EPA/600/A-95/036
A Systematic Study on tbe Control of Lead in a New Building
Darren A Lytle, Environmental Engineer,
Michael R Schock, Research Chemist,
and Thomas J. Sorg, Branch Chief
U.S. Environmental Protection Agency
RREL, DWRD, l&PCB
Cincinnati, Ohio 45268
Background
In early 1991, a federal agency requested the assistance of the U.S.
Environmental Protection Agency's (EPA's) Drinking Water Research Division
(DWRD) to solve a problem of high lead in the drinking water of a research
facility built to house approximately 1000 employees. Drinking water samples
taken from various sites in their newly constructed facility contained extremely
high levels of lead. The facility, whose construction was completed in 1986, had
been unoccupied, with the exception of a few building maintenance employees,
due to a series of structural defects and repairs.
About six months before the building was scheduled to be opened, the
water in the building was tested for a range of regulated inorganic contaminants.
Samples where taken from a variety of locations including water coolers,
bathroom faucets, laboratory faucets, and sample bottle washer units. Results
indicated that the water exceeded the lead maximum contaminant level (MCL)
which at that time was 0.05 mg/L. The lead test results prompted a number of
follow-up studies to establish the degree of lead contamination in the building's
drinking water. Between December, 1990 and February, 1991, two more
sampling studies were conducted. A wide variety of locations throughout the
building were sampled and the samples were split between two laboratories. Test
results indicated extremely high lead levels in many locations (as high as 2.4
mg/L). However inconsistencies among lead levels seen in samples taken from
the same site on multiple sampling periods and split sample results from multiple
laboratories raised questions regarding the testing and sampling protocols.
Because of high lead levels and the inconsistency of the sampling results,
the EPA's Drinking Water Research Division was requested to visit the facility
and re-sample the facility. DWRD sampled 19 of the previously tested locations
using a slightly different sampling protocol than was used before. The new
protocol consisted of taking 250 mL samples following a 10 minute flushing
period and after a 24 hour standing time. This procedure was based on EPA's
protocol for testing lead in drinking water in schools and buildings." The
protocol sets 20 #ig Pb/L in the 250 mL standing sample as a guide for remedial
action. Further, visual inspection of the facility suggested that the sources of high
lead concentrations were brass faucets and valves, and lead-based solder.
345
-------
Therefore, sequential samples were also taken al several locations. Sequential
sampling involved taking a 250 mL sample from the site after a 10 minute flush
and a sequence of two 60 mL samples and ten 125 samples following a 24 hour
standing time. The samples were analyzed for lead, copper, and zinc.
Results from the DWRD sampling confirmed high lead levels in the
drinking water at many locations in the building. Once again the levels were
inconsistent from past results at the same site. Results from the sequential
sampling typically showed high lead levels in the first two samples falling off to
levels observed in flushed samples then peaking again at about the fifth or sixth
sample in the sequence (example shown in Figure I). Profiles for zinc, a major
component of brass, typically showed the highest levels in the first and second
samples, dropping rapidly to background levels. These test results suggested that
the sources of lead were brass fixtures and lead-tin solder.
Visual inspection of the faucet, faucet connection, and plumbing material
beneath sinks leading up to the faucet in each location sampled showed that the
basic plumbing was generally the same for laboratories in the building. Typical
cold water plumbing in the laboratories (Figure 2) had several potential sources
of lead; solder joints, a brass bolt, and brass faucet. Slight differences in the
number of solder joints existed among rooms.
Potential solutions
Buildings, such as hospitals, office complexes, apartments, and schools
have the same plumbing materials used in homes (e.g. copper pipes, brass faucets
and fixtures, solder joints, etc..). Publicized problems related to high lead and/or
copper in the tap water may likely have a significant negative impact on
occupancy and may raise political issues. While plumbing systems in buildings
are typically larger and more complicated in design than a home, water quality
parameters that affect the solubility of lead and copper and treatment strategies
to reduce the metal levels are the same. Issues that could complicate treatment
in a building are more related to engineering, and flow and usage patterns of the
system.
Based on the sampling results, DWRD concluded that the primary sources
of lead were the solder joints and the brass faucets. To reduce lead
concentrations in the drinking water, three options were provided: (1) remove all
soldered copper plumbing, (2) install point-of-use (POU) devices at the taps, or
(3) install a chemical treatment system. Removing the copper plumbing would
eliminate the lead leached from the solder but not the lead from the brass fixtures.
Secondly, such an option would be expensive, inconvenient, and time consuming.
The disadvantages of POU devices were the cost of purchasing and installing
devices at hundreds of faucet sites, maintenance, and the need for constant
monitoring. Additionally, in-line systems would not solve the faucet lead
problem. The installation of a chemical treatment system raised many questions
by the employees. The major concern expressed was that adding chemicals
346
containing zinc, silicate, or phosphate to the water supply could potentially impact
their research studies.
Water usage study
Building water was supplied by the Patuxent Water Treatment Facility,
Patuxent, Maryland. The source water is river water treated by alum
coagulation. Water quality parameters as measured at the treatment facility are
shown in Tabic 1. The water quality would normally be considered relatively
non-aggressive, however, testing results indicated aggressiveness towards new
lead surfaces
New plumbing systems are more susceptible to corrosion attack than older
systems. Flux from the installation of solder joints, and oils and residues from
the manufacturing of copper pipe and brass plumbing components can promote
corrosion of the metal surfaces they contact. Metal particulate debris left on
plumbing materials following manufacturing and installation practices can
dislodge during system usage, contributing to tap water lead and copper levels.
However, with time and usage or system "aging", these materials will be reduced
or removed through dissolution and physical mechanisms. In addition, water
usage will enhance the development of protective superficial and passivating films
on the plumbing material, reducing metal diffusion from the metal surface.
Therefore, the first proposal made by the DWRD was to evaluate the impact of
water usage on metal levels throughout the new building's water supply. The
DWRD designed a water usage study as well as offered to perform the necessary
water sample analysis at the USEPA's Andrew W. Breidenbach Environmental
Research Center (AWBERC), Cincinnati, Ohio.
The building plumbing system is split into 2 distinct halves, the
"laboratory" and the "animal" sections. The laboratory section was designed for
general lab experimentation while the animal section was designed as an isolation
wing to prevent escape of biohazards to the outside Based upon accessibility and
plumbing considerations, the laboratory section was selected for the water usage
study. The laboratory section was comprised of four levels (ground, 1", 2"4, and
3" floors). Conveniently, the water lines feeding each floor could be isolated
from the rest of the building. Each floor was comprised of two wings, each
consisting of 9 rooms or laboratories, each with at least 1 faucet, that can be
further isolated.
Two wings were selected for the water usage study; one on the ground
floor and one on the third floor. A contractor was hired to supply a technician
to turn on and off the faucets in the rooms. Flow meters were installed at the
front of each wing to monitor water usage.
Beginning in early May, 1991, all faucets in nine laboratory rooms on
each wing were opened for a total of 2 hours per day, 5 days per week. Faucets
were opened four times a day for 1/2 hour with approximately 1-1/2 hour
stagnation time between flow periods. Flow rate was approximately 1 L/min.
347
-------
Sampling
Initially, sampling of the faucets was performed on every Tuesday and
Friday of each week. Generally each room had two faucets, although some had
one or three faucets. As a rule, one 250 mL sample was collected from the
faucet furthest from the utility chase in each room prior to the first flow period
on the specified days. This guaranteed at least a 12 hour standing time. The
samples were preserved with ultrapure nitric acid (0.15% v/v) and immediately
sent back to DWKD in Cincinnati, Ohio for analysis. Tuesday's samples were
analyzed for lead, copper, and zinc, and Friday's samples were analyzed for lead
only. Flame and graphite furnace methodologies were used for metals analysis.
Friday sampling was eliminated in mid-October, 1991 to reduce laboratory
workload after starting the chemical treatment portion of the study.
After approximately 100 days, two "flushed" 250 mL samples from two
rooms of each wing were taken during the first flush period on Tuesdays. One
sample was preserved with nitric acid (0.15% v/v) and analyzed for lead, copper,
zinc, iron, calcium, potassium, and magnesium. The other sample was not
preserved and was analyzed for alkalinity, phosphate, silicate, ammonia, sulfate,
nitrate, and chloride. In addition, a hand held pH meter was used to monitor pH
daily in the field.
Chemical treatment study
By late Oc.ober, 1991, six month test data showed that although the lead
levels had decreased slightly, the levels were still significantly above the desired
maximum level of 20 ^g'L, At that time, a decision was made to evaluate
chemical treatment. Based upon water quality and plumbing material
considerations, three treatment chemicals were proposed for the study; (1) a 1:3
Zn:P04 (as mg/L) ratio zinc orthophosphate formulation, (2) "generic" alkali
metal orthophosphate formulation, and (3) type "N" sodium silicate. Three
different building wings, one on the I" floor and two on the 2°* floor, were
chosen to evaluate the corrosion inhibitors.
Chemical feed system
A relatively simple, low maintenance, and inexpensive chemical feed
system was designed and installed in the utility chase of each test wing. The feed
system, shown in Figure 3, consisted of covered 100 gallon Nalgene® feed tank
with mechanical stirrer. A Milton Roy A7f feed metering pump was installed on
the cover of the tank to feed chemical inhibitor to the existing 1-1/4" cold water
line. The rate at which the chemical was fed was set by a Milton Roy RFP
Scries programmable flowmeter/puisar apparatus. An in-line static mixer was
included to insure sufficient chemical mixing. Chemical feed settings were
adjusted to deliver approximately 3.0 mg/L PO,' in the orthophosphate wings and
' Milton Roy, Acton, MA
348
30 mg/L Si02 in the silicate wing. The materials cost for each feed system was
approximately $1300.
Feed chemicals
A general description of the treatment chemicals used in the study is
presented in Table 2.
Sodium silicate solution was supplied by the PQ Corporation, Chester,
PA.. N-type sodium silicate with a 1:3.2 ratio of Si02/Na2O was used. The
solution contained 28.7% Si02 The manufacturer recommended using a dose of
24 mg/L SiOj in the system during the first 30-90 days, falling back to a
maintenance dose of 4-8 mg/L Si02. Because of a lack of field data on the use
of silicates for corrosion control and an understanding of corrosion control
mechanisms by which silicates reduce lead and copper levels, a higher start-up
dose of 30 mg/L was used. After approximately 70 days, the dose was decreased
to a maintenance dose of 15 mg/L, Dosage recommendations were derived from
research described in AWWARF's (1985) Internal Corrosion of Water
Distribution Systems.3 The sodium silicate feed tank was covered because sodium
silicate exposed to the atmosphere becomes viscous, which may lead to clogging
problems within the feed system. The solution is very basic, and would increase
the pH of the treated water significantly because of the source water's limited
buffering ability at the high Si02 level. The concentrated solution was diluted 1:2
with distilled water in the storage tank to accurately meter the basic feed solution.
The "zinc" orthophosphate formulation, product name SLI-939, was
supplied by Shannon Chemical Corporation, Malvern, PA.. Commonly named
zinc orthophosphoric acid, it's chemical formula is proprietary. However, it was
identified by the manufacturer to contain 8% zinc and 24% phosphate as PO/ or
a zinc to phosphate ratio of 1:3. Based upon manufacturer recommended dosage,
a desire to quickly minimize lead solubility, previous experience4"7, and water
quality, the dosage applied was 3 mg/L as P043. The concentrated solution was
diluted 1:30 with distilled water in the storage tank.
The generic orthophosphate, product name SL1-1226, was also supplied
by Shannon Chemical Corporation, Malvern, PA,. SLI-1226 is a proprietary
blend of alkali metal orthophosphate salts in acidic solution and does not contain
polyphosphate, silicate, or zinc. The orthophosphate content was 36% phosphate
as PO,''. The dose used was also 3 mg/L as PO/ for the same reasons given
previously and the solution was diluted 1:30 with distilled water in the storage
tank,
Sampling
On November 25, 1991, the chemical inhibitor test program began. All
faucets in 9 laboratory rooms on each test wing were opened for a total of 2
hours per day, 5 days per week. Faucets were opened four times a day for 1/2
hour with approximately 1-1/2 hour stagnation time between flow periods. Flow
349
-------
rate was approximately 1 L/min. This protocol was identical to that followed in
the water usage study.
Samples were take in the same manner as in the water usage study.
Additionally, an unprescrved 125 mL flushed sample was taken from the first
(closest to the chemical feed system) and the last room (farthest from the
chemical feed system) in each wing to verify chemical dosing on Tuesday and
Friday. These samples were analyzed for silicate or phosphate only.
The pH of the wafer was monitored daily on site using a hand held pH
meter. Portable test kits were used on site to monitor consistency of phosphate
or silicate dosing in the first and last room in each wing on a daily basis.
Analytical melbods and reagents
Water s?..iples analyzed for metals were preserved on site by adding
0.15% v/v ultra pure reagent grade HNOj' in accordance with EPA
recommendations for preserving metals in drinking water samples.1
The analytical techniques used in this study are listed in Table 3. Method
detection limits are also presented. Lead was analyzed with a Perkin Elmer
model 4000 atomic absorption spectrophotometer" equipped with a model HGA-
400 graphite furnace and model AS40 autosampler. AJ1 other metal
determinations were made on a Perkin Elmer model 5000 flame atomic absorption
spectrophotometer and model AS50 autosampler.
Alkalinity, chloride, and sulfate were measured with a Metrohm E-636
titroprocessorm. Nitrate, phosphate, and silicate were measured on a Alpkem
RFA/2 autoanalyzerMH,
Silica and phosphate determinations were made on site using a Hachtrm
SI-5 and a Hach PO-19A test kits, respectively. Sample pH was measured with
a Cole-Parmer"1™ model 5985-75 pH meter and accompanying electrode.
Particulate material was analyzed with a Link EXL Energy Dispersive X-
Ray System (EDXA)'""'' mounted on a JEOL 5300 Scanning Electron
Microscope (SEM)t+tmT* at 30 kV.
' Ultrex, J T. Baker Chemical Comp . Phillipsburg, NJ
" Perkin-Elmer Corp., Norwalk, CT
m Metrohm, Switzerland
^ Alpkem Corp., WilsonviUe, OR
Hach, Loveland, CO
m+t* Cole-Parmer, Chicago, IL
tmm Analytical, Madison, WI
ttmlM J.E.O.L., Peabody, Mass
350
Results
Water Usage study
The water usage study began on May 1, 1991 and was terminated on
February 4, 1992, lasting about 9 months. During that time, nearly 260,000
gallons of water (1500 gallons/day) and 106,000 gallons of water (600
gallons/day) were flushed through the ground and third floor wings, respectively.
Water usage was greater on the ground floor wing because, as it was later
discovered, two of the nine rooms on the wing had been used prior to and during
the study by the building's janitorial and maintenance employees. In addition,
towards the end of the study, employees began to gradually move into the
building. The nature of their work required the employees to lock unattended
laboratories. During the final month of the study, it was often impossible to enter
locked laboratories and proceed with water usage study.
Results from the usage study indicated no apparent reduction in lead levels
with water usage over the time period of this study. Variability in the data
because of frequent lead spikes tended to statistically suggest that water usage had
no effect on lead levels. For example, a room would have low lead values (< 20
jig/L) during a sequence of daily sampling, then a few very high values (>50
pg/L). The sporadic occurrence of high lead values occurred regularly among
nearly all of the rooms sampled in the usage study. The lead distributions for the
rooms on the third floor wing is shown in Figure 4. Some lead spikes were
greater than 1000 fig/L. Two rooms on the ground floor (Figure 5) appeared to
be exceptions. As mentioned, these two rooms were used prior and during the
study by the janitorial employees. The additional water usage probably
contributed to the lower overall lead concentrations and reduced the occurrence
of sporadic lead spikes, suggesting a very long time of usage may reduce the
problem after all.
The occasional unusually high lead levels in samples, and a general non-
Gaussian pattern of the lead levels, arc in accord with many previous laboratory
and field studies of lead corrosion., s*14 In addition to occurring as dissolved
aqueous ions and complexes, lead can be present as, or associated with, various
colloids or particulates. These solids can originate as non adherent corrosion
deposits, eroded pieces of plumbing material, or be present in the background
water in the building or municipal distribution system. Many kinds of particles
have been shown to be effective scavengers of lead in natural and potable waters.
Notable among the solids having a high affinity for sorbing or incorporating lead
are hydrous iron oxides121417, humic substance colloids (with and without
associated iron)"1"1, calcite particles", and various stream sediments.10 In
several cases, the sorption of lead on particle surfaces can be explained by either
a "surface complexation" approach" or by a cation-exchange process. In either
of these two cases, lead should usually be readily removed from the particles after
sample acidification. Thus, the lead originally associated with the particles would
be detected in the same manner as dissolved lead by conventional graphite furnace
351
-------
atomic absorption spectroscopy (GFAAS) or inductively coupled plasma
spectroscopy (ICP).
Often, visual examination of the acid-preserved (0.15% HNO,) water
samples, showing abnormally high lead concentrations, revealed thai metal
particulates were present at the bottom of the sample. On several occasions,
when samples showing unusually high lead concentrations were reanalyzed on
later days, the concentrations tended to be inconsistent, and slightly increased to
some eventual!v-consistent level. Originally, the spikes were thought to be an
instrumental problem, before the samples were more thoroughly examined. The
observations were later found to coincide reasonably with occasions when samples
were analyzed within several hours of receipt from the field site, reflecting 1-2
days of total elapsed time after sampling and immediate acidifications.
Ten samples (from rooms in the two usage wings) analyzed with high lead
concentrations and identified as containing particulates were filtered through 0.2
polycarbonate filters. The particles retained on the filters were examined by
an energy-dispersive X-ray analysis (EDXA) system attached to a scanning
electron microscope (SEM) in an attempt to characterize the source of lead. In
all cases, the individual particles examined showed EDXA elemental spectra
containing tin, but lead or lead-containing particles were not found. When tin
particles were found, high lead levels were likely caused by the corrosion of the
lead-tin solder. Lead or lead-containing particulates were likely to coexist with
tin or tin-containing particles in the original samples. Therefore, preferential
dissolution of lead phases by the acid preservative was strongly implicated.
These observations have been discussed in detail in another paper."
Copper levels did not follow the same sporadic peak trend set by lead
levels (shown in Figure 6 for the third floor). There were a few occasional
variations in copper levels, however, typically several rooms experienced these
variations or small peaks on the same day. This suggests the influence of water
quality parameter changes (e.g. pH and chlorine residual) rather than particulate
material. Previous experience has demonstrated that copper can be very sensitive
to small pH and chlorine fluctuations.35 Visually, there did not appear an obvious
overall trend, although it could be argued that copper levels increased slightly
over time. Copper levels were relatively low, nearly all were less than the 1.3
mg/L action level set by the Lead and Copper Rule34". Although the Rule
specifies 1 L samples, it stands to reason that I L samples taken from the rooms
would remain below 1.3 mg/L.
Zinc levels more closely followed the leaching patterns described by
copper levels than lead levels (shown in Figure 6 for the third floor). There were
a few more random zinc peaks, some of which could be more related to
particulate material than water quality changes. Two zinc spikes, in room 3301
(at approximately 100 days) and in room 3329 (at approximately 50 days), also
have corresponding lead peaks on those days. This suggests that brass
particulate, possibly coming from mechanical or hydraulic abrasion of brass
surfaces in the faucet, was present in those samples. However, with only two
35?
standout cases of zinc peaks, brass does not appear to be a significant source of
lead spikes.
Background water quality measurements represented by flushed samples
and field pH measurements were not initiated until late into the usage, study as
previously mentioned. Table 4 shows pH and other water quality parameters of
the background water in flushed samples taken from the treated wings.
Background water quality agreed well with water quality leaving the Patuxent
Water Filtration Plant (Table 1).
Chemical treatment
Chemical treatment began on November 25, 1991 and ended on April 14,
1992, lasting just over 4 months. Prior to the start of chemical treatment, one
set of "baseline" samples was taken from the faucets in the 3 wings (shown in
Table 5). Only one set could be collected because the immediate necessity of the
chemical treatment evaluation limited the length of time such an evaluation could
take place. The absence of a statistically viable baseline precluded the later
application of objective statistical procedures for objective evaluation of
comparative treatment performance. However, based upon building construction
records, plumbing materials used, visual examination of the plumbing, baseline
results, and previous sampling results, it is reasonable to assume that the wings
used in the treatment study would behave similarly to those wings in the usage
study with respect to metal leaching trends. Also, instrumental malfunctions
associated with the silicate and phosphate analysis conducted by the EPA resulted
in eliminating nearly all phosphate and silicate laboratory analysis. Therefore,
silicate and phosphate concentrations monitored in the field were the most reliable
determinations of chemical feed doses.
Zinc orthophosphate
During the study period, 101,277 gallons of water was passed through the
plumbing of the nine rooms in the zinc orthophosphate test wtng. Phosphate
levels were maintained at approximately 3.3 mg 1'043/1, (see Table 6). It
appeared that phosphate deposition between the first room (room closest to the
chemical feed system) and last room in the wing accounted for a 0.2 mg/L
phosphate concentration drop. Zinc orthophosphate reduced the pH by about 0.5
pH units, from approximately 7 7 to 7.2. Other than zinc, which increased by
approximately 1.25 mg/L, no other water quality parameter was notably changed
from background water quality using zinc orthophosphate inhibitor (sec Table 4).
Zinc orthophosphate effectively reduced lead levels in the wing. Lead
levels among the majority of rooms in the wing dropped rapidly and stabilized at
less than 5 pg/L by 80 days (shown in Figure 8). In addition, the occurrence of
random lead spikes among samples was nearly eliminated.
353
-------
Zinc orthophosphate addition reduced copper levels. As with lead, copper
levels appca ed to stabilize after about 80 days of treatment. Copper levels
remained below 0.15 mg/L after stabilization.
Logically, zinc levels in the water treated with zinc orthophosphate
became much higher than in water treated with other treatment chemicals. Zinc
levels increased approximately 1.25 mg/L. Zinc concentrations correspond well
to phosphate concentrations in that the ratio of zinc to phosphate was
approximately 1:3; the mean zinc concentration was 1,3 mg/L and the mean
phosphate concentration was 3.3 mg PQ4'/L (background zinc and phosphate
levels were insignificant).
Generic orthophosphate
Over the study period, 92,676 gallons of water was passed through the
plumbing in the eight rooms of the calcium orthophosphate test wing. Phosphate
levels were maintained at approximately 3.3 mg P04'/L (see Table 6). It
appeared that phosphate deposition between the first room (room closest to the
chemical feed system) and last room in the wing accounted for about a 0.1 mg/L
phosphate concentration drop. Orthophosphate dosing reduced the pH of the
rooms in the 2300 wing by about 0.4 pH units, from approximately 7.7 to 7.3.
No other water quality parameter was notably changed from background water
quality using calcium orthophosphate (see Table 4).
1 ..nd levels in the water were effectively reduced by the orthophosphate
addition and appeared to drop and stabilize more rapidly with generic
orthophosphate than with zinc orthophosphate. Lead levels generally dropped to
and stabilized at about 5 jig/L after only 25 days (shown in Figure 9). However,
there, were exceptions. Room 2305 took almost 100 days to stabilize at 10 jig/L.
This was probably related to the plumbing in the room; perhaps there were more
solder joints used in the room's plumbing. Room 2303 stabilized after 25 days
but increased dramatically after 50 days. This behavior is inconsistent with
solubility behavior and normal orthophosphate passivation behavior, suggesting
perhaps the plumbing was disturbed in some way. The sporadic occurrence of
lead peaks was generally reduced among rooms.
Copper levels were also reduced and appeared to stabilize after about 50
days at less than 0.2 mg/L by orthophosphate dosage. There were only a few
small random copper peaks of almost 1 mg/L. The appearance of the peaks is
consistent with behavior of particulate material, however lack of vcnfiability of
odd trends is a problem with remote field sampling. There were no
corresponding lead or zinc peaks which might suggest the source of copper to be
copper pipe particles rather than brass.
Zinc levels were reduced almost immediately by the addition of the
generic orthophosphate. Zinc levels were maintained below 0.2 mg/L. No
significant peaks appeared in the data.
354
Sodium silicate
Over the study period, 96,476 gallons of water was passed through the
plumbing in the nine rooms of the sodium silicate test wing. Silicate dose started
at about 32 mg Si02/L, and after 71 days was dropped to a maintenance dose of
16 mg SiOj/L (see Table 6). It appeared that silicate deposition between the first
(room closest to the chemical feed system) and last room in the wing accounted
for about 1 0 mg/L silicate concentration drop for both doses. The addition of
sodium silicate at the start-up dose raised the pH of the treated water by more
than 15 pH units, to about 9.5. The pH dropped to a range of 8.8-9.1 after
reducing to the maintenance silicate dose. Sodium levels resulting from sodium
silicate addition increased sodium levels by only about 4 mg/L over background.
Lead levels dropped rather rapidly, stabilizing after about 25 days at
approximately 10 ng/L (shown in Figure 10), In addition, the occurrence of
sporadic lead peaks was greatly reduced. An exception was room 2408 that
stabilized at about 25 fig/L, The higher lead levels may be an artifact of the
plumbing; perhaps there were more exposed solder joints or the workmanship
related to the soldering was poor Also, room 2416 was very peculiar in that the
lead levels continuously climbed. This is difficult to explain and would require
closer examination of the plumbing in that particular room for a full analysis.
Dropping to the maintenance silicate dose appeared to make no impact on lead
levels. As mentioned, the silicate doses used were greater than manufacturer
recommended doses.
Copper was almost immediately reduced (shown in Figure 11). Sodium
silicate appeared to be the best of the chemical inhibitors at reducing copper
levels, maintaining levels below 0.07 mg/L. Also, there were no sporadic copper
peaks except for one exception, room 2402. Copper levels appeared to stabilize
at about 0.25 mg/L, far above normal solubility levels of any oxide, hydroxide,
or basic carbonate solids of Cu'+. Lead values in this room however, were not
noticeably different than the other rooms in the wing.
Zinc levels were also reduced almost immediately by sodium silicate
addition. Levels stayed below 0.05 mg/L with the exception of room 2402 which
was sporadic and about 0.5 mg/L. This room also had high copper levels which
might indicate excess brass plumbing or high erosion of plumbing material.
Discussion
The results of the water usage study indicated that high lead levels in
water from a new building may take a long time to drop under conditions of
"normal" water usage; in the case of the agency building, greater that 8 months.
Lead levels of the water samples taken bi-weekly showed overall high and
inconsistent lead levels with the occurrence of random lead spikes. Sources of
lead were identified as solder joints and brass fixtures. Extraordinary high lead
concentrations or spikes in that data were probably due to lead-containing
355
-------
particulate material. This type of occurrence is not uncommon. At this building,
a major contributor to the particulate material causing the erratic lead levels was
found to be lead-tin solder. Lead containing particles could also be coming from
brass, however solder was discovered indirectly to be the cause because of the
insoluble tin that was sometimes left in the samples with high lead concentrations.
Zinc orthophosphate, generic orthophosphate, and sodium silicate
treatment effectively and quickly reduced lead levels and the occurrence of lead
spikes. All treatments also reduced copper concentrations. Differences in
treatment performances may strictly be due to variations in plumbing
configurations such as the number of solder joints or the workmanship of those
joints. Results arc strictly based on observations and would be difficult to prove
statistically with any certainty. Every wing had at least one room in it that was
an exception 10 these general niles. Exceptions may be due to differences in
plumbing configuration or usage patterns. However, all of the treatment
chemicals reduced lead in the water to acceptable levels.
To further reinforce the proceeding observations, box plots were
constructed for each treatment wing (Figures 12-15) Box plots offer the ability
lo visually and statistically describe data sets. Box plots were used to identify
general characteristics of the distribution of lead levels in each wing over time.
In other words, the effect water usage and corrosion inhibitors have on the lead
distributions of all the rooms of a wing can be tracked. Each box shows various
distribution percentiles and the mean of the lead values for the 8 or 9 rooms in
the wing for each day. Plots clearly show that lead levels and the sporadicness
are reduced (percentile ranges are reduced) by all chemical treatments.
In the absence of a long enough period of pre treatment monitoring to
establish a true baseline, statistical tests could not be made to quantify relative
differences caused by the treatments. The variability among rooms in the usage
study as well as within treatment wings also argues that more sites would
probably be needed for controlled evaluations.
Conclusions
The following conclusions from the corrosion control study conducted at
the agency facility can be made:
1. General water usage, as described by the test protocol in this paper, did not
appear to reduce lead levels in the drinking water during the duration of the
study. Results indicated that lead reductions by continuous water usage in a
building may take more than 8 months and probably years.
2. Water samples taken during the water usage study showed inconsistent lead
levels among daily samples, often with sporadic lead concentration spikes.
Particulate material, most likely Pb:Sn solder, was the major contributor of the
larger random lead spikes.
356
3. A simple, economic, low maintenance chemical inhibitor feed system can be
designed and installed to add inhibitor to an existing building plumbing system
that will produce consistent application of a variety of liquid chemical inhibitors .
4. Zinc orthophosphate, alkali metal orthophosphate, and sodium silicate
corrosion control inhibitors all reduced lead concentrations to acceptable levels.
Although copper was not an issue, they all reduced copper levels as well, at the
dosages and respective pH ranges employed.
5. Zinc orthophosphate, alkali metal orthophosphate, and sodium silicate
corrosion control inhibitors all reduced and nearly eliminated the occurrence of
random lead spikes in daily monitoring
The success of this study was based on theoretical considerations and
previous knowledge and experience of using silicate- and phosphate- based
corrosion inhibitors in drinking water systems. An understanding of the water
quality conditions (ie. pH, DIC, etc.) that favor their usage and proper dosage
rates is essential if success in reducing lead and copper solubility is to be
achieved.
Acknowledgements
The authors would like to thank the following members of the USEPA,
DWRD, IPCB staff for their contributions: Keith Kelty, James Doerger, James
Caldwell, and Louis Trombly for performing the many water analysis'*, and Herb
Braxton for graphical assistance and water analysis. We would also like to thank
Greg George, Roger Rickabaugh, Steve Harmon, and John Dammann of
Technology Applications Incorporated for their technical assistance.
Disclaimer
Mention of specific trade names or instrument models is for explanatory
putposes only, and does not constitute an endorsement by the U.S. Environmental
Protection Agency.
357
-------
References
1 Office of Water. "Suggested Sampling Procedures to Determine Lead in
Buildings Other Than Single Family Homes", U.S. Environ. Protection
Agency, Washington, D.C. (June, 1988).
2. Office of Water. "Lead in School Drinking Water", EPA 570/9-89-001.
U.S. Environ. Protection Agency, Washington, D.C. (Jan., 1989).
3 American Water Works Association Research Foundation and DVGW
Forschungsstelle. Internal Corrosion of Water Distribution Systems. A,
Denver, CO. (1985)
4. Schock, M R. Understanding Corrosion Control Strategies for I-ead.
Jour. AWWA, 81:7:88 (1989).
5. Colling, J.H., Whincup, P.A.E., and Hayes, C R The Measurement of
Plumbosolvency Propensity to Guide the Control of Lead in Tap Waters.
Jour. hut. Water and Environ. Mgmt., 1(3):263-269 (1987).
6 Gregory R , and Jackson, P.J. Central Water Treatment to Reduce Lead
Solubility. In Proc. AWWA Annual Conf, Dallas, TX. A, Denver, CO.
(1984).
7. American \»ater Works Association Research Foundation. Lead Control
Strategies. AWWA, Denver, CO. (1990).
8. Environmental Monitoring and Support Laboratory. Handbook for
Analytical Quality Control in Water and Wastewater Laboratories. EPA-
600/4-79-019. U.S. Environ. Protection Agency, Cincinnati, OH (1979).
9. Schock, M. R. Causes of Temporal Variability of Lead in Domestic
Plumbing Systems. Environ. Monit & Assessment. 15:59, (1990).
10. Schock, M. R. et. al. The Significance of Sources of Temporal
Variability of Lead in Corrosion Evaluation and Monitoring Program Design.
Proc. AWWA WQTC, St. Louis, MO (1988).
11. Britton, A. & Richards, W. N. Factors Influencing Plumbosolvency in
Scotland. Jour. Inst. Water Engr. & Scientists, 35:4:349 (1981).
12. Hulsmann, A. D. Particulate Lead in Water Supplies. Jour. Inst. Water
Engnr. & Mgmt., 4:2:19 (1990).
13. Breach, R. A., et. al. A Systematic Approach to Minimizing Lead
Ixvels at Consumers Taps. Proc. AWWA Annual Conf., Philadelphia, PA
(1991),
358
14. Harrison, R. M. & Laxen D. P, Physicochemical Speciation of Lead in
Drinking Water. Nature, 286:791 (Aug. 1980).
15. de Mora, S. J. & Harrison, R. M. Lead in Tap Water: Contamination
and Chemistry. Chem. Britain, 20:900, (1984).
16. de Mora, S. J. et. al. The Effect of Water Treatment on the Speciation
and Concentration of Lead in Domestic Tap Water Derived from a Soft
Upland Source. Water Res., 21:1:83, (1987).
17. dc Lurdes Simoes Gongalves, M , et. al. Voltammetric Methods for
Distinguishing between Dissolved and Particulate Metal Ion Concentrations in
the Presence of Hydrous Oxides. A Case Study on Lead(ll). Environ. Sci. cJ
Techno!., 19:141 (1985).
18. Salim, R. Adsorption of Lead on the Suspended Particles of River
Water. Water. Res, . 17:4:423 (1983).
19. Wouters, L. C. et. al. Discrimination between Coprecipitated and
Adsorbed Lead on Individual Calcite Particles Using Laser Microprobe Mass
Analysis. Anal. Chem., 60:2218 (1988).
20. Hem, J. D. GeochemicaJ Controls on Lead Concentrations in Stream
Water and Sediments. Geochim. Cosmochim. Acta, 40:599 (1976).
21. Osaki, S. et. al. Adsorption of Fe(lH), Cofll) and Zn(U) onto Particulates
in Fresh Waters on the Basts of the Surface Complexation Model. II.
Stabilities of Metal Species Dissolved in Fresh Waters. Sci. Tot. Environ.,
99:115 (1990).
22. Lytic, D.A., Schock, M.R., Dues, N.R., and Clark, P.J. investigating
the Preferential Dissolution of Lead From Solder Particulates. Jour. AWWA,
7:85:104-110 (July 1993).
23. Lytle, D.A., Schock, M R., and Tackett, S. Metal Corrosion Coupon
Study Contamination, Design, and Interpretation Problems. In Proc. AWWA
Water Quality Technology Conf., Toronto, Ontario, Canada (1992).
24. Lead and Copper. Final Rule. Fed. Reg., 56:26460 (June 7, 1991).
25. Lead and Copper. Final Rule Correction. Fed. Reg., 56:135:321132
(July 15, 1991).
359
-------
«
*
*
<
<
a.
yj
£
<
<
a.
u.*
V5 "D 5®
c 5 !
| £" &
3 9, j3
' s
¦6
i 5*
3 s
s -s
s- £
s 1
E
O
a.
O
a.
r* £ "O
OC
E
£
e
o a
Vi o
c ^
n: .2
0 g 0"
1 S 2
Z 5 #
- .3 »
^ Q c
3 •• 5
•SX c
R " .0
HZll
„g
o* f
a. £
:i I
i Z 2
i «• §
' « u
c «
9
S- *
v Q S
•5 <¦»
H o
0 "
Z U
tz
m
360
Table 2 Source water analysis from the Piitmcnl Water Filtration
Plant (1990).
Parameter
Minimum Maximum
Unit of Yearly Monthly Monthly
Measure Average Average _
Physical
mg/L"
32
Alkalinity
Dissolved Solids, Total
mg/L
105
Hardness
mg/L
52
pH
Units
8.0
Turbidity
NTU'
016
25
98
46
7.5
0.10
39
114
56
8.2
0.19
Metals .
Aluminum mg/L
Arsenic
Banuiu -
Cadmium ng/L
Calcium mt/L
Chromium »g/L
Copper rag/L
Iron mg/L
Lcad ™g/L
Magnesium mg/L
Mangancsc mg/L
Mircurv mg/L
Potassium mg?L
Seleoium
-------
Table V Analytical techniques used lo measure water quality parameters.
Paramctci
Calcium
Copper
Iron
I.cad
Magnesium
Manganese
Potassium
Sodium
Zinc
PH
Chloride
Sulfate
Alkalinity
Silicate
Nitrate
Phosphate
Phosphate
Total chlorine
Free chlorine
Method
Number
Reference
Detection Limit
(mg/I.)
7140
7210
7380
7-121
7450
7460
7610
7770
7950
9040
9252
9038
3101
A303-S220-13
A303-S173-00
A303 —S200-02
8048
8167
8021
EPA'
CPA*
F.PA*
EPA'
EPA*
EPA'
EPA'
EPA'
EPA"
EPA"
EPA"
EPA'
EPAh
Alpkem'
Alpkem'
Alpkem'
Hachd
Ilach11
Hachd
1.0
0 02
0.05
0 002
2.0
001
0.07
0.75
0.01
10 (as CI)
7.0 (as S04)
1.0 (as CaCO,)
0.4 (as SiOJ
0.02 (as N)
0.1 (as POJ
0 7 (as Clj)
0.7 (as Cl2)
a Tea Methods for Evaluating Solid Wastes " (SW846) KKFI_ U S EPA, September, 19R6
b Methods for Chemical Analysis of Water and Wastes " U S hFA. Revised 1981
c "KFA/2 Metltods Manual" AJpkenj. April, 1991
d 'Hach DR/2000 Spectrophotometer Methods Manual." Hach Company. Lovrland. CO (1992)
362
I I
C4 'O — V"}
O o o
o o v d
r- r- o
f«4 — & TT
-do«
oooo
ri K
c c c c
.
.
n
n
n
<0
o
r»
r-
r.
o
r»
r-
Tj-
c\
—
v;.
V~l
o
o
W-l
«o
c-i
o
r4
r»
r>»
d
r)
ri
ri
o
fvj
r4
rj
o
ri
r<
IT.
O
cN
O
n
o
r»
©
r-,
O
O
CN
o
<3
o
s
ri
O
NO
f.
O
s
r«
o
n
o
o
o
d
o
o
o
o
d
d
d
o
o
d
o
o
V
V
V
V
o
*r
n
n
w
ro
<*
n
<9
n
(9
&¦
pK
wn
c
C
c
C
c
c
C
n
c
c
c
c
"""
p~.
o
f-.
o
(r,
>©
n-
OQ
*->
o
r~.
Wl
wt
00
r4
o
K
wS
*-<
v-i
—
•—
o
n
N r. <
^ O h i
7.2SR
^ o « a
,82;
,85#
! o r-' K
^ 3 ;
2 co .
5"®
IS:
a
<9
¦g.
I
£
3
€
O
£
V s
p
*8 iz
i ill
Q- « C >
? 1 g 2
« 2 o
• re
~ « c
«/>
363
-------
Table 5. Baseline lead levels in chemically treated building wings.
Room Number
Lead (
-------
41
g
B
v
a
v
4)
<«
4)
a
B
«
Cfl
*0
«
J
~
V
J3
w
3
o
S
o
o
t-
o
.*-
3
w
v
i*
_«
"a
S
«
c/5
u
o
©
.2 63
c -a
&> B
3 3
cr o
V fc"
C/5 OD
JM
»H 0£
£ •-
3 2
.2f» 3
fa jo
1/8n 'peaq
366
Connects directly
to lab faucet.
Brass Bolt
1/4" ID
Cu
Tee (4) ©
Reducer
Cu
Elbow
Figure 2. Typical cold water plumbing connection to brass
faucet. Numbered locations represent visible soldered connections
(drawing not to scale).
367
-------
Frogrtnjaubie
divider
1 1/4* cold
wuer Uac
Mixer
Figure 3. Chemical feed system installed in treatment wings
(not to scale).
368
1000
800
600
400
200
I | V | I- | ¦?
° Rm, 3301
I ' I ' I ' I 1 )
Rm. 3305
«.t1C8» f) I (Ti
Q —i
1000
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Time, days Time, days
I ' i 1I
Rm. 3307
800 --
600 --
400 --
-1—|—'—j—t—p—i—[—I—]—r
Rm 3309
¦¦oternitiiit;
1000
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Time, days Time, days
800 --
Rm. 3315-
400 --
1 i ' i 1
Rm 3317
tm-
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Time, days Time, days
1 i ' i ¦ t 1 i 1 i
Rm. 332
' i ' I—¦ I ' i ¦—I
Rm. 3325
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Time, days Time, days
1000 "> ||i 'q'i 1 i ' i ¦ i 1 i
Rm. 3329 :r Figure 4, Lead levels in the
3000 wing treated by water
usage.
0 50 100 150 200 250 300
Time, days
369
-------
, 1000
^ 800
00
3 600
¦o
ea
J
Rm. G402 ::
in ft'
» ¦ I f—-J
Riti. G404
oi)
3
-a
pj
¦u
50 100 150 200 250 300
Time, days
_r
Rm. (;4()6
0 50 100 150 200 250 300
Time, days
• Rm. G408
%
- J- \r!It J •'
1000
0 50 100 150 200 250 300
Time, days
0 50 100 150 200 250 300
Time, days
w>
3
¦a
a
v
J
Rm G4IO
Rin G4I2
50 100 150 200 250 300 0 50 100 150 200 250 300
Oil
3
-d
n
it
_l
1000
800
l ime, days
1 i 1 i ¦ r
1 i ¦
Rm (1414
Time, days
400
200
0
j!
0 50 100 150 200 250 300
l ime, days
" i ' i
Rm G416
OA
3
¦a
m
u
T
1000
800 -J- Rm G4)8
600
400
200
0 -
rx-
0 50 100 150 200 250 300
Time, days
Figure 5 Lead levels in the
ground wing treated by water
usage.
0 50 100 150 200 250 300
Time, days
i;n
1 • r ¦ | r j ¦
Rm. 3301 -H.5
( » | I {
Rm. 3305 T
-4.0 --
tWH%. <1
«%>- $ 5
' ' '0.0
50 100 150 200 250 300
Time, days
50 100 150 200 250 300
Time, days
BO
e
4>
Cl
XX
O
U
I ' I
1.5 + Rm. 3307
1.0
0.5
0.0
-•ojj1
:4.5 --
- 4.0
i-0.5
-J—K).0
t ¦ i—r—r
Rm 3309
50 100 150 200 250 300
Time, days
50 100 150 200 250 300
Time, days
Rm. 3317
Rm. 3315
50 100 150 200 250 300
Time, days
50 100 150 200 250 300
Time, days
| t | f |
Rm. 3325
Rm 3321
J
"So
u
Cl
O.
O
U
2.0
1.5
1.0
0.5
0.0
50 100 150 200 250 300
Time, days
-i—¦—i—i—r
Rm. 3329
"TO. I .
50 100 150 200 250 300
Time, days
Figure 6. Copper levels in the
3000 wing treated by water
usage.
50 100 150 200 250 300
Time, days
3T1
-------
o
c
N
-J
ah
E
'j
c
N
2.0
1.5
10
0.5
0.0
-T—i—|—' | ' I • 1
Rm. 330:
5.0
.5
0
-0.5
€.0
-i.
' i ' i 1 i 1 "i • i
Rm. 3305
0 50 100 150 200 250 300
Time, days
Rm. 3307
.5
,0
-0.5
<>.0
0 50 100 150 200 250 300
Time, days
1 i ¦ i ¦ i ¦ i ¦ i
Rm. 3309 ""
_d«a2£<
0 50 100 150.200 250 300
Time, days
0 50 100 150 200 250 300
Time, days
I ' I 1 I 1 ! ' I
Rm. 3317
Rm 3315
itOliU
oil
E
o
c
M
2.0
1.5
1.0
0.5
0.0
0 50 100 150 200 250 300
Time, days
Rm 3321 - -1.5
HO
- 6.5
%0
»r^JOO. i—<—i—i—i—i—i—i—
oh
1.5 :
Rm.3329 --
c
10 :
S'
'J
IZ
0.5 -
A
N
0.0
0 50 100 150 200 250 300
Time, days
Figure 7. Zinc levels in the
3000 wing treated by water
usage.
0 50 100 150 200 250 300
Time, days
m
T3
n
iML
I 1 I ' I
Rm 1414
0 25 50 75 100 125 150
Time, days
r-200
150
-400 --
50
0
0 25 50 75 100 125 150
Time, days
T T | T- y T T f | I
Rm. 1416
Rm. 1418
0 25 50 75 100 125 150
Time, days
Figure 8. Lead levels in the
1400 wing treated by zinc
orthophosphate addition.
0 25 50 75 100125 150
Time, days
373
-------
100
75
50
25 %
0
-i—¦ i 1 i '"i 1 i
Rm. 2301
Rm. 2303
p^'i
iwrmatc, i ..a
0 25 50 75 100 125 150 0 25 50 75 100 125 150
Time, days Time, days
1f, i 1 i
Rm. 2305
Rm. 23
WJI'T
0 25 50 75 100 125 150 0 25 50 75 100 125 150
Time, days Time, days
4
75
50
25
0
1 i 1 i ' i ¦ i 1 i 1
Rm. 2313
iOO
75
--50
25
0
i ¦ i ' i ¦ i ' i '
Riii 2315
jfawnrr^r
100
75 *
50 -i
0 25 50 75 100 125 150 0 25 50 75 100 125 150
Time, days Time, days
1 i ¦ i ¦ i 1 i ¦ i «
Rm. 2319
25 ft, l
0 Tiln'ffmarjiffli
'M&mmtemsiL
[00
¦75
--50
--25
) * i 1 i 1 i ' i ¦
Rm 2321
0 25 50 75 100 125 150 0 25 50 75 100 125 150
1'imc, days Time, days
Figure 9. Lead levels in the 2300 wing treated by
alkali metal orthophosphate addition.
Rm.2402
Rm. 2404
0 25 50 75 100 125 150 0 25 50 75 100 125 150
Time, days Time, days
Rm.2406
11 1 i ' i
Rm. 2408 : f
0 25 50 75 100 125 150 0 25 50 75 100 125 150
Time, days Time, days
t ' i ' i 1 i 1 i
Rm. 2412
Rm. 2410
0 25 50 75 100125 150 0 25 50 75 100 125 150
Time, days Time, days
100
75 :1
50 '[
25:Sj.
I ¦ I
0
Rm. 2414
"") ' I ¦ I I i
Rm. 2416
0 25 50 75 100 125 150 0 25 50 75 100 125 150
Time, days Time, days
Figure 10. Lead levels in the
2400 wing treated by sodium
silicate addition.
0 25 50 75 100 125 150
Time, days
Rm. 2418
17ft
-------
Copper. mg/L
© © — — ro
o ui o ui o
Copper. mg/L
o c — — fO
O In C V> b
Copper, mg/L
C O >— ~ to
o U o w c
U I 1 ' I
o o © a-m
o In o o
!2, Tl
O 3* J1.
CL ^ CFO
5' ^ E
3 ft s
o —
P? 2. r 3'
M = o J*
« O c.
tl -t2 S3
a g 3 c/>
Si* o —
5^2
~ cr ^
c/T
O O CJ-'to
o in o <-a o
LA ) I [ I | I | I
I'''
Copper, mg/L
© p •— — to
b u o vi b
o fc-t-4
Copper, mg/L
C O - - N)
o Ln © in b
o ©
b <-n o
Opo^^-M
o Ln o w> o
M ¦ 1 h
H-H-H
3
•i
95th percentile
75th percentile
Mean
50th percentile
25th percentile
5th percentile
Time, days
Figure 12. Box plot lead distribution of rooms in the water usage wins
of the building.
-------
1 19 36 54 71 89 106 124
Time, clays
Figure 13. Box plot distribution of lead of rooms in
wing treated with zinc orthophosphate.
300
250
1 17 35 52 71 87 105 122
Time, days
Figure 14. Box plot distribution of lead of rooms in the
wing treated with alkali metal orthophosphate.
379
-------
300 -
250
200
1 26 42 61 78 96 113 120 141
Time, days
Figure 15. Box plot distribution of lead of rooms in the
wing treated with sodium silicate.
380
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before comple
1. REPORT NO. 2.
EPA/600/A-95/036
3.
4. TITLE AND SUBTITLE
A Systematic Study on the Control of Lead in a New
Building
S. REPORT DATE
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Darren A. Lytle, Michael R. Schock & Thomas J. Sorg
8, PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME ANO AODRESS
USEPA, DWRD, I&PCB, RREL
26 W. Martin Luther King Dr.
Cincinnati, OH 45268
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND AODRESS
USEPA
26 W. Martin Luther King Dr
Cincinnati, OH 45268
13. TYPE OF REPORT ANO PERIOD COVEREO
Published Paper
14. SPONSORING AGENCY CODE
EPA/600/14
15. supplementary NOTES 1994 AWWA Annual Conference Proceedings, 6/18-23/94, New York City
NY, P:345-380
Project Officer = Darren Lvtle
16. ABSTRACT
A new building was identified as having high lead levels in its drinking water.
Through a detailed sampling protocol, the sources of lead were identified as brass
plumbing fittings and fixtures, and Pb:Sn solder. A study was performed in two iso-
lated sections of the building plumbing system to determine if the lead levels could
be reduced naturally with time by simply using the water. Significant reductions in
lead levels were not achieved following 8 months of water usage. A second study was
performed to evaluate the effectiveness of three chemical corrosion inhibitors: zinc
orthophosphate, calcium orthophosphate, and sodium silicate, to reduce the lead levels.
Three economic, simple, low maintenance chemical feed systems were designed and
installed in three different isolated sections of the building's plumbing system.
The chemicals were fed into the building sections for approximately 4 months. Results
showed that all inhibitors effectively and rapidly reduced lead and copper levels.
17 K£Y WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c, COSATI Field/Group
CORROSION CONTROL
BUILDING CORROSION
LEAD CORROSION
COPPER CORROSION
ORTHOPHOSPHATE BASED CORROSION INHIBITORS
SILICATE BASED CORROSION INHIBITORS
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. Of PAGES
37
20. SECURITY CLASS (This page!
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (R«v. 4-77) previous edition is obsolete
------- |