 1 year	Not specified
Masters and Edwards (2015)	Investigate the effect of corrosion control strategies for low-alkalinity, high-pH waters on lead leaching	Synthetic water representative of a utility w/ pH adjustments	New lead pipes (6″)	2–3 days	3× per week (weekly composites of all 3 water changes)	Conditioning: 4 weeks Study: 24 weeks	Triplicate
Clark et al. (2015)	Examine the effects of PLSLRs on lead leaching	Tap water w/adjustments (copper added)	New lead pipes (0.75″ ID, 4″ long)	2–3 days	3× per week (weekly composites of all 3 water changes)	24 weeks	Triplicate
		Synthetic water representative of a high CSMR water (w/ high copper)	New lead pipes (0.75″ ID, 1 OD, 12″ long)	2–3 days	3× per week (weekly composites of all 3 water changes)	24 weeks	Triplicate

Note: This list is not exhaustive and follow-up studies using similar set-ups were excluded for brevity.

Clarifications: Synthetic water: water made in the lab by adding chemicals to deionized water to simulate finished drinking water; Tap water: water from the tap in the lab (this differs from finished water which comes directly from the treatment plant and does not come in contact with any plumbing material within the distribution system).

Conversions: 1″ =25.4 mm.

Abbreviations: ID, inner diameter of pipe; OD, outer diameter of pipe; w/, with; w/out, without.

^aThis table includes studies (historical to 2022) with enough level of detail to fill each column.

^bDuration accounts for pre-testing and conditioning (if mentioned) as well as actual testing phases.

TABLE 4.

Overview of recirculating pipe loops reviewed.

Studya	Intent	Water chemistry	Plumbing material	Flow regime	Samplingb	Durationc	Replicates
Schock and Gardels (1983)	Study effect of pH and low total inorganic carbon (TIC) on lead leaching	Finished water with adjustments (pH, alkalinity, & free chlorine)	New lead pipes (0.75″ ID)	Stagnation: >7 h Flow: 0.9–1.5 gpm (5 pumping cycles per day—15–30 min long)	Not specified	Not specified	Not specified
Kirmeyer et al. (1994)	Evaluate effectiveness of various treatment options in controlling lead and copper leaching	Finished water from 6 utilities w/ adjustments (pH, inhibitor)	New lead pipes and copper pipes (0.5″ ID) with lead solder	Stagnation: 8–16 h Flow: 1–3 gpm	1–5 × per week • Flowing sample (after 1 h of flow) • Standing sample (after 8 h of stagnation)	> 1 year	Not specified
Tang et al. (2006)	Investigate the effects of blending water sources on lead leaching	Source water from utility treated in lab to represent current and potential treatments	Copper tubing and pure lead coupon (3″ × 0.5″)in a corrosion loop fed by simulated pipe distribution system (PDS) which did not include lead	Corrosion loop: flushed with 7.6 L of water (from PDS) every morning	1 × per day from corrosion loop after 6 h stagnation	Conditioning: 4 months Study: 1 year	Not specified
Nguyen et al. (2010)	Evaluate effects of chloride-to-sulfate mass ratio (CSMR) on lead leaching	Finished water from a utility	New lead, bronze, and copper pipes (0.75″ ID); 50:50 Pb/Sn solder	Transitioned between • Continuous flow • Long stagnation (5 min flow every 8 h) • Short stagnation (5 min stagnation every 8 h) • Long stagnation	1× per week • Standing sample from metal pipe loop sections (after 8 h stagnation) • Flowing and standing sample from recirculating reservoirs	4.5 months	Not specified
Bae et al. (2019)	Examine effects of adding phosphate to water before switching from free chlorine to monochloramine on lead leaching	Synthetic water representative of historical, current, and potential water chemistries at a utility	New lead pipes (0.75″ ID)	Stagnation: 8 h Flow: 16 h  recirculation period: 1.10–1.45 gpm	• 1 × per week after first 8 h stagnation period • 1 × per week (end of week) from recirculated tank (i.e. weekly composite sample)	Initial conditioning: 66 weeks Conditioning w/ P04 (3 pipe loops w/ P04; 3 w/out):14 weeks Study: >30 weeks	Triplicates
Bae et al. (2020)	Examine impact of PO4 on lead	Synthetic water representative of a water	Harvested lead pipes	Stagnation: 8 h Flow: 16 h			Triplicates
	leaching from lead service line (LSL) in lowalkalinity high-pH water	utility with and without corrosion control			• 1 × per week after first 8 h stagnation period • 1 × per week (end of week) from recirculated tank (i.e. weekly composite sample	Dump-and-fill conditioning: 5 months Pipe loop conditioning: 45 weeks Study: 30 weeks

Note: This list is not exhaustive and follow-up studies using similar set-ups were excluded for brevity.

Clarifications: Finished drinking water: water directly from the drinking water treatment plant that has not come in contact with any plumbing within the distribution system; Synthetic water: water made in the lab by adding chemicals to deionized water to simulate finished drinking water.

Conversions: 1 gpm = 63 mL/s; 1 ft/s = 0.3 m/s.

Abbreviations: gpm, gallon per minute; ID, inner diameter of pipe; w/, with; w/out, without.

^aThis table includes studies (historical to 2022) with enough level of detail available from the study to fill each column.

^bStagnation time prior to sampling is listed for studies that provided this information.

^cDuration accounts for pre-testing and conditioning (if mentioned) as well as actual testing phases.

TABLE 5.

Overview of once-through pipe rigs reviewed.

Studya	Intent	Water chemistry	Plumbing material	Flow regime	Samplingb	Durationc	Replicates
Treweek et al. (1985)	Evaluate the effects of corrosion mitigation on lead leaching	Finished water from utility	Copper with lead/tin solder	Stagnation: 3–8 h Flow: 10 min six times per day (5 gpm)	1 × per month (8 h stagnation)	18 months	Not specified
MacQuarrie et al. (1997)	Examine effects of corrosion inhibitors on lead leaching	Finished water from a water utility w/ adjustments (pH, alkalinity, inhibitor)	Cast iron, lead, copper with lead/tin solder, and brass faucets	Stagnation: 3–8 h Flow: 2 h four times per day (1.6–4.0 gpd)	1 × per week: (electrical resistance measuring) Stagnant samples measured regularly	12 months	Not specified
Churchill et al. (2000)	Evaluate corrosion control treatments to reduce lead leaching	Finished water from a utility with adjustments (pH, alkalinity, inhibitor)	Cast iron, lead, copper with lead/tin solder, and brass faucets	Stagnation: 8 h (first 7 months) and 16 h (last 5 months) Flow: 6 h per day (Vel: 0.45 and 0.79 m/s)	Every 3 weeks (after 8 h stagnation for first 7 months; after 16 h stagnation for last 5 months)	12 months	Not specified
Cartier et al. (2012)	Examine effects of flow rate and service line configuration on lead leaching	Tap water (filtered through 5 μm and 1 μm pore size filters)	New copper and lead pipes (0.75″ ID)	5 days per week (M-F): Stagnation: 16 h and 5 h Flow: 100 min of twice per day (0.34gpm per rig)	3 × per week (after 16 h stagnation) at low flow (1.3 LPM), med flow (8 LPM), and high flow (32 LPM) At least 3 grab samples taken during each flow condition	7 months	Triplicates
Cartier et al. (2013)	Examine the impact of pH, phosphate, and CSMR on particulate and lead leaching from fully and partially replaced LSLs	Finished water from utility with adjustments (pH, PO4, and CSMR)	Aged lead pipes (0.75″ ID)	Weekdays: Stagnation: 16 h Flow: 8 h (1.3 gpm) Weekends: Stagnant	lx per week Under medium flow after 16 h stagnation	Conditioning: 8 months Phase 1 and 2: 11 weeks each Stabilization: 3 weeks Phase 3: 9 weeks	Triplicates
Parks et al. (2014)	Develop standardized set-up to assess corrosion control effectiveness on lead leaching	Finished water from five utilities	New Brass, copper with lead solder, and lead pipes (0.75″ ID)	Stagnation: 8 h Flow: 10 min (2.64 gpm per test piece) every 8 h	1 × per month (collected after at least 6 h stagnation)	Pre-testing (dump and fill w/ 12 h stagnation × 3): 36 h Conditioning: 1 month Study: 1 year	Triplicates
Cantor (2017)	Examine impact of phosphate on lead and copper leaching	Finished water from eight utilities	New metal plates of copper and lead (2″ × 2″ × 1/16″).	Stagnation: 23 h Flow: 1 h per day (0.13 gpm)	2 × per month (collected after at least 6 h stagnation)	1.5–3.3 years	None; 1 chamber for lead and 1 for copper(16 metal plates each)
Williams et al. (2018)	Examine effects of corrosion control treatment (CCT) on lead leaching	Finished water from a water utility	Aged lead service line extracted from distribution system (0.75″ ID)	Stagnation: 2.5–12 h (typically five 2.5 h and one 8 h stagnation period) Flow: 1 60 and 5 30 min periods per day (1 gpm)	1–2 × per month (collected after 6–12 h (typically 8 h) stagnation period)	15 months	Quadruplicates
Masters et al. (2022)	Examine the effect of pH and orthophosphate on lead leaching	Finished water from two treatment plants, with adjustments (pH, inhibitor)	Harvested lead service lines from distribution system. (0.75″ ID)	Stagnation: three 5 h stagnation periods per day Flow: 1 gpm	Weekly or bi-weekly (collected after stagnation period)	Conditioning: 4– 7 months Study: 1–2 years	Triplicates

Note: This list is not exhaustive and follow-up studies using similar set-ups were excluded for brevity.

Clarifications: Finished drinking water: water directly from the drinking water treatment plant that has not come in contact with any plumbing within the distribution system; Synthetic water: water made in the lab by adding chemicals to deionized water to simulate finished drinking water; Tap water: water from the tap in the lab (this differs from finished water which comes directly from the treatment plant and does not come in contact with any plumbing material within the distribution system).

Conversions: 1 m/s = 3.3 ft/s; 1 LPM = 15.9 gpm.

Abbreviations: gpm, gallon per minute; LPM, liter per minute; w/, with.

^aThis table includes studies (historical to 2022) with enough level of detail to fill each column.

^bStagnation time prior to sampling is listed for studies that provided this information.

^cDuration accounts for pre-testing and conditioning (if mentioned) as well as actual testing phases.

3 |. REVIEW OF LEAD CORROSION BENCH TOP STUDIES

Bench top studies have been used to evaluate lead corrosion and potential corrosion control treatments dating back to the 1940s (Hatch, 1941; Moore & Smith, 1942). These studies typically consist of either new metal coupons submerged in a glass jar full of water (termed coupon test herein) or pipes full of water with stoppers on each end (termed pipe test herein). Bench top studies typically follow a simple dump-and-fill protocol, where water in the jar/pipe is manually emptied periodically (typically every 2–3 days) and replaced with fresh drinking water (CDPHE, 2019; Nguyen et al., 2010). Lead release is evaluated by directly measuring the lead concentration in the water that has been emptied from the jar/pipe, either after single stagnation events or after each test week in composite samples. Composite samples are larger cumulative water samples that combine samples collected from several stagnation events throughout the week into a single sampling container for metal analysis. Different terminologies have been used for bench top studies through the years, including dump and fill testing, fill and draw testing, immersion testing, coupon testing, and static testing (CDPHE, 2019). The Colorado Department of Public Health and Environment (2019) summarized key design considerations for lead and copper corrosion bench top studies in their guidance manual.

3.1 |. Historical bench top (coupon and standing pipe) studies (pre-1991)

Moore and Smith (1942) conducted standing pipe tests using old lead and lead-lined pipes extracted from the Boston and Cambridge water systems to test the effects of various amounts of sodium hexametaphosphate (a polyphosphate corrosion inhibitor) on lead leaching. These tests were conducted using a simple dump-and-fill technique, where extracted pipes were filled with water and left to stand overnight. Variable test results were attributed to pipe exposure to the open air for long periods of time during transportation, causing them to dry out before the study began (Moore & Smith, 1942).

Patterson and O’Brien (1979) conducted studies using lead coupons and coupons of lead soldered to copper that were submerged in waters with varying levels of carbonate for 9 days and 30 days. A short-term lead pipe study was also conducted using 500 mm (20 in.) pipe segments cut from a 5 m (15 ft), 12.7 mm (0.24 in.) diameter pipe extracted from the Boston drinking water distribution system. Eight pipes were conditioned for about 3 weeks with water containing no corrosion control treatment, using the dump-and-fill method every 24 h until the variability of lead corrosion for each pipe was within about 20% of its mean lead concentration after the 24 h stagnation period. After the conditioning period, duplicate tests were run for four water conditions, where the dump-and-fill method continued every 24 h for 60 days (Patterson & O’Brien, 1979).

3.2 |. Coupon tests (1991–2022)

In the last three decades, coupon tests have been commonly used in the US to study the effects of potential changes in water treatment, water source, or addition of corrosion inhibitors on lead leaching into water (Table 2). These tests are the simplest to perform, presumably adding to their popularity. Although glass jars have been typically used, some studies used Teflon cells (Lytle et al., 1992), PVC pipe chambers (Cornwell Engineering Group, 2022), or other unspecified plastic containers (Lin et al., 1997). Most of the studies reviewed herein (Table 2) used finished drinking water from treatment plants with adjustments made in the laboratory (i.e., pH, disinfectant, and corrosion inhibitor dose) to represent potential changes in treatment (CDPHE, 2019; Cornwell & Wagner, 2019; Lytle et al., 1992; Nguyen et al., 2010; Tang et al., 2018, Masters et al., 2022). For these studies, water was either shipped from water utilities (Cornwell & Wagner, 2019; Lin et al., 1997; Nguyen et al., 2010; Tang et al., 2018) or taken from the taps in the lab (Lytle et al., 1992).

The metal sheets and pipe coupons in these studies consisted of either single leaded materials, such as lead, brass, and galvanized iron pipes (Masters et al., 2022; Arnold et al., 2021; CDPHE, 2019; Cornwell & Wagner, 2019; Lin et al., 1997; Lytle et al., 1992; Roth et al., 2018; Tang et al., 2018) or copper with lead/tin solder (Arnold et al., 2021; Bradley, 2018; CDPHE, 2019; Lin et al., 1997; Lytle et al., 1992; Nguyen et al., 2010). The coupons were either suspended in the water by plastic zip ties or other non-metallic material from the lid (Figure 3a), sitting at the bottom of the glass jar (Figure 3b), or epoxied to the bottom of the glass jar (Figure 3c). These tests typically used new leaded materials, but one study used galvanized iron pipes harvested from an actual distribution system. For these harvested pipe tests, epoxy was used to coat the outer surfaces of the pipe to prevent their exposure to the test water in the jar and only allow internal surface exposure (Tang et al., 2018). Coupon studies using new materials cannot simulate the pipe scale buildup that occurs within the water distribution system, even when a conditioning phase is included in the study, because of the amount of time it would take for the pipe scale to build up (Cornwell & Wagner, 2019). Among other constraints, the short study timeframes essentially prohibit the significant development of pipe corrosion solids as they would naturally occur in the field after decades or a century of pipe use.

FIGURE 3.

Examples of bench top study set ups with (a) new lead soldered copper pipe coupon suspended in a jar of water (Seidel, 2020), Reprinted with permission; (b) new lead soldered copper pipe coupon immersed in a jar of water (Roth et al., 2021), ©2021 AWWA; (c) harvested galvanized iron pipe epoxied to bottom of jar (Roth et al., 2021), ©2021 AWWA; (d) new lead pipe segments filled with test water (Nguyen et al., 2010), Reprinted with permission, © The Water Reserch Foundation; and (e) harvested lead pipes filled with test water (Sharp et al., 2009), Reprinted with permission.

Information on the conditioning of fresh material surfaces prior to testing was not available for most bench top studies identified. Studies that use harvested pipes still require a conditioning period to recover pipe scale after transport, and one such study (Tang et al., 2018) conditioned harvested pipes for 4 weeks before testing. During the conditioning phase, all coupons were exposed to water from the same well, following the standard dump-and-fill protocol every Monday, Wednesday, and Friday (Tang et al., 2018).

There appeared to have been no standard water volume or standard metal coupon dimensions used for the coupon studies. Some studies used a small volume of water ranging from 25 to 50 mL (Bradley, 2018; Lytle et al., 1992; Nguyen et al., 2010), whereas others used larger volumes of water ranging from 250 mL to 1 L (Cornwell & Wagner, 2019; Lin et al., 1997; Roth et al., 2018). Use of a smaller water volume allowed for observation of minor changes in lead leaching, but limited the number of analyses that could be performed as a larger volume of water was needed to run certain analyses (Lytle et al., 1992). Some studies took composite samples at the end of each week (Bradley, 2018; CDPHE, 2019; Masters et al., 2022; Nguyen et al., 2010; Tang et al., 2018), which provided a greater volume of water to be analyzed and presumably simplified the study.

Coupons of various shapes (e.g., rods, sheets, and small segments of excavated pipe) and dimensions have been used in coupon tests. Metal sheets ranged in size from 12.7 to 25.4 mm (0.5–1 in.) × 50.8–101.6 mm (2–4 in.) with a thickness of 0.79–3.2 mm (1/32–1/8 in.) (Cornwell & Wagner, 2019; Lin et al., 1997; Lytle et al., 1992; Roth et al., 2018). Pipe coupons were generally 6.4 mm (0.25 in.), 12.7 mm (0.5 in.), or 25.4 mm (1 in.) in diameter and ranged from 9.7 mm (0.38 in.) to 25.4 mm (1 in.) in length (CDPHE, 2019; Nguyen et al., 2010; Tang et al., 2018).

The volume of water held within the jar relative to the exposed surface area of the leaded coupon that it contacts should represent the volume to surface area of the plumbing component of interest found within a typical drinking water system. However, the multitude of plumbing components and the various models and designs for each make establishing a reference surface area to volume ratio for comparison exceptionally hard. This ratio was not reported in most studies, but Nguyen and colleagues found that the volume to surface area ratio was relatively small in their brass fixture experiment compared to the ratio typically found in home faucets (Nguyen et al., 2010). A lower volume to surface ratio for testing can result in higher concentrations of lead leaching than would be observed in an actual drinking water system.

The vast majority of bench top studies followed a simple dump-and-fill protocol, where water was manually poured out of the glass jars every 2–4 days and refilled with fresh water. In some cases, however, test coupons were suspended in plastic cells/sleeves with valves at the top and bottom. The valves allowed fresh water to be gravity-fed into the cell from the top, as water samples were instantaneously collected at the bottom (Cornwell Engineering Group, 2022; Lytle et al., 1992). This plastic cell gravity-fed system allowed for headspace-free conditions throughout the study, whereas the typical dump and fill jars had brief moments where coupons were exposed to the air during sampling events. For most coupon studies with glass jars, water remained stagnant between each dump and fill, but some studies placed jars on shaker tables to mix between 25 and 100 rounds per minute (Bradley, 2018; Masters et al., 2022; Tang et al., 2018). Composite samples were typically collected each week and analyzed for lead and other water quality parameters. The weekly lead samples were used to determine trends in lead leaching over time. There was a wide range of durations (1 month–1 year) for the studies reviewed in Table 2. Shorter studies are more convenient and can provide quick results, but longer studies tend to allow more time for stabilization.

Coupon tests were typically conducted in duplicate or triplicate to account for the inherent variability in lead leaching from the coupon materials used. One study highlighted the importance of having at least duplicate tests for every condition in order to have more confidence in the results (Nguyen et al., 2010). The single coupon study that used harvested iron coupons employed a 4-week pre-testing/conditioning phase to assess 420 coupons as “good” and “bad” by measuring the turbidity of the water in each jar after exposure to the same ground water using the dump-and-fill method every Monday, Wednesday, and Friday (Tang et al., 2018). After this phase, the 50 coupons with the highest turbidity were considered “bad” coupons and the 50 coupons with the lowest turbidity were considered “good” coupons. Fifteen water conditions were tested using three “bad” coupons, and three “good” coupons for each condition to represent extremes that could be seen in an actual drinking water system.

Because bench top studies are so variable, the State of Colorado recently standardized the process for optimal corrosion control treatment for drinking water systems assumed to have no lead service lines (LSLs) (Arnold et al., 2021; CDPHE, 2019) (Table 2). The standardized test includes three test materials for each water condition (50:50 lead/tin solder galvanically connected to copper pipe, leaded brass alloy, and copper pipe). The State of Colorado estimated in 2019 that it would cost $5400–$14,800 to test two to eight water qualities, respectively. Most of the cost was attributed to laboratory analysis with only $1000–$1800 for materials. Estimates were also made for operational costs if certified operators were contracted to perform the tests ($5000–$15,000) or if engineering consultants were contracted to perform the tests ($15,000–$25,000) for a 6-week duration. Estimated costs for the standardized bench top studies increased substantially ($50,000–$150,000) if an engineering consultant was hired for the entire project (i.e., from performing desktop studies to development, execution, and analysis of the bench top study results).

3.3 |. Pipe tests (1991–2022)

Pipe tests at the bench top seem to have generally been conducted in research laboratories to investigate the effects of potential water chemistry scenarios on lead leaching, according to the available peer-reviewed literature (e.g., Edwards & Dudi, 2004; Masters & Edwards, 2015; Triantafyllidou et al., 2012). Pipe tests have been used to evaluate uniform corrosion of single lead pipes (Figure 3d,e) (Nguyen et al., 2010; Sharp et al., 2009) and galvanic corrosion of lead pipes electrically connected to copper pipes (Clark et al., 2015; Nguyen et al., 2010; Triantafyllidou & Edwards, 2011), which is a more complicated configuration.

The pipe test studies reviewed herein (Table 3) consisted of mostly new lead pipe segments that ranged from 0.10 m (4 in.) to 0.3 m (12 in.) in length. However, some studies also tested fixtures other than pipes, including new brass hose bibs (Edwards & Dudi, 2004) and brass ball valves (Triantafyllidou et al., 2012). Unlike the submerged coupon tests, these pipes/fixtures were filled with water, as would typically occur if they were in service, and sealed with rubber stoppers or other means to contain standing water within them. The standard dump-and-fill protocol was used, with typical stagnation times of 2–3 days. Just as with the coupon tests, composite samples were taken weekly and analyzed for lead and other water quality parameters. Most of the studies used synthetic water created in the lab, to simulate known real-world water chemistries of interest. However, one study used tap water with pH adjustments and copper additions (Clark et al., 2015).

Pipe tests were typically conducted in triplicate (Clark et al., 2015; Edwards & Dudi, 2004; Masters & Edwards, 2015; Nguyen et al., 2010; Triantafyllidou et al., 2012), although some conditions were only tested on a single pipe or in duplicate (Edwards & Dudi, 2004; Triantafyllidou et al., 2012). While there were no standardized pipe lengths for these studies, lead pipes had standard inner diameters (IDs) of 12.7 mm (0.5 in.) or 19 mm (0.75 in.) (Clark et al., 2015; Nguyen et al., 2010; Triantafyllidou & Edwards, 2011). Due to the nature of the dump-and-fill protocol, the water was stagnant in the pipes for a longer amount of time than would typically be seen in a premise or other building plumbing system. These conditions represent a “worst-case” scenario that could occur in buildings that are not used over the weekend or over other prolonged absences.

3.4 |. Benefits and limitations of bench top studies

The low cost, small area (bench top) occupied, simple set up and operation, and the small volume of water needed for each bench top test make it easier to run multiple conditions at once. These studies also generally run for a shorter amount of time (typically 4–6 months), which is beneficial in getting results quickly, but does not capture what could be occurring in a real system over time. Additional challenges with bench top studies include the unrealistic flow regime (water is not flowing through the pipes/coupons) and longer than typical stagnation times (often 2–4 days, presumably for practical reasons and weekend breaks for study participants). These tests also typically use new plumbing materials that cannot form realistic corrosion scales even when conditioned, due to the short duration of the studies.

These unrealistic conditions generally create higher lead concentrations in water samples than would be observed with aged materials under typical hydraulic conditions. Pipe disturbance during the manual dump-and-fill process could also affect results, but disturbance might be reduced if pipes are held in a stand and can be emptied by gently removing the stopper from the bottom rather than inverting the pipe to dump the water. These bench top setups and conditions are often qualified as “worst-case” scenarios for lead release (Triantafyllidou & Edwards, 2011). They may be used for a simulated comparative bench top assessment among different water conditions (Table 1), but the exact lead leaching data are not meant to be directly extrapolated.

In addition to the peer-reviewed bench top studies of Tables 2 and 3, numerous other peer-reviewed bench top studies followed adapted similar set-ups and/or protocols which were excluded from the tables for brevity (Edwards et al., 2021; Edwards & Triantafyllidou, 2007; Pieper et al., 2016; Zhou et al., 2015). Other non-peer-reviewed studies may have been performed by water utilities or engineering consultants to test potential changes to water treatment and/or for regulatory compliance. Some of those have been presented at conferences (Seidel, 2020; Corwin, 2020; DeLorenzo & Klayman, 2013; Friedman, 2020; McFadden et al., 2009; Sharp et al., 2009; Vosa & Mofidi, 2021), published in reports (Black & Veatch, 2018; Cornwell Engineering, 2022), and Master’s theses (Arnold, 2011; Bradley, 2018). Such studies did not always contain enough information to be included in Table 2 or Table 3. Lastly, although not intended as corrosion assessment studies per se, industry standards under NSF/ANSI 61 (“Drinking water system components-Health effects”) for plumbing manufacturers to certify the safety of their plumbing products are essentially bench top tests under specific synthetic water exposure, sampling, and statistical analysis protocols for those products to obtain certification (NSF, 2016).

4 |. REVIEW OF LEAD CORROSION PILOT STUDIES

The use of pilot studies to investigate lead leaching from plumbing materials and to evaluate corrosion treatment options began in the 1940s (Hatch, 1941; Moore & Smith, 1942). Pilot studies consist of either single metal pipes (e.g., lead, leaded brass/bronze, and galvanized iron) to assess uniform corrosion or galvanic connections of multiple pipe segments (e.g., lead with copper) and/or leaded brass plumbing fixtures (e.g., faucets, valves, and meters) to incorporate galvanic corrosion. Pilot setups have typically been built for individualized needs and have not been commercialized on a large scale. There is only one US commercially available pilot setup mentioned in the literature (Cantor, 2009), that consists of new lead and copper plates instead of actual pipes.

In pilot studies, water can either recirculate through the system (termed recirculating pipe loop herein) or pass through the system once (termed once-through pipe rig herein) to a drain (Giani & Hill, 2017; USEPA, 2016). Pilot studies are complicated in design, because they incorporate a multitude of plumbing components other than leaded test pieces to complete the setup (e.g., pumps, time controllers, solenoid valves, pressure regulating valves, pressure gauges, PVC, flexible tubing segments, water meters, flow meters, sampling valves, automated feeding systems to alter water chemistry, a rack to hold the pilot, etc.). PVC pipes and components are used for these set ups to avoid contributing to lead leaching from anything other than the specific lead material(s) being evaluated. Different terminologies have been used for pilot studies through the years, including pipe loops (Kirmeyer et al., 1994), closed loop systems (Reiber et al., 1996), pipe racks (Kirmeyer et al., 1994; Parks et al., 2014), and pilot-scale distribution systems (Eisnor & Gagnon, 2003). Eisnor and Gagnon summarized key design considerations and findings for pilot-scale distribution systems developed between 1980 and 2002 (Eisnor & Gagnon, 2003).

4.1 |. Historical pilot (recirculating and once-through) studies (pre-1991)

Hatch (1941) conducted a series of tests where water was pumped through a lead wool column (Figure 4a, left) to determine the effects of sodium hexametaphosphate on lead leaching into water at a range of pH values. The lead wool was packed into a 32 mm (1.25 in.) diameter glass tube. The amount of lead wool used was calculated to be equivalent to the internal surface area of a 6 m (19.6 ft) long, 6.4 mm (0.25 in.) diameter standard lead pipe (Hatch, 1941). The glass wool leached more lead than would be typical of a pipe with a similar surface area; therefore, an additional test was run using 2.7 m (9 ft) of 6.4 mm (0.25 in.) diameter coiled lead tubing. Moore and Smith (1942) used a similar lead wool setup (Figure 4a, right), but a different water source and sodium hexametaphosphate dose. Tests with aged lead pipes were also conducted using the same setup as the lead wool experiments, with the lead pipes replacing the wool. For these experiments, water flowed at a rate of 200 mL/min during the day and was stagnant at night and over the weekend. Water samples were taken each morning after stagnation and every day after 8 h of steady flow.

FIGURE 4.

Schematic set ups of historical pilot studies for evaluating lead corrosion using (a) a continuous flow apparatus with lead wool (Hatch, 1941; Moore & Smith, 1942), © 1941 AWWA; (b) a pipe loop (Schock and Gardels (1983), © 1983 AWWA; and (c) a once-through pipe rig commonly referred to as the AWWARF pipe rack developed in 1990 (EES, 1990), Reprinted with permission, © The Water Reserch Foundation.

Beginning in 1978, once-through pipe rig studies were performed in the United Kingdom (UK) to test potential corrosion treatments (Carruthers & Jackson, 1981). The pipe rigs contained 2 m (6.6 ft) long lead pipe sections carefully excavated from the distribution system. High alkalinity water was pumped through the pipe rigs using a peristaltic pump controlled by a timer that operated at a cycle of 8 h on, 16 h off. The setup also allowed for the dosing of chemicals (NaOH, Na₂HPO₄, H₃PO₄, and sodium silicate) for corrosion control treatment options. Pipes were conditioned for an undefined amount of time before corrosion control treatments were tested. Water samples were collected weekly after 30-min, 3-h, and 16-h stagnation times for analysis (pH, temperature, alkalinity, lead, and orthophosphate concentrations).

In 1981, once-through pipe rigs were set up at two stations in Portland, OR to test the effects of different disinfectants (Treweek et al., 1985). Free chlorine was used at the Bull Run Reservoir Headworks, whereas chloramine was used at Sandy River Station. Six new pipe materials were used for the study (black iron, galvanized steel, copper, lead-tin-solder-coated copper, lead, and asbestos-cement). The study lasted for 18 months, with the lead concentration in the water reaching equilibrium at about 6–9 months of operation.

Schock and Gardels (1983) conducted a pipe loop study to examine the effects of pH on lead solubility. The study used Cincinnati, OH tap water and adjusted the pH using sodium hydroxide (NaOH) and hydrochloric acid (HCl). Water was held in a 1890 L (500 gal) reservoir and pumped through a 31 m (100 ft) long lead pipe loop (Figure 4b) five times a day for 15–30 min each time, at a flow rate of 0.06–0.09 L/s (0.9–1.5 gpm). Water samples were collected after various stagnation periods (typically greater than 7 h) for metal analysis (lead, copper, zinc, etc.) as well as alkalinity, pH, and free chlorine residuals. Water samples were also filtered through a 0.4 μm pore size vacuum filter to determine the dissolved lead concentration.

In anticipation of the 1991 LCR, the AWWA Research Foundation (AWWARF) developed a once-through pipe rig in 1990 (Figure 4c) (Giani & Hill, 2017; Reiber et al., 1996). The initial setup included one or more harvested lead or new copper pipe(s) that could hold 1 L of water. This setup used a stagnation time of 6 h, as is the minimum required stagnation time for LCR monitoring. This pipe rig was set up at the Lowell Water Treatment Plant in Lowell, MA in 1993 to test potential corrosion control treatment options (MacKoul et al., 1995). For this study, three pairs of pipe loops were constructed to test three different corrosion control scenarios (no treatment and two doses of zinc orthophosphate with pH adjustment). Each pipe loop pair had one loop with lead tubing and one loop with copper tubing sections soldered together using 50/50 Pb/Sn solder. The AWWARF pipe rig was also set up in Providence, RI in 1993 to fulfill regulatory requirements and improve corrosion control in the system (Yannoni & Covellone, 1998). Four pairs of pipe loops were used in RI, each containing one 9.1 m (30 ft) copper pipe loop of 12.7 mm (0.5 in.) diameter with 16 50/50 Pb/Sn soldered joints and one 9.1 m (30 ft) lead pipe of 12.7 mm (0.5 in.) diameter excavated from the system.

4.2 |. Recirculating pipe loops (1991–2022)

Recirculating pipe loops (Table 4) have been more commonly used in laboratory experiments, presumably because they require significantly less water than once-through pipe rigs. Some pipe loop setups are small enough to fit on bench tops in the lab (e.g., Bae et al., 2019, 2020; Nguyen et al., 2010), which can be beneficial for smaller utilities with limited resources and space than those required for larger pipe rig setups.

Pipe loops have generally used new lead pipes (Figure 5a) (Bae et al., 2019; Kirmeyer et al., 1994; Nguyen et al., 2010; Welter et al., 2015) or lead pipes harvested from drinking water systems (Bae et al., 2020) with standard inner diameters of 12.7 or 19 mm (0.5 or 0.75 in.). Some studies also used other plumbing materials, such as new bronze and copper pipes with lead solder (Nguyen et al., 2010; Zhang et al., 2012), aged lined and unlined cast iron pipes, and aged galvanized steel pipes excavated from drinking water systems (Tang et al., 2006). Pipes were conditioned for studies that used new pipes (Bae et al., 2020) as well as harvested pipes (Bae et al., 2020; Tang et al., 2006).

FIGURE 5.

Examples of pilot study set ups for (a) recirculating pipe loops with new lead and PVC pipes (Welter et al., 2015), ©2015 AWWA; (b) once-through pipe rig with harvested lead pipes (Williams et al., 2018), ©2018 AWWA; (c) once-through pipe rig with excavated brass meters, copper pipes, and lead goose necks (Friedman, 2020), Photo source: Tacoma Water and HDR Engineering, Reprinted with permission; (d) Process Research Solutions (PRS) monitoring station with new lead and copper plates held in chambers (Cantor, 2017), Reprinted with permission.

Studies used new lead pipes for simplicity and reproducibility, but results may not be as realistic as from excavated pipes that have scale build up from years of use. Bae et al. (2019) conditioned new lead pipes with synthetic water similar to the water in Washington, DC, prior to the disinfectant switch in 2000. These pipes were conditioned for an unusually long time (66 weeks) with the goal of producing Pb(IV) (i.e., PbO₂) scale on the pipe surface, as demonstrated by low lead leaching (<5 μg/L) after conditioning, since harvested pipes from the system were not available. Pipe scale analysis confirmed that a 15 μm-thick scale of crystalline PbO₂ had formed on the pipe after the conditioning phase (Bae et al., 2019). It is still unclear how close that scale was to a PbO₂ scale formed and maintained in the field. Studies that used harvested pipes also had a conditioning phase ranging from 4 months (Tang et al., 2006) to over a year (Bae et al., 2020) in order to restabilize the pipe scales after transport to the labs.

Most pipe loop studies used finished drinking water shipped from the water utilities with adjustments made in the lab to test the effects of pH, alkalinity, disinfectants, and corrosion inhibitors (Kirmeyer et al., 1994; Nguyen et al., 2010; Tang et al., 2006). Due to the lower volume of water required for pipe loop studies, synthetic water can also be made in the lab to represent real world waters (Bae et al., 2019, 2020). Water in the reservoirs (typically 10–20 L) was exchanged with fresh water weekly for most studies (Bae et al., 2019, 2020; Nguyen et al., 2010).

To simulate real world situations, the pipe loop studies had intermittent flow cycles with a typical stagnation period of 6–8 h (Bae et al., 2019, 2020; Kirmeyer et al., 1994; Schock & Gardels, 1983) although some studies had stagnation periods over 24 h (Schock & Gardels, 1983; Tang et al., 2006). Schock and Gardels (1983) found that after a stagnation period of 7–9 h, nearequilibrium lead concentrations were reached, and there was not a significant difference in concentration if the stagnation period was increased to 21–88 h. Pipe loop tests tried to mimic real world situations with flow rates ranging from 0.06 L/s (0.9 gpm) to 0.19 L/s (3 gpm) during the recirculation period. Water samples were generally taken 1–2 times per week after the stagnation period. These samples were analyzed for lead and other water quality parameters. The weekly/bi-weekly lead samples were used to determine trends in lead leaching over time. While most of the studies did not specify the number of tests run for each condition, the studies that did had triplicate tests (Bae et al., 2019, 2020).

4.3 |. Once-through pipe rigs (1991–2022)

Once-through pipe rigs (Table 5) are generally set up at water treatment plants, where they can be connected to an unlimited supply of water (Churchill et al., 2000; MacQuarrie et al., 1997; Parks et al., 2014; Williams et al., 2018). One study also set up rigs at pumping stations downstream of the treatment plant in the distribution system (Parks et al., 2014). However, some rigs have been set up in laboratories, where they are connected to the building’s tap water supply (Cartier et al., 2012, 2013; St. Clair et al., 2016).

Some rigs were composed of excavated lead pipes (Cartier et al., 2013; Masters et al., 2022; Williams et al., 2018), while others used new lead pipes (Figure 5b) (Cartier et al., 2012; Churchill et al., 2000; MacQuarrie et al., 1997; Parks et al., 2014). Some pipe rig setups also used new copper pipes with lead solder (Cantor et al., 2000; Churchill et al., 2000; Parks et al., 2014; Treweek et al., 1985), as well as leaded brass plumbing fixtures (Figure 5c) (Churchill et al., 2000; MacQuarrie et al., 1997; Parks et al., 2014). Although not all studies reported the inner diameter of the lead pipes used in each rig, those that did had a diameter of 19 mm (0.75 in.).

Instead of actual pipes, the Process Research Solutions (PRS) Monitoring Station (Cantor, 2009) consisted of new lead and copper metal plates inside two separate test chambers (Figure 5d). Water was intermittently flowing through the PRS Monitoring Station, and the lead and copper surface area to volume ratio was similar to that of a 45 mm (1.77 in.) diameter pipe (Cantor, 2017). Although not technically a pipe rig due to the metal plates, the PRS monitoring station’s once-through flow regime was similar to that of pipe rigs and was thus classified here.

These once-through pipe rig studies tried to mimic real world situations with stagnation times typically ranging from 5–16 h per day (Cartier et al., 2012, 2013; Churchill et al., 2000; Masters et al., 2022; Parks et al., 2014; Williams et al., 2018). However, some studies had longer stagnation periods of 23 h (Cantor, 2017), or 72 h over the weekend (Cartier et al., 2013). Samples were typically collected 1–2 times per month after a stagnation period (Churchill et al., 2000; Parks et al., 2014; Treweek et al., 1985; Williams et al., 2018), but were sometimes collected more frequently (Cartier et al., 2012, 2013; Masters et al., 2022). The PRS monitoring station had a 23 h stagnation period each day and a flowrate of 0.032 L/s (0.5 gpm) per test chamber for 1 h per day, and samples were taken bi-weekly after a 6 h stagnation period for metal analysis (Cantor, 2017).

These studies lasted 7–44 months (0.6–3.7 years), including conditioning periods that lasted from 0 to 8 months. Cartier et al. (2013) conditioned harvested lead pipes from Montreal for 8 months, which included water passing through the pipe rig that consisted of only PVC and lead pipes. Before the conditioning phase, the pipes were stored in the rig setup with Montreal water inside for 5 months under stagnant conditions. After the conditioning phase, there was a baseline sampling phase where the PVC pipe was replaced with copper pipe, the water chemistry remained the same, and samples were collected weekly to ensure reproducibility within the pipe rigs. Cartier et al. (2012) reported the short-term results from 11 weeks of the study, while Doré et al. (2019a) reported the results of the long-term lead release from the 155-week (3-year) study using the same pipe rigs. Parks et al. (2014) used new plumbing materials for the pipe rigs to aid in reproducibility among triplicate tests. This study initially prepared 40 new test pieces for each plumbing material (brass, copper with lead solder, and lead pipes) for the rigs and pre-tested all the pieces using a protocol based on NSF Standard 61, Section 9 (NSF/ANSI, 2007). After measuring lead and copper concentrations during the pre-test, 30 test pieces were chosen to ensure the pieces would provide the most reproducible lead results. Each of the 10 pipe rigs constructed had three test pieces for each plumbing material so that triplicate tests could be run for each water condition. The rigs were then conditioned for 1 month prior to the start of the study. After the study, it was recommended that the conditioning phase be extended in the future. That would allow for additional passivation time for the plumbing materials, which could help with the representativeness of the study (Parks et al., 2014).

There is limited information available on the cost of pilot studies. Only Parks et al. (2014) estimated the material cost of their pilot rig at $3100–$3300 at the time of reporting. The cost accounted for triplicate tests of three plumbing materials (copper/solder test pieces, brass test pieces, and lead/copper test pieces), but did not estimate pilot operation, water sample metal analysis, and other costs.

In addition to the peer-reviewed pilot studies of Tables 4 and 5, numerous other peer-reviewed pilot studies adapted similar set-ups and/or protocols which were excluded from the tables for brevity (Aghasadeghi et al., 2021; Doré et al., 2019a, 2019b; Kogo et al., 2017; Li et al., 2021; Trueman & Gagnon, 2016; Wang et al., 2012; Woszczynski et al., 2013, 2015; Gagnon & Doubrough, 2011). Once-through pipe rigs have presumably been more commonly used by water utilities for regulatory compliance, rather than for academic research. As a result, many studies may have been performed through the years, which may not have been published in peer-reviewed journals or elsewhere to be incorporated here. Limited information was available in some conference proceedings as well as utility reports on studies that used similar setups to the studies published in peer-reviewed journals using LSLs (Atassi et al., 2004; Atassi & Putz, 2010; Boyd et al., 2007; Cantor, 2008; Ingels & Poncelet-Johnson, 2020; Kwan et al., 2006; Welter et al., 2015; Wysock et al., 1995), copper pipes with lead solder (Doubrough, 2009), and galvanized iron pipes (McFadden et al., 2009) as well as new plumbing faucets and new/used water meters containing leaded brass (Maynard et al., 2008).

5 |. SUMMARY

Bench top and pilot lead corrosion assessment studies are two of many corrosion assessment tools (Figure 1) that can help water utilities better understand their current situation, determine the potential effects of changes in water treatment, and meet regulatory requirements if they are in violation of the LCR. They have also been used to improve scientific understanding of lead release by academic groups or for industry certification of plumbing products under NSF/ANSI 61. A historical review of such corrosion assessment studies demonstrated a progression from lead wool or metal sheets to actual lead pipe or other plumbing segments, and from electrochemical measurements of corrosion rate to direct metal release quantification.

This review classified modern bench top studies into two general categories, namely coupon tests and pipe tests. It also identified two general categories of pilot studies, namely recirculating pipe loops and once-through pipe rigs. Between and within these general categories, there are many different designs (e.g., lead pipe or other leaded plumbing material type, age, number of replicates) and operating conditions (e.g., water chemistry, flow regime, stagnation time, sampling frequency, study duration) (Table 6) to be aware of before making a choice.

TABLE 6.

Summary of design considerations for lead corrosion bench top and pilot studies based on reviewed literature.

Parameter	Considerations	Bench top study (coupon test or pipe test)	Pilot study (recirculating pipe loop or once- through pipe rig)
Pipe/plumbing fixture materiala	Select leaded materials of concern: Lead service lines, or other leaded materials for systems that removed or never had lead service lines	Lead Lead solder with copper Brass (leaded) Copper Galvanized iron	Lead Lead solder with copper Brass (leaded) Copper Galvanized iron
Pipe/plumbing fixture age	Aged pipe is more representative but can affect variability	New (typically)—conditioned for 0–1 year (typically <4 weeks) Aged - Excavated from drinking water system (conditioned for 4 weeks)	Aged—excavated from drinking water system (conditioned for 4–8 months) New—conditioned for 0–15 months
Water chemistry	Finished water represents real-world seasonal changes and other variabilities which can affect reproducibility	Synthetic drinking water Finished water from utilities (Shipped from utilities or at the utility) or tap water in lab	Finished water from utilities (studies set up at water treatment plants) Synthetic drinking water
Reproducibility	Replicate test conditions are preferable to account for sources of variability	Typically: triplicates Range: 1–3 coupon(s)/pipe(s) per condition	Typically: triplicates Range: 1–4 pipe(s)/plumbing fixture(s) per condition
Flow regime	Intermittent flow is more representative but also more complicated than dump and fill	Stagnant (dump and fill)	Intermittent flow (typically 1–3 gpm)
Stagnation time	Longer stagnation can cause unrealistic results due to more lead leaching	Typically: 2–3 days Range: 1–4 days	Typically: 6–8 h Range: 6 h–5 days
Sampling frequency	Frequent sampling can help show trends in data but is also more time consuming	Typically: 3 times per week for weekly composite sample Range: 1× per month-3× per week	Typically: 1–2 × per month Range: 1× per month-3 × per week
Quantification of lead	Lead leaching is the main analysis after weight- loss fell out of favor	Lead concentration in water sample (coupon weight loss previously)	Lead concentration in effluent water after stagnation Pipe scale analysis before and/or after study
Study length (including conditioning)	Longer study length allows for longer conditioning, stabilization and testing with results generally more realistic	Typically: 4–6 months Range: 1 month-1 year	Typically: 1.3 years Range: 1–3.7 years

Note: Clarifications: Finished drinking water: water directly from the drinking water treatment plant that has not come in contact with any plumbing within the distribution system; Synthetic water: water made in the lab by adding chemicals to deionized water to simulate finished drinking water.

Abbreviation: gpm, gallon per minute.

^aFocus of this review is lead but other material pipes may be tested (i.e., cast iron or ductile to represent water mains).

Because there is no standardized pilot or bench top design for lead corrosion assessments, it can be challenging to compare results between different studies. However, within each study, a variety of different water conditions can be tested to give an idea of the relative change that a treatment can cause. While variability within a study is unavoidable, proper design and operating conditions can achieve reasonable control over water quality and material variables. Schock and Lytle (1994) highlighted the importance of proper measurements and control over certain water quality parameters (pH and dissolved inorganic carbon-DIC) that have a significant impact on lead leaching over the course of a study.

In order to account for variability within a study, triplicate tests are typically run for both bench top and pilot studies when enough resources are available, ideally incorporating some combination of pre-testing, conditioning, and stabilization. The reader is directed to Wysock et al. (1995) and Kimeyer et al. (1996) for statistical analysis considerations that could not be covered here. Worth noting is that the majority of current pilot studies use lead pipes (both new and harvested) to represent LSLs. Once LSLs are removed from drinking water systems, the focus will inevitably switch to other remaining leaded plumbing materials such as leaded brass plumbing fixtures (e.g., faucets, valves, and meters), copper pipes with lead/tin solder, and galvanized iron pipes.

Table 7 summarizes the key differences between bench top studies and pilot studies. Due to their relatively low cost and simpler set up, bench top studies can be used to test multiple conditions. Results of these studies can then be used to determine the condition(s) to test in a pilot study with a more realistic flow regime, with cycles of flow typically followed by stagnation periods of 6–16 h. Pilot studies that used pipes harvested from the distribution system, can allow the metal release behavior of pre-existing scales on actual distribution system material under typical hydraulic conditions to be observed. These factors help pilot studies more closely approximate real-world complexities. However, they also increase costs, take a longer amount of time to start up and operate, and create greater variability within the study due to the complex nature of pre-existing scale. For example, Wysock et al. (1995), showed that there was variability in lead leaching in pipe rig studies using harvested lead pipes after a 162-day conditioning period, even when the pipes were excavated from the same street. While there was limited information available for the cost of both bench top and pilot studies, looking at material cost alone, one pilot study was estimated to be at least six times more expensive per water condition (CDPHE, 2019; Parks et al., 2014).

TABLE 7.

General comparison of bench top and pilot studies based on reviewed literature.

	Bench top study (coupon test or pipe test)	Pilot study (recirculating pipe loop or once-through pipe rig)
Pros	• Low cost	• More realistic intermittent flow regime
	• Simple set up (small footprint, thus “bench top”) & operation	○ Cycles of flow followed by stagnation • Shorter stagnation time (6–16 h)
	• Small water volumes needed	• Typically pipes harvested from the distribution system
	• Easy to run multiple conditions at once	○ More realistic aged materials
	• Typically less variability between replicates (except galvanic connections)	• Longer testing period (can capture seasonal impacts and generate more realistic results)
	• Shorter time-frame (in the order of months; this is also a con)	• More closely approximates real-world complexities and mimics full-scale conditions
Cons	• Unrealistic flow regime	• Significant time and cost
	• No flow (stagnant samples) • Longer stagnation time (3–4 days)	• Can be difficult to maintain pressure and flow conditions throughout
	• Typically new plumbing materials  ○ Study duration too short for realistic scales to form, even when conditioned	• Can take months for system to stabilize and may never stabilize to same level prior to pipe excavation (e.g., in case of Pb(IV) scale)
	• Higher lead concentrations than aged materials under premise plumbing hydraulic conditions	• Variability among replicates even after stabilization • Longer time-frame (1–2 years; this is also a benefit if it can be afforded)

It can also be difficult to maintain the pressure and flow conditions throughout the pilot study, and it can take months for the system to stabilize. Pilot studies tend to run for a longer period of time, and while it can take longer to obtain results, they can be more realistic. Because of their large startup cost, pilot studies are sometimes kept in place after the end of a study, but they are not always subsequently maintained or sampled. Large drinking water utilities with pilot rigs previously set up have an advantage in that they can monitor their current situation and could theoretically assess an upcoming treatment or other water quality change with their existing rig, with appropriate maintenance and conditioning. For drinking water systems starting from scratch, the time and monetary investment could conceivably be quite high.

While most peer-reviewed literature focused on either bench top or pilot studies, Masters et al. (2022) reported both coupon and pipe rack studies to examine the effects of pH adjustment and orthophosphate addition on lead leaching for two water sources in Denver, CO. Although the water chemistries were the same for both study types, the plumbing materials, stagnation time, sampling, and duration were different (Tables 2 and 5). The average percent lead reduction results from these studies showed surprisingly similar findings for the corrosion control treatments tested, although pipe rack lead results more closely matched actual water system sampling results. The authors warned not to generalize these results to other systems. Many more comparative studies like this would be needed to understand how coupon studies, pipe studies, and field water lead sampling compare to each other in various situations.

Overall, the complexity of corrosion assessment studies progressively increases as we move from simpler coupon tests to more complicated once-through pipe rigs (Figure 6). Generally, increasing the complexity of the study will produce more realistic results but will also increase the time, resources, and space (footprint) needed. This can be especially difficult for smaller utilities that may not have the same resources as large utilities. Bench top stagnant coupon tests are the easiest and simplest tests to pre-screen test conditions, whereas once-through pipe rigs offer the most complex and realistic study conditions if resources are available. Recirculating pipe loops could be considered a reasonable middle ground overall. Pipe loops can simulate the intermittent flows that bench top studies do not, but they are not as complex as once-through pipe rigs and require much less water and potentially less space, with some studies even being small enough to fit on bench tops (e.g., Bae et al., 2019, 2020; Nguyen et al., 2010). Whatever the choice, any surrogate test system that is selected cannot replace an actual system’s water sampling program.

FIGURE 6.

Increasing complexity of bench top and pilot studies from simpler coupon tests to more complicated once-through pipe rigs. As complexity increases, the time, resources/cost, and footprint also increase.

This literature review focused on bench top and pilot studies that were available in journals, conference proceedings, or reports. It would be impossible to capture all published studies herein, whereas practical knowledge and experience may not have always been passed on through publications in the first place. As a result, other designs or modifications may have been used that are not captured here. Similar to bench top and pilot studies, other corrosion assessment tools have usefulness and limitations (Table 1). Combining different tools can provide a better scientific understanding of corrosion control treatment options, depending on the specific needs and constraints.

Article Impact Statement.

Comparison of bench top and pilot designs (lead test piece age/dimensions/configuration, flow pattern, duration, etc.) can allow selection of the appropriate lead corrosion study depending on intent.

DATA AVAILABILITY STATEMENT

Data sharing not applicable to this article as no new datasets were generated or analyzed during the current study.