CHAPTER TWO A CRITICAL EVALUATION OF THE NSF STANDARD 61/ SECTION 9 TEST WATER
Simoni Triantafyllidou and Marc Edwards Civil and Environmental Engineering Department 418 Durham Hall; Virginia Tech; Blacksburg, VA 24061
The ANSI/NSF Standard 61/ Section 9 protocol is critically evaluated from the perspective of test water chemistry and lead leaching propensity. Overall, the test water is roughly representative of a typical water supply. However, some lower pH and lower alkalinity, chloraminated waters can be more aggressive than the existing section 9 water, especially at longer exposure times, which might cause problems with first draw compliance for lead even for products that are section 9 certified. For public water supplies, the aggressiveness of the lower pH and lower alkalinity waters is often countered by addition of orthophosphate corrosion inhibitors. However, Section 9 devices are also recommended for use in the approximately 16% of private water supplies and cistern-type systems, which can include waters that are much more aggressive than tested herein. To account for these issues, it is recommended to tighten the Section 9 pass/fail criterion, and to add product information about limitations in corrosivity of the products relative to private water supplies.
KEYWORDS: ANSI/NSF 61 Section 9, brass, lead leaching INTRODUCTION
ANSI/NSF Standard 61 was developed in order to “ establish minimum requirements for the control of potential adverse human health effects from products that contact drinking water” (NSF, 2001). It was intended for use by utilities, regulatory agencies, and/or manufacturers as a basis of providing assurances to consumers that adequate health protection exists for certified products.
Section 9 of NSF 61 Standard is aimed at protecting the public from harmful levels of contaminants derived from plumbing devices installed in the last one liter of water volume in plumbing systems. These plumbing materials are termed “end-point” devices and can include kitchen faucets, stop valves, and drinking water fountains. The one liter demarcation is of particular interest, because this is the volume of water collected after an overnight stagnation under the United States Environmental Protection Agency (US EPA) lead and copper rule (LCR). Plumbing components covered in Section 9 are those that might contact the water collected during stagnation in the LCR and the Lead Contamination Control Act (LCCA) to determine compliance with lead standards for potable water. Any leaching of lead (or copper) from endpoint devices can strongly impact compliance with these regulations.
Lead, amongst other contaminants that are considered in the NSF 61 standard, poses great concern due to increasing evidence of harm from low -level lead exposure. Lead and copper are also the only contaminants which have drinking water standards determined at the tap after overnight stagnation. Moreover, lead leached from brass to first draw water samples collected under the LCR and LCCA is often considered a significant cause for action level and LCCA exceedences. There is anecdotal evidence that in at least some rare cases, new devices passing Section 9 can still leach lead to potable water at concentrations above the standards of 20 ppb Pb for the LCCA (200 mL sample first draw) and 15 ppb Pb for the LCR (1 liter first draw) . This work examines possible reasons why the section 9 test might be failing to perform as expected in some of these situations.
The evaluation of products covered by Section 9 requires their exposure to one synthetic extraction water under a specified protocol, followed by statistical analysis of the lead leaching results which allows “safe” products to pass the test and be certified (Figure 2-1). Products above the standard pose a potential public health or regulatory concern and do not get the certification. The corrosivity of the test water is of obvious concern, since some waters have a high propensity to leach lead from end-point devices and other waters have a low propensity to leach lead. If devices were to be
tested in a water that had a low propensity to leach lead, the values might be lower than those which would occur in real waters, and the test would not be protective of public health.
When examining the test water chemistry used for Section 9 certification, the relatively high alkalinity, high pH and free chlorine level are of concern due to their influence on lead leaching propensity. Specifically, the test water has high alkalinity of 500 mg/L (± 25 mg/L) as CaCO3, a relatively high pH of 8.0 (± 0.5) and free chlorine at a concentration of 2.0 mg/L (± 0.05 mg/L). Due to the addition of hydrochloric acid (HCl) for pH adjustment, the extraction water also contains chloride (Cl-) at an estimated concentration of about 3.5 mg/L. The source of the high alkalinity is sodium bicarbonate, which adds dissolved inorganic carbon (DIC) at a concentration of 120 mg/L (± 5.0 mg/L).
In a practical study of 90%’ile lead levels at about 400 large US water utilities, Dodrill and Edwards (1995) found that waters with alkalinity above 174 mg/L and pH above 7.81 had a 0% frequency of exceeding the lead action limit (Figure 2-2). This is consistent with known benefits of higher pH and higher alkalinity (or DIC) in reducing lead leaching propensity (Schock, 1989). The highest alkalinity for utilities in the Dodrill study was 433 mg/L as CaCO3 (median alkalinity was 60 mg/L as CaCO3). Therefore, use of a test water with 500 mg/L alkalinity at pH 8.0 in the Section 9 testing is not representative of aggressive conditions encountered in practice. For example, the Dodrill study found that waters of pH 7.4-7.8 and alkalinity below 30 mg/L had a 65% likelihood of exceeding the 15 ppb lead action limit (Figure 2-2).
The importance of pH on lead release from brass was also emphasized upon in other studies. An in-depth laboratory investigation of several brass types with varying lead content determined that lowering the water pH from 8.5 to 7.0 always increased lead leaching, in some cases to as much as 100 ppb more (Lytle and Schock, 1996). That study also assessed the effect of orthophosphate on lead leaching from brass and concluded that when the inhibitor was dosed at 3.0 mg/L it caused lead levels to drop rapidly and stabilize for all brass types tested (Lytle and Schock, 1996).
Recent work (Dudi, 2005) has also heightened concern regarding other factors that might increase lead leaching. Specifically, chloramine is often slightly more aggressive in relation to lead leaching from brass than chlorine (e.g., about 15% higher lead), and chloramines can cause much more lead leaching from certain types of brass. It was recently suggested that the presence of fluoride and chloramine was a particularly aggressive combination for lead leaching from brass (Allegood, 2005). Galvanic connections between copper and brass were also demonstrated to increase lead leaching from brass in some cases. In addition, recent studies by Korshin et al. (2000) demonstrated that the presence of natural organic matter (NOM) in water can increase lead leaching from the surface of leaded brass, over both the short and long-term. But other studies with NOM have shown the opposite effect. The Section 9 water does not contain any NOM, fluoride or chloramines (instead it contains chlorine). Likewise, galvanic connection of the products to copper is not required in the NSF Section 9 test protocol.
In terms of other possible factors that might influence the difference in aggressiveness of real waters versus the Section 9 test water, the Cl- level and the Cl-:SO4-2 ratio might also be important. The effect of chloride has been studied with respect to the dezincification of brass, a typical form of brass corrosion. When the chloride concentration is high, dezincification can more readily occur under some conditions (Lytle and Shock, 1996). Dezincification is sometimes correlated to lead dissolution from the brass surface (Kimbrough 2001, USEPA 1993, Mariñas et al., 1993). The ratio of Cl-:SO4-2 also plays a poorly understood role in lead leaching (Edwards et al., 1999). In general, higher Cl-:SO4-2 ratio can increase difficulties in meeting the EPA Lead and Copper Rule Action limit. Dudi (2005) demonstrated that increasing the Cl-:SO4-2 ratio causes increased corrosion and lead leaching from galvanically connected lead bearing materials and copper. The NSF water has Cl- levels that are about 90% lower than terrestrial waters in the US (Davies and DeWiest, 1966) and the chloride to sulfate ratio is infinite. These factors might also play a role in either increasing or decreasing lead leaching from brass in the test.
The goal of this research was to examine leaching of lead from one common type of brass under a range of circumstances, in an attempt to isolate key factors that might be desirable in a standardized test to make it more representative of aggressive conditions that can be encountered in practice. Other work examined other types of brass and the role of phosphate inhibitors.
MATERIALS AND METHODS
Evaluation of the leaching solution for C36000 Brass. One of the most common types of brass used in faucets is C36000 (60% Cu, 3% Pb, 37% Zn). In this test small C36000 coupons were exposed to the Section 9 test water, as well as to eight modifications of that water. With exceptions that are detailed below, all tests were performed following Section 9 procedures, with metals leaching evaluated over the 19-day test.
Machined brass rods of 0.64 cm (0.25 in) diameter and 0.97 cm (0.38 in) height were epoxied to the bottom of a 46 mL glass vial. This vial was filled all the way to the top with test water and air was excluded. For testing of galvanic connections between brass and copper, the small brass sample was connected to the inside surface area of a copper pipe using lead-free solder. The bottom of the brass in this case was covered with epoxy, as occurred in the case without the connection to copper pipe. The C36000 brass samples were in an “as extruded” form and were provided by the Copper Development Association (CDA).
The brass surface area to water volume ratio achieved was 8.0 x 10-3 in2/mL in the case of the samples not galvanically connected to copper. This ratio is at the lower end of those typically encountered in NSF testing, assuming ¾ in2 surface area is typical for a brass faucet (Weed, 2005) and that a typical brass faucet can hold a water volume of 100 mL (Gardels and Sorg, 1989). The exposed surface area to volume ratio was decreased by only about 15% in the case of galvanic connections to the copper, due to the solder used to attach the brass coupon to the copper pipe. A lower surface area to volume ratio would tend to decrease the final concentration of lead in the water.
The nine different test water chemistries were as follows:
A) NSF Section 9 test water. This water has high alkalinity of 500 mg/L (± 25 mg/L) as CaCO3, pH of 8.0 (± 0.5) and free chlorine dosed at a concentration of 2.0 mg/L (± 0.05 mg/L).
B) NSF Section 9 test water modified with 0.5 mg/L natural organic matter. The NOM is reference Suwannee River material obtained from the International Humic Substance Society.
C) NSF Section 9 test water with 4.0 mg/L chloramine (Mass ratio Cl2:NH3 of 4:1) instead of chlorine.
D) NSF Section 9 test water with 4.0 mg/L chloramine and 1 mg/L fluoride added as NaF.
E) NSF Section 9 test water with a solder connection to a small piece of 1/2″ copper tube.
F) NSF Section 9 test water but with lower alkalinity of 10 mg/L as CaCO3 and pH of about 7.4.
G) NSF Section 9 test water with additional 30 mg/L chloride added as NaCl.
H) Water with the galvanic connection and all modifications from B to G. This condition examined the cumulative result of all modifications.
I) Real Blacksburg tap water, which is considered non-aggressive.
In addition, control conditions were run, including 1) Epoxy control (a drop of epoxy glued to the bottom of the glass vial without the brass), and 2) Galvanic Control (copper with the solder inside the glass vial but without the brass).
The brass samples were exposed to the different water chemistries following the Section 9 test protocol. That is, they were exposed for 19 days using a static “fill and dump” procedure. Water samples were collected after 16-hour dwell times on days 3, 4, 5, 10, 11, 12, 17, 18 and 19 and analyzed for metals. Lead, zinc and copper dissolution were quantified using Inductively Coupled Plasma Mass Spectrometry (ICP-MS).
Each test was conducted in triplicate to support statistical analysis of the results. With the 9 different water conditions and 2 sets of controls, 33 tests were conducted for the C36000 brass.
Evaluation of the Leaching Solution for Brasses of Different Lead Content and Using Different Inhibitors. Evaluation of tests from the preceding section raised new questions that were addressed in a separate phase of experiments. Three brass types with a lead content of 2% (C35300), 3% (C36000) and 5% (C83600) were tested, using the same general procedures as in part 1 for a limited range of waters.
In addition, galvanic samples were prepared by physically squeezing brass into a hole machined in the side of a copper tube, thereby eliminating the need for solder.
The experimental conditions tested in part 2 are as follows:
A) NSF Section 9 test water, with high alkalinity of 500 mg/L (± 25 mg/L) as CaCO3, pH of 8.0 (±0.5) and free chlorine dosed at a concentration of 2.0 mg/L
B) Synthetic water with lower alkalinity of 10 mg/L as CaCO3, lower pH of about 7.4 and monochloramine instead of chlorine dosed at a concentration of 2.0 mg/L (± 0.05 mg/L)
C) Water B but with 3.0 mg/L PO4-3
D) Water C but with added zinc at 0.5 mg/L Zn+2
Exposure of the brasses to water was via a static“fill and dump” protocol three times per week (Monday/Wednesday/Friday), for four weeks. The water from each test condition was collected every Friday, after a 48-hour dwell time, and the sample was analyzed for metals. Even though the exact exposure protocol of Section 9 was not followed in part 2, the results of this analysis allow for drawing conclusions about the Section 9 water’s aggressiveness, as it is compared head to head with other water chemistries and under similar exposure conditions. All glass vials were kept out of light throughout the testing period and each test was performed in triplicate. Metals analysis was performed via Inductively Coupled Plasma Mass Spectrometry (ICP-MS).
Evaluation of the leaching solution for C36000 Brass. Lead dissolution was much higher for the low alkalinity-low pH water compared to the standard Section 9 water. Though absolute levels should be viewed with caution due to the very low surface area to volume ratio, the levels in the low alkalinity and low pH water were above the LCR Action Level of 15 ppb for most of the exposure period (Figure 2-3). Moreover, lead leaching was on an uptrend with time in this more aggressive water, contrary to the trend observed for the other conditions. For the condition at low alkalinity, zinc dissolution was also significantly higher compared to the standard extraction water (Figure 2-4). The brass sample that was exposed to the low alkalinity water became more red in color by the end of the 19-day testing (Figure 2-5), consistent with higher dezincification.
Although no primary (health-based) MCL exists for zinc in drinking water, higher zinc could be indicative of general brass corrosion and is also partly responsible for higher lead. That is, as the zinc in the alloy dissolves, more of the trace lead in the alloy is likely to dissolve or detach to the water. Kimbrough (2001) and various researchers (US EPA, 1993) also suggest that lead is mobilized along with zinc during the dezincification process, and that lead leaching might be reduced if dezincification is controlled.
In the case of the low alkalinity-low pH water, copper released to the water from brass was low, and at the same time zinc and lead released were the highest (Figure 2-3, 2-4, 2-6). The reddish color of the low alkalinity-low pH brass coupon due to enrichment of copper in the brass as zinc and lead were selectively leached out (Figure 2-5) is also characteristic of dezincification.
At first glance, lead dissolution was either not significantly different, or was even lower, for all other water conditions examined (besides the low alkalinity-low pH case), compared to the standard extraction water (Figure 2-3). However, by pooling the 17, 18, and 19-day data for lead for each test condition and applying a t-test at 95% confidence, statistical comparisons can be made (Figure 2-7). The addition of NOM did not make a significant difference in lead leaching at 95% confidence. However, the presence of chloramine instead of chlorine in the water resulted in a 48% increase in lead. The combined effect of chloramine and fluoride was a 19% increase versus the standard water, although this difference was not significant at 95% confidence. If anything, addition of fluoride tended to decrease lead leaching in this particular instance, since the combination of chloramine and fluoride leached 29% less lead than did chloramine alone. Chloride addition to the standard water decreased lead leaching by about 20%. The highly non-aggressive tap water from Blacksburg, VA leached 80% less lead than the section 9 water (Figure 2-7).
For all conditions examined, the copper leached to the water was significantly lower than the LCR Action Level of 1,300 ppb for the non-galvanic samples (Figure 2-6). For the galvanic connections, the copper leached in the standard section 9 condition, was higher and close to that level (Figure 2-8). It is uncertain whether most of the copper came from the brass or from the copper pipe.
The non-aggressive real tap water using a zinc orthophosphate inhibitor leached much lower lead, copper and zinc compared to the section 9 test water (Figure 2-3, 2-4, 2-6 and 2-7). While there are many other differences between the real water and the section 9 water, it is possible that the presence of a zinc orthophosphate inhibitor has a dominant effect. Nationally, about 56% of public water utilities currently add phosphate-based corrosion inhibitors to counter the corrosivity of water (McNeill and Edwards, 2002).
Surprisingly, the samples in which brass was galvanically connected to copper were either not significantly different or even lower in lead leaching, relative to the same sample without galvanic connection (Figure 2-3, 2-9). The galvanic samples also had high variability of results between triplicates for the same water condition. The lower lead using a galvanic connection was surprising, since earlier work by Dudi (2005) demonstrated that galvanic connections to copper tended to increase lead leaching, which is expected given that brass is typically anodic to copper. One tentative important difference between this study and that of Dudi (2005), is that this work used a 95% tin and 5% antimony solder to attach the brass to the copper pipe. It is possible that the small amount of tin solder used to connect the copper to the brass may have been a major influence. Tin is anodic to both copper and brass, and if it were sacrificed, it would reduce corrosion of both copper and brass. In some cases the water in the test vials was white with a solid (Figure 2-10) and there were very high concentrations of tin detected by ICP-MS. This supports the idea that tin solder was being sacrificed. The amount of white solid in the water was highly variable, depending on the test water chemistry.
Effect of Brass Lead Content and Inhibitors. The tests aiming at understanding the role of lead content and inhibitors involved changing the water 3 times per week and used a 48 hour stagnation event before quantifying metals. Using this modified protocol, the lower pH, lower alkalinity chloraminated water was still more aggressive than the Section 9 water by a factor of about 2 times at the end of a 4 week exposure period (Figure 2-11). This difference was not as large as was observed for the same brass type during the earlier test rounds, due to the modified sampling protocol followed here compared to the standard section 9 protocol followed in the earlier rounds.
Addition of orthophosphate at a concentration of 3.0 mg/L to that water, was able to counter the negative effects of low alkalinity, low pH and chloramines, consistent with the observations of others (e.g. Lytle and Schock, 1996). Addition of zinc orthophosphate was also effective in reducing lead leaching (Figure 2-11). In fact, by the fourth week of the experiment, ortho-P reduced lead leaching by a factor of almost 70 times, and zinc ortho-P by a factor of 40 times for the 3% leaded brass (Figure 2-11).
Table of Contents
List of Figures
List of Tables
CHAPTER ONE: CURRENT UNDERSTANDING OF NON-LEADED BRASS
Description of the composition, lead content, and mechanical properties of new nonleaded
Mechanical strength properties
Potential health effects of non-leaded brass alloys
Potential of non-leaded brasses to release lead and copper and other metals (bismuth and
selenium) into drinking water
ASTM specifications for new non-leaded brasses
ASTM standard specifications
Summary of current state experience with non-leaded materials
Utility motivations and experience with non-lead brass components
Los Angeles Department of Water and Power
East Bay MUD
Bangor Water District
Greater Cincinnati Water Works
Seattle Public Schools
San Francisco Public Utility Commission (SFPUC)
Potential impacts to manufacturers
Supply and demand for bismuth and selenium
Component manufacturing issues
Health and environmental impacts
Impacts to plumbing industry
Standards and plumbing codes
Installation and inspection
Impacts to utility operations and maintenance procedures
Alternative materials for fittings and components
Other materials for fabrication
Appendix: Evaluation of the Leaching Potential of Non-Leaded Brasses
CHAPTER TWO: A CRITICAL EVALUATION OF THE NSF STANDARD 61/ SECTION 9 TEST WATER
Materials and methods
Evaluation of the leaching solution for C36000 brass
Evaluation of the leaching solution for brasses of different lead content and using
Evaluation on the leaching solution for C36000 brass
Effect of brass lead content and inhibitors
CHAPTER THREE: PARTICULATE LEAD IN DRINKING WATER
Materials and methods
Laboratory Simulation of Particulate Lead Occurrence in Drinking Water
Real World Results
Behavior of simulated particles in EPA sampling protocol and simulated gastric
Real world sampling results
Behavior of lead particles involved in childhood lead poisoning from Greenville NC
and Durham NC
Discussion and synthesis
CHAPTER FOUR: ROLE OF CHLORIDE TO SULFATE MASS RATIO IN LEAD LEACHING FROM SOLDERED JOINTS AND BRASS
Galvanic corrosion of solder/brass and copper connections
Chloride to sulfate mass ratio (CSMR)
Part 1: Bench scale experimental study
Materials and methods
Results and discussion
Role of galvanic connection and exposure time
Effect of CSMR and inhibitors on lead leaching from solder galvanically connected
Effect of CSMR and inhibitors on lead leaching from brass galvanically connected
Mechanistic insights via pH microelectrode measurements
Effect of disinfectant type on lead release
Effect of orthophosphate dose on lead release
Part 2: Real world implications/case studies
Durham, North Carolina
Greenville, North Carolina
GET THE COMPLETE PROJECT