EPA/600/A-93/119
CORROSION CONTROL PRINCIPLES AND STRATEGIES FOR REDUCING
LEAD AND COPPER IN DRINKING WATER SYSTEMS
Michael R. Schock
and
Darren A. Lytle
Drinking Water Research Division
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
LNTRODUCTION
The newly promulgated Lead and Copper Rule (CFR 141.82, May, 1991) will
force substantial changes in the way water utilities of all sizes treat their water and
control corrosion. The corrosion process is fascinating in that the more learned about
it the more it is realized that there are very few generalities. This paper emphasizes
concepts associated with sampling, and the trade-offs in water quality associated with
various methods of corrosion control. The point-of-entry water treatment industry in
general will have a substantial opportunity to grow in the area of corrosion control.
There are things that central water treatment can't accomplish. Having an awareness
of the complexities involved can put the water treatment industry in a good position to
develop applications in the future that can be quite helpful.
There are two parts to this paper. First, an overview of the plumbing and
corrosion issues for both building and domestic systems is presented. The United
States Environmental Protection Agency (USEPA) is conducting research on building
lead corrosion problems. The second part of this paper will discuss a joint project
between the Water Quality Association (WQA) and USEPA, involving the impact of
domestic-type water softeners on corrosivity. Naturally soft waters tend to be more
corrosive than naturally hard waters; however, there are many exceptions. The soft
water effect is probably more related to the pH of the water than the hardness. Most
naturally soft waters tend to have a relatively low pH, for example, in the 5.5-7.0
range.
Suggested References
Several good general texts and discussions covering drinking water corrosion
treatment considerations are available for more background on the subject. One such
book: Internal Corrosion of Water Distribution Systems', was jointly produced by the
American Water Works Association Research Foundation (AWVVARF) and the Engler-
Bunte Institute from Karlsrue, Germany, and was published in 1985. The document
will be revised this year by most of the original authors of the first edition, and the
target publication time is the last half of the year. Hopefully, this revision will include
much of the experience from Britain and the Scandinavian countries as well. The book
contains chapters dealing with all types of drinking water and potable water corrosion,
plus discussions of inhibitors, inhibitive mechanisms, and sampling and testing
1

-------
protocols. A general overview of corrosion is also presented. Many case studies and
references are included in the book.
Another book, Water Quality and Treatment. Fourth Edition2, published by
American Water Works Association (AWWA), deals predominantly with different types
of water treatment and corrosion control. A third book available is an older
publication particularly useful to the smaller water systems. It was originally
developed by USEPA and was entitled Corrosion Manual for Internal Corrosion of
Water and Distribution Systems.3 AWWA took the same document and published it
under the name of Corrosion Control for Operators." The repackaged version,
including color photographs, is still available from AWWA. Although parts of the
book are outdated, a lot of practical information about chemical feed systems and
alternatives associated with setting up small system corrosion control projects are
presented. A fourth, more specific book is Lead Control Strategies5 which came out in
1989 and was published by AWWARF.
PART I. CORROSION CONTROL ISSUES AND STRATEGIES
Secondary Impacts of Corrosion Control
One issue that complicates the development of new regulations is that
considerations of health effects and concern on the part of Congress are driving the
EPA to regulate contaminants at levels that water treatment scientists and engineers
sometimes doubt that central water treatment can control. When many simultaneous
regulations are imposed, chances of chemical incompatibilities increase. Prioritizing
regulatory choices then involves making a trade-off between one risk and another. For
instance, attempting to enhance lead corrosion control through pH adjustment may
adversely effect trihalomethane (THM) formation. THM's are potential carcinogens
that are also being simultaneously controlled by another regulation. The ultimate result
is that while trying to optimize treatment for one contaminant (lead levels), another
situation may be made worse. States and utilities will have to deal with this problem in
the future as more and more regulations are passed. Other water quality parameters
that could be adversely affected by lead and copper control could be things like
disinfectant effectiveness, formation of a variety of disinfection byproducts, and iron
and manganese control. Another example is that many of the optimum conditions for
sequestration of iron and manganese by treatment chemicals are precisely the conditions
that would mobilize lead and copper.
Another secondary impact of corrosion control is wastewater discharge
compliance for water containing phosphate and zinc. Many of the well-known
corrosion inhibitors used to control lead and copper corrosion are based on phosphate
compounds, and many of the inhibitors may also contain zinc. Several states are
setting restrictions on zinc discharge into wastewaters, zinc in sludges, and phosphate
2

-------
loading of treatment plants. Certainly there is a potential conflict between corrosion
control in drinking water and the fields of water and solid waste management.
Corrosion control may also have an impact on industrial processes. Each
industry has some particular new requirements for the water chemistry it has to
process. Adding different chemicals, or manipulating the water quality for corrosion
control, may force various industries to modify or completely change their current
treatment process. One example is the fairly critical pH control on water required by
textile industries. If a water utility is to suddenly implement corrosion control and
raise their pH from 6.5 to 8.0, the textile industry needs to be informed of that so they
have time to investigate potential impacts, and make allowances for modifying their
process, if necessary. Major local industries should always be informed of potential
municipal water treatment changes, so that they can make plans to adjust to the
changes.
Water conservation in itself may increase corrosion problems. In areas where
water usage is reduced by installation of low-flow and flow-restrictive devices, the
stagnation time of water in contact with the pipe will be increased. The amount of
water and the time required to flush that system of stagnated water will be increased.
The problem may be worse for buildings with many lead-soldered joints or lead service
lines. Flushing to clear unpreventable (through treatment) contamination will result in
wasting water where water conservation is attempted.
Assessing and Locating Corrosion Problems
There are several different issues that are involved in assessing and locating
domestic plumbing corrosion. One issue focuses on the characteristics and principles
of corrosion control, and what corrosion control strategies utilities are going to have
available to them. Another critical issue in this field is diagnostic sampling. Sampling
is not only a science, but an art that can be used for many different purposes. One of
the problems with sampling is that depending on how the sampling program is set up,
sample bias can show almost any level of lead or copper desired. While trying to solve
a metal contamination problem, both an accurate identification of what the problem was
in the first place, and an assessment of how good any control measures are operating
are critical. In one case study, a conscientious state agency thought there was a wide-
spread lead problem in some school systems. The problem turned out to be primarily a
sampling artifact. Once a very systematic and reproducible sampling program was
applied, it was found that the high lead concentrations were seen to be an idiosyncrasy
of when they chose to do the sampling, the size of the samples, and other related
factors. There is a wide variety of materials in a distribution system and for each
material there are water chemistry conditions that can be either corrosive or
noncorrosive. Typically, older systems have some combination of lead pipe, lead
goose necks, lead service lines, and copper pipe with plumbing having soldered joints.
Until several years ago, normal solder used in drinking water systems was 50/50 or
3

-------
60/40 Pb:Sn. Subsequently, in 1988, solder used in drinking water systems was
restricted to 0.2% lead. Most brasses both used in the past and in the present contain 2
to 8% lead and various water chemistry conditions can cause leaching of this lead.
Much of the application of brass materials are beyond the control of the utility (i.e.
water fixtures). However, because the utility's compliance with parts of the new Lead
and Copper Rule is going to be governed by tap water sampling, the brass plumbing
fixtures enter into the picture. Copper is currently the material of choice now for most
metallic plumbing systems. Other major materials currently in use are galvanized pipe,
uncoated iron mains, cement-mortar lined pipe, asbestos/cement pipe, and plastic pipe.
Many utilities have a major mix of all or many of the various pipe materials. These
utilities will have a problem on their hands to come up with a compromise treatment
that will help control contaminant levels from some pipe while simultaneously protect
and prolong the lifetime of the other plumbing materials.
Principles of Corrosion
Source Materials
The principles of corrosion are based on several phenomena. One phenomenon
is the oxidation of the metal which is the transformation of the metal from the base
state to a chemical form where it can be mobilized. Metal solubility governs how
much of the metal stays in the water and whether it forms a passivating film that will
retard future corrosion. Speciation of the metal is important in governing the mobility
of the metal, how it reacts to treatments and removal processes, and the relationship to
the source of the material. For example, lead pipe will have the potential of producing
or leaching lead indefinitely. However, lead leaching will follow a relatively uniform
response to water quality after the initial films are built up on the surface of the pipe
(normally five or ten years). Soldered joints on the other hand, are a inhomogeneous
material and involve competing dissolution mechanisms. There is some passivation on
the surface of the solder by chemical reaction with the water, but physical removal also
takes place. Eventually, much of the surficial lead actually dissolves away.
Sometimes lead contamination is associated with the overuse of flux during
soldering. The flux preferentially dissolves the lead, spreads it out over the pipe
surface, and separates it from the rest of the solder. Eventually, the lead contained in
the flux film will be partially removed. A similar reduction of leaching with time is
observed for brass. Lead is disseminated fairly evenly throughout the brass. The lead
is not really a part of the brass alloy, but is used for machinability purposes. After
being exposed to water over time, the lead on the brass's external surface in contact
with the water will be dissolved away. Lead is still present, but it is deep enough in
the brass and some diffusion barriers exist on the brass surface to prevent further lead
leaching. There is no universal guideline for the time frame these processes to work.
The processes may take five months or they may take ten years. The reactions depend
on the water quality, and there is no way of predicting it at this time. Predicting
4

-------
contamination depends strongly on the nature of the source of the lead or other
contaminant.
Electrochemical Oxidation
Base metal lead exposed to drinking water can react with oxygen to eventually
form lead and hydroxyl ions in the water. These reactions will cause an increase in pH
in water standing in lead pipe over an extended period of time (overnight to days). In a
chlorinated drinking water, such as disinfected municipal drinking water, hypochlorous
acid (HOCl°) and/or hypochlorite ion (OC1) are present, depending on pH. In
drinking water systems, these species will become the dominant oxidants (more than
oxygen) driving lead ion formation. Example reactions are shown in Figure 1 for
dissolved oxygen and hypochlorous acid. These reactions are the initial driving forces
that convert lead into a form it can be introduced into the water. Similar reactions can
occur with copper. The copper reactions are slightly more complicated because copper
can exist either as copper metal in the Cu(I) form (cuprous ion, Cu+) or in the Cu(II)
form (cupric ion, Cu2+).
Galvanic Corrosion
Another corrosion process is galvanic corrosion, which results from a coupling
of dissimilar metals. Commonly observed examples of a potential for "galvanic
corrosion" are metal screws in a dissimilar metal. The result is often a lot of rust,
which is created because the two metals are electrochemically incompatible. In
drinking water applications, galvanic corrosion is the principle behind the corrosion
protection qualities of galvanized pipe. In the case of galvanizing, zinc is sacrificed or
dissolves preferentially relative to the steel or iron. An empirical series often referred
to as a "galvanic series" can be developed. The galvanic series places metals in order
from most sacrificed to most protected metals. Figure 2 illustrates one galvanic series
based on the work of Larson in the early 1970's.6 In the distribution system, lead-tin
solders and brasses are typically coupled with copper. Based on the relative placement
of lead, lead-tin solder, brass, and copper in Figure 2, electrochemically the solder and
brass will dissolve preferentially to copper. This driving force helps mobilize the lead
from the brass and soldered joints.
Water Quality
Distinguishing the difference between alkalinity and inorganic carbon is an
important consideration in corrosion control. Alkalinity is normally measured in
drinking water to control treatment processes, and to measure the water's ability to
neutralize acids and bases. The amount of inorganic carbon is the fundamental variable
of interest from a corrosion and metal solubility point. The form of carbonate present
in the water will change with pH. At low pH, dissolved C02 or carbonic acid
[C02(aq)-t-H2C03°] may be present. At intermediate pH, the bicarbonate ion (HC03),
5

-------
is available and at high pH the carbonate ion (C032 ) predominates. In a pure water
containing carbonate as the only important weak acid species, total alkalinity (TALK) is
defined as:
TALK=2[C032] + [HC03] + [0H]-[H+]
where [ ] indicates concentrations in mol/L units, and TALK in eq/L. TALK can be
converted to the conventional mg CaC03/L by multiplying by 50044. There is a
theoretical linear relationship between alkalinity and the inorganic carbon content.
Figure 3 shows this general relationship for a water at 25°C, with an ionic strength of
0.005. From the figure for example, if the alkalinity is 25 mg/L at pH 6, then there is
18.6 mg/L inorganic carbon (as C), but at a pH of 10 it represents 3.4 mg/L inorganic
carbon. Many variables get confused because of the pH effect. Alkalinity and
dissolved inorganic carbon (DIC) are related, but they are not identical. Figure 4 is a
three dimensional projection of the effect of pH and alkalinity on lead solubility in
water at 25°C. The solubility of lead is quite complicated. If there is very little
alkalinity and essentially no inorganic carbon, the lead solubility is very high.
Introducing a little inorganic carbon (producing a little more alkalinity) into the system
will reduce the lead solubility. However, introducing too much inorganic carbon can
cause an increase in lead solubility. The complexity of lead solubility was not fully
understood until about 10 years ago. Preliminary work shows that copper behaves in a
similar fashion.
The effectiveness of corrosion inhibitors, such as orthophosphate-based
inhibitors to reduce metal levels, are dependent on several water quality parameters.
Some of the more obvious parameters include: pH, dissolved inorganic carbon
concentration, hardness, and temperature. Figure 5 illustrates the effect of
orthophosphate addition on lead solubility at pH 7.5. Waters having a high alkalinity
will require high levels of orthophosphate (POJ before a lead solubility decrease is
observed. Whereas, in a water with a lower alkalinity and at the same pH, a great
benefit from the first 0.5 or 1.0 mg/L of orthophosphate is achieved. The benefits
taper off with increasing orthophosphate addition. The important concept to understand
is that inhibitor dosages and general solubility controls, such as pH and alkalinity, are
interrelated. Some utilities purchase an inhibitor and apply it without examining the
entire water chemistry and are surprised when it doesn't work. An example is the
widely reported failure of a "zinc orthophosphate" chemical to control lead when
applied to the city of Boston's system in the late 1970's.7 The unadjusted pH was
around 6.0-6.5, and the inhibitor formulation was acidic. Using the dosage applied in
the 1970's, additional pH adjustments to around pH 7.5 would have been required for
the inhibitor to work correctly. Because pH adjustment was not done, the failure of the
treatment was assured from the start.
Figure 6 is a solubility diagram for zinc. The diagram shows the effect of
carbonate on solubility is not as pronounced for zinc as for lead. A similar trend exists
6

-------
where zinc solubility decreases as initial carbonate is added. However, higher levels of
carbonate are required in the pH range normally encompassed by drinking waters to
achieve a level where solubility is no longer reduced. In this case again, a little
carbonate is good, but more is not necessarily better.
Figure 7 illustrates, for pH 7.5, two interesting points to be made about
orthophosphate treatment chemicals containing zinc. First, there is a certain region at a
particular pH, especially with lower carbonate concentrations, in which a large impact
on zinc solubility can be made by a relatively small dosage of orthophosphate.
However, under those conditions it is very possible to end up with a turbid water
because the alkalinity and pH are too high to hold the zinc in solution. The phosphate
might remain dissolved in the water, but zinc can precipitate as basic zinc carbonate,
Zn5(C03)2(0H)6. In many cases zinc is not essential to the performance of the inhibitor
and in these cases it should not be used. Clearly, zinc solubility interrelates with lead
control treatment. These differences in zinc and orthophosphate solubility, because of
the interactions with other constituents in water, demonstrate that direct measurement
of orthophosphate is the only viable way to monitor inhibitor dosages within the
distribution system. Regardless of the ratio of zinc to orthophosphate in the treatment
chemical at the time of dosage, the relationship between P04 and Zn concentration is
not constant in the system itself. Similarly, when "blended" or polyphosphates are
used, both orthophosphate and acid hydrolyzable (or "total") phosphate must be
monitored.
Metal Speciaiion
The general simplified view of metal in water is that the metal is either
dissolved in the water, such as Pb2+, or it is as a base metal in the pipe. In actuality,
lead in a water with inorganic carbon present is available in many forms, depending
upon pH and temperature. Figure 8 shows the major lead forms present in water, at
different pH's. In the intermediate pH range, lead is mostly present as PbC03° which
is an uncharged, dissolved form (but it is not Pb2+, which is what many tend to
believe). As more carbonate is added to the water, the speciation is changed or driven
away from the free metal and hydroxide forms into forms where lead is complexed to
carbonate species instead. This illustrates that there are many aqueous chemical factors
governing the performance and contamination potential of plumbing materials. Lead
can exist as a negatively charged form at high pH's; it can exist neutrally in
intermediate pH's; and, it can exist as a positively charged form at lower pH's. There
is a lot of previously unexplained behavior that can be explained by the complexities of
the chemistry. These characteristics are especially pertinent to adsorptive and ion-
exchange metal removal systems.
7

-------
Characteristics of Lead Sources
Plumbing Workmanship
Poor plumbing workmanship can be a contributor of lead in drinking water.
Even well-made soldered joints by a professional plumber leave some exposed solder
and lead. Ironically, one study showed that more lead was exposed in a solder joint by
the experienced plumbers than by the inexperienced ones.5 The interesting
characteristic about lead and lead/tin solder is that the lead is basically a diluent and
allows solder to flow readily. In fact, the really doesn't bind with the copper. In
actuality, copper forms an intermetallic alloy with the tin. If too much flux is used, or
if an acid based flux is used, lead can preferentially run with the flux and spread out
over the inside surface. This demonstrates how workmanship influences the availability
of lead.
Faucet Characteristics
Lead, copper, and zinc from brass faucets present many problems caused by the
lack of uniform flow through the faucet, unlike the flow in a distribution main in the
middle of the city. A faucet is exposed to an intermittent flow pattern of pulsing
turbulence in a variety of areas. It is difficult to build a firm, homogenous passivating
film on the inside of faucet when there is vibration from the valve opening and closing.
Physical examination of cut-open faucets confirm the absence of a uniform coating.
Lead Pipe Characteristics
The interior surface of used lead pipes look different depending on the type of
water the pipe to which it was exposed. The pipe can be protected in different ways
depending on water quality and water treatment practices. For example, a lead
gooseneck from a midwest city using lime/soda softening to treat a hard water, was
coated with a unique film. The film consisted primarily of a mixture of calcium
carbonate, silica, and some basic lead carbonate [(Pb3(C03)2(0H)2]. In this case, the
city had implemented a good corrosion control treatment. Pipe from a system that has
been using a zinc orthophosphate compound for lead corrosion control will show a film
that is much thinner and is different in color in different places. This film was
composed of various layers with the outer surface composed primarily of lead
orthophosphate [Pbs(PO)3OH, hydroxpyromorphite]. There is not any zinc compound
in this coating. The nature of the protective film will differ from system to system and
is important to understand.
8

-------
Corrosion Treatment Strategies for Utilities
Selection Criteria
When utilities are trying to choose a lead control strategy, there are a variety of
considerations they must take into account. The considerations may be briefly
summarized as: 1) mix of materials, 2) initial water quality, 3) related requirements,
4) reliability in meeting compliance goals, and 5) cost.
The mix of materials involves understanding the materials that comprise the
distribution system. If one type of pipe isn't present in the distribution system, any
peculiar conditions that may pertain to the treatment of that material can be neglected.
The initial water quality is an important consideration because it relates both to
corrosion control and to what other treatment is needed to meet drinking water
regulations. There also could be an engineering concern in that some plant process
may affect the ability to alter treatment parameters or effects later on in the process
train.
Reliability is an issue in the case where a treatment theoretically works very
well, but, it is only possible to control the process or keep the hardware working 70%
of the time. In this case, the process would not be a worthwhile procedure for
compliance. For example, many small systems have trouble with lime feeders used to
keep a constant pH in order to maintain stabilized film formation. However, the feeder
can only be attended by an intermittent operator and there are many hardware problems
with clogging of the valves, etc. The treatment system selected in this case is not a
good way to perform pH adjustment and a different corrosion control strategy should
be considered. There are many operational requirements that will help determine what
a utility can and cannot do to address the corrosion issue.
Cost is also a factor when choosing a lead control strategy. In order to give
comparable performance in regulatory compliance, the least expensive choice is
preferred, if feasible. There are generally three broad approaches to control cost. One
of these approaches is materials replacement. If the material of concern is replaced and
thus not in the system, it won't create a problem. Point-of-use is another way to
control costs. The point-of-use devices would provide an additional opportunity to
assure protection where central water treatment could not meet the regulations
adequately. The third cost control approach and the starting point for corrosion
control, is the chemical water treatment strategies that can be used by utilities.
Control Strategies for Lead and Copper
There are two broad ways to accomplish lead and copper control by chemical
treatment. The first way is to basically form a superficial coating consisting of either
natural or induced diffusion barriers. A common example is pipe with a coating
9

-------
consisting of some organic material (usually mixed with iron or manganese). This
coating works very well with material such as asbestos-cement pipe. Lead and copper
pipe can frequently have organic coatings (typically mixed with naturally-occurring
silica) providing a barrier to corrosion.
Another way to develop a protective film is by forming calcium carbonate
coatings. The calcium carbonate coating has been widely discussed in water treatment
literature over the last 50 years. Although utilities and researchers have attempted to
form these coatings, few have really been able to achieve it. Another way to put down
a protective film would be to use sodium silicate or other types of silicate products.
The mechanism with silicates is probably a surface film barrier formation, or
converting an existing corrosion film to a less-permeable form, more so than reacting
with the metal on the pipe itself.
Calcium carbonate stabilization is commonly mentioned and requires some
focus. Many states and utilities strive to maintain a positive "Langelier Index" to
prevent corrosion.1'2 However, there are many other complexities to consider. One
consideration is that the bulk solution pH and alkalinity of the water are not necessarily
exactly the same as at the pipe surface. The pH difference is seen when patches of
calcium carbonate are deposited on the surface resulting in the pipe being unevenly
coated. One reason for the uneven coating is that localized corrosion cells form,
causing conditions at these sites to favor precipitation. Because the inside of the pipe is
not covered by an even coating, this process would not be considered a good corrosion
control measure. This uneven coating is also influenced by the pH, dissolved inorganic
carbon, and alkalinity relationships previously discussed. Combine these parameters
with the hardness of the water and the real issue becomes the available mass of CaC03
that can be precipitated. If the Langelier Index is positive, but there isn't enough
calcium and carbonate in the water and the pH isn't favorable, the potential to form a
coating does not exist. The Langelier Index dependency is a cause of problems in
many water treatment plants. Water utilities try to control CaC03 deposition
throughout the system, but find that the deposits in the mains exist for only a short
distance from the plant. Plant filter clogging is also a problem. The further reaches of
the distribution system stay undersaturated with CaC03, resulting in no benefit from
this corrosion control strategy.
There are two main computational approaches that can be employed when using
calcium carbonate precipitation to control corrosion. The most common method is to
calculate the Langlier Index. A more obscure but more accurate method is "calcium
carbonate precipitation potential" (CCPP). CCPP is very amenable to calculation on
personal computers and thus it should be usable by more workers in the future. The
third way is an empirical test, like a "marble test". The marble test has gone out of
fashion, but in many respects is still the most reliable method.
10

-------
Many factors can confound the Langelier Index, or calcium carbonate
precipitation potential. For example, anything that will inhibit precipitation and growth
of calcium carbonate such as magnesium, zinc, or orthophosphate can be confounding
compounds. These compounds will stick to the surface and distort the CaC03 crystal
growth so that it can't continue to grow and nucleate correctly. The action of
polyphosphates is a good example of this effect. The polyphosphates are generally
added to the water to protect filters from post-deposition at the final phase of water
softening in central water treatment plants. Polyphosphates can form calcium
complexes, and can poison carbonate crystal growth by adsorbing on the surface of
growing crystal nuclei. This action invalidates calculations of either the Langelier
Index or the calcium carbonate precipitation potential by the conventional equations. In
other words, if polyphosphates are being added to control post deposition, calculating
the Langlier Index or CCPP would be totally futile. In this case, the only way to get
an assessment of the conditions is to use an empirical procedure like the marble test.
Figure 9 is a comparison of the Langelier Index and the CCPP for a hard and a
soft water. The Langlier Index is a thermodynamic driving force indicating if a water
has a potential to form calcium carbonate based upon fundamental chemistry principles.
The two waters have widely differing pH, hardness, alkalinity, and total dissolved
solids (TDS) values, but have the same Langelier Index. However, one of them has a
CCPP of 0.4 mg/L as CaC03 and the other 15 mg/L as CaC03. If a utility was trying
to form a calcium carbonate film throughout their distribution system, the case where
there is more mass of calcium carbonate (in this example, the hard water), would
provide a much greater opportunity to deposit a film than the softened water. This is a
good example of why the Langelier Index alone does not give an accurate picture of the
potential to form a protective CaC03 film.
Another issue that complicates the whole picture is that cast iron mains and
galvanized pipe frequently lack calcium carbonate films. The deposits that exist are
often a combination of calcium and iron carbonate. Research has shown that there
really is not a good correlation between any of the corrosion indices, the metal levels in
the water, and the weight loss of the pipe materials in the absence of actually putting
down a film.
Another type of barrier film is formed through silicate addition. Silicate
addition is not understood very well, as it has not been studied systematically to any
great extent. The actual silicate dosage probably depends upon the pH of the water and
the hardness of the water. There is probably an interaction between the silica and the
calcium in the water; however, it is presently hard to quantify and predict. The
reactions that do occur would be slow, which is quite important. If tests are being
conducted in a pilot plant or pipe loop setting and there is a deadline to have some
corrosion control treatment be in place in 18 months, only about six months might be
reserved for testing and evaluating treatments. Silicate addition may be a good
11

-------
technique and 5 years down the road it may achieve excellent results, but benefits may
not show up very well in a short time-frame.
Another silicate question is that they may require a preexisting film to work
properly. Many researchers believe that the silicates bridge the gap between corrosion
products already on the pipe. The existing films allow some diffusion of the oxidants
into the pipes and migration of metals away from the pipe. Silicates might react with
the existing film and "seal" it like grouting tile. If silicate chemicals are tested on
brand-new plumbing, they may not work. If silicates require a preexisting film on the
copper pipe, they will not work nearly as well in new homes as in older homes that
have some scale built up in them.
The other general chemical treatment approach is to create a passivating film.
A "passivating" film involves reacting the metal that is coming off of the pipe with
some constituent in the water to form a film in place that helps to immobilize that
metal. Undissolved particulate iron and manganese frequently cause a tremendous
problem for consumers by forming "red" and "black" water, clogging filters, and
creating other obnoxious aesthetic effects. Similarly, an insoluble phosphate or silicate
compound can be formed by chemical treatment. However, if the compound will not
adhere to the pipe surface and remain there, merely dispersing the compound into the
water will not solve the problem. If a water sample is taken and measured for lead or
copper, the colloidal forms of the metals are usually included in the analysis. What
must really be done is to immobilize the metal. Solubility is not the only issue.
Getting the film to stick uniformly to the pipe is also important. Generally, there are
four specific chemical treatment approaches a utility can follow to achieve lead and
copper control through the formation of a passivating film. The four approaches are:
1) pH adjustment; 2) pH/alkalinity/DIC adjustment; 3) orthophosphate addition, and 5)
"blended" phosphate addition.
The issues involved in pH and DIC/ alkalinity adjustment are based on the
sensitivity of the metal solubility. Figures 6-8 show that in certain ranges some
solubility control can be achieved by controlling these parameters. Often, solubility
can be reduced quite a bit by forming a passivating film. Solubility control is not
equally sensitive to pH and alkalinity, or other inhibitive anions. Optimizing treatment
involves considering these relative sensitivities.
Another factor to consider when adjusting pH and alkalinity is the buffer
intensity of the water. The buffer intensity of the water is basically the resistance of
the water to pH change. It is desirable to have a good buffering intensity at the surface
of the pipe where it meets the water, as that will keep the conditions at the pipe surface
as close to the bulk water conditions as possible. The uniform pH will prevent the
formation of localized spots with really high pH that would tend to cause types of
pitting corrosion. A particularly bad pH from this standpoint is about 8.3, which is the
12

-------
buffer intensity minimum for the carbonate system.12 The buffer intensity is effected
both by pH and DIC.
To control lead or copper by pH control, a pH of 8 to 9 and above is often
necessary. However, calcium carbonate precipitation may limit the ability to get the
pH high enough. One problem inherent here is that even though calcium carbonate
precipitates, a protective film may not be formed.
Orthophosphate addition requires several considerations and does interrelate to
pH and inorganic carbon. The orthophosphate dosage can also be limited by the
calcium concentration. Octacalcium phosphate or other insoluble orthophosphate
compounds may be formed, causing either turbid water or removing desirable
phosphate from being available to form passivating films inside the pipe metal.
Zinc compounds are limited by both pH and DIC, so the ratio of zinc to
phosphate in the chemical becomes important in many cases in preventing the
previously mentioned turbid water or phosphate depletion. There is some data in the
British literature suggesting that zinc compounds may be helpful for controlling brass
corrosion, but this data is not definitive in all cases. Little is really known about
beneficial effects of zinc in film formation with orthophosphate.
Blended phosphates have all the same considerations of zinc orthophosphate
because orthophosphate is critical in achieving the control of the metal pipe corrosion.
One of the reasons blended phosphates were developed was to attempt to enable a
utility to address more than one problem at the same time. For example, blended
phosphates may be used to inhibit tuberculation in iron mains or to inhibit red water
caused by iron in the source water, while at the same time attempting to provide some
control of lead, copper, zinc or some other metals in the system. One important
parameter that must be considered when using blended phosphates is the polyphosphate
to orthophosphate ratio. The amount of orthophosphate available for forming the
passivating films will govern the metal solubility. The speciation of the polyphosphate
component is also important, because the different kinds of polyphosphates (eg.
tripolyphosphate, pyrophosphate, hexametaphosphate, etc.) have different affinities for
metals. Some polyphosphate species favor complexing calcium very strongly and iron
only weakly, while others complex iron and manganese more strongly than calcium.
There are many intricacies in polyphosphate chemistry, and little has been published in
the objective scientific literature. Depending on what the background water quality is
and to what extent the lead or copper must be controlled, the particular polyphosphate
can be important. A good corrosion control situation would be when a polyphosphate
blend is found that adequately sequesters or stabilizes iron and manganese, but does not
have a residual complexing capacity to attack copper and lead. An orthophosphate
could then be used to control the lead and possibly the copper corrosion. If a utility
uses a type of polyphosphate that complexes trace metals very strongly and it isn't
13

-------
otherwise occupied with hardness ions or iron and manganese, the polyphosphate will
increase the lead and copper corrosion. There are many combinations that need to be
tested and considered, and no single product or blend is even remotely universally
applicable.
Diagnostic Sampling
An important issue that is often highly overlooked is sampling. There is a high
level of normal variability in sampling.910 Every time a sample is taken, the sample
is going to be somewhat different from those taken before and after it. Therefore,
there is no such thing as "the lead level" of a house. A question exists as to how
different are each of the water samples. One possible scenario exists where water
probably flowed through a sequence of materials. Prior to the distribution system, the
water contains no lead. The distribution system may be comprised of iron mains or
cement-mortar lined mains, for example. Possibly, there could be joints in the mains
that were filled in with a lead-containing compound which used to be a common
practice. Later on, the water then arrives at the service line containing lead pipe. The
household plumbing could be galvanized steel, copper, or even all lead. In some cases
there is no lead at all in contact with the water until it reaches the faucet. The isolation
of these potential sources through sampling is important, because the best treatment for
corrosion may be dependent on the source of metal. For example, brass corrosion
control may not be the same best treatment as for controlling copper pipe corrosion, or
lead pipe corrosion. It is necessary, and possible to a great degree, to be able to
precisely sample so that the source of the problem can be identified.
The determination of "baseline" metal levels is another task that is very
important. Many investigators take a couple of samples, decide they have a problem,
and immediately implement some kind of control program. This strategy does two
things, one of which is that they can never then really be sure there was a general
problem. Even if the "problem" goes away, it's not clear how much of the problem
has been addressed, since it wasn't correctly quantified in the first place. Because of
the variability in normal sampling, it is difficult (if not statistically impossible) to tell
whether or not there is any difference between implemented treatment practices,
whether evaluated by pipe rig tests, or directly in field studies. It is very important to
determine a baseline metal level because a utility embarking on a corrosion control
program is going to be regulated based on its effectiveness as reflected in tap water
sampling.
Variability obscures and complicates detection of treatment trends. If a pipe
loop or pilot plant system is used to evaluate five or six different kinds of treatments
side by side, one or two of them will not always consistently be lower than the others.
There is a lot of variability in metal level concentrations created by corrosion and
14

-------
corrosion controls. Sometimes curves connecting the points for a treatment will cross
over due to natural variability, making it very hard to tell one treatment from the other.
The new regulations will require "optimization" of corrosion control for lead
and copper. It may be very difficult to determine a one true "optimization" for any
system. To illustrate this point, Figure 10 shows a test run at the Illinois Water
Survey.5 In this test run, there were three identical lead loops (3 equal lengths cut
from a 100 foot section of pipe), from the same manufacturer, running simultaneously
in the same rig. At the beginning of the run, the lead levels were fairly consistent.
However, as time went on, discrepancies (30 to 40 jtxg/L) arose among them. If a
comparison of relative treatment effectiveness is the goal of rig studies, simultaneous
replication is absolutely necessary regardless of the metal involved. If only one pipe is
present for each treatment, one treatment may appear to be better than another
treatment, but this relationship may be an artifact of the behavior of that particular
pipe. Enough replication and sampling over time must be done to ensure a basis for
taking into account this type of variability.
Figure 11 shows flushed sample data taken from pipe loop experiments
conducted at the USEPA' under an identical sampling protocol. Some of scatter can be
attributed to minute invisible colloidal particles, and some due to dissolved lead. The
essential point is that there is a lot of scatter under very reproducible and controlled
experimental conditions. This scatter would be more out of control when sampling is
done at a consumer's house. Figure 12 displays pipe loop data for lead samples after
different standing times.1 After the first few hours, under the same sampling
conditions and with the same length of standing time, there is still 50 percent or more
variability in the lead concentrations. The scatter decreases as the standing times used
approach what is needed for chemical equilibrium. Once again, under reproducible
experimental conditions there is still considerable variability in the data. A final
example is shown in Figure 13, where a set of standing samples are taken from three
lead-soldered copper loops.5 Three copper loops with lead soldered joints (as
identically made as possible), have a relative standing water lead level variability by at
least a factor of 2. The lead levels in this case are near the quantification and detection
limits of the measuring instrument. Here, there is much analytical variability factored
in with actual physical variability of the operation of the system. It is difficult to make
any important interpolations with pipe loops having soldered joints, because the whole
workmanship issue is built into it. From statistical considerations well beyond the
scope of this paper, it follows that the more accurate small differences that need to be
detected, the number of samples to be taken becomes quite large.5 Practically
speaking, differences in real systems can only be determined imprecisely. Very
expensive treatment and design decisions may be made based on this kind of test data,
so careful effort and consideration to understand the limits of data interpretation must
be made.
15

-------
Isolation of Sources
Some logic and common sense will serve investigators well in isolating the
source of contamination. A "first-draw" water sample can also be thought of in terms
of the feet of pipe represented. For example, a liter of water represents approximately
25 feet of lead pipe, with 0.5 inch inside diameter (ID). In other words, one foot of
pipe will hold 40 mL of water. Galvanized and copper pipe are slightly different than
lead pipe because of the definition of ID. For "half inch" ID copper pipe, the internal
diameter is not exactly one-half inch, but the volume of water available can easily be
calculated from the correct ID.
Knowing the length of pipe or fixture represented by a water sample becomes
important when trying to determine if a problem is coming from the brass faucet, the
soldered joints, or some inner section of the building or distribution system. For
example, there may be a feed line and a tee coming off into a building wing with a
concentration of soldered joints. This area may be 15 or 16 feet prior to the sampling
point. If a 125 mL sample is taken, it represents essentially the faucet. With a 250
mL first draw sample, the faucet and about 3 feet of pipe (0.5-inch ID) are
represented. A 1 L sample represents the faucet and 20 feet or more of the pipe. In a
consumer's house, a first draw 1 L sample may not reach the service line. To
illustrate this point, Figure 14 shows data from a field study that tried to assess the
impact of some water quality parameters and water quality adjustments on tap water
lead levels from houses that had lead service lines. Schematically, this shows that 1 L
of water was in contact with some of the internal plumbing, plus the faucet. To obtain
samples in contact with the service lines, water was wasted so that samples would
"intercept" water coming from the desired source, the lead service lines. The trouble
with remote sampling is that there is turbulent flow and mixing in the system. If there
are not high lead levels, or if the water isn't very corrosive, the ability to intercept and
detect that lead becomes very difficult. From a consumer's standpoint, that is very
good because it means that in a fairly noncorrosive water, a lead service line will not
cause any adverse effects at the tap. When examining and comparing studies of
corrosion control and the sampling schemes are not consistent, the data is not
necessarily equivalent in terms of conditions and sources.
Britton and Richards did an interesting study in Scotland" where they were
taking samples from the same tap and varied the flow rate. They found that when the
flow rate was very slow, there is relatively more contact time with the brass material
and was more lead. As the flow was increased, there was relatively less lead pickup.
However, when the flow went beyond a certain point, turbulence resulted and erosion
from the flow was created and higher lead levels appeared again. This is another
factor contingent upon the variability and reproducability of sampling and the meaning
of measured lead levels. It has been shown that mixing during flow reduces the
concentration at the end of the water parcel in contact with the pipe. More information
is given in reference 5.
16

-------
PART 2: THE WQA/USEPA JOINT STUDY
A WQA/USEPA joint project was an outgrowth of a new program under the
Federal Technology Transfer Act. The program allows the government to accept
money from "private industry" under terms that are agreed upon by mutual consent.
All components of the project including how the data is disseminated and a work plan
must be agreed upon by both parties. The focus of the study is to obtain some initial
information on whether or not there are impacts from water softeners on corrosivity.
The Water Quality Association donated $50,000 to USEPA to help begin this
investigation. There has been much speculation, but little systematically generated data
about the effects of water softeners. The reason the softening issue is complex and
somewhat controversial, is that whether or not softening will have any effect at all
depends very much on the water quality.
Possible Impacts of Softening
Why might softeners impact corrosion, and why might they not? There are
many issues to be considered here. One possibility is that softening could reduce the
water scaling tendency. If the pipes of a system are protected by surface films (for
example mixed calcium- and magnesium- based films), removal of the calcium and
magnesium from the water by softening will have an adverse impact on that film.
Some utilities that have a high pH and very low hardness, typically produce films
composed of basic copper carbonate, copper oxide, zinc hydroxy carbonate, basic lead
carbonate, etc. Here, the calcium and magnesium don't play an integral role in scale
formation. Therefore, the softening really shouldn't have an impact on corrosivity.
Another question is the effect softening has on chlorine, chloramines, and
dissolved oxygen. These are oxidants that greatly effect the ability to mobilize the
metals, and especially corrode copper. The fate of these species in water passed
through domestic water softeners has not been documented. One of the objectives of
this project will be to measure small changes in concentration of potentially important
chemical parameters.
o
It is not known whether or not there is an interaction between the softeners and
corrosion inhibitors. If a utility doses a zinc orthophosphate and there is a case where
the zinc does play a role in the formation of the protective film, removal of the zinc by
the softener will change the nature of the film formed on the pipe after the water
softener.
The effects of changes in scaling potential on corrosivity may depend upon the
temperature. Larson did some work where he pointed out that, calcium carbonate
becomes less soluble as temperature goes up.6 Scaling is often observed in hot water
systems. However, this increase in scaling only takes place in certain ranges of
hardness and alkalinity. A water that is not very hard and has a low alkalinity may
17

-------
become less scale-forming at higher temperatures.2 What happens is that as the water
is heated, the intrinsic pH of the water goes down as the temperature goes up. In a
closed system, like a household plumbing system, the water is still under pressure in
the pipe, and the pH will go down by itself just because of the carbonate chemistry
involved.2'5 In harder waters, and in waters with more alkalinity, that effect does not
occur very much because the water is buffered.
Formation of Mixed Solids
As mentioned previously, metals can participate in the scale formation. In hard
waters, calcium exists in many forms other than in just Ca2+. Often 10 to 20 percent
of the calcium can be in the form of either bicarbonate or carbonate ion complexes
(called "ion-pairs). If the water quality is drastically changed by removing calcium and
magnesium by substituting sodium, the carbonate equilibrium could become slightly
perturbed because the sodium and potassium forms of these complexes are much
weaker. That is real subtle and probably doesn't have much of an effect, but it is
something that should be considered.
Other Indirect Effects
Another indirect effect of softening is lack of binding of ligands. Ligands, such
as polyphosphate, tie up and bind metals. If for example, assume iron or manganese is
being controlled in the source water, or post deposition of calcium is being controlled
with a polyphosphate, the polyphosphate's complexing ability is filled. However, if a
home water softener is used, the metal is stripped from the polyphosphate ligand, and
the polyphosphate passes through. In this case a greater potential to dissolve the lead
and the copper may be created because now an aggressive chemical is present that
wasn't aggressive in its previous state because it was chemically associated with other
metals. There have been no studies that show the exact chemistry of these indirect
effects of softeners along with the other effects mentioned earlier, such as the ion
pairing difference and the fate of chlorine and chloramines. These studies should be
done.
Experimental and System Design
A set of test loops much like the ones used and produced for the AWWA's Lead
Control Strategies manual5 has been designed for the joint project. These loops are
also very similar to the ones used for the corrosion inhibitor studies currently being
conducted at the USEPA research facility, Cincinnati, OH. The loops basically consist
of five materials including lead/tin solder, copper pipe, copper tubing (unjoined copper
tubing), galvanized pipe, and lead pipe. Lead pipe was included because the Cincinnati
Water Works is interested in seeing softening effects on the worse case scenario. Two
brass faucets were also included to see if there is impact on the lead and copper
contributions to the water from the faucets.
18

-------
The basic design of the system is a uniform pipe diameter of 1/2" ID and
uniform flow rate of one gallon per minute, which is comparable to domestic systems.
Each of the loops will be 50 ft in length. There will be an intermittent flow and an 8
hour stagnation time during the day to simulate domestic use. One hundred gallons of
water per day will be used for each loop, which is less than what most typical houses
probably use. But, because water quantity must be conserved, a balance must be
struck. All of the materials to support the system will be lead and copper free plastics.
The system will be set up at Cincinnati's Bolton Water Treatment Plant, which utilizes
a ground water source. The Bolton Plant treats the ground water by lime softening,
with a low level of polyphosphate addition (to protect the filters), chlorination, and
fluoridation before distribution.
The Bolton plant offers a location in which the raw, untreated water, and the
finished, treated water can be used. The raw and treated water quality is shown in
Figure 15. In Ohio and most of the Midwest, the kind of water Cincinnati normally
distributes is considered hard enough by many consumers to justify using water
softeners. The water coming off the main line will be split, one side feeding an ion
exchange softener and then the pipe loops. The other side feeds a set of identical loops
directly. There will be duplicate loops, two of each, used because triplicate loops
(although desirable) would require too many samples and too much water. Operating
in duplicate will improve statistics on differentiating trends over single loop studies.
Sampling ports will be placed immediately before and after the softener to evaluate the
chemistry effects of the softener. Each loop will have a flow control and sampling port
at the end of it so standing and running samples can be taken from each loop. Figure
16 represents a schematic of the pipe loop system.
After six months, the input water lines are going to be crossed. That is, the
loops in contact with softened water will be switched to unsoftened water, and vica
versa. Total chemical analysis will be done on the samples of the water immediately
before and after the softeners. In the loops themselves, metal levels will be monitored.
In summary, several short term results will come from this study. Statistical
differences between softened and unsoftened waters will be determined, taking into
account the appropriate comparison statistics will be applied. An understanding of the
chemical characteristics in metal leaching, and the chemistry of the corrosion process
will be gained. If these results show some intriguing information and trends, it would
be valuable to eventually evaluate systems operating in the field. It may be possible to
locate homes that have had softeners, and have comparable temperatures to those of the
lab pilot study, and collect data from them. The field data could then be compared to
the data acquired from the lab and pilot study to see if there is any correspondence.
19

-------
REFERENCES
1.	American Water Works Association Research Foundation and DVGW
Forschungsstelle. Internal Corrosion of Water Distribution Systems. Denver,
AWWA Research Foundation, 1985.
2.	American Water Works Association. Water Quality and Treatment. 4th edition.
New York, McGraw Hill, 1990.
3.	USEPA. Corrosion Manual for Internal Corrosion of Water and Distribution
Systems. EPA 570/9-84-001, Washington, D.C., April. 1984.
4.	American Water Works Association. Corrosion Control for Operators.
Denver, AWWA, 1986.
5.	American Water Works Association Research Foundation. Lead Control
Strategies. Denver, AWWA Research Foundation, 1990.
6.	Larson, T.E. Corrosion by Domestic Waters. Bulletin 59, Illinois State Water
Survey, Urbana, IL (1975).
7.	Karalekas, P. C., Jr. "Control of Lead, Copper and Iron Pipe Corrosion in
Boston", Jour AWWA. 75:2:92 (1983).
8.	APHA-AWWA-WPCF Standard Methods for the Examination of Water and
Wastewater. 17th Ed. (1989).
9.	Schock, M. R., Levin, R., & Cox, D. "The Significance of Sources of
Temperal Variability of Lead in Corrosion Evaluation and Monitoring Program
Design", Proc. AWWA Water Quality Technology and Conference, St. Louis,
MO. (1988).
10.	Schock, M. R. "Causes of Temporal Variability in Lead in Domestic
Plumbing Systems, Environmental Monit. & Assessment.
11.	Britton, A. & Richards, W. N. "Factors Influencing Plumbosolvency in
Scotland". Jour. Inst. Water Engr. and Scientists. 35:4:349 (1981).
20

-------
2Pb + 02 + 2H20 = 2Pb2+ + 40H"
Pb + H0C10 + H+ = Pb2+ + CI" + H20
Figure 1. Oxidation reactions for lead in water.
Abbreviated Empirical
Galvanic Series
Corroded End (Anodic^
Zinc
Steel or Iron
Cast Iron
Chromium-iron
Lead/Tin Solders
Lead
Tin
Brass
Copper
Silver Solder
Silver
After Larson (1975), ISWS Bulletin 59
Figure 2. Galvanic series for common plumbing materials
used in the drinking water distribution system.
21

-------
Alkalinity/DIC Relationship
-25°C, 1=0.005
o 120:
100-
I Vo *J20
^ - 18.6
3.4
mg C/L Inorganic Carbon
Figure 3. Alkalinity /DIC relationship for various values at 25 C.
Lead Solubility, 1=0.01, 25 deg. C
log mg PWL
1£
igure 4. Lead solubility as a function of alkalinity.and pH at 25 C

-------
Lead Solubility at Different
Alkalinities, 25°C, 1=0.01
-I.OOO
pH = 7.5
X
mg CaCOj/L
Alkalinity
200
0.1 oo
n
o_
O)
E
100
O.OIO
30
O.OOI
o
2
1
3
4
5
mg PO4/L
Figure 5. Lead solubility as a function of phosphate concentration at
different alkalinities and pH 7.5.
Zinc Solubility
I = 0.01, 25°C
100
10 11
Inorganic
Carbon
' '1 mg/L
— 5 mg/L
~'20 mg/L
" ' 80 mg/L
Figure 6. Zinc solubility as a function of pH at 25 C.
23

-------
Zn Solubility, pH = 7.5, 1 = 0.01, 25°C
10.00
2. i-;
w 1.00
0.30
0.20
0.10

L-J-J
10.00
5.00
3.00
2.00
1.00
0.50
0.30
0.20
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
mg PO4/L Orthophosphate
Figure 7. Zinc solubility as a function of
orthophosphate concentration at 25°C.
0.10
1 mg C/L
2.5 mg C/L
5 mg C/L
10 mg C/L
20 mg C/L
40 mg C/L
60 mg C/L
80 mg C/L
.2
•3
-6
€
i*
r

-7
A
-9
•10
e
12
6
7
B
9
10
11
PM
Figure 8. Lead speciation as a function of pH at 25°C.
24

-------
LI versus CCPP
Same Langelier Index
Concentration	Soft Water	Hard Water
mi!L	High pH	Low pH
Temp (°C)
15
15
Alkalinity (mg/L
as CaC03)
25
350
Calcium (mg/L)
17
130
TDS (mg/L)
75
750
pH (units)
8.90
7.03
LI (units)
0.10-
0.10
CCPP (mg/L)
0.40
15
Figure 9. Comparison of the Langilier Index versus
the Calcium Carbonate Precipitation Potential for two waters.
Lead Pipe Loop
400-
350-
325-
O Loop 1
250-
Loop 2
Loop 3
225-
5 10 15 20 25
Elapsed Time (Days)
igure 10. Lead pipe leaching data from a study conducted at the
Illinois State Water Survey.^
25

-------
0.06
MCL
0.05
D	
-j
1
0.03
0.02
m
00
a PHASE 1
A PHASE 2
« PHASE 3
0.01
9.0
82 . 84
86
88
8.0
9.2
9 4
7.6
96
7.B
Figure 11. Lead concentrations of flushed samples taken from
pipe loop experiments conducted at USEPA.
0.20
T	"T

0 ~


0.15








Jnh D



/ ® ~



/ ~§


o>
E 0.10
-/ "G...


xT
a.
0
(
o
o


/ — ° °
0.05 -
o
~
Phase 2 (theoretical)
Phase 1 (theoretical)
Phase 2 (data)
Phase 1 (data)
_L
_L
_L
-L
X
-L
_L
5 10 15 20 25 30 35 40 45 50
HOURS
Figure 12. Lead concentrations of samples taken from
pipe loop experiments at various standing times.
26

-------
-fl.O
7.0
•.0 -
8.0 -
t
C
£ 4.0
I „
o
o
2.0
1.0
0.0
•1.0
• • Loop 1
6 Loop 2
O ...
A - — - A Loop 3
"Q
• °.
6 Q
A, °o
/ .~
n
i \
9, \
• • v . °.
o r. v - ° . •• °
V 1
/ \

H	1	I	1	1	(	1	1	1		1	1	1	H
H	1	1	1	1	1	1	1	»-
10
25
30
35
. 15	20
Tatting Intarvil, diyt
Figure 13. Lead concentrations of standing samples taken from three identical
as can be made lead solder/copper tube sections.
PLUMBING REPRESENTED BY SAMPLES
1-litre first draw
-wz
FAUCET
t /r
SOLDERED
JOINTS
WASTE VOLUME
HOUSEHOLD
PLUMBING
0 METER

250 mL samples
12 3 4
t f T t
—r^ss:
pH/TlC vials
1-litre
MAIN
SERVICE LINE
Figure 14. Household plumbing represented by samples.
27

-------
Water Quality
Bolton Plant
Constituent
mg/L
Raw
Treated
PH
7.4
9.2
Calcium
90
25
Magnesium
25
23
Sodium
19
19
Alkalinity
254
82
Sulfate
58
60
Chloride
35
36
Nitrate
3
4
Total-P04
<0.3
>0.5
Silica
9
10
Figure 15. Raw and treated ground water quality from
Cincinnati's Bolton Treatment Plant.
Duplicate Loops
Raw Water
Control Loops
Lead Cu Cu Galv.
Pipe Tube Pipe Pipe
/An/An /An /An /An /An /An /An
Raw
Water
A




v
\j
V
V
Brass
Faucets
DRAIN
¦>
Softened Raw Water Test Loops
Background /7^/A\/A\/A\/A\/A\/AN/AN
Sampling
Ports
-------