Review of the Interim Optimal Corrosion Control Treatment for
Washington, D.C.
Final Report
March 30, 2007
Prepared for:
George Rizzo, Work Assignment Manager
U.S. Environmental Protection Agency Region III
1650 Arch Street
Philadelphia, PA 19103-2029
Contract Number 68-C-02-069
Work Assignment Number 47
Prepared by:
The Cadmus Group, Inc.
1600 Wilson Boulevard
Suite 500
Arlington, VA 22209

-------
Acknowledgements
This report was completed for US EPA, Region III, with direction from Rick
Rogers, George Rizzo, and Jennie Saxe. Special thanks is given to Jennie Saxe for
compiling and providing compliance data for analysis and to Rich Giani, Maureen
Donnelly, and Will Keefer of the D.C. Water and Sewer Authority (DCWASA) for
providing additional data.
This report was prepared by The Cadmus Group, Inc., with assistance from its
subcontractor, Dr. Steve Reiber of HDR Engineering.

-------
Table of Contents
1.	Introduction and Background	1-1
1.1	Purpose and Scope	1-1
1.2	Description of the DC Water System	1-1
1.3	History of OCCT Designation	1-4
1.4	Response to Elevated Lead Levels and Change in OCCT	1-5
2.	Summary of Research Relevant to the D.C. Lead Issue	2-1
2.1	Desktop Corrosion Control Study	2-1
2.2	Lead Pipe Scale Analysis	2-2
2.3	DCWASA Circulation Pipe Loop Studies	2-4
2.3.1	Purpose of Study	2-4
2.3.2	Summary of Key Findings	2-5
2.3.3	DCWASA Pipe Loop Protocol	2-5
2.3.4	Results	2-9
2.4	Washington Aqueduct Flow-Through Pipe Loop Studies	2-17
2.4.1	Purpose of Study	2-17
2.4.2	Summary of Key Findings	2-17
2.4.3	Washington Aqueduct Pipe Loop Protocol	2-18
2.4.4	Results	2-21
2.5	Studies Related to Partial Lead Service Line Replacement	2-30
2.5.1	Purpose of the Studies	2-31
2.5.2	Summary of Key Findings	2-32
2.5.3	DCWASA Pipe Cutting Study	2-32
2.5.4	EPA/HDR Galvanic Corrosion Study	2-34
3.	Review of Relevant Water Quality Data	3-1
3.1	Description of Dataset	3-1
3.2	Orthophosphate Treatment and WQP Monitoring	3-4
3.2.1	Orthophosphate Levels	3-5
3.2.2	pH Levels	3-9
3.2.3	Nitrite, Nitrate, and Free Ammonia as Nitrogen	3-9
3.3	Results from Lead Profiling	3-15
3.3.1	Lead Profiling Procedure	3-15
3.3.2	Lead Profiling Results	3-17
3.4	LCR Monitoring Data	3-28
3.4.1	Description of LCR Dataset	3-28
3.4.2	LCR Monitoring Results	3-30
3.5	Analysis of Bacteriological Activities in the Distribution System	3-37
3.5.1	Total Coliforms	3-39
3.5.2	Heterotrophic Plate Count (HPC) Results	3-40
3.5.3	Nitrification Parameter Monitoring	3-44
IOCCT Review
Final Report
l
March 2007

-------
4.	Conclusions and Recommendations	4-1
5.	References	5-1
Appendices
Appendix A: IOCCT Designation Letter, USEPA 2004
Appendix B: Oxidant/Disinfectant Chemistry and Impacts of Lead Corrosion (Schock
and Giani 2004)
Appendix C: Galvanic Corrosion and Grounding Effects Study (Reiber 2006)
Appendix D: Lead Profile Results
Appendix E: Technical Memorandum from Dr. Anne Camper, July 30, 2004
IOCCT Review
Final Report
ii
March 2007

-------
Acronyms
AL
Action Level
AWWA
American Water Works Association
CCPP
Calcium Carbonate Precipitation Potential
CDCP
Centers for Disease Control and Prevention
CFU/mL
Colony-forming units per milliliter
DBP
Disinfection Byproduct
DC.
Washington, D.C.
DCWASA
D.C. Water and Sewer Authority
DOH
D.C. Department of Health
EC
Electrochemical
EPA
U.S. Environmental Protection Agency
HPC
Heterotrophic Plate Count
IOCCT
Interim Optimal Corrosion Control Treatment
LCR
Lead and Copper Rule
LSL
Lead Service Line
MCL
Maximum Contaminant Level
MPY
Mils per year
mg/L
Milligrams per liter
NE
Northeast Quadrant of D.C.
NW
Northwest Quadrant of D.C.
OCCT
Optimal Corrosion Control Treatment
ORD
U.S. EPA Office of Research and Development
ORP
Oxidation Reduction Potential
PL SLR
Partial lead service line replacement
ppb
Parts per billion
ppm
Parts per million
SDWA
Safe Drinking Water Act
SE
Southeast Quadrant of D.C.
SW
Southwest Quadrant of D.C.
TCR
Total Coliform Rule
TDS
Total Dissolved Solids
TOC
Total Organic Carbon
TTHM
Total Trihalomethanes
TEWG
Technical Expert Working Group
U.S. EPA
U.S. Environmental Protection Agency
WA
Washington Aqueduct
WQP
Water Quality Parameter
WQTC
Water Quality Technology Conference
WTP
Water Treatment Plant
IOCCT Review
Final Report
iii
March 2007

-------
This page intentionally left blank.

-------
1. Introduction and Background
1.1	Purpose and Scope
The purpose of this document is to provide a technical review of the effectiveness
of the Interim Optimal Corrosion Control Treatment (IOCCT) for Washington D.C.
Specifically, we reviewed findings from various research studies as they relate to
corrosion control treatment. A significant part of this report is a review of water quality
data collected from the distribution system from January 2003 through December 2005.
Reduction of lead concentrations in drinking water is the primary measure of
corrosion control treatment success. Reduced microbial activity in the distribution
system is a second potential benefit of the orthophosphate treatment implemented in
August 2004 and is assessed in this report.
It is important to note that this report evaluates all water quality data available,
including data not used to calculate compliance with drinking water standards.
Therefore, analyses are not meant to evaluate compliance with Safe Drinking Water Act
(SDWA) regulations.
1.2	Description of the DC Water System
The Washington Aqueduct (WA), the D.C. Water and Sewer Authority
(DCWASA), and U.S. EPA Region III are responsible for providing safe drinking water
to D.C. residents. Owned and managed by the U.S. Army Corps of Engineers, WA
draws water from the Potomac River, treats the water, and sells it to three consecutive
systems: DCWASA, the City of Falls Church, VA, and Arlington County, VA. WA
provides all water treatment; no additional treatment is provided by DCWASA, Falls
Church, or Arlington. DCWASA, a private, semi-autonomous municipal utility,
distributes drinking water throughout all of D.C. EPA Region III in Philadelphia, PA is
the primacy agency for D.C.'s water system and thus provides regulatory oversight and
system supervision for both WA and DCWASA. A service map of the D.C. water
distribution system is provided in Exhibit 1.2.1.
WA treats Potomac River water at two plants, Dalecarlia and McMillan.
Dalecarlia and McMillan perform pre-sedimentation, coagulation/flocculation, primary
disinfection with free chlorine, and secondary disinfection with chloramines. The plants
use alum for coagulation and add polyaluminum chloride as a filtration aid. Since
August 23, 2004, WA has added orthophosphate as a corrosion inhibitor at doses
generally above 3.5 mg/L. WA continues to use lime for pH control as part of its
corrosion control treatment regime. Exhibit 1.2.2 shows finished water quality
parameters (WQPs) for 2005 as reported by WA in their Report of Water Analysis for
2005.
IOCCT Review
Final Report
1-1
March 2007

-------
Exhibit 1.2.1 Service Map of DCWASA and its Consecutive Systems
/ WASHINGTON AQUEDUCT
GREAT FALLS DAM AND INTAKES
r,GREAT FALLS
V
\v
MacARTHUR BOULEVARD
RAW WATER CONDUITS ,
CABIN JOHN
BRIDGE & SIPHON
GLEN ECHO
s\
N
LITTLE FALLS RAW WATER'
PUMPING STATION
FAIRFAX COUNTY
/
• $

/
/
FALLS*
CHURCH
ARLINGTON COUNTY

PENTAGON
REAGAN
NATIONAL
AIRPORT "
ALEXANDRIA
/
//
/
/

5 miles
Source: DCWASA (2005).
Exhibit 1.2.2 Water Quality Parameters for Washington Aqueduct Finished Water
(2005) for McMillan and Dalecarlia
Parameter
Units
Treatment
Plant
Average
Maximum
Minimum
Total Organic
Carbon (TOC)
ppm
McMillan
1.8
2.3
1.3
Dalecarlia
1.7
2.5
1.2
Alkalinity
ppm
McMillan
74
100
43
Dalecarlia
86
112
60
pH
pH
McMillan
7.7
7.7
7.7
Dalecarlia
7.7
7.7
7.7
Total Chlorine
mg/L
McMillan
3.7
3.8
3.7
Dalecarlia
3.7
3.8
3.7
Orthophosphate
mg/L
McMillan
3.20
3.34
3.08
Dalecarlia
3.18
3.51
2.92
Source: WA Water Quality Analysis Report for 2005.
(http://wasliingtonaqueduct.nab.usace.amiv.mil/AnnualReports/2005WaterAnalvsisReport.pdf)
IOCCT Review
Final Report
1-2
March 2007

-------
The distribution system under DCWASA's management consists of four pumping
stations, five reservoirs, four elevated storage tanks, 1,300 miles of pipes, 36,000 valves,
and 8,700 hydrants. WA also owns and operates three finished water storage facilities
and one finished water pumping station in D.C. In all, these facilities deliver water to
DCWASA's more than 550,000 customers,1 who consume an average of 120 million
gallons a day.
DCWASA's distribution system is separated by valves into seven major pressure
zones, which vary by elevation and are served by different storage and pumping facilities.
Exhibit 1.2.3 summarizes these pressure zones, the ground elevation served, and
overflow elevation accommodated by each zone.
Exhibit 1.2.3 DCWASA Service Area Information
Service Zone
Ground Elevation Served (ft)
Overflow Elevation (ft)
West of Anacostia River
Low
0-(50) 70
172
1st High
(50) 70-140
250
2nd High
140-210
335
3rd High
210-(330) 350
424
4th High (East and West)
(330) 350 +
485
East of Anacostia River
Anacostia—Low
0-(50) 70
172
Anacostia—1st High
(50) 70-170
258
Anacostia—2nd High
170 +
382
Source: DCWASA distribution system map
At 150 years old, the D.C. water distribution system is comprised of pipelines of
widely varying age and composition. While pipelines range in age from 30 years to over
a century, most transmission and distribution lines date to the first half of the twentieth
century. Distribution mains and lines range in size from 4 to 54 inches in diameter. Most
of DCWASA's distribution system is constructed from cast iron pipe (87%), ductile iron
pipe (8%), steel pipe (2.5%), and pre-stressed concrete pipe (2.5%). The predominance
of iron pipe in the distribution system is an important factor to be considered in reviewing
D.C.'s corrosion control treatment. Iron pipe has been associated with bacteriological
growth, and high bacterial counts have been a frequent problem for the D.C.
While most of D.C.'s large water mains consist of iron, service pipes to customers
are composed primarily of lead, copper, and brass. Following the Lead and Copper Rule
(LCR) lead action level (AL) exceedance reported in 2002, DCWASA was required by
EPA to replace 7% of its lead service line (LSL) inventory and to revise its inventory of
lines of unknown composition each year, at least until it is at or below the LCR lead AL.
In 2004, DCWASA committed to replacing all LSLs under its control by the year 2015.
In many cases, DCWASA is limited to replacing the portion of the service line that it in
the public space between the water main and the property line. The portion of the line
1 Based on the 2005 U.S. Census estimate of D.C.'s population. Does not include Virginia customers.
IOCCT Review
Final Report
1-3
March 2007

-------
from the property line to the building is in private space and owned by the customer.
Although DCWASA initiated an aggressive customer participation program, the
proportion of residents that are simultaneously paying to have their portion of the LSL
replaced is low.
1.3 History of OCCT Designation
In 1991, EPA promulgated the LCR to reduce lead exposure via tap water, setting
an AL for lead of 15 parts per billion (ppb) based on the 90th percentile of all tap water
samples taken by a water system. In other words, the Federal AL is exceeded if more
than 10 percent of tap water samples contain more than 15 ppb of lead. An exceedance
of the lead AL triggers requirements for additional monitoring, public education, LSL
replacement, and corrosion control treatment until lead levels return to below the AL.
Corrosion control treatment embodies the method or group of methods used by a
water system to prevent tap water from corroding metals such as lead and copper from
distribution pipes. Because each system differs in water chemistry, treatment regime, and
pipe makeup, its "optimal" corrosion control treatment (OCCT) is unique and can be
determined by assessing several WQPs such as pH and alkalinity.
Upon promulgation of the LCR, all large water systems—including D.C.'s—were
required to conduct corrosion control studies regardless of their LCR compliance status.
(In contrast, small and medium systems were typically only required to initiate CCT
research if they exceeded the AL). Based on these studies, the primacy agency would
then approve OCCT for a system, and that system was required to operate within a
specified range of WQPs to maintain optimized corrosion control. If a system, after
implementing OCCT, again exceeds the AL, it may have to revisit its corrosion control
strategy and implement new treatment. This was the case for D.C.
Since the promulgation of the LCR, WA and DCWASA have used pH adjustment
to control the corrosion of lead. EPA Region III first submitted official approval for
interim OCCT in 1997, on condition that WA and DCWASA conduct further study of
potential corrosion control treatment methods for the D.C. system. High pH levels can
result in the formation of less soluble lead compounds, meaning that they will not
dissolve into the drinking water. Historically, WA has added small amounts of lime
(calcium oxide) to the water to maintain a high pH and thereby control lead corrosion.
Corrosion control studies conducted by WA and its contractor recommended that lime be
used to maintain a pH range of 7.4 to 8.5. While one EPA contractor recommended that
WA try to achieve the highest pH possible, excessive application of lime results in the
accumulation of calcium carbonate.
EPA's original IOCCT approval depended on WA and DCWASA's commitment
to further investigate sodium hydroxide (caustic soda) for pH adjustment as a corrosion
control option. The use of caustic soda would allow them more leeway in raising water
pH without exceeding the Total Trihalomethane Maximum Contaminant Level (TTHM
IOCCT Review
Final Report
1-4
March 2007

-------
MCL)2 and without causing excessive precipitation of calcium carbonate. As part of this
investigation, WA was required to evaluate the costs as well as economic impacts of pH
adjustment strategies, as well as the feasibility and costs of introducing a non-zinc
orthophosphate corrosion inhibitor. After two studies were submitted reviewing caustic
soda application as well as corrosion inhibitors, it was recommended that D.C. continue
using pH adjustment with lime as its OCCT. In 2000, EPA granted OCCT approval for
pH maintenance using lime, later modifying pH goals for OCCT in 2002. Exhibit 1.3.1
summarizes the corrosion control studies and actions undertaken by DCWASA, WA, and
EPA Region III from 1994 to 2002 to determine the optimal treatment strategy.
Exhibit 1.3.1 Corrosion Control Treatment History for D.C. (1994-2002)
Action
Date
Corrosion Control Study (DCWASA) recommends pH adjustment using
lime for OCCT
June 1994
EPA Region III Approves interim OCCT of pH adjustment w/ lime
December 1996
Caustic Soda Feasibility Study
January 1998
Corrosion Inhibitor Study for Dalecarlia and McMillan Treatment
Plants
May 1998
EPA Grants OCCT for pH adjustment
February 2000
WA switches to chloramines
November 2000
EPA Adjusts Approved OCCT pH Range
May 2002
1.4 Response to Elevated Lead Levels and Change in OCCT
Lead levels in D.C. tap water remained low through 2001 until DCWASA
reported an exceedance for the monitoring period of July 2001 to June 2002. The 90th
percentile level of lead went from 8 ppb the previous monitoring period to 75 ppb, with
more than half the samples exceeding the AL. The following monitoring period, ending
June 30, 2003, DCWASA again exceeded the AL (40 ppb), suggesting that the elevated
lead levels represented a trend in the D.C. distribution system. To evaluate potential
causes of the elevated lead levels and identify potentially useful research approaches,
EPA commissioned an evaluation by Virginia Tech professor Marc Edwards. Professor
Edwards is a national expert on corrosion of drinking water system materials.
Professor Edwards reviewed the available research on lead corrosion, analyzed
DCWASA and WA's historical water quality data, and evaluated sampling protocols.
Based on this work, he identified the switch to chloramines as the probable cause of
corroded lead in the distribution system, noting that samples taken during the July 2000-
June 2001 monitoring period may have possibly under-represented the effect of
chloramine on lead levels during the time the switch took place. Dr. Edwards also found
that elevated nitrates in chloraminated water may exacerbate lead corrosion, while greater
turbidity may affect particulate lead release.
2 A type of DBP, TTHMs form in highly basic conditions or when lime exists in the system.
IOCCT Review
Final Report
1-5
March 2007

-------
Dr. Edwards issued several recommendations:
1.	Compare sampling protocols for the seven utilities treating Potomac River
water;
2.	Initiate a corrosion study of brass, pure lead, and lead-solder coupled to
copper;
3.	Conduct an analysis to help determine the source of particulates;
4.	Conduct a nitrification study of DCWASA and other systems obtaining water
from the Dalecarlia Plant at different times in the year;
5.	If possible, research any relationship between the switch to chloramines and
zinc, copper, and lead loads in the sewage treatment plant to characterize
when the lead problem started; and
6.	Examine whether if the annual switch from chloramine to chlorine (the
"chlorine burn") is detrimental to lead control. Examine the basic rationale for
switching disinfectants.
In addition, Dr. Edwards recommended that WA initiate a pipe loop study
simulating the distribution system and using pipe extracted from the system in order to
better study lead corrosion in D.C. See Chapter 2 for further discussion of WA's pipe
loop research. Dr. Edwards also worked with DCWASA to develop a method for "lead
profiling." Section 3.3 of this report details the lead profile procedure and DCWASA's
use of it to analyze lead at several homes before, during, and after the corrosion control
treatment changes.
As EPA and DCWASA continued to investigate reasons for the exceedance and
revisit possible corrosion control options, another AL exceedance reported in December
2003 (63 ppb) confirmed the need to address this growing problem. In January of 2004,
the Technical Expert Working Group (TEWG) convened to facilitate comprehensive
research toward a solution to the lead problem in the D.C. distribution system. Since the
TEWG's inception, its members (DCWASA, WA, EPA, and other stakeholders) have
conducted various studies to determine the source of the lead problem and to identify a
solution. These studies are summarized and discussed in Chapter 2 of this report. Based
on the results of this research, the TEWG recommended that orthophosphate treatment be
initiated as soon as possible to reduce lead levels in drinking water.
Following a successful partial system application, which began in D.C.'s 4th High
Service Area in June 2004, EPA Region III officially approved a system-wide application
of orthophosphate in a letter dated August 3, 2004. The interim OCCT designation letter
(Appendix A) established interim requirements for various WQPs related to
orthophosphate treatment. According to the letter, EPA will revise WQPs once it
determines that the D.C. distribution system has been passivated.
IOCCT Review
Final Report
1-6
March 2007

-------
2. Summary of Research Relevant to the D.C. Lead Issue
As noted in the previous chapter, the TEWG was formed in early 2004 in
response to the elevated lead levels in D.C. drinking water. The TEWG consists of
representatives from EPA, WA, DCWASA, the D.C. Department of Health (DOH), the
Centers for Disease Control and Prevention (CDCP), Arlington County, Virginia, and
Falls Church, Virginia. Its primary mission was to develop a plan to reduce the
corrosivity of treated water in D.C. In their Action Plan (U.S. EPA 2004) the TEWG
identified the following seven priorities for their work:
1.	Communicate actions and progress to the community on a regular basis;
2.	Choose a revised OCCT based on desktop analysis and verified through
partial system application and DCWASA and WA pipe loop studies;
3.	Consider demonstration of revised OCCT in a partial system test;
4.	Leave open the possibility of immediate full system implementation;
5.	Obtain EPA interim and final approval of selected re-optimization of
corrosion control treatment;
6.	Execute full system operations; and
7.	Use ongoing pipe loop studies to refine chemistry and determine the cause of
the elevated lead levels.
A complete copy of the TEWG action plan is available on-line at
http://www.epa.gov/region03/Action Plan to Reduce Pb 3 10 04.pdf.
The TEWG has accomplished the first six of the priorities listed above. They also
made significant steps in identifying the cause of the elevated lead levels, although
research is ongoing. The group continues to coordinate research and meet on a regular
basis to evaluate and refine the corrosion control treatment.
The purpose of this chapter is to provide a brief summary of the research done by
various TEWG members to identify the cause of elevated lead levels and select the best
corrosion control treatment option for D.C. For each section, we highlight key findings
of each research effort. The reader is referred to other reports (or they are included by
way of Appendices) for detail on the various study protocols and other findings.
Conclusions based on the combined research findings and water quality data are
presented in Chapter 4 of this report.
2.1 Desktop Corrosion Control Study
In April of 2004, the Washington Aqueduct and its contractor, CH2MHILL,
completed a Desktop Corrosion Control Study (WA and C2HMHILL, 2004). The study
summarized findings of recent engineering reports dealing specifically with corrosion
control treatment and alternative methods of pH and alkalinity control. The study
reviewed the feasibility of a wide range of corrosion control options including pH and
IOCCT Review
Final Draft
2-1
March 2007

-------
alkalinity adjustment, calcium adjustment to precipitate a layer of calcium carbonate
(CaC03), phosphate inhibitor addition, and silicate inhibitor addition. A copy of the full
desk top study report is available on-line at
http://www.epa.gov/dclead/CorrosionControl.pdf.
Key findings of the desktop study are summarized below.
•	Although an earlier report suggested adjusting pH to 8.8 for corrosion control,
subsequent studies and mathematical modeling results concluded that
adjustment of finished water pH by 8.5 or greater using the existing lime
treatment would cause excessive precipitation of calcium carbonate in
DCWASA's distribution system. This is a very undesirable side effect,
reducing the carrying capacity of pipes and causing water to have a white or
cloudy appearance. Thus, pH control to 8.5 or greater is not a viable
corrosion control option.
•	Based on solubility models and experience of other similar systems,
phosphoric acid appeared to be the best corrosion inhibitor for D.C.
2.2 Lead Pipe Scale Analysis
By characterizing the nature of the lead scales, TEWG researchers gained
important insights into understanding how the corrosion scales formed and why lead
began dissolving into the water after the conversion from free chlorine to chloramines.
When corrosion occurs (i.e., a metal undergoes chemical oxidation), oxidized
metal either goes into the solution or segregates itself into a different mineral form on the
corrosion surface. The product of this segregation is generally referred to as a corrosion
scale. Under aqueous chemistry typical of drinking water systems, almost all corroding
metals will form a corrosion scale, and, frequently, this corrosion scale will consist of
multiple mineral forms of the same oxidized metal. When corrosion scale forms, the pipe
can lose structural integrity and, if enough scale accumulates, lose carrying capacity.
Corrosion scales themselves can undergo dissolution, releasing oxidized metal into the
water.
To understand the nature of the lead corrosion scale in DCWASA's system,
researchers from EPA's Office of Research and Development (ORD) performed x-ray
diffraction on extracted LSLs from the D.C. distribution system. The majority of tests
discussed here were performed in early 2004. A description of the lead scale analysis and
detailed discussion of findings was presented at a special workshop held during the
American Water Works Association (AWW A) Water Quality Technology Conference
(WQTC) in November, 2004 (Schock and Giani, 2004). A copy of the paper from this
conference is included as Appendix B to this report for reference.
IOCCT Review
Final Draft
2-2
March 2007

-------
In summary, the x-ray diffraction method involves first scraping the inside of the
LSL samples to remove the corrosion scale and then bombarding ground scale material
with a series of x-rays. The diffraction patterns of the x-rays are analyzed to identify the
mineral content of the scale. Other analytical methods were used to confirm the x-ray
diffraction findings. (For a description of other methods, see Appendix B.)
Results of several analyses showed that the corrosion scale on LSLs in the
DCWASA distribution system consists primarily of lead oxide (PbCte) compounds,
specifically plattnerite and scrutinyite. The presence of lead oxide scales is associated
with water of consistently high oxidation reduction potential (ORP). A high ORP can be
caused by several mechanisms, one of which is very high levels of free chlorine. WA
and DCWASA had historically maintained a high free chlorine residual for biofilm
control. The working theory developed by ORD researchers based on pipe scale analysis
is that when the ORP level dropped with the conversion from free chlorine to
chloramines, the lead corrosion scale began slowly dissolving into the water.
There has been considerable research into the occurrence and chemistry of lead
oxide scales since the paper by Schock and Giani was presented at the AWWA WQTC
workshop in 2004 in support of these findings (Giani, Donnelly et al., 2005; Korshin
2005; Lytle and Schock 2005; Schock 2005; Vasquez et al. 2006). Many studies
continue today.
The DCWASA experience has also spurred the development and application of
new analytical tools. One tool now being applied is Raman Spectroscopy, which can
simultaneously assess both the topology and crystallography of corrosion scales. Exhibit
2.2.1 is a Raman spectrograph of a DCWASA LSL corrosion scale destabilized by
exposure to chloramines (Maynard and Mast, 2005).
IOCCT Review
Final Draft
2-3
March 2007

-------
Exhibit 2.2.1 Raman spectrograph of a transforming LSL showing the
spatial distribution of Pb(IV) and Pb(ll) mineralogy
Hydrocerussite
PbO Massicot
PbO Litharge
400 -1Q200 ¦ 10000 -9800 -9600 -9400 -9200
Source: (Maynard and Mast, 2005).
2.3 DCWASA Circulation Pipe Loop Studies
In early 2004, DCWASA initiated a set of bench/pilot scale studies to assess lead
(and in some cases copper) corrosion rates and metals release. The combination of
bench/pilot scale studies consists of two circulation loop testing apparatus:
1)	Stagnation loop testing focusing primarily on metal release from relatively
long LSL sections.
2)	Electrochemical (EC) pipe loop testing focusing on corrosion rates and EC
parameters measured on short sections of LSL and domestic copper tubing.
These complementary analyses measure corrosion rates and metal release under
simulated distribution system conditions. Study protocols and test conditions were
designed to overlap such that one set of tests could provide confirmatory evidence for the
others. Both protocols tested similar water quality regimens.
The studies were performed at DCWASA's Fort Reno facilities beginning in
March 2004, and have continued with minor modifications since that time (Reiber et al.
2004).
2.3.1 Purpose of Study
The purpose of these studies was to quickly screen several potential corrosion
control strategies. A secondary objective of the testing program was to help determine the
IOCCT Review
Final Draft
2-4
March 2007

-------
cause of the lead corrosion problem. After selection and implementation of the new
corrosion control treatment for D.C. in August 2004, the loops have been used to evaluate
the impact of changes in water quality on lead release and continue to be used today.
Because a final report of findings is not currently available, all discussions and
conclusions are based on interim reports from DCWASA.
The focus of this testing is primarily on LSLs, although copper tubing was also
included in the program to provide an EC reference point and to ensure that any adverse
potential corrosion impacts on other plumbing surfaces were fully investigated. The LSL
test specimens were pipe sections that had been removed from the DCWASA system and
reflected the history and scaling conditions of in-situ DCWASA LSLs. The basis for
much of the bench and pilot scale testing was a relative comparison, meaning that a
baseline condition (corrosion rate or metal release rate) was established; then the water
chemistry is changed, and the effect on the test parameter is observed over time.
2.3.2	Summary of Key Findings
•	Pipe loop data (i.e., total lead levels after stagnation) correlated well with the
maximum lead concentrations observed in the lead profiles (lead profile results are
presented in Section 3.3), indicating that the circulation studies are a useful tool for
predicting optimal corrosion conditions in the DCWASA system.
•	Multiple pipe loops operated for 2 years show that orthophosphate is highly effective
at controlling corrosion of LSL surfaces in the D.C. system.
•	Other corrosion control strategies, such as raising the pH with lime and using
ortho/polyphosphate blends, were not successful in reducing lead leaching from
DCWASA LSLs.
•	Adjustments in orthophosphate dose and changes in oxidant type (from chloramines
to free chlorine back to chloramines) may have a very slight impact on lead release,
although it is too early to make firm conclusions.
2.3.3	DCWASA Pipe Loop Protocol
EC testing was performed in accordance with the methodology for "Pipe Section
Flow Cells" contained in "Internal Corrosion of Water Distribution Systems," second
edition (AwwaRF, 1996). The electrochemistry testing protocols were considered useful
because they estimate the actual corrosion rate (electron exchange) on the test specimens,
which in turn provides an indication of how water quality conditions are influencing
overall corrosion conditions. They are, however, very sensitive to changing redox
potential, and can vary substantially as disinfectant concentrations vary, as was the case
in this testing protocol. It is important to note that EC measurements of corrosion rate are
a crude indicator of the corrosion processes occurring on the test specimens.
IOCCT Review
Final Draft
2-5
March 2007

-------
The text, "Internal Corrosion of Water Distribution Systems" also contains a
description of pipe-loop stagnation testing of the type carried out for DCWASA. While
the principal advantages of the EC approach are its speed and ease of measurement, the
stagnation loops are more time-consuming, and require substantially more effort, but
yield a result that is more reflective of actual lead release.
A total of 12 circulation loops, comprising six EC loops and six stagnation loops,
were constructed to test the various corrosion control strategies. Exhibit 2.3.1 is a
photograph of the EC pipe loops. Exhibit 2.3.2 is a schematic of the stagnation loops.
At the beginning of the study, all pipe loops were run with distribution system
water (with chloramines) prior to testing in order to condition the pipe specimens and to
establish a baseline metal release rate. After the conditioning period, the sample water
was dosed with the appropriate chemical regime. The stagnation loop test consisted of 8
hour stagnation periods, representing conventional operation of in-service lines, followed
by a 16-hour circulation period. EC loop tests consisted of a 24 hour circulation. Water
quality analyses were performed for multiple parameters, as shown in Exhibit 2.3.3.
IOCCT Review
Final Draft
2-6
March 2007

-------
Exhibit 2.3.1 Photograph of Electrochemical Polarization Cell Used in the EC Pipe
Loop System
Source: EPA. http://www.epa.gov/dclead/corrosion.htm
Source: Reiberet al. (2004).
Exhibit I
2 Stagnation Loop Schematic
IOCCT Review
Final Draft
2-7
March 2007

-------
Exhibit 2.3.3 Water Quality Parameters Evaluated During the DCWASA Study
WQP
EC Loop Testing
Stagnation Loop Testing

Pre-24 hr
circulation
Post-24 hr
circulation
Pre-8 hr
stagnation
Post-8 hr
stagnation
Post 16 hr
circulation
PH
V
V
V
V
V
temperature
V
V
V
V
V
alkalinity
V
V
V

V
hardness
V

V


free chlorine
/
V"
V"
V"
/
total chlorine
V
V
V
V
V
Total dissolved
solids (TDS)
V

V


ORP

V


V
lead

V

V
V
copper

V



phosphate (as
applicable)

V"


/
Source: Reiber et al. (2004).
Although the testing conditions were modified as the testing proceeded, the
original strategies evaluated in both types of loops are summarized in Exhibit 2.3.4.
These were the original lead mitigation strategies, and the first six months of testing
focused largely on establishing their relative effectiveness. As it became clear that certain
strategies were going to be ineffective relative to lead release, the equipment devoted to
these efforts was reassigned to explore other options. Testing beyond the original scope
included evaluation of corrosion mitigation strategies as well as the impact of various
distribution system operational conditions. Some of these additional testing efforts
included:
•	Application of stannous chloride as a supplemental inhibitory in conjunction
with phosphoric acid;
•	Evaluation of periodic, short-term free chlorine "burn out" on lead release;
•	Assessment of changes in orthophosphate dose.
These and other strategies and operational criteria continue to be tested as needed
by DCWASA
IOCCT Review
Final Draft
2-8
March 2007

-------
Exhibit 2.3.4 Original Corrosion Control Strategies Tested by DCWASA
Strategy
Description
Rationale
1
Chloraminated (3.5 mg/L)
finished water without any
additional chemical treatment.
Control. Represents the finished water
discharged from the Dalecarlia Plant.
2
Add lime to raise the calcium
carbonate precipitation to 1
mg/L, maintain chloramines
concentration at 3.5 mg/L
Initiate the development of supposedly protective
calcite scales.
3
Add monosodium phosphate
(MSP) at 10 mg/L, adjust pH to
7.5-7.8
Test effectiveness of orthophosphate under high
chloramine residual. Given the short duration of
the testing, a high phosphate dose is necessary in
order to rapidly passivate the lead surfaces and
assess whether phosphate has likely value as a
corrosion inhibitor.
4
Add 50/50 blended
ortho/polyphosphate at a total
phosphate dose = 1-2 mg/L.
Test the effectiveness of polyphosphates.
5
Add monosodium phosphate
(MSP) at 10 mg/L, adjust pH to
7.5 - 7.8. Decrease chloramine
residual to 1 mg/L.
Test the effectiveness of strategy 3 under low
chloramine conditions.
6
Add lime to raise the calcium
carbonate precipitation to
lmg/L, reduce chloramine
concentration to 1 mg/L.
Test the effectiveness of strategy 2 under low
chloramine conditions.
Source: Reiber et al. (2004).
2.3.4 Results
A factor that helped make the DCWASA pipe loop testing a useful assessment
tool is that it was demonstrated early on that the test rigs were able to simulate the lead
release process from LSLs in the distribution system. Lead levels measured in the pipe
loops closely approximated the highest lead levels recorded during the household lead
profiling events (generally greater than 100 ppb total lead, see Section 3.3 for detailed
results of DCWASA's lead profiling program). This relatively high initial level of lead
made it easier to do comparative evaluations of the various test strategies.
Using these test protocols, DCWASA collected a substantial body of data relative
to the various mitigation strategies. One of the most important findings of this study has
been the reduction in corrosion rate and lead release resulting from application of
orthophosphate. Exhibit 2.3.5 displays a substantial amount of pipe loop data from
March 2004 through March 2006 for four stagnation loops treated with orthophosphate.
IOCCT Review
Final Draft
2-9
March 2007

-------
All loops show substantial and continuous reductions in lead leaching resulting from the
orthophosphate treatment. Results from the EC loops were similar and showed a
comparable although slightly delayed initial reduction in EC corrosion rate with the
orthophosphate treatment. Exhibit 2.3.6 is an example of the corrosion rates measured for
Pipe Loop Number 1.
A second important finding regarding the screening of corrosion control strategies
is that other corrosion strategies did not reduce lead release rates under chloramines
conditions. Exhibit 2.3.7 shows total lead concentration for Pipe Loop 2, which was
treated with lime to raise the calcium carbonate precipitation potential. This potential
strategy did not appear to have any impact on lead release. Exhibit 2.3.8 shows total lead
concentration for Pipe Loop 4, which was initially fed a 50/50 ortho/polyphosphate blend
at between 1 and 2 mg/L. Lead release in this loop fell slightly, but very little in
comparison to the orthophosphate loops. In 2006, Stannous Chloride was tested in Pipe
Loop 5. Preliminary results show an increase in total and dissolved lead concentrations.
In addition to evaluating inhibitors and other mitigation strategies, the DCWASA
bench and pilot-scale efforts were useful in evaluating operational and dosage regimen
criteria. In Pipe Loop 3, DCWASA switched to free chlorine in late November 2005 to
simulate a free chlorine burn then switched back to chloramines in the first part of
January 2006. Lead levels exhibited a very slight increase, although conclusions are
difficult due to the inherent variability in the lead data. Pipe Loops 1,3, and 6 have
recently undergone a reduction in orthophosphate dose from 3.5 mg/L for 2.5 mg/L.
Exhibit 2.3.5 shows a very slight increase in lead concentration following this change in
all loops, but there are too few data points to draw any firm conclusions regarding the
potential impact of this change.
IOCCT Review
Final Draft
2-10
March 2007

-------
Exhibit 2.3.5 Results from Four DCWASA Stagnation Loops Showing Reduction in Lead Release as a Function of Orthophosphate
Treatment
Pipe Loop 1 Final (Control Loop)
3.5 - 4.0 mg/L Chloramines
2.5 mg/L Orthophosphate
100
90
80
—
3.5 - »• ^ v cy >v cy -w <«•  ^ «•
70
•Q
Q.
Q.
r 60
C
01
O
c
o
O
¦a
ro
at
50
40
30
20
10
Date
Source: Distributed to the TEWG by DCWASA.
IOCCT Review
Final
2-11
March 2007

-------
Exhibit 2.3.5 (continued)
Pipe Loop 3 Final
100
3.5 mg/L Chloramines
2.5 mg/L Orthophosphate
3.0 - 4.0 mg/L Chloramines
3.5 mg/L Chloramines
2.5 mg/L Orthophosphate
1.5 mg/L MSP Addition
~ Tota Lead
Fi tered Lead
3.5 - 4.0 mg/L Chloramines
10 mg/L Orthophosphate
4.0 mg/L Free Chlorine
3.5 - 4 q mg/L
Free Chlorine
< >;<
3.5 - 4.0 mg/L Chloramines
10 mg/L Orthophosphate
3.0 r 3.5 mg/L
pH = 8.00
orine
10 mg/L Orthophosthate
10 mg/L
Ortljophosphata
3.0 - 4.0 mg/L Chloramines
3.5 mg/L Orthophosphate
'V A 'V A NN 
y 
-------
Exhibit 2.3.5 (continued)
3.5 mg/L Chloramines
Pinploon R Final	2.5 mg/L Orthophosphate
ripeioop 0 rma^ 5 mg/L Ch|oramineS	Stannous Chloride Addition
3.0 - 4.0 mg/L Orthophosphate
Stannous Chloride Addition 12/7/03
3.5 mg/L Chloramines
2.5 mg/L Orthophosphate
1.5 mg/L Stannous Chloride
Addition
3.5 - 4.(j mg/L
Free Chlorine
1 mg/L Chloramines
10 mg/L Orthophosphate
~ ~

* \

~ ~
~ ~~ VT
<
!	~~.Art*
0 -I—:	1	j	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1 i i i i i i i i i i i i i . i	i i -r^^i i
<°--0'cv-(yK>-(ycv-cv^-CVv-"0'^
^VI^^^^*svc\>	^  *s *s
*v *v	is.	is.	*s.	*s.
4/5/04 5/1'2/04
• I I I I	I I I I
T *"	TOO/^
11/21/05	2/17/06
Source
Distributed to the TEWG by DCWASA
Date
IOCCT Review
Final
2-13
March 2007

-------
Exhibit 2.3.5 (continued)
350
300
Pipe Loop 6 Final
Lime
1 mg/L Chloramines
~ Total Lead
I Filtered Lead
3.5 mg/L Chloramines
2.5 mg/L Orthophosphate
3.5-4.
Free Cjp
l
¦ 10 mg/L
P mf"- 5 mg/L i phinrin® ^^ rng/L Free Chlorine
Drlne Frie Crflorine	10 mg/L Qrthophosphate
»**< >!	to v cv> J rw- ^	J  A^'n> V ^	Q* <£ Q*
QJ |N \N| 9*	^ ts. ^ K-V ^
I	I''
* if
5/20/04 6/24/04
*
10/13/04
Date
,? K
I
^ ^ 
-------
>-
Q.
s
E
0)
CS
a:
c
o
"t7>
2
L.
o
o
Exhibit 2.3.6 Change in Corrosion Rate (Measured in mils per year, MPY) for EC Loop Number 1
3.5-4.0 mg/L
Free Chlorine
800
700
600
500
400
300
200
100


<	>'
3.5 - 4.0 mg/L Chloramines
~
<	>
3.5 - 4.0 mg/L Chloramines
3.5 mg/L Orthophosphate





~


~	Coupon 1a
~	Coupon 1a+

~
~
~
~


* ~ '
1 ~~
~ ~
~ ~ . ~ ~
*> ' t

~~
1 ~ *J
t
\ ~ * *~
'~ ~# V

1 ~
J ~ %
% ~
v _ #
~ V4** ^ ~ . ,


*~ 3*7«

xcr or. «J' 
-------
Exhibit 2.3.7 Results from Pipe Loop 2 (March 2004 - July 2004)
/icn

Pipeloop 2 Final




• Pb

400
350
—«300
¦Q
a 250
"g 200
0)
150
100
50
n






I
I


















1
1








•








I
I

















I


















i



















I
S

•
•
%
•
•
•
•
Ki
# I





•


I
r



9
•










i


















I
I











u
"s
C
0
c
UU/Zlt
03/31 -
04/07 -
04/14 -
04/21 -
04/28
05/05 -
05/12 -i
05/19
05/26
06/02
06/09 -
06/16 -
06/23
06/30 -
07/07 -
07/14 -
07/21 -
Source: Distributed to the TEWG by WA.
Exhibit 2.3.8 Results from Pipe Loop 4 (March 2004 - July 2004)
/inn
Pipeloop 4 Final




• Ph

Lead (ppb)
—^ ^ N> N> CO CO J
cnocnocnocnc
DOOOOOOOC
I I I I I I I






I
•



















l




•













I
I


















I
A

•



•









I


















I



• i
* 1

•
r
# «
>









I

















I











c
?
c
03/31 -
04/07 -
04/14 -
04/21 -
04/28
05/05 -
05/12 -
05/19
05/26
06/02
06/09 -
06/16 -
06/23
06/30 -
07/07 -
07/14 -
07/21 -
Source: Distributed to the TEWG by WA.
IOCCT Review
Final Report
2-16
March 2007

-------
2.4 Washington Aqueduct Flow-Through Pipe Loop Studies
This section describes the flow-through lead pipe loop testing conducted by WA
and its contractor, CH2MHH1 at the Dalecarlia Water Treatment Plant. The study began
in January 2005 and complements the DCWASA recirculation loop testing described in
Section 2.3. Several experiments have been completed, but others are ongoing at the
time of this report. Because a final report of findings is not currently available, all
discussions and conclusions are based on interim reports from WA and its contractors.
2.4.1	Purpose of Study
The WA pipe loop study was designed to answer the following four main
questions:
•	Does zinc orthophosphate perform better or worse than orthophosphate?
•	What is the optimum long-term dose of phosphoric acid for the DCWASA
distribution system?
•	What is the approximate timeframe for reducing lead concentrations in the
DCWASA distribution system?
•	Does periodically switching from chloramines to free chlorine have any
impact on lead concentrations? If so, what is the impact and approximately
how long does the effect last?
2.4.2	Summary of Key Findings
Based on preliminary results presented by WA and CH2MHill to the TEWG, key
findings of the WA pipe loop study are as follows:
•	Orthophosphate was very effective at reducing lead release from LSLs.
•	The study found no additional benefit of zinc orthophosphate over
orthophosphate (added as phosphoric acid). In fact, addition of zinc may serve
to destabilize corrosion scales.
•	Lead release appears to be sensitive to changes in orthophosphate dose.
•	A temporary change to free chlorine then back to chloramine may cause slight
increases in metal release from LSLs.
Most findings of the WA study were similar to the DCWASA study, although
there are some differences that warrant mention. Overall lead levels were lower in the
WA loops, although the particulate portion of the lead was higher compared to
DCWASA circulation loop study results. Also, the operational time needed to observe
meaningful lead reduction was longer for the WA loops, although this may be attributable
to the high initial dose of 10 mg/L used by DCWASA to accelerate results. Other
differences may be attributable to different loop fabrication techniques or handling of the
LSL specimens prior to fabrication, as discussed in the next section.
IOCCT Review
Final Report
2-17
March 2007

-------
2.4.3 Washington Aqueduct Pipe Loop Protocol
WA uses finished (i.e., potable) water as source water for testing. Additional
treatment chemicals are added to the filtered water flow stream to "simulate" finished
water quality under a variety of conditions. All pipe loops in the test are operated with
chloramines or a combination of free chlorine and chloramines (i.e., no free chlorine pipe
loops were used).
The WA pipe loops were initially tested under a total of seven different operating
conditions (i.e., racks), as defined in Exhibit 2.4.1. Three replicate pipe loops were
provided in each of the seven racks to ensure that the test results were statistically
significant and reproducible. Twenty-one lead pipe loops were provided in all. One rack
(set of three loops) was assigned to evaluate lead release associated with the finished
water produced at the Dalecarlia WTP. No additional chemical conditioning of the
finished water was performed prior to testing this flow stream. This set of loops served as
a control for the study.
Salvaged LSLs excavated from the DCWASA distribution system were used to
construct the pipe loops. Note that this is the same pool of LSL pipe sections from which
the DCWASA pipe loops were fabricated. Each loop included two or three separate
sections of LSL, for a total length of 13 feet, of 3/4-inch-diameter pipe. This length was
selected because it yields a total sample volume of 1.1 liters per pipe loop. Exhibit 2.4.2
shows how the pipe loops were constructed using salvaged LSLs from the distribution
system. LSLs used in the WA study were allowed to dry out before the pipe loop study
began. This contrasts with the DCWASA LSL portions, which remained wet.
At initialization, all of the pipe loops were conditioned with finished water for a
period of one month to allow for the scale on the pipes to reach a common baseline.
After the one-month conditioning period, the pipe loops (with the exception of the control
loop) were fed "filtered" water with chemicals added according to the Test Plan
presented in Exhibit 2.4.1.
IOCCT Review
Final Report
2-18
March 2007

-------
Exhibit 2.4.1 Operating Conditions, Objectives and Rationale for the Seven Pipe
Racks used in WA Study
Pipe Rack
Number
Rack Name
Water
Source
Chemicals to be
Added to Water
Chemical Dose (mg/l)
Pipe Rack
PH
Question to be Addressed by this Rack
1
High Chloramines
with Zinc Ortho,
Decrease Zinc
Ortho Dose over
Time
filtered
water
zinc
orthophosphate
3.5 mg/l as phosphate, ramp
down once lead levels drop below
action level
7.7
1.	What dose of zinc orthophosphate should be
used to control lead levels in the distribution
system once the system has been passivated?
What is the lowest effective dose that will still
ensure compliance with the LCR lead action
levels?
2.	How does zinc orthophosphate performance
compare with phosphoric acid (i.e., compare
Rack 1 and 2 results).
sodium hydroxide
as needed for pH control
sodium
hypochlorite
as needed to maintain 3.5 mg/L
chloramine concentration
ammonia
as needed to maintain 3.5 mg/l
chloramine concentration
fluoride
1.0 mg/l
2
High Chloramines
with Phosphoric
Acid, Decrease
Phosphoric Acid
Dose over Time
filtered
water
phosphoric acid
3.5 mg/l as phosphate, ramp
down once lead levels drop below
action level
7.7
1.	What dose of phosphoric acid should be used
to control lead levels in the distribution system
once the system has been passivated? What is
the lowest effective dose that will still ensure
compliance with the LCR lead action levels?
2.	How does zinc orthophosphate performance
compare with phosphoric acid (i.e., compare
Rack 1 and 2 results).
sodium hydroxide
as needed for pH control
sodium
hypochlorite
as needed to maintain 3.5 mg/l
chloramine concentration
ammonia
as needed to maintain 3.5 mg/l
chloramine concentration
fluoride
1.0 mg/l
3
Switch Between
Free Chlorine and
Chloramines with
Constant
Phosphoric Acid
Dose
filtered
water
phosphoric acid
3.5 mg/l as phosphate, no
change overtime
7.7
1.	How are lead levels impacted by periodically
swinging back and forth from free chlorine to
chloramines in the presence of a corrosion
inhibitor?
2.	Does switching disinfectants inhibit the
effectiveness of phosphoric acid for some period
of time? An item to be resolved here involves
whether to initially condition this loop with free
chlorine or chloramines???
sodium hydroxide
as needed for pH control
sodium
hypochlorite
3.5 mg/l +/-, or as needed to
achieve distribution system
microbial goals
ammonia
as needed to maintain 3.5 mg/l
chloramine concentration
fluoride
1.0 mg/l
4
High Chloramines,
No Corrosion
Inhibitor
filtered
water
sodium hydroxide
as needed for pH control
7.7
1.	What lead levels can be expected with
chloramines in the absence of a corrosion
inhibitor?
2.	How do chloramine lead levels compare with
and without orthophosphate (i.e., compare racks
1, 2, and 4)?
sodium
hypochlorite
as needed to maintain 3.5 mg/l
chloramine concentration
ammonia
as needed to maintain 3.5 mg/l
chloramine concentration
fluoride
1.0 mg/l
5
Low Chloramines
with Constant
Phosphoric Acid
Dose
filtered
water
phosphoric acid
3.5 mg/l as phosphate, no
change overtime
7.7
1. How do lower chloramine concentrations
impact lead concentrations in the presence of a
corrosion inhibitor (I.e., compare racks 5 and 6)?
sodium hydroxide
as needed for pH control
sodium
hypochlorite
as needed to maintain 1.0 - 2.0
mg/l chloramine concentration
ammonia
as needed to maintain 1.0 - 2.0 mg/l
chloramine concentration
fluoride
1.0 mg/l
6
High Chloramines
with Constant
Phosphoric Acid
Dose
filtered
water
phosphoric acid
3.5 mg/l as phosphate, no
change overtime
7.7
1. How do lead levels compare if phosphoric
acid concentrations are lowered over time after
passivation versus maintained at a constant
concentration after passivation (I.e., compare
racks 2 and 6)?
sodium hydroxide
as needed for pH control
sodium
hypochlorite
as needed to maintain 3.5 mg/l
chloramine concentration
ammonia
as needed to maintain 3.5 mg/l
chloramine concentration
fluoride
1.0 mg/l
7
Finished Water
Control Rack
finished
water
phosphoric acid
full scale plant dose during test
period (3.5 mg/l dose anticipated)
7.7
1.	Control loop - finished water conditions during
the pipe loop test period.
2.	Lead-containing faucets will be installed in a
separate pipe loop on this rack.
lime
full scale plant dose as needed
for pH control during test period
sodium
hypochlorite
full scale plant dose during test
period (3.5 mg/l chloramine
concentration anticipated)
ammonia
full scale plant dose during test
period (3.5 mg/l chloramine
concentration anticipated)
fluoride
full scale plant dose during test
period (1.0 mg/l dose anticipated)
Source: Distributed to the TEWG by WA.
IOCCT Review
Final Report
2-19
March 2007

-------
Exhibit 2.4.2 Photograph of the pipe rack elements installed at the Dalecarlia
Water Treatment Plant
Source: Distributed to the TEWG by WA.
The pipe loops were operated on a 16-hour flow period, followed by an 8-hour
stagnation period. Unlike the DCWASA system, water was not recirculated within the
loop and exchanged 011 a daily basis. Instead, all WA racks were a single pass-through
operation. The flow-through pipe loop facilities were operated seven days per week.
Periodic water quality samples were collected Monday through Friday and delivered to
the WA Laboratory for analyses. The standard analysis set consisted of the following:
•	Total and dissolved lead
•	pH
•	Alkalinity (as CaCCb)
•	Calcium (as Ca)
•	Total Dissolved Solids (TDS)
•	Calcium Carbonate Precipitation Potential (CCPP)
•	Turbidity
•	Dissolved Inorganic Carbon (DIC)
•	Periodic Heterotrophic Plate Counts (HPCs)
•	NHa (as N)
•	Nitrite and Nitrate (as N)
IOCCT Review
Final Report
2-20
March 2007

-------
2.4.4 Results
The pipe racks became operational in January 2005. All seven racks operated
continuously for approximately twelve months. In February of 2006 it was believed that
sufficient data had been collected from some of the racks to allow conclusions to be
drawn and resolve the questions they were designed to answer. Operation of the
following racks has been halted:
•	Rack 1. High chloramines, zinc orthophosphate addition;
•	Rack 4. High chloramines, no orthophosphate inhibitor; and
•	Rack 5. Low chloramine dose, constant orthophosphate feed.
The data set collected by the WA research team is substantial. Exhibits 2.4.3a
through 2.4.6b graphically portray findings from the WA pipe loop study through early
March 2006. The "a" exhibits contain three graphs: total lead over time, dissolved lead
over time, and a combination of pH, alkalinity, and temperature over time for each of
three loops in the Rack. The "b" exhibits are a larger version of dissolved lead over time
with notations as to treatment regime changes. Key findings as they relate to the OCCT
follow these exhibits.
IOCCT Review
Final Report
2-21
March 2007

-------
Exhibit 2.4.3a Results for Rack 2, Decrease Orthophosphate over Time
WA Pipe Loops
Total Lead Concentration Vs. Time - Rack2
160
140
_i 120
O
3
100
C
o
80
15
c

-------
Exhibit 2.4.3b Dissolved Lead Results for Rack 2: Decrease Orthophosphate Over Time
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
2/17/05
WA Pipe Loops
Dissolved Lead Concentration Vs. Time - Rack 2

¦
~

~
Switched from 3 mg/L P04 to 2 mg/L P04

•
¦
~
. v


~ - . _
¦
~ ~
~ . y «
«« ¦ • . * *«¦ i , ~ a 4

»Vv ***." *
c ¦ • ~
' ¦ ~
: v ~' - - *¦ v <"\ *
H ¦
~ k\
~ H ~ "V ¦ "V ~
"* ! V i " r" " " ft
K ~*>
' rv
O)
£
o
*•>
c
0)
o
c
o
O
3/19/05
4/18/05 5/18/05
~ Loop 2A ¦ Loop 2B Loop 2C
6/17/05	7/17/05 8/16/05 9/15/05 10/15/05 11/14/05 12/14/05
Date
1/13/06
2/12/06
3/14/06
Source: Distributed to the TEWG by WA.
IOCCT Review
Final Draft
2-23
March 2007

-------
Exhibit 2.4.4a Results for Rack 3, Orthophosphate with Simulated Chlorine Burn
160
140
_ 120
J
O)
B 100
c
O
f 80
§ 60
3 40
20
WA Pipe Loops
Total Lead Concentration Vs. Time - Rack 3

2/17/05 3/29/05 5/8/05 6/17/05 7/27/05 9/5/05 10/15/05 11/24/05 1/3/06 2/12/06 3/24/06
~ Loop 3A
iLoop 3B
Loop 3C
Date
WA Pipe Loops
Dissolved Lead Concentration Vs. Time - Rack 3
~ 30
o 20
v ' ¦ ' , T T-	n- ^ ~
2/17/05 3/29/05 5/8/05 6/17/05 7/27/05 9/5/05 10/15/05 11/24/05 1/3/06 2/12/06 3/24/06
« Loop 3A
iLoop 3B
Loop 3C
Date
35
30
£ 25
D
TO
| 20
E
0)
¦- 15
TO
WA Pipe Loops
pH, Alkalinity, Temperature Vs. Time - Rack 3 - Loop 3A


Mm* U
Aii . .
* ¦ ¦ *
r. ft ¦
¦
•f- :
if ¦ ¦
;
¦ ¦"
I-
¦
¦ ¦

¦¦


140
120
100
(0
<
40
20
2/17/05 3/29/05 5/8/05 6/17/05 7/27/05 9/5/05 10/15/05 11/24/05 1/3/06 2/12/06 3/24/06
~ pH Temperature ¦ Alkalinity
Source: Distributed to the TEWG by WA.
Date
IOCCT Review
Final Draft
2-24
March 2007

-------
Exhibit 2.4.4b Dissolved Lead for Rack 3, Orthophosphate with Simulated Chlorine Burn
30
28
26
24
.	. 22
_l
O) 20
3
"—' 18
c
o
(0
16
14
c
a) 12
o
c 10
o
o 8
6
4
2
WA Pipe Loops
Dissolved Lead Concentration Vs. Time - Rack 3









Switched from chloramines to chlorine





¦
*
Switched from chlorine

<
to chloramines

u ~
N







*:¦ V- ~ ' -'•¦ a'"?
R

~ %~ *' » ~; a -
w

V
m "
*
¦
• ***.

	T	T	T	I	1	1	


2/17/05 3/19/05 4/18/05 5/18/05 6/17/05 7/17/05 8/16/05 9/15/05 10/15/05 11/14/05 12/14/05 1/13/06 2/12/06 3/14/06
» Loop 3A ¦ Loop 3B Loop 3C
Source: Distributed to the TEWG by WA.
Date
IOCCT Review
Final Draft
2-25
March 2007

-------
Exhibit 2.4.5a Results for Rack 6, High Chloramine Dose
WA Pipe Loops
Total Lead Concentration Vs. Time - Rack 6
^ 120

2/17/05 3/29/05 5/8/05 6/17/05 7/27/05 9/5/05 10/15/05 11/24/05 1/3/06 2/12/06 3/24/06
~ Loop 6A ¦ Loop 6B Loop 6C
Date
WA Pipe Loops
Dissolved Lead Concentration Vs. Time - Rack 6
ra 40

2/17/05 3/29/05 5/8/05 6/17/05 7/27/05 9/5/05 10/15/05 11/24/05 1/3/06 2/12/06 3/24/06
~ Loop 6A ¦ Loop 6B Loop 6C Date
WA Pipe Loops
pH, Alkalinity, and Temperature Vs. Time - Rack 6


2/17/05 3/29/05 5/8/05 6/17/05 7/27/05 9/5/05 10/15/05 11/24/05 1/3/06 2/12/06 3/24/06
~ pH
i Alkalinity	Temperature|
Source: Distributed to the TEWG by WA.
IOCCT Review
Final Draft
2-26
March 2007

-------
Exhibit 2.4.5b Dissolved Lead for Rack 6, High Chloramine Dose
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
WA Pipe Loops
Dissolved Lead Concentration Vs. Time - Rack 6






¦

1

¦
Switched from 3 mg/L PO4

~ ¦ ~
to 1 mg/L P04

¦
*
~

¦ # ~
* _ _j ¦

¦ ¦ ¦
j ~ ¦ ~~ ~ ** . >

:> ' v* 1 ~•v: v . '


~
f ~

?>*• *"
~
* " *
it '



IV1 ¦
O)
3^
C
O
CB
+¦>
C

-------
Exhibit 2.4.6a Results for Rack 7, Finished Water Control
160
140
_ 120
_i
O)
3. 100
c
o
| 80
§ 60
c
o
° 40
20
0
WA Pipe Loops
Total Lead Concentration Vs. Time - Rack 7
U U B
.? ¦¦ ¦	
~A *
.v .W - ^

2/17/05 3/29/05 5/8/05 6/17/05 7/27/05 9/5/05 10/15/05 11/24/05 1/3/06 2/12/06 3/24/06
~ Loop 7A ¦ Loop 7B Loop 7C
Date
WA Pipe Loops
Dissolved Lead Concentration Vs. Time - Rack 7
*' •VVAiAJ^.
2/17/05 3/29/05 5/8/05 6/17/05 7/27/05 9/5/05 10/15/05 11/24/05 1/3/06 2/12/06 3/24/06
~ Loop 7A ¦ Loop 7B Loop 7C
Date
WA Pipe Loops
pH, Alkalinity, Temperature Vs. Time - Rack7 - Loop7A
35
30
3 25
n3
£
H 15
"O
TO 10
I

jiV


k
¦
1
¦ i
¦
i

¦ ¦
v;


¦v







5
0
2/17/05 3/29/05 5/8/05 6/17/05 7/27/05 9/5/05 10/15/0 11/24/0 1/3/06 2/12/06 3/24/06
5 5
140
120
100
80 "E
15
60 ^
<
40
20
0
~ pH Temperature
Alkalinity
Date
Source: Distributed to the TEWG by WA
IOCCT Review
Final Draft
2-28
March 2007

-------
Exhibit 2.4.6.b Dissolved Lead for Rack 7, Finished Water Control
WA Pipe Loops
Dissolved Lead Concentration Vs. Time - Rack 7
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
VL ***¦¦ V ' ~
~ ~ ~ ¦r" - "i «
.	y
• »g-		


<«*
l^-V
2/17/05
3/19/05 4/18/05
5/18/05
6/17/05
« Loop 7A
Loop 7B
Loop 7C
7/17/05 8/16/05 9/15/05 10/15/05 11/14/05 12/14/05
Date
1/13/06
2/12/06
3/14/06
Source: Distributed to the TEWG by WA
IOCCT Review
Final Draft
2-29
March 2007

-------
The results from Racks 2, 3, 6, and 7 confirm the effectiveness of orthophosphate
treatment. The time needed to observe meaningful reductions in lead was longer than for
the DCWASA circulation loops, although this may be attributable to the high
orthophosphate dose of 10 mg/L used to accelerate results in the DCWASA loops. Total
and dissolved lead both before and during orthophosphate treatment are generally lower
in the WA loops compared to the DCWASA loops. However, the relative proportion of
particulate lead in the WA loops was substantially greater than in the DCWASA loops.
Only after substantial treatment with orthophosphate did the proportion of the particulate
lead in the WA loops diminish. The presence of the higher particulate fraction may
suggest a difference in loop fabrication techniques or handling of the LSL specimens
prior to fabrication. Also, the pipe loops used in the WA study were allowed to dry out,
while DCWASA's pipe loops remained wet. This may account for differences between
the study results.
The WA data strongly suggests that there is no additional benefit associated with
application of a zinc orthophosphate inhibitor as compared to the addition of a simple
orthophosphate corrosion inhibitor. Moreover, the addition of the zinc component may
serve to destabilize lead corrosion scales by accelerating the formation of friable lead
particulates that are easily shed from the surface of the LSL.
The WA testing shows that LSL lead release levels are sensitive to changing
dosage levels of orthophosphate. An abrupt decrease in orthophosphate addition in early
March 2006 resulted in rapid lead release. Also, results for Racks 3 and 6 show a
measurable increase in lead when orthophosphate was reduced, although the dissolved
lead concentrations remained below approximately 11 (J,g/L.
As in the case of DCWASA, Rack 3 of the WA Pipe Loop study demonstrates
that a change back to chloramines after a simulated free chlorine burn can result in minor
increases in lead release, although lead levels in Rack 3 stabilized fairly quickly.
2.5 Studies Related to Partial Lead Service Line Replacement
The configuration of a typical DCWASA water service line is shown in Exhibit
2.5.1. For LSL replacements, the LCR requires that DCWASA replace the public space
portion and offer to replace the portion of the LSL on private property at cost. (When
replacing an LSL on private property, DCWASA may only charge the property owner a
price that reflects the cost of replacing the LSL, without adding any fees to the price.)
Because of issues related to mechanical durability, the LSL is always replaced with a
copper line. If the property owner elects not to have their portion of the LSL replaced,
and DCWASA replaces all of the lead piping in the public space, the service line is
considered to be a partial LSL.
IOCCT Review
Final Draft
2-30
March 2007

-------
Exhibit 2.5.1 Typical Water Service Line in the DCWASA System
Property
Line
Shutoff
Water Valve
Service Line
Water Main
Meter
Private Property
i	~
Public Space
Source: Reiber, Keefer et al. (2004).
Partial LSL replacement raises another corrosion issue, namely, the possibility
that the coupling (or near coupling) of a partial LSL to the replacement copper line could
create a galvanic corrosion cell that may accelerate corrosion on the remaining portion of
the LSL. The concern is that coupling of the dissimilar metals may create a localized
condition with the potential to elevate overall lead release rates above pre-replacement
levels, which would not only defeat the intent of the replacement program but also
exacerbate the situation.
2.5.1 Purpose of the Studies
Several technical questions were raised, both mechanical and EC, as to whether a
partial LSL replacement may accelerate lead release from the remaining portion of the
LSL, negating any benefit associated with the partial replacement. Two studies were
undertaken to help address these questions:
•	DCWASA performed a series of field tests that examined lead release in
individual homes, both before and after partial LSL replacement.
•	As subcontractor to EPA's contractor, The Cadmus Group, Inc., HDR
engineering conducted a series of laboratory based studies to examine the EC
issues associated with partial LSL replacement to determine if replacing a
portion of a lead pipe with copper piping might cause accelerated lead release.
IOCCT Review
Final Draft
2-31
March 2007

-------
2.5.2 Summary of Key Findings
•	The potential mechanical disruption of lead corrosion scales on the remaining
portion of the LSL is not a serious threat if reasonable care is taken in the
cutting and removal process.
•	Vigorous flushing alone following partial LSL replacement is sufficient to
remove lead particulates generated in the cutting process.
•	Well-aged DCWASA LSL specimens - including those that have been
exposed to an orthophosphate inhibitor - are exceptionally well passivated
and highly resistant to electrical perturbations of any kind.
•	When a well-passivated LSL is coupled to a new length of copper tubing (as
in a partial LSL replacement) the area of galvanic influence is very limited.
The actual reach of the galvanic current is partially a function of the water
quality, but is likely limited to the first inch of the LSL.
•	A conventional plumbing dielectric junction removes even the minor
corrosion risks associated with galvanic coupling. Any break in electrical
continuity between the copper and LSLs effectively eliminates the potential
for significant galvanic effect.
•	A chlorine residual (free or combined) does elevate the galvanic effect on the
LSL/copper couple by accelerating the cathodic current exchange process.
The impact overall, however, is largely limited to the galvanic influence on
the copper service line. The overall impact on the LSL surface is nearly
imperceptible. Interestingly, water conductivity has a more important effect
on the galvanic process than chlorine residual.
2.5.3 DCWASA Pipe Cutting Study
Early in the LSL replacement program, some homes that had undergone partial
LSL replacement experienced high (>1000 ppb), albeit brief, tap water lead levels. The
galvanic corrosion issue was raised as a potential cause, but it was also recognized that
the method used to cut the lead pipe combined with insufficient flushing to remove lead
particles derived from the installation process (the disturbance/exposure of the "cut"
joint) may have contributed to the high lead levels. Out of concern over these post-partial
replacement lead levels, the DOH ordered DCWASA to cease partial LSL replacements
performed by cutting lead pipes, but allowed replacement of the LSL to the first threaded
joint, usually to the water meter. Most of these replacements, however, did not meet the
requirements of the LCR and could not be counted toward compliance with the LCR. To
investigate the issue of partial LSL replacement, DCWASA put into place a program to
assess lead release in a variety of homes undergoing partial LSL (Wujek, 2004).
IOCCT Review
Final Draft
2-32
March 2007

-------
The actual mechanics of the replacement process in this study involved a
sequence of materials verification, homeowner notification and education, and lead
profiling, followed by excavation and re-plumbing. Approximately 15 feet of lead service
piping remained from the property line to the building face at each of the addresses.
Three methods of cutting the existing service lines were used: hacksaw, tube cutter, and
pipe lathe. It was also decided that as part of this study, the contractors performing the
partial LSL replacement would vigorously flush the new line at for at least 15 minutes
immediately following the replacement of the lead service.
Lead profiling was conducted before the partial lead service replacement,
immediately after cutting and flushing, and regularly for a period of 14 days after the
service lines had been replaced. As will be discussion in Section 3.3, lead profiling
involves collecting and analyzing consecutive 1-liter samples from a kitchen tap
following a 6-hour or longer stagnation period. Eighteen (18) pre-partial LSL
replacement sampling profiles were conducted at seven addresses. Forty five (45) post-
partial LSL profiles were obtained at these same addresses. A total of almost 700
individual lead measurements were conducted in this testing. Overall, the average results
of the pre- and post-partial replacement sampling performed are shown in Exhibit 2.5.2.
Exhibit 2.5.2 Pre- and post-lead profiles for partial replacement sampling results
70
60
50
40
30
20
10
0
0
2
4
6
8
10
12
Sample
Source: Wujek, K. (2004).
¦	Avg Re
¦	Avg Fbst
The sampling results indicate that flushing immediately following a careful partial
lead service replacement can reduce lead levels delivered to the household tap. The
sampling also showed that the disturbance/exposure of the existing LSL where it is cut
and connected to new copper piping does not necessarily increase lead levels in the
delivered water. Analysis of the different cutting-method data suggests little difference in
IOCCT Review	2-33	March 2007
Final Draft

-------
the final lead levels based on the manner in which the pipe was cut. Moreover, many of
the problems previously attributed to partial LSL replacement (high transient lead levels
immediately following replacement) can be avoided by vigorous flushing immediately
following the replacement.
2.5.4 EPA/HDR Galvanic Corrosion Study
In theory, it is conceivable that replacing a portion of a lead line with a new
copper service line could create a strong galvanic couple with an initial Cu/Pb
electromotive difference in the 400 - 500 mV range (Reiber, 1991). If a significant
portion of the remaining section of LSL were shifted in the anodic direction by even a
fraction of this amount, there should be a substantial acceleration of the corrosion rate
and associated metal release rates.
A study was initiated in late 2004 to assess the potential effects of both external
currents and dissimilar metals contact on corrosion from LSLs. The final report of study
findings is in Appendix C. The next several sections focus on the analysis of potential
galvanic corrosion effects resulting from partial LSL replacement in D.C.
Research Protocol
At the core of this study was the search for the substantial EC impacts that,
theoretically, should be associated with the galvanic and impressed currents imposed on
the LSLs. The principal measure of these impacts would be a significant shift in the EC
potential of the interior surface of the LSLs away from the freely corroding surface
potential. This research did not attempt to create laboratory conditions that exactly
replicate field conditions. Instead, the goal was to demonstrate whether or not extremes
of grounding currents or galvanic coupling could affect the LSL electrochemistry.
Testing was generally short-term, and designed to answer the question, "Can grounding
and/or galvanic currents under a worst-case scenario meaningfully contribute to lead
corrosion and metals release?"
The study used a series of EC cells which allowed the mounting of sections of
LSLs under flow conditions and the placement of electrodes capable of quantifying shifts
in surface potential. The surface potential measurement is sensitive, easy to use, and
allows speedy measurements, but its principal advantage is that it is influenced only by
the electrochemistry of the metal surface and the water in contact with that surface. It
reflects the corrosion conditions of the underlying metal, which, in this case, is the factor
most directly influenced by application of the galvanic and/or impressed currents in
question.
The galvanic coupling research utilized polarization cells in which individual
sections of LSLs and copper tubing were mounted. These cells could be connected in a
hydraulic series, with the electrical connections between the individual cells manipulated
at will. Because the pipe specimens of each cell were not in direct contact, these cells
were referred to as indirectly coupled. The importance of the indirectly coupled cells
IOCCT Review
Final Draft
2-34
March 2007

-------
relates primarily to the ability to control cathode/anode ratios. It is critical to the
appreciation of the galvanic-couple concern to understand that it is not the contact of
dissimilar metals, per se, that creates the corrosion risk. Rather, corrosion risk is created
by the fact that the cathodic surface (the more electropositive metal), if present in
abundance, can affect a shift in the surface potential of the anodic surface, and any
meaningful shift in the anodic surface in a more positive direction generates a higher
corrosion rate on that surface. A second approach to galvanic testing utilized longer
segments of LSLs and copper pipe coupled together in a manner similar to an actual
partial LSL replacement. Because these pipe specimens are in direct contact, this type of
testing is referred to as directly coupled pipe specimens. This form of testing yielded the
most useful results about the nature of the galvanic couple formed between copper and
LSL sections. The schematic presented in Exhibit 2.5.3 shows the arrangement of the
individual cells, hydraulics and electrical connections for the indirectly coupled cells.
Exhibit 2.5.4 presents a schematic showing the same for the directly coupled cells.
Exhibit 2.5.3 Schematic of a typical pipe rig configuration using indirectly coupled
cells
Potentiostat
Cathodic
Coupling
(parallel)
20L
Reservoir
Cathodic
r-B-jN	l Coupling
^	—(parallel)
Reference Electrode
Soldered Pigtail
Pump
Source: Galvanic Corrosion and Grounding Effects Study (Appendix C).
IOCCT Review
Final Draft
2-35
March 2007

-------
Exhibit 2.5.4 Schematic of a test rig showing the direct coupling of LSL and
copper pipe sections
Reference Electrode
Electrode
Insertion Port
Soldered Pigtail
Sample Port
Cu Service Line (3/4' dia.)
Pb Service Line
Pump
Potentiostat
Source: Galvanic Corrosion and Grounding Effects Study (Appendix C).
A third type of test cell loop consisting of DCWASA LSL segments, new copper
tubing, water reservoir, flow control and pumping hardware coupled to an AC/DC current
generator and potentiostat was employed for the assessment of impressed currents on
partial LSLs (see schematic in Exhibit 2.5.5). As in the previous cells, the LSL segments
were modified to accept high impedance reference electrodes penetrating the pipe wall at
multiple locations along its length. The electrodes monitor surface potential on the
interior of the pipe relative to the electrolyte, yet allow for pipeline pressurization.
Internal surface potential along the pipeline was monitored, while different current forms,
amperages, voltages and grounding scenarios were applied to the test pipes.
IOCCT Review
Final Draft
2-36
March 2007

-------
Exhibit 2.5.5 Schematic illustration of impressed current test rig
AC and DC
Current
Source
1
Current Flow
Potentiostat
Reference Electrode
Soldered Pigtail
Electrode
Insertion Port
Sample Port
Headpiece
Pb Service Line
/"Calomel
X Ref, Elec.
Rubber
Stopper
Pipe Wall
mm gap
Source: Galvanic Corrosion and Grounding Effects Study (Appendix C).
Using this equipment and its unique approach to simulating LSL corrosion, the
study went on to investigate the impact of a variety of WQPs relative to galvanic action,
including: pH, conductivity, disinfectant concentration and disinfectant type. It also
looked at mechanical fabrication issues associated with partial LSLs, including the ratio
of lengths of connected lead and copper pipe left in place following a partial LSL
replacement, as well as the use of plumbing dielectrics to electrically isolate the lead and
copper sections. Finally, the study investigated the impact of impressed currents, both
alternating and direct, shunted through LSLs to ground.
Study Observations and Conclusions
In general, the study report shows that grounding and/or impressed currents
moving along LSLs, and eventually leaving the pipe to ground, have no meaningful
impact on internal pipeline corrosion and do not likely contribute to metals release.
Secondly, while the study found that galvanic impacts can be substantial on unpassivated
lead surfaces (freshly exposed surfaces), the magnitude of the impact on aged and
passivated LSL surfaces (as well as on copper service lines) is so minimal as to be
inconsequential. The study provides a strong basis for discounting claims and concerns
IOCCT Review
Final Draft
2-37
March 2007

-------
relative to accelerated metal release associated with partial LSL replacement. Moreover,
it shows that the long-debated controversy about the impacts of grounding currents is
largely a non-issue. In short, partial LSL replacements and impressed currents are not
meaningful concerns relative to optimizing distribution system corrosion control.
Specific observations presented in the final report in Appendix C are as follows.
•	Passivation. Lead is a highly electroactive metal, and in pure form oxidizes
extremely rapidly. An unsealed lead surface, even under natural
environmental conditions, has an exceedingly high initial corrosion rate. Lead
also passivates strongly and quickly. Observations in this study suggest that
meaningful passivation on LSLs can be achieved within a matter of days.
Well-aged DCWASA LSL specimens - especially those that have been
exposed to an orthophosphate inhibitor - are exceptionally well passivated
and highly resistant to electrical perturbations.
•	Lead Electrochemistry. Passivated LSL specimens are highly polarization
resistant - meaning that it takes an exceptional surface perturbation to affect
the underlying corrosion rate. The actual degree of polarization resistance
expressed as a Tafel Value is in excess of 500 - 600 mV per decade of current
shift. Overall, this explains, at least in part, why the galvanic coupling has
little apparent effect on passivated lead surfaces.
•	Area of Galvanic Influence. When coupled to a new length of copper tubing
(as in a partial LSL replacement) the area of galvanic influence on a well
passivated LSL is likely limited to less than the first inch of LSL pipe in the
immediate vicinity of the coupling. The galvanic area of influence on an
unpassivated LSL specimen is larger, but likely limited to the first few inches
of pipe in the vicinity of the coupling. As the LSL passivates, the area of
galvanic influence decreases rapidly. The period of transition can be as short
as a few days under normal distribution system conditions. A potential reason
why galvanic impacts do not generate a more significant corrosion response
relates to the respective geometries of the anodic and cathodic surfaces of the
pipeline couple. Because sequential pipelines (LSL to copper tubing) are
connected at only a single location, only a small portion of the LSL is
polarized by the galvanic current. And, given the relatively rapid rate at which
both copper and lead surfaces passivate, the duration of the polarization is
relatively brief. Hence, even the meager galvanic effect is short-lived.
•	Cathodic Effect of Copper Pipe. The cathode/anode ratio on a well
passivated LSL surface is unimportant relative to the galvanic effect. This
means that even an exceptionally long length of copper pipe connected to a
partial LSL does not elevate the galvanic effect. (It had been argued that long
lengths of copper service line connected to short LSL sections would
exacerbate the galvanic effect.)
IOCCT Review
Final Draft
2-38
March 2007

-------
•	Water Quality and Galvanic Impacts. A free-chlorine residual does elevate
the galvanic effect by accelerating the cathodic current exchange process.
Conversely, chloramine has a lesser galvanic impact than free chlorine. The
impact overall, however, is largely limited to the galvanic influence on the
copper service line. The overall impact on the LSL surface is nearly
imperceptible. Interestingly, water conductivity has a more important effect
on the galvanic process than chlorine residual. The area of galvanic influence
on the LSL specimen is marginally expanded as the conductivity of the
electrolyte (water) increases, while the area of influence on the copper service
line is substantially expanded. This is because the higher conductivity lessens
the resistance of the electrolyte circuit (water), expanding the "reach" of the
galvanic current. DCWASA distributes a low conductivity water (< 200
microSiemens), which, in part, explains the minimal galvanic impacts
observed.
•	Dielectric Effects. While galvanic impacts relative to DCWASA partial LSL
replacements are likely minimal, any break in electrical continuity between
the copper and LSL lines effectively eliminates the potential for a galvanic
effect. In short, a conventional plumbing dielectric junction removes even the
minor corrosion risks associated with galvanic coupling.
•	Impressed Current Effects. Impressed currents (AC or DC) on LSLs and
copper service lines, including grounding type currents, have no impact
whatsoever on the internal corrosion of the household service lines (or any
other plumbing appurtenance for that matter). There is likely no acceleration
of corrosion associated with the conventional practice of electrical system
grounding to household water systems.
IOCCT Review
Final Draft
2-39
March 2007

-------
This page intentionally left blank.

-------
2. Summary of Research Relevant to the D.C. Lead Issue	2-1
2.1	Desktop Corrosion Control Study	2-1
2.2	Lead Pipe Scale Analysis	2-2
2.3	DCWASA Circulation Pipe Loop Studies	2-4
2.3.1	Purpose of Study	2-4
2.3.2	Summary of Key Findings	2-5
2.3.3	DCWASA Pipe Loop Protocol	2-5
2.3.4	Results	2-9
2.4	Washington Aqueduct Flow-Through Pipe Loop Studies	2-17
2.4.1	Purpose of Study	2-17
2.4.2	Summary of Key Findings	2-17
2.4.3	Washington Aqueduct Pipe Loop Protocol	2-18
2.4.4	Results	2-21
2.5	Studies Related to Partial Lead Service Line Replacement	2-30
2.5.1	Purpose of the Studies	2-31
2.5.2	Summary of Key Findings	2-32
2.5.3	DCWASA Pipe Cutting Study	2-32
2.5.4	EPA/HDR Galvanic Corrosion Study	2-34
Exhibit 2.2.1 Raman spectrograph of a transforming LSL showing the spatial distribution
of Pb(IV) and Pb(II) mineralogy	2-4
Exhibit 2.3.1 Photograph of Electrochemical Polarization Cell Used in the EC Pipe	2-7
Exhibit 2.3.2 Stagnation Loop Schematic	2-7
Exhibit 2.3.3 Water Quality Parameters Evaluated During the DCWASA Study	2-8
Exhibit 2.3.4 Original Corrosion Control Strategies Tested by DCWASA	2-9
Exhibit 2.3.5 Results from Four DCWASA Stagnation Loops Showing Reduction in
Lead Release as a Function of Orthophosphate Treatment	2-11
Exhibit 2.3.6 Change in Corrosion Rate (Measured in mils per year, MPY) for EC Loop
Number 1	2-15
Exhibit 2.3.7 Results from Pipe Loop 2 (March 2004 - July 2004)	2-16
Exhibit 2.3.8 Results from Pipe Loop 4 (March 2004 - July 2004)	2-16
Exhibit 2.4.1 Operating Conditions, Objectives and Rationale for the Seven Pipe Racks
used in WA Study	2-19
Exhibit 2.4.2 Photograph of the pipe rack elements installed at the Dalecarlia	2-20
Exhibit 2.4.3a Results for Rack 2, Decrease Orthophosphate over Time	2-22
Exhibit 2.4.3b Dissolved Lead Results for Rack 2: Decrease Orthophosphate Over Time
	2-23
Exhibit 2.4.4a Results for Rack 3, Orthophosphate with Simulated Chlorine Burn	2-24
Exhibit 2.4.4b Dissolved Lead for Rack 3, Orthophosphate with Simulated Chlorine Burn
	2-25
Exhibit 2.4.5a Results for Rack 6, High Chloramine Dose	2-26
Exhibit 2.4.5b Dissolved Lead for Rack 6, High Chloramine Dose	2-27
Exhibit 2.4.6a Results for Rack 7, Finished Water Control	2-28
Exhibit 2.4.6.b Dissolved Lead for Rack 7, Finished Water Control	2-29
Exhibit 2.5.1 Typical Water Service Line in the DCWASA System	2-31
Exhibit 2.5.2 Pre- and post-lead profiles for partial replacement sampling results	2-33

-------
Exhibit 2.5.3 Schematic of a typical pipe rig configuration using indirectly coupled cells
	2-35
Exhibit 2.5.4 Schematic of a test rig showing the direct coupling of LSL and	2-36
copper pipe sections	2-36
Exhibit 2.5.5 Schematic illustration of impressed current test rig	2-37

-------
3. Review of Relevant Water Quality Data
The purpose of this chapter is to evaluate distribution system water quality data to
answer the following questions:
•	Is the orthophosphate treatment being implemented consistent with EPA's
WQP goals?
•	Has the orthophosphate treatment been successful at reducing lead in drinking
water?
•	Has the orthophosphate treatment had positive or negative impacts on
bacteriological growth (with respect to both biofilm growth and nitrification)
in DCWASA's distribution system?
Section 3.1 describes the water quality data obtained and reviewed for this report;
Sections 3.2, 3.3, and 3.4 discuss the three questions above, respectively. This report
reviews all available data (including non-compliance data) and, thus, does not attempt to
confirm DCWASA's compliance with any Maximum Contaminant Level (MCL) or
WQP defined by EPA. Conclusions based on the combined research findings for Chapter
2 and water quality data in this chapter are presented in Chapter 4 of this report.
3.1 Description of Dataset
Water quality data from January 2003 through December 2005 were obtained
from EPA Region III and DCWASA. The data, originally organized in MS Excel files
and summary reports, were uploaded into an MS Access database and organized by three
main data types: "Monitoring Programs," "Sites," and "Sample Results." Each data type
is described below. Once uploaded into the database, the files were checked to ensure no
duplicate entries were created. Parameter names and units were standardized to facilitate
ease of use in writing queries.
In addition to water quality data, we reviewed data from DCWASA's lead
profiling program. All lead profiling data were obtained from DCWASA and organized
in an MS Excel workbook. Section 3.3 provides a description of the lead profiling
program and data types reviewed for this report.
Monitoring Programs
The data we received were generally associated with one of three DCWASA
compliance monitoring programs:
•	Total Coliform Rule (TCR) monitoring. DCWASA has approximately 65
TCR monitoring sites throughout District. It collects data from each site
IOCCT Review
Final Draft
3-1
March 2007

-------
approximately once per week, usually generating between 250 and 300
samples per month, under an EPA approved monitoring plan;
•	Supplemental monitoring. DCWASA used their hydraulic model to identify
more than 25 supplemental monitoring sites, which are generally located in
dead-end and low flow areas of the distribution system. The start of
supplemental monitoring coincided with the start of orthophosphate treatment
in August 2004. DCWASA monitored at supplemental sites per the
monitoring plan submitted pursuant to the interim OCCT designation. They
collect data at hydrants as well as inside tap locations (e.g., a restroom sink) at
each site at least twice per month; and
•	LCR monitoring. DCWASA has listed more than 100 tier 1 sites1 in its
EPA-approved LCR monitoring plan. LCR samples are collected by
customers according to an instruction sheet provided by DCWASA.
Customers collect a "first draw" sample from a cold water tap after at least a
6-hour stagnation period (either in the morning or after returning home from
work). Then they collect a "second draw" sample after flushing the tap until
the water becomes cold.
DCWASA and EPA also provided water quality data at alternative distribution
system sites, which were often sampled in response to customer complaints, water quality
problems, etc. These sites and samples were labeled "no program" in our database.
Sites
Site location is important in assessing the spatial (or locational) variability of
water quality in the distribution system. The District of Columbia is divided into four
geographic quadrants: Northwest (NW), Northeast (NE), Southwest (SW) and Southeast
(SE), as shown in Exhibit 3.1.1. As described earlier in this report, DCWASA's
distribution system consists of seven service areas, or pressure zones, based on elevation.
These are the Low, 1st High, 2nd High, 3rd High, 4th High, (East and West), Anacostia 1st
High, and Anacostia 2nd High. General correlations between quadrant and service area
are as follows:
•	NW, the most populated quadrant, is served primarily by the 1st, 2nd, and
3rd High pressure zones. The downtown portion of the quadrant is
primarily served by the Low service area. NW contains the only 4th High
lines, and shares the 2nd and 3rd High lines with NE;
•	NE is served primarily by 1st High, 2nd High, and Low;
•	SW is the smallest geographic area and is served primarily by the Low
pressure lines; and
•	SE is served primarily by Anacostia 1st High and Anacostia 2nd High.
1 Tier 1 sites are defined in the LCR as 1) single family homes containing copper pipes with lead solder that
were installed after 1982 or single family homes containing lead pipes, and/or 2) homes that are served by
an LSL.
IOCCT Review
Final Draft
3-2
March 2007

-------
Exhibit 3.1.1 Geographic Quadrants in D.C.
WASHINGTON
DC
NW
NE
VIRGINIA
sw
The primary information for all sample sites is the mailing address. In addition to
address, both TCR and supplemental sites have unique ID numbers that begin with the
service area designation (e.g., sites in the Low service area are designated "L-l," "L-2,"
and so on). Supplemental sites use the same numbering system as TCR sites except the
ID number includes the acronym "BKJV." ("BKJV" is short for the name of the
DCWASA contractor who selected these sites.) The LCR monitoring locations do not
have ID numbers and thus do not include pressure zone information. Because service area
information was not available for all sites, all spatial analyses of distribution system
monitoring data were done according to city quadrant rather than service area.
IOCCT Review
Final Draft
3-3
March 2007

-------
Sample Results
Various WQPs were analyzed for each site. At supplemental and TCR sites,
DCWASA generally collected total chlorine, pH, and temperature data on-site. The WA
laboratory conducted analyses for coliforms (TCR sites only), HPCs, orthophosphate,
free ammonia, nitrite, nitrate, and many other supplemental parameters required by EPA
in their August 3, 2004 OCCT letter (see Appendix A).
If results exceeded an internal AL for a given WQP, DCWASA usually flushed
and then re-sampled that same day until the parameter was below acceptable limits.
When repeat samples occurred at a site on a single day, results were averaged to produce
one value for that day.
For LCR compliance samples, the WA laboratory analyzed for total lead and total
copper. Dissolved lead cannot be determined for these samples since the procedure
requires immediate filtration.
3.2 Orthophosphate Treatment and WQP Monitoring
In August 2004, EPA approved the full-distribution-system application of
orthophosphate in D.C., setting requirements for orthophosphate dose as well as other
WQPs important to the maintenance of OCCT and monitoring of the distribution system
for adverse effects, including pH, free ammonia nitrogen, and combined nitrite and
nitrate nitrogen. Exhibit 3.2.1 summarizes WQP requirements and WQP optimal goals
set by EPA for D.C.'s OCCT for both DCWASA and WA.2 Under the interim OCCT
designation, DCWASA is required to monitor for these parameters at identified TCR
sites as well as at supplemental monitoring sites selected to represent the areas of the
distribution system most likely to experience water quality problems.
Exhibit 3.2.1 WQPs and WQP Goals for DCWASA and WA

DCWASA
WA
Interim
WQPs
WQP Goals
Interim
WQPs
WQP
Goals
PH
7.7 + 0.3
7.7 + 0.1
7.8-7.9 + 0.3
7.7 + 0.1
Orthophosphate residual
in tap samples
1.0-5.0 mg/L
3.0 mg/L
1.0-5.0 mg/L
3.0 mg/L*
Free Ammonia nitrogen
0.5 mg/L
0.2 mg/L

Nitrate/ nitrite nitrogen
0.5 mg/L
<0.1 mg/L
* Dose needed to reach this residual in tap samples.
Source: EPA IOCCT Letter (Appendix A).
2 Virginia Department of Health set similar WQPs for Arlington County and Falls Church, VA.
IOCCT Review
Final Draft
3-4
March 2007

-------
The purpose of this section is to assess how well the corrosion control treatment is
meeting the WQPs and WQP goals by reviewing DCWASA monitoring data. To gauge
any temporal and/or geographic trends, all parameters are analyzed by month as well as
by quadrant (NW, SW, SE, and NE). Charts comparing per-quadrant averages for
ammonia, and nitrite/nitrate did not show significant spatial trends and are not included in
this section.
3.2.1 Orthophosphate Levels
Because WA made adjustments to the orthophosphate feed during the last several
months of 2004, we evaluated orthophosphate levels for 2005 only. The dataset consists
of more than 1,700 orthophosphate sample results from the distribution system during
this time period, representing nearly 58 TCR and supplemental monitoring locations
sampled once or twice per month at hydrants, taps, or both. Specifically, there were 30
TCR sites (all tap samples) and 28 supplemental monitoring sites (for most, samples were
collected at both inside taps and hydrants). Only TCR sites (tap samples only) were
subject to the interim OCCT WQPs.
Exhibit 3.2.2a shows monthly average, minimum, and maximum values for all
geographic quadrants for TCR sites, tap samples only. Exhibit 3.2.2b shows monthly,
maximum, and minimum orthophosphate values for taps and hydrant samples taken at
both TCR and supplemental sites (supplemental sites are considered to be representative
of the areas least likely to meet water quality goals). Exhibit 3.2.3 shows monthly
averages by geographic quadrant. Exhibits 3.2.4 and 3.2.5 show average, minimum, and
maximum levels by sample site for TCR and supplemental sites, respectively.
In general, the D.C. distribution system has consistently met the WQP goal for
orthophosphate of 3.0 mg/L. Average orthophosphate levels for TCR sites were slightly
above 3.0 mg/L for every month of 2005 and remained within the 1.0 - 5.0 mg/L range
required by EPA's OCCT designation letter. Even when the "worst case" supplemental
monitoring sites are considered along with TCR tap samples—as shown in Exhibit
3.2.2b—average orthophosphate levels remained consistently above 3.0 mg/L throughout
2005. D.C. experienced its lowest individual sample results for orthophosphate in June
and October 2005 in the NW and SE quadrants of the city. This appears to be a result of
a few isolated cases of very low orthophosphate readings, between 0.8 and 1.7 mg/L at
two TCR and four supplemental sites. Exhibit 3.2.6 presents the lowest and second-
lowest orthophosphate values for each of these sites. For all of these sites, the second
lowest orthophosphate reading exceeded 2.5 mg/L, indicating that the chronic low
orthophosphate levels were not a problem at these sites. For June 2005, repeat samples
and samples taken at the same sites on the same day exhibited normal orthophosphate
levels. For October 2005, other sites sampled on the same day reflected normal
orthophosphate levels, while repeat samples at the same sites continually experienced low
levels. These results could be caused by instrument or sampling error, but they also
could represent a slug of water with low orthophosphate concentration.
IOCCT Review
Final Draft
3-5
March 2007

-------
Exhibit 3.2.2a Maximum, Minimum, and Average Orthophosphate Concentration
by Month, TCR sites, Tap samples
WQP Range Upper Bound
|WQP Goal Range
I Maximum Concentration
I Minimum Concentration
A Average Concentration
a. 2
WQP Range Lower Bound
2005
Source: TCR sites, tap samples only.
Exhibit 3.2.2b Maximum, Minimum, and Average Orthophosphate Concentration
by Month, supplemental and TCR sites, tap and hydrant samples
| Maximum Concentration
I Minimum Concentration
| A Ave ra ge Co nee ntrati o n |
Source: Supplemental and TCR sites, taps and hydrants.
IOCCT Review
Final Draft
3-6
March 2007

-------
Exhibit 3.2.3 Monthly Average Orthophosphate Concentration by Quadrant
Source: Supplemental and TCR sites, hydrant and tap samples.
Exhibit 3.2.4 Maximum, Minimum, and Average Orthophosphate Concentrations
by TCR Site (2005)
Maximum P04
Minimum P04
~ Average P04
WQP Range Upper Bound

m
WQP Range Lower Bound
CO	CO
CO "d- CO
(N^iDcoaiO'^'^-CNj^r
III
IIICMCMCOCOCOCOCOCOl^
I I I I
t- T- CM CM
< < < <
Source: TCR sites, tap samples.
IOCCT Review
Final Draft
3-7
March 2007

-------
Exhibit 3.2.5 Maximum, Minimum, and Average Orthophosphate Concentrations
by Supplemental Site (2005)
5.5 -
5 -
? 4.5-
U)
E
r 4-
o
TO
£ 3.5 -
• Maximum P04
| Minimum P04
A Average P04
o 1.5 -
1 -
0.5 -
t - " t •
" I " t i " t
>
>
>
>
>
—>
>
—>
>
—>
>
—>
>
—>
>
—>
>
—>
>
—>
>
—>
>
—>
>
—>
>
—>
>
>
>
>
>
>
>
>
>
m
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO




X
X
X
X
X
X
X
X
X
X
X
X
LLJ
i
X
X
X
X
X
X
X

















X
"3-
<
<
<
<
<
<
<
Source: Supplemental monitoring sites, tap samples only.
Exhibit 3.2.6 Sites Experiencing the Lowest Orthophosphate Results for 2005

Lowest
Date
Next
Lowest
Date
A2H-2 BKJV
0.8
20-Jun
2.55
18-Jul
L-2 BKJV
0.86
20-Jun
2.63
17-Nov
3H-1
1.01
20-Jun
2.52
14-Jun
4HE-1 BKJV
1.13
24-Oct
2.54
2-Sep
3H-4 BKJV
1.16
24-Oct
2.69
21-Nov
A1H-4
1.69
15-Jun
2.9
22-Nov
Source: Supplemental and TCR monitoring sites, tap samples only.
IOCCT Review
Final Draft
3-8
March 2007

-------
3.2.2 pH Levels
Orthophosphate effectiveness depends on maintaining a fairly narrow pH
operating window, which is reflected in the WQP goals established for DCWASA.
Studies suggest that a sustained pH level substantially below 7.5 degrades the
effectiveness of the inhibitor. Conversely, at pH levels substantially greater than 8.5 there
is little evidence to suggest a meaningful benefit to orthophosphate addition (AWW A,
Internal Corrosion in the Water Distribution System, 2nd edition, 1998).
DCWASA measures pH in the field at TCR and supplemental sites (both hydrant
and inside tap sites). Our dataset consists of more than 4,000 pH readings taken between
September 2004 and December 2005, with between 100 and 350 taken each month
throughout the distribution system. Only tap samples at TCR sites are subject to WQPs
(supplemental sites are meant to represent "worst case" conditions). The data presented in
this analysis represent 59 TCR sites (2,750 tap samples and 11 hydrant samples) and 38
supplemental sites (608 tap samples and 576 hydrant samples).
Exhibit 3.2.7a shows overall maximum, minimum, and median pH values per
month for TCR sites, tap samples only. Exhibit 3.2.7b shows maximum, minimum, and
median pH for TCR and supplemental sites, tap and hydrant samples. Exhibit 3.2.8
shows median pH for each month by geographic quadrant. April 2005 had the lowest
median pH value. To further investigate these data, individual pH measurements for the
month of April were plotted in Exhibit 3.2.9. Levels appear to be low throughout the
system in April. A review of WA finished water by EPA Region III showed that finished
water pH levels at both the Dalecarlia and McMillan plants were very stable, with an
average pH of 7.7 and minimum values of 7.6 at both plants. The lower distribution
system pH values in April could be related to a reduction in raw water alkalinity that is
common for that time of year. Exhibits 3.2.7 and 3.2.8 show a very slight downward
trend in pH over the time frame shown, particularly during the last three months of 2005.
Exhibits 3.2.10 and 3.2.11 show maximum, minimum, and median pH for TCR
and supplemental monitoring sites, respectively. Data from all sites exhibit a fairly wide
range of pH values, with average values for both TCR and supplemental sites remaining
close to with average values for both TCR and supplemental sites. These exhibits do not
reveal any identifiable pattern relative to pH spatial variation across the service area.
3.2.3 Nitrite, Nitrate, and Free Ammonia as Nitrogen
EPA's WQPs include goals for free ammonia calculated as nitrogen, as well as
the combination of nitrite and nitrate calculated as nitrogen. Exhibits 3.2.12 and 3.2.13
show the maximum, minimum, and average values by month for free ammonia as total
nitrogen and nitrite/nitrate as total nitrogen, respectively. Note that nitrate was only
measured if nitrite results for the site exceeded 0.1 mg/L or if free ammonia results either
exceeded 0.4 mg/L or were below 0.2 mg/L. Average values were above the WQP goals
every month except one for free ammonia, and most months for combined nitrate/nitrite.
Further evaluation of ammonia, nitrite, and nitrate and their implications for nitrification
are discussed in Section 3.5 of this report.
IOCCT Review
Final Draft
3-9
March 2007

-------
Exhibit 3.2.7a Maximum, Minimum, and Median pH Level by Month, TCR sites, Tap
Samples Only



|pH WQP range upper bound |
















1











k










A
k
k
k




k
i
L
k


L






i

























A













3 ra





fpHWQ
nge lower bound 1

Maximum pH



| Minimum pH
~ Average pH

















September
October
NJ
o
o
November
December
January
February
March
April
(0
5
June
NJ
o
0
01
July
August
September
October
November
December
Source: TCR sites, tap samples only.
Exhibit 3.2.7b Maximum, Minimum, and Median pH Level by Month, Tap and
Hydrants Samples, TCR and Supplemental Sites
8.2	-
8.1	-
8 -
7.9 -
7.8
7.7 -
7.6 -
7.5 -
7.4 -
7.3	-
7.2	-
7.1 -
7 -
































































i

k












i













































Maximum
Minimum
Median p
PH





|
~
)H
H

















September
October
O
o
November
December
January
February
March
April
May
June
o
0
01
July
August
September
October
November
December
Source: TCR and supplemental sites, tap and hydrant samples.
IOCCT Review
Final Draft
3-10
March 2007

-------
Exhibit 3.2.8 Monthly Median pH Levels by Quadrant
pH WQP upper bound
NE
NW
SE
O SW
I pH WQP Goal Range
£ 7.7
V
7.5
I pH WQP upper bound
7.4
2004
Source: TCR and supplemental sites, tap and hydrant samples.
2005
Exhibit 3.2.9 pH Values by Site for the Month of April 2005 (Month Reporting the
Lowest Median pH Value)
I pH WQP range upper bound
I pH WQP Goal Range
£ 7.7
I pH WQP range
lower bound
7.5
7.4
CO

i:
i:

-------
Exhibit 3.2.10 TCR Sites, pH Maximum, Minimum, and Median Values by Site ID
(September 2004 to December 2005)

pH WQP
?ange






V
VQPGoal Range 1

































l
/

























k
























I
























>



k
k





»





k





L
I
ii

L
L

k
k


.

1
i
i




A

¦i
.





i

>
L

L

k



k
k



»
A












A











L

L
I
































































































/G
























pH V\
P Range
~
| Maximum pl-
1 Minimum pH











ower bound
~ Average pH









T-	™ ™	<<<<<<<<
Source: TCR sites, tap samples only.
Exhibit 3.2.11 Supplemental Sites, pH Maximum, Minimum, and Median Values by
Site ID (September 2004 to December 2005)



































A

k


































A

k

k
i


L A








L


, i

k
'
~ A
k







A

k

k

k

k
i
k
k


A










































A











Maximum pH
Minimum pH
Median pH




A


>>>>>>>>>>>>>>>> Q->>>f->>>>>>>>>>>>>^"g;io
~)~3~3~3~3~3~3~3~3~3~3~3~3~3~3~3	S S

-------
Exhibit 3.2.12a Maximum, Minimum, and Average Free Ammonia Concentration
(as Nitrogen) by Month, TCR sites, Tap Samples Only
E 06
2 0.4
WQP
I Maximum
Minimum
A Average
WQP Goal
Source: TCR sites, tap samples only.
Exhibit 3.2.12b Maximum, Minimum, and Average Free Ammonia Concentration
(as Nitrogen) by Month
E
E
< 0.2
Maximum
Minimum
l Aw rage
Source: TCR and supplemental sites, tap and hydrant samples.
IOCCT Review
Final Draft
3-13
March 2007

-------
Exhibit 3.2.13a Maximum, Minimum, and Average Total Nitrate/Nitrite
Concentration as Nitrogen by Month, TCR sites, Tap Samples Only
.9
Maximum N
Minimum N
.8
0.7
[WQP
.6
0.5
.4
.3
IWQP Goal
.2
1
0
2004
2005
Source: TCR sites, tap samples only.
Exhibit 3.2.13b Maximum, Minimum, and Average Total Nitrate/Nitrite as Nitrogen
by Month, Supplemental and TCR sites, Hydrant and Tap Samples
¦ Maximum N
Minimum N
a A\«rage N
_i
CT>
E
WQP I
z
to
c
o
c
©
c
o
o
WQP Goal
z
z
0.2
2004
2005
Source: TCR and supplemental sites, tap and hydrant samples.
IOCCT Review
Final Draft
3-14
March 2007

-------
3.3 Results from Lead Profiling
3.3.1 Lead Profiling Procedure
DCWASA initiated lead profiling at customers' homes in late 2003, focusing on
homes with LSLs. The primary goals of the program were to 1) identify the primary
sources of lead in drinking water, i.e., was lead leaching from the service lines, household
plumbing, brass faucet fixtures, etc., 2) examine the water for co-occurring constituents
that might provide insight into why lead corrosion began increasing in 2002, and 3)
determine if the elevated lead levels were in a particulate or dissolved form. Particulate
forms may indicate that lead scale is detaching from the pipe wall, while dissolved lead
may indicate dissolution through chemical or biochemical mechanisms (Giani et al 2004).
DCWASA also hoped to use this program to help track the performance of any
new lead reduction treatment.
Lead profiling consists of seven main steps:
1.	Document the material, diameter, and length of customer plumbing from the
water main to the kitchen tap. Calculate the volume of water in each pipe
section;
2.	Collect a one-liter baseline sample after high water use (typically in the
morning);
3.	Stop all tap water use in the home for at least 6 hours;
4.	Collect consecutive one-liter samples from the kitchen tap. The number of
samples is based on the total volume of water in the customer plumbing from
the water main to the kitchen tap;
5.	Collect a water hammer sample by first fully opening and closing the faucet
several times over a one-minute period, then running the faucet for 30 seconds
prior to collecting a one-liter sample;
6.	For each one-liter sample (including baseline and water hammer samples),
filter 300 milliliters using a 0.45 micron filter, saving the filtrate for dissolved
lead analysis in the laboratory; and
7.	Analyze the remaining 700 milliliters for temperature, pH, free chlorine, and
total chlorine for each sample on-site. Preserve and save a portion of the
sample for laboratory analysis for lead (total and dissolved), iron, aluminum,
zinc, copper, and HPC (Giani et al., 2004).
DCWASA created a standard graphic to represent lead profile results. They used
a bar chart to plot total and dissolved lead for each profile, with lead concentration in ppb
on the y-axis and liter of water on the x-axis. Using vertical dashed lines, they show
which liters of water represent in-house plumbing, the LSL, and the water main. Exhibit
3.3.1 shows a typical lead profile graph developed by DCWASA. Information on how to
interpret the graphs follows the exhibit. The next section presents and discusses results
from the lead profiling program.
IOCCT Review
Final Draft
3-15
March 2007

-------
Exhibit 3,3.1 Typical Lead Profile Graph Developed by DCWASA
House # 3
3-2-04
200
Main
Lead Service Line
160
160
¦I Lead
¦ Lead Filtered
140
120
Q- 100
80
60
40
20
0
m
Oi
co
m
o
00
CO
+
X
o
CN
Interpreting Lead Profile Graphs:
The lightly shaded bar is total lead. The more darkly shaded bar is dissolved lead. Particulate lead can be
derived by subtracting dissolved lead from total lead values. The x-axis lists the liter number that was
collected from the tap. "1" represents the first liter drawn from the tap after the 6-hour stagnation period.
Liter "2" represents the second liter taken, "18" represents liter 18, and so on. A. number with a phis sign
followed by a second number (e.g., 27 + 3) represents the liter number followed by the amount of minutes
after the last sample was collected. For example, 27 + 3 would represent a sample collected 3 minutes
after the 27th liter. "A" represents water hammering. In order to obtain "X", the faucet was open fully
and closed rapidly several times over a one-minute period. Then the sample was allowed to rim for 30
seconds prior to collection. "0" represents the baseline sample taken in the morning.
IOCCT Review
Final Draft
3-16
March 2007

-------
3.3.2 Lead Profiling Results
DCWASA conducted a total of 46 lead profiles between December 2003 and
January 2006. Exhibit 3.3.2 lists each profile organized by date. Note that all except
"Profile a" were taken at homes with LSLs. Appendix D contains lead profile graphs
showing both total and dissolved lead per liter for all 46 profiles. The Appendix is
organized chronologically and profiles are numbered sequentially for easy reference.
Most profiles were conducted in the NW quadrant of D C. Repeat profiles were
done for seven homes in the District, as indicated by a note in the last column of Exhibit
3.3.2. Repeat profiles were done at two homes before and after partial LSL replacement.
WAS A conducted lead profiling at 12 homes during its temporary conversion from
chloramines to free chlorine for residual disinfection (or "chlorine burn" period) from
April 2 through May 7, 2004. Many of these profiles were repeats of profiles conducted
prior to the chlorine burn.
DCWASA lead profile data can be used to evaluate three different aspects of the
lead corrosion problem in D.C. First, lead profiles provide critical information in
assessing the effectiveness of the orthophosphate treatment. Second, profiles done before
and after partial LSL replacements are useful in assessing the effectiveness of that
program. Lastly, profiles done before and after the chlorine burn as well as profiles done
at homes with and without LSLs support our understanding of the causes of elevated lead
levels in D.C.
Findings Related to Cause of Lead Problem
Appendix D shows that prior to the orthophosphate treatment, peak total and
dissolved lead almost always occurred in the LSLs. The profile for a home with a copper
service line resulted in very low lead levels in the water (<4 ppb). Exhibit 3.3.3
compares a typical profile for a home with an LSL to the profile from the home with a
copper service line. These two findings suggest that lead was leaching predominantly
from the service lines.
Peaks in LSL samples were composed predominantly of dissolved lead, indicating
that a chemical or biological reaction was most likely causing the lead to leach from the
service lines, rather than a physical removal of scale material. Dissolved lead levels
frequently exceeded 100 ppb in the service line portion of the profile prior to
orthophosphate treatment.
As noted previously in this report, it was suspected that the November 2000
conversion from free chlorine to chloramines for secondary disinfection led to increased
lead corrosion. Lead profiling conducted during the chlorine burn supports this theory.
Profiles 15 through 26 all have fairly low total and dissolved lead levels, as shown in
Appendix D. With the exception of a particulate lead spike in the first 1-liter sample in
Profile No. 22, the peak total lead concentration was always below 80 ppb during the
chlorine burn, and peaks continued to fall throughout the burn period.
IOCCT Review
Final Draft
3-17
March 2007

-------
Exhibit 3.3.2 Summary of Lead Profiles Conducted by DCWASA from December
2003 through January 2006
Profile
No.
Date of
profile
Profi e
Profile Address
Quadrant
7/7/2004
12/8/2003
12/15/2003
1/5/2004
1/13/2004
2/9/2004
2/24/2004
3/2/2004
3/9/2004
3/16/2004
3/24/2004
3/24/2004
3/30/2004
3/31/2004
4/1/2004
4/5/2004
4/6/2004
4/6/2004
4/13/2004
4/26/2004
4/27/2004
4/29/2004
4/30/2004
5/3/2004
5/3/2004
5/7/2004
5/18/2004
6/28/2004
7/6/2004
7/16/2004
11/30/2004
12/6/2004
1/6/2005
1/25/2005
2/22/2005
3/30/2005
4/29/2005
5/16/2005
6/1/2005
6/7/2005
7/25/2005
9/28/2005
10/5/2005
11/29/2005
12/12/2005
1/27/2006
Special Profile Conditions
No LSL
Repeat of Profile 1 following partial LSL replacement
Repeat of Profile 3 following partial LSL replacement
Affected by CI Burn
Affected by CI Burn
Affected by CI Burn
Affected by CI Burn
Second repeat of Profile 3 following par ial LSL replacement; Affected by CI Burn
Affected by CI Burn
Second repeat of Profile 1 following par ial LSL replacement; Affected by CI Burn
NW
NW
Repeat of Profile 12; Affected by CI Burn
Affected by CI Burn
Affected by CI Burn
NW
NE
Repeat of Profile 13; Affected by CI Burn
Repeat of Profile 5; Affected by CI Burn
NW
NW
Second repeat of Profile 13; Affected by Partial System Application
Affected by System-Wide Orthophosphate TMT
Affected by System-Wide Ortho phosphate TMT
Repeat of Profile 11; Affected by System-Wide Ortho phosphate TMT
NW
NW
Affected by System-
Affected by System-
Wide Ortho phosphate TMT
•Wide Ortho phosphate TMT
Affected by System-Wide Ortho phosphate TMT
NW
NW
NW
Affected by System-Wide Ortho phosphate TMT
Affected by System-Wide Ortho phosphate TMT
Affected by System-Wide Ortho phosphate TMT
Affected by System-Wide Ortho phosphate TMT
NW
NW
Affected
Affected
by System-
by System-
Wide Ortho phosphate TMT
•Wide Ortho phosphate TMT
Affected by System-Wide Ortho phosphate TMT
Affected by System-Wide Ortho phosphate TMT
NW
NW
Affected
Affected
by System-
by System-
Wide Ortho phosphate TMT
Wide Ortho phosphate TMT
Notes:
Chlorine burn was conducted from April 2, 2004 through
the change in oxidants.
Profile 29 was done approximately 7 weeks after start of
CI = chlorine; LSL = lead service line; TMT = treatment
May 7, 2004. Although the Profile 26 was conducted after the burn, lead leaching was likely still impacted by
partial system application in the 4th high service area.
IOCCT Review
Final Draft
3-18
March 2007

-------
Exhibit 3.3.3 Comparison of Lead Profiles for Homes with Lead and Copper
Service Lines
5-18-04 (Profile No. 26)
ln-house Plumbing
Lead Serivce Line
Main


i-i













I—1


























I—1





¦ —|—, —

—

















¦¦¦¦¦»>
1 2 4 5 6 7 8 9 11 13 16 19 19+3 19+10 0 X
Liter
~ Total Lead DI Dissolved Lead
7-7-04 (Profile a)
60
Copper Service
Line
In House Plumbing
Main
50
40
20
10
o
o
2
3
4
5
7
9
13
15
18
21
24 24+3 24+10 X 01(15)07(15)
~ Total Lead ¦ Dissolved Lead
IOCCT Review
Final Draft
3-19
March 2007

-------
During the chlorine burn period, DCWASA repeated profiles at three locations3.
Exhibit 3.3.4 shows the peak total and dissolved lead in the LSLs prior to and during the
chlorine burn period for comparison. Note that except for Profile No. 22, peak total and
dissolved lead were substantially less during the chlorine burn period than during the
regular chloramine conditions.
Exhibit 3.3.4 Comparison of Peak Lead Concentration Before and During the
Chlorine Burn
Address of Profile
Lead Concentration (ppb) Prior to
the Chlorine Burn
Lead Concentration (ppb) During the
Chlorine Burn
Profile No.
and Date
Peak
Total
Lead
Peak
Dissolved
Lead
Profile No.
and Date
Peak
Total
Lead
Peak
Dissolved
Lead
Location 1
12. 3/30/04
27
11
22. 4/30/04
1101
3
Location 2
13. 3/31/04
110
101
25. 5/7/04
10
8
Location 3
5. 2/9/04
82
75
26. 5/18/04 2
48
38
Notes:
1.	A lead peak of 110 ppb, predominantly particulate lead with almost no dissolved lead, occurred in the
first liter sample. The next highest total lead peak during the profile was 6 ppb. This profile also
showed elevated particulate lead in the water hammer sample (48 ppb). It is suspected that the peak of
110 ppb is either a sample error or potentially corrosion scale abrasion on the valve surfaces of the
faucet.
2.	This profile was done approximately 11 days after the chlorine burn ended on May 7, 2004. Because
changes in lead scale can occur slowly, this profile may have still been impacted by the chlorine burn
event.
Findings Related to PARTIAL LSL REPLACEMENT
Based on information provided by DCWASA, two locations in NW D.C. were
profiled before and after an LSL replacement. The first home was profiled on December
8, 2003 (Profile No. 1). Approximately half of the lead service line was replaced with
copper immediately after the profile was taken. This location was profiled a second time
on January 13, 2004 (Profile No. 4). Lead levels in Profile No. 4 are about half as much
as lead levels in Profile No. 1. Exhibit 3.3.5a shows these profiles on the same page for
comparison.
The second home was profiled on January 5, 2004 (Profile No. 3) and again after
replacement of all but 1 foot of the LSL on February' 24, 2004 (Profile No. 6).
Consistent with an almost complete reduction in LSL, the lead levels decreased from a
peak of approximately 110 ppb (Profile No. 3) to less than 7 ppb in all samples (Profile
No. 6). Exhibit 3.3.5b shows these graphs on the same page for comparison.
3 Two additional sites were repeat profiled during the chlorine burn (see subsection titled Findings Related
to PARTIAL LSL REPLACEMENT). However, DCWASA had already replaced a portion of the lead
service lines at these sites prior to the chlorine burn so they were not included in the analysis.
IOCCT Review
Final Draft
3-20
March 2007

-------
Exhibit 3.3.5a Comparison of Profiles Done Before and After Partial LSL
Replacement At the Same Residence
12-08-03 (Profile No. 1)

180 -r

160 --

140 --

120 -
.Q

Q.
o
o
LL
¦a
80 --
to
0)
-I


60 --

40 --

20 --

0 --
¦
ln-house Plumbing
LSL
Main
9
Liter
13
17
21
25
25+3
~ Total Lead ¦ Dissolved
1-13-04 (Profile No. 4)
180
ln-house
Plumbing
Q. 100
Copper
Replacement
Main
m m m m m l~h
13
Liter
17
21
25
45
~ Total Lead n Dissolved Lead
IOCCT Review
Final Draft
3-21
March 2007

-------
Exhibit 3.3.5b Comparison of Profiles Done Before and After Partial LSL
Replacement At the Same Residence
1-5-04 (Profile No. 3)
ln-house Plumbing j
Lead
Se
•vice Line
Main






	










, n_ i




—


llmmm






Tl m m m m
1	2	4	5	6	8	10	12	14	16 16+3 X	0
Liter
~ Total Lead ~ Dissolved Lead
2-24-04 (Profile No. 6)
ln-house Plumbing
LSL
Copper
Replacement
Main




















1 2 4 5 6 8 10 12 14 16 16+3 16+5 16+7 16+9 16+11 X 0
Liter
~ Total Lead ~ Dissolved Lead
IOCCT Review
Final Draft
3-22
March 2007

-------
Findings Related to the Effectiveness of the Orthophosphate Treatment
To assess the effectiveness of the orthophosphate treatment, we compared profiles
done after the system-wide orthophosphate treatment began on August 23, 2004, to
profiles conducted before that time. Comparisons of peak lead concentrations as
measured in the service line or household plumbing samples and peak lead in water
hammer sample provide useful information. As described in the previous section, LSL
replacement also reduces lead levels in water. Hence, profiles conducted after LSL
replacement (as indicated by a note in the fourth column in Exhibit 3.3.2) were removed
so as not to influence the analysis of orthophosphate effectiveness.
Exhibit 3.3.6 shows peak total and dissolved lead concentrations in the service
lines, averaged for those profiles conducted before and those profiles conducted after the
initiation of orthophosphate treatment. Exhibit 3.3.7 shows similar statistics for water
hammer samples. Exhibits 3.3.8 and 3.3.9 graphically depict the reduction in total and
dissolved lead that occurred in profile samples after the start of orthophosphate treatment.
One home was profiled both before and after the orthophosphate treatment. Exhibit
3.3.10 shows that the total and dissolved lead concentrations at this location decreased
after orthophosphate treatment.
These exhibits collectively show the success of the orthophosphate treatment in
reducing both total and dissolved lead concentrations in drinking water. The average
total and dissolved lead in service lines or household plumbing for profiles conducted
after the initiation of the orthophosphate treatment (15 ppb and 7 ppb, respectively) are
much lower than averages for profiles conducted before the treatment (105 ppb and 94
ppb, respectively). While the average total lead in water hammer samples is not
significantly different before and after the orthophosphate treatment, Exhibit 3.3.9 shows
that the concentration of particulate lead in the water hammer samples is decreasing as
the orthophosphate treatment progresses. This finding suggests that the orthophosphate
treatment may be enhancing the physical stability of lead scales.
It is important to recognize that analyses in Exhibits 3.3.6 through 3.3.9 consider
primarily different homes profiled before and after orthophosphate treatment. The
magnitude of the change, particularly in Exhibit 3.3.8, however, is substantial and
unlikely to be caused by differences in sample sites alone. As noted above, data exists
for one home profiled before and after the orthophosphate treatment. Exhibit 3.3.10
shows a substantial decrease in total and dissolved lead in the profile done after the
treatment, supporting findings in Exhibits 3.3.6 through 3.3.9.
IOCCT Review
Final Draft
3-23
March 2007

-------
Exhibit 3.3.6 Peak and Dissolved Lead Concentration in Service Lines or In-House
Plumbing, Average for Profiles Done Before and After Orthophosphate
Treatment


Average of Peaks in LSLs or


Household
Plumbing

No. of
Total Lead
Dissolved Lead
Profile Category
Profiles
(ppb)
(ppb)
Prior to Orthophosphate Treatment
15
105
94
After Start of Orthophosphate
Treatment
16
15
7
Total
31


Notes: Does not include profiles conducted during the chlorine burn (4/2/04 - 5/7/04), after LSL replacement, or
Profile No. 27, which was done after partial system application of orthophosphate in the 4th high service area.
Exhibit 3.3.7 Lead Concentrations in Water Hammer Samples, Average for
Profiles done Before and After Orthophosphate Treatment


Average of Peaks in Water


Hammer Samples


Total Lead
Dissolved Lead
Profile Category
No. of Profiles
(ppb)
(ppb)
Prior to Orthophosphate Treatment
15
19
12
After Start of Orthophosphate
Treatment
16
22
2
Total
31


Notes: Does not include profiles conducted during the chlorine burn (4/2/04 - 5/7/04), after LSL replacement,
or Profile No. 27, which was done after partial system application of orthophosphate in the 4th high service
area.
IOCCT Review
Final Draft
3-24
March 2007

-------
Exhibit 3.3.8 Peak Lead Concentration in Service Lines or In-House Plumbing for
Individual Profiles by Date
1
Chlorine
Burn V>
» I


• 1
1
D
1
i


1
i
~ !
~
~ ^ I

Orthophosphate treatment (initiated 8/23/04)
r
d i
i
i
~ ~~ ~
~


8 *i
jj
~°.

'~ .
	i ft
~ 3
Q
~
*Sa*ncifli 8 £ a
^ # # # # # # ^ <# # # ^ <# <# # # #
Sampling Date
~ Total Lead ~ Dissolved Lead
Note: Total of 41 profiles shown, see Exhibit 3.3.2 for listing. Excludes profiles done after LSL
replacement (Profile Nos. 4, 6, 19, and 21).
IOCCT Review
Final Draft
3-25
March 2007

-------
Exhibit 3.3.9 Lead Concentrations in Water Hammer Samples for Individual
Profiles by Date
Chlorine
Burn
\


Orthophosphate treatment (initiated 8/23/04)
N


~




~










~





»


~~
m ^ X

~
~
0 »
6°
S
~~
~ ~~ ^ ~
i a ,~ n n 0 0 Off , , , ^ ,n
Profile Date
~ Total Lead	~ Dissolved Lead
Note: Total of 41 profiles shown, see Exhibit 3.3.2 for listing. Excludes profiles done after LSL
replacement (Profile Nos. 4, 6, 19, and 21).
IOCCT Review
Final Draft
3-26
March 2007

-------
Exhibit 3.3.10 Comparison of Profile Conducted Before and After Start of the
Orthophosphate Treatment At the Same Residence
3-24-04 (Profile No. 10)
ln-house
Plumbing
LSL
Main
J2
q.
q.
-o
(0
Q)
40
20
10 12 14 17 20 23 26 30 0 30+3 30+10 X
1 2 3 4 6
~ Total Lead ¦Dissolved Lead
1-6-05 (Profile No. 32)
ln-house
Plumbing
LSL
Main



















pi

i-i












¦	m	— —	—			m	^	„	i i—¦


1 2 3 4 6 8 10 12 14 17 20 23 26 30 0 30+3 30+10 X
Liter
~ Total Lead II Dissolved Lead
IOCCT Review
Final Draft
3-27
March 2007

-------
3.4 LCR Monitoring Data
3.4.1 Description of LCR Dataset
Lead data from January 2003 through December 2005 were evaluated to assess
the effectiveness of the orthophosphate treatment and identify potential spatial and/or
temporal trends in peak lead occurrences. Our dataset comprises a total of 1,662 lead
samples taken during the three years: 833 first-draw and 829 second-draw samples. Data
are for total lead only (assessing the dissolved fraction is difficult for the standard LCR
compliance sample because it requires immediate filtration of the sample, which is
problematic when done by residents). It is important to note that the data analyses in this
report may include more samples than were approved by EPA for use in calculating the
90th percentile LCR compliance value.
Our dataset contains results for more than 450 homes in the District, most with
full LSLs. DCWASA collected data from 100 to 281 homes per 6-month period,
indicating that the sampled homes changed from one 6-month period to the next.
Participants can request to be removed from the LCR program or are removed by
DCWASA if they undergo a full LSL replacement.
Service area information was not provided for the LCR sites; therefore, quadrant
information was used as a proxy to evaluate spatial variability. The samples are not
allocated evenly among the four quadrants, as illustrated in Exhibit 3.4.1. More than
two-thirds of the samples were taken in NW, which is expected since this quadrant
contains a disproportionate amount of the LSLs known to exist in D.C.
Sample collection dates are not spread evenly over a given year. Exhibit 3.4.2
shows the number of first-draw samples collected each month for 2003 through 2005.
DCWASA collected a significantly higher number of LCR compliance samples in
December 2003 and February and March 2004 compared to other months: these months
represent an intense period of research on the D.C. lead issue. In 2005, samples were
collected in the first several months of each 6-month period, with zero or very few
samples collected in May, June, and December.
Exhibit 3.4.1 Number of LCR Samples in Dataset by Geographic Quadrant
(January 2003 - December 2005)

First
Second
Total
Percent of Total
Quadrant
Draw
Draw
Samples
Samples
Northeast (NE)
145
145
290
17%
Northwest (NW)
576
572
1,148
69%
Southeast (SE)
107
107
214
13%
Southwest (SW)
5
5
10
1%
TOTAL
833
829
1,662

IOCCT Review
Final Draft
3-28
March 2007

-------
Exhibit 3.4.2 Number of First Draw LCR Samples in Dataset by Month
¦.'.2003
¦ 2004
2005
Note: Total number of first draw samples from January 2003 - December 2005 = 833. Results for second
draw samples are almost identical.
IOCCT Review
Final Draft
3-29
March 2007

-------
3.4.2 LCR Monitoring Results
Exhibits in this section are presented to demonstrate changes both in the
magnitude and variability of lead concentrations before and after the orthophosphate
treatment (initiated on August 23, 2004). The first set of exhibits track changes in all
lead sample results (both first draw and second draw) from year to year. The last exhibit
evaluates changes in the subset of homes that were sampled both before and after the
orthophosphate treatment.
Analyses of all LCR data
Exhibit 3.4.3 shows the cumulative graphical distribution of all lead results (from
both first draw and second draw samples) for each six-month time period starting in
2003. The distribution of high lead levels is much lower for the first and second halves
of 2005 as compared to 2003 and 2004. The graph also shows that peak values have
decreased significantly in 2005 compared to prior years. The peak total lead
concentrations in the first and second 6-months of 2005 were 51 ppb and 102 ppb,
respectively. 2005 results can be compared to very high peak results of 364 ppb and 265
ppb in 2003 and 2004, respectively.
Similar trends are shown in Exhibit 3.4.4, which compares the percent of all LCR
samples (considering both first and second draw samples) over total lead threshold levels
of 15 ppb, 30 ppb, and 50 ppb for 2003, 2004, and 2005. The exhibit shows consistently
and significantly lower percentages of samples above the threshold levels in 2005
compared to 2003 and 2004 in each category. The exhibit also shows a decrease in the
overall average total lead concentration for all samples from 2003 (14.6 ppb) and 2004
(14.7 ppb) to 2005 (6.9 ppb).
Exhibits 3.4.5a through 3.4.5d expand on Exhibit 3.4.4 by showing the percent of
samples over total lead threshold levels per month for 2003, 2004, and 2005. These
graphs show that occurrence of high lead levels fluctuated more month-to-month prior to
2005. Since orthophosphate treatment began on August 23, 2004, the occurrence of peak
lead concentrations in first and second draw samples has been generally lower for all
threshold categories.
Exhibits 3.4.6a and 3.4.6b display the monthly averages and peak value for first
draw samples, respectively, by each quadrant (data from SW was not considered when
comparing the quadrants due to the low sample size). Second draw samples show similar
results and thus, are not displayed here. The quadrant analysis was conducted to examine
any spatial trends in lead levels. Prior to the orthophosphate treatment, the data show a
slight trend of higher average and peak lead concentrations in NW compared to the NE
quadrant. The highest averages in the fall of 2005 occurred primarily in NW, although
this is likely due to its larger sample size. Overall, there does not appear to be any
meaningful anomaly in the spatial distribution of LCR monitoring results.
IOCCT Review
Final Draft
3-30
March 2007

-------
Exhibit 3.4.3 Cumulative Percent of Total Lead in First and Second Draw Samples
by Monitoring Period
220
200
	2003 First Half (N=208)
	 2003 2nd Half (N=216)
	2004 1st Half (N=565)
	2004 2nd Half (N=250)
2005 1st Half (N=215)
— 2005 2nd Half (N=208)
180
160
140
B 120
100
75%
80%
85%
90%
95%
100%
Cumulative Percentile
Source: LCR Monitoring Data
Notes: Not shown on graph - 2003 Second Half (100% = 364 ppb); 2004 Second Half (100% = 265ppb)
Exhibit 3.4.4 Comparison of Lead Results for 2003, 2004, and 2005
Year
Total Number of
Average Total
Percent of Samples with Total Lead

LCR Samples (first
Lead for all

Greater Than


and second draw)
Samples (ppb)
15 ppb
30 ppb
50 ppb
100 ppb
2003
424
14.6
26%
14%
8%
2%
2004
815
14.7
69%
38%
20%
5%
2005
423
6.5
13%
2%
1%
0%
IOCCT Review
Final Draft
3-31
March 2007

-------
Exhibit 3.4.5a Percent of Peak Lead Values Greater than 15 ppb
Each Month for 2003, 2004, and 2005
70%
60%
50%
-¦~--2003
—•—2004
—A—2005
re 30%
20%
10%
0%

Month
Exhibit 3.4.5b Percent of Peak Lead Values Greater than 30 ppb
Each Month for 2003, 2004, and 2005
70%
60%
50%
o 40%
--~--2003
—•—2004
—6—2005
re 30%
20%
10%
~- -
0%

Month
IOCCT Review
Final Draft
3-32
March 2007

-------
Exhibit 3.4.5c Percent of Peak Lead Values Greater than 50 ppb
Each Month for 2003, 2004, and 2005
70%
60%
50%
o 40%
- - ~ - - 2003
—¦—2004
—£—2005
re 30%
20%
10%
0%

Month
Exhibit 3.4.5d Percent of Peak Lead Values Greater than 100 ppb
Each Month for 2003, 2004, and 2005
70%
60%
50%
g 40%
--~--2003
—¦—2004
—6—2005
30%
20%
10%
0%
Month
IOCCT Review
Final Draft
3-33
March 2007

-------
Exhibit 3.4.6a Monthly Average Total Lead Concentration in First Draw Samples
by Geographic Quadrant (January 2003 - December 2005)
120
100
» NE (N=145)
—*—NW (N=576)
--•¦-¦SE (N=107)
Began Orthophosphate Treatment
on August 23, 2004
Si
Q.
Q.
T3
TO
O
—I
TO
¦*—i
o
H
2003
2004
2005
IOCCT Review
Final Draft
3-34
March 2007

-------
Exhibit 3.4.6b Monthly Maximum Total Lead Concentration in First Draw Samples
by Geographic Quadrant (January 2003 - December 2005)
400
350
Began Orthophosphate Treatment
on August 23, 2004
» NE (N=145)
—*—NW (N=576)
¦¦•¦¦¦SE (N=107)
300
n 250
1 200
H 150
100
50
0
2003
2004
2005
IOCCT Review
Final Draft
3-35
March 2007

-------
We assessed the magnitude of the elevated lead levels after orthophosphate
treatment by reviewing individual results for those samples collected in 2005 with total
lead concentration greater than 15 ppb. Exhibit 3.4.7 shows that most of the values are
less than 30 ppb, with only two first draw samples greater than 30 ppb at 64 ppb and 51
ppb. Additional analysis revealed that approximately one half of the first draw results
greater than 15 ppb are associated with a second draw sample result greater than 15 ppb.
The rest of the samples were followed by second draw sample results of less than 15 ppb
Exhibit 3.4.7 Results for Samples with Total Lead concentration >15 ppb in 2005,
Ranked High to Low
Sample Type
Total No. ot
Samples Taken
in 2005
t otal No. ot
Samples with Total
Lead >15 ppb
Concentration (ppb) for Samples >15 ppb, Ranked High to Low
First Draw
211
17
64 51 28 27 26 25 24 24 22 20 20 17 17 17 16 16 15
Second Draw
212
22
102 39 37 32 28 26 26 23 22 21 20 18 18 18 17 16 16 16 16 16 15 15
Analysis ofLCR data for Homes Sampled Before and After Orthophosphate Treatment
Analyses ofLCR monitoring data is complicated by the fact that different homes
are sampled during different times of the year. Only a fairly small subset of the more
than 450 homes in our dataset were sampled more than once from January 2003 through
December 2005. Approximately 95 homes were sampled both before the start of the
orthophosphate treatment on August 23, 2004 and after January 1, 2005 when the
treatment process had stabilized.
Exhibit 3.4.8 compares results for homes sampled before and after the start of
orthophosphate treatment. In all cases, the percent of samples with high total lead
concentrations was significantly reduced following the implementation of the
orthophosphate treatment.
IOCCT Review
Final Draft
3-36
March 2007

-------
Exhibit 3.4.8 Comparison of Lead Levels for Subset of Homes Sampled Before
and After Start of Orthophosphate Treatment
Sampling Time
Frame
Sample
Type
Total
Number
of LCR
Samples2
Average
Total Lead
for all
Samples
(ppb)
Perc
Tota
ent of
Lead
Samples with
Greater Than
15
ppb
30
ppb
50
ppb
100
ppb
Prior to
Orthophosphate
Treatment
1st Draw
127
27.8
63%
32%
14%
2%
2nd Draw
127
27.9
55%
31%
15%
5%
After
Orthophosphate
Treatment1
1st Draw
148
7.1
9%
1%
1%
0%
2nd Draw
149
6.4
10%
1%
1%
1%
because adjustments were still being made to the treatment process, we excluded data from August 24,
2004 through December 31, 2004 from this subset.
2A11 samples are from a total of 95 homes.
3.5 Analysis of Bacteriological Activities in the Distribution System
Although the primary goal of orthophosphate addition in D.C. is to reduce lead
levels, orthophosphate treatment can have a beneficial impact on microbiological activity
in the D.C. system. There was a concern that orthophosphate could increase
bacteriological activity and exacerbate the nitrification problem, since phosphate acts as a
nutrient for some microorganisms. However, phosphate-based corrosion inhibitors
combined with appropriate disinfection have been shown to reduce biofilms growth.
A review performed by Dr. Anne Camper of past studies and of D.C. distribution
system conditions suggests that orthophosphate indeed may help alleviate
microbiological activity in the long term. (See Appendix E for a copy of Dr. Camper's
review memo.) In systems with a considerable amount of unlined iron pipe—such as
D.C.'s—biofilm activity is apt to be even greater when humic (organic) substances
interact with corroded iron oxides. By reducing the corrosion of iron pipes,
orthophosphate can help eliminate some of the favorable conditions for biofilm growth,
thus improving the biological stability of drinking water. Initial reactions with pipe
linings, however, may cause bacteria to slough off, resulting in a temporary spike in
HPCs and total coliforms. As shown later in this section, D.C. did experience a peak in
HPCs and total coliforms in September 2004, soon after the orthophosphate initiation.
Flushing was recommended and implemented in response to HPC spikes in the
distribution system.
Dr. Camper did not believe it was likely that phosphate would increase
bacteriological activity, as phosphate was likely not the limiting nutrient to bacterial
growth in this system. Rather, carbon is the limiting nutrient for biofilms in D.C.'s
system.
IOCCT Review
Final Draft
3-37
March 2007

-------
Corrosion control treatment such as orthophosphate has been shown to reduce
nitrification by allowing for increased chlorine residuals in areas of historically high
water age. Because ammonia-oxidizing and nitrite-oxidizing bacteria (AOB and NOB,
respectively) perform nitrification in drinking water systems, reduced microbial activity
in distribution systems often has direct implications for nitrite and nitrate levels.
To assess the impact of the DCWASA orthophosphate treatment on
microbiological activity in the distribution system, we evaluated the following data:
•	Percent positive total coliform samples by month for 2003 - 2005;
•	HPCs; and
•	WQPs related to nitrification (nitrite, nitrate, free ammonia, and total
chlorine).
WQPs were reviewed by date and by site to evaluate potential spatial and
temporal trends. Tracking results by location can be particularly helpful because both
nitrification and biofilm growth are highly localized water quality problems that depend
on water age (disinfectant residual) and pipe composition. It is important to note that
DCWASA revised its unidirectional flushing procedures in mid-2004, confounding
potential conclusions regarding the impact of orthophosphate on bacteriological
conditions
Exhibit 3.5.1 summarizes the water quality data assessed in this section of the
report.
IOCCT Review
Final Draft
3-38
March 2007

-------
Exhibit 3.5.1 Number of Samples in the Dataset

2004
2005
Total
TCR
Supplemental
TCR
Supplemental
HPC
2,146
155
285
847
3,433
Total Chlorine
2,556
183
2,363
479
5,581
Nitrite
95
215
305
943
1,558
Nitrate
58
94
67
399
618
Free Ammonia
87
220
305
922
1,534
Notes: TCR = DCWASA's TCR Compliance Monitoring Program sites.
Supplemental = DCWASA's Supplemental Monitoring Program sites.
3.5.1 Total Coliforms
Historically, HPC and total coliform monitoring results for the D.C. distribution
system follow a seasonal pattern, with the highest results occurring in the spring and
summer. Serving a population of approximately 550,000, DCWASA is required to
collect 210 TCR samples per month, although they routinely collect many more samples
than this number.
Exhibit 3.5.2 depicts monthly results for positive total coliforms in 2003, 2004,
and 2005. In 2003 and 2004, DCWASA recorded very similar results for total coliforms,
in fairly similar patterns. During each of these years, the D.C. distribution system
experienced the fewest positive total coliforms in February, and it experienced peaks in
both May and in late summer months (August and September), separated by a drop
positive total coliforms in early summer months (June and July). This drop may be
attributed to the annual, spring-time conversion (in 2003 and 2004) of chloramines to free
chlorine for residual disinfection (i.e., the chlorine burn).
IOCCT Review
Final Draft
3-39
March 2007

-------
Exhibit 3.5.2 Percent Positive TC Samples
8%
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
MCL	2003 - - - .2004 	 2005
In August of 2004, WA initiated the full-system application of orthophosphate.
Soon after, for the month of September 2004, DCWASA experienced a total coliform
peak in excess of the monthly limit of 5% positive total coliform samples. This resulted
in a TCR violation, and was followed by aggressive flushing by DCWASA in problem
areas. By October 2004, total coliform levels were in compliance and began a downward
trend in keeping with seasonal patterns. Because orthophosphate has been known to
loosen iron and biofilm deposits that accumulate along pipe walls (as pointed out in Dr.
Anne Camper's memo in Appendix E), the initiation of this corrosion inhibitor is the
suspected cause of the TCR violation.
Although the initiation of the orthophosphate corrosion inhibitor may have caused
a spike in total coliform levels, orthophosphate may eventually reduce bacteria levels by
eliminating distribution system conditions that facilitate biofilm growth. As shown in
Exhibit 3.5.2, D.C. experienced generally lower positive total coliform results following
the system-wide orthophosphate addition. As with previous years, total coliforms were
lowest in February of 2005 and were highest in the spring and summer. Unlike previous
years, 2005 results peaked in June, a month where the system typically observed a drop
in positive total coliforms. To minimize interference with the orthophosphate
application, WA skipped the annual chlorine burn in 2005, possibly explaining the total
coliform peak in June. D.C. also experienced a lesser peak in September 2005, but, true
to historical patterns, levels continued to decline into the fall and winter months.
3.5.2 Heterotrophic Plate Count (HPC) Results
Similar to total coliforms, HPC results reflect biological conditions throughout the
distribution system. Unlike total coliforms, HPC results, measured in colony-forming
units per mililiter (CFU/mL), include a wide range of bacteria types and, thus, are
generally higher in distribution systems and better illustrate changes in bacterial quality.
IOCCT Review
Final Draft
3-40
March 2007

-------
DCWASA measured HPC at its TCR sites and supplemental sites. In 2004,
DCWASA analyzed a total of 2,301tap and hydrant samples at these sites; DCWASA
analyzed approximately 1,132 samples in 2005. Exhibit 3.5.3 depicts monthly average
HPC results for all samples from January 2004 through December 2005. The number of
samples taken per month ranged from 60 to 370. For this analysis, if multiple samples
were taken at a site at the same location (hydrant or tap) on the same date, an average was
taken to create a single value per site per location per day. Also, values found to be <1 or
>5,700 were noted as 1 and 5,700, respectively.
Exhibit 3.5.4 compares average hydrant and tap results for 2004 and 2005. Note
that hydrant and tap samples exhibited similar HPC results, with the exception of higher
hydrant results for July 2005.
DCWASA's HPC counts—like its total coliform results—follow a seasonal
pattern, with peaks in the warmer spring and summer months and lows in the fall and
winter.
•	The D.C. water system experienced its highest average HPC in September
2004 (404 CFU/mL), the same month that DCWASA exceeded the total
coliform MCL;
•	Average monthly HPCs for the first half of 2005 (ranging from 34 to 129
CFU/mL) were slightly higher than for the first half of 2004 (ranging from 7
to 119 CFU/mL); and
•	Average monthly results for the second half of 2005 (90 to 296 CFL/mL)
were lower than the same time period in 2004 (87 to 404 CFL/mL). The high
results for early 2005 may be related to mild temperature conditions in the
winter of 2005.
Slightly lower HPC results for spring and summer of 2005, compared to spring
and summer 2004, may be evidence of the effect of orthophosphate in making system
conditions less favorable to bacterial growth, although differences are not substantial
enough to draw firm conclusions.
Because microbial activity in distribution systems is highly localized, HPC results
for the D.C. system were also analyzed by site. Exhibits 3.5.5a and 3.5.5b show average
HPC results for TCR sites and supplemental monitoring sites, respectively, for the period
of September 2004 through December 2005. The HPC data do not appear to exhibit a
strong pattern by site or by pressure zone. The sites with the highest results (>5,700
CFU/mL)—sitelDs L-5 BKJV, A1H-2 BKJV, and A1H-4 BKJV, all tap samples-
experienced these peaks at different times throughout the September 2004 to December
2005 period of analysis, suggesting these were isolated events.
IOCCT Review
Final Draft
3-41
March 2007

-------
Exhibit 3.5.3 Average Monthly HPC Results for 2004 and 2005
450
400
350
300
E 250
150
V -
100
*0 -
2004
2005

(D
c
sz
o
0)
_Q
0)
_Q
0)
.Q
0)
_Q
CL
3
ro
ro
Source: TCR and supplemental monitoring sites, tap and hydrant samples, including downstream and
upstream samples. Note that the number of samples taken per month varies.
Exhibit 3.5.4 Average Monthly HPC Results for Tap and Hydrant Samples
(Jan. 2004- Dec. 2005)
450
- -O- - Hydrant
400
350
-r 300
250
LL
~ 200
x 150
100
Q.
Q.
O)
O)
Q.
Q.
2004
2005
Source: TCR and supplemental monitoring sites, tap and hydrant samples, including downstream and
upstream samples. Note that the number of samples taken per month varies.
IOCCT Review
Final Draft
3-42
March 2007

-------
Exhibit 3.5.5a Average HPC Results for TCR Sites (Sept. 2004- Dec. 2005)
2500
E
5
1500
500
S CO O)
4 ~ +
S CO O)
First High
Second High
ii
Third High
CN

CN
CO

m
r-~
X
X
X
X
X
X
X


<
<
<
<
<
4th
Anacostia 1st
High

High

A.
2nd
High
Source: TCR Monitoring Sites, taps and hydrants, no upstream or downstream samples.
Note: The number of samples taken per site varies.
Exhibit 3.5.5b Average HPC Results for Supplemental Sites (Sept. 2004- Dec. 2005)
11
uu
is is
is is
is is
^ ^ ^
First High
Second High
Third high
Fourth High Anacostia First High
> >
2 2
m m
V ^
X X
A. 2nd
High
Source: Supplemental monitoring Sites, taps and hydrants, no upstream or downstream samples.
Note: The number of samples taken per site varies.
IOCCT Review
Final Draft
3-43
March 2007

-------
3.5.3 Nitrification Parameter Monitoring
Nitrification is the biochemical process by which bacteria convert ammonia to
nitrite and nitrate. EPA regulates both nitrite and nitrate in finished water (water entering
the distribution system) because of their impacts on human health. Nitrite's MCL is 1
mg/L as nitrogen, while nitrate's MCL is 10 mg/L as nitrogen.
Nitrification most often occurs in low-flow and remote areas of distribution
systems where disinfectant residuals are lower and bacteria counts are higher. It is a
particularly vexing problem for chloraminated systems (such as D.C.'s) when excess
ammonia accumulates and helps accelerate the nitrification process. With growing
concentrations of nitrate and nitrite, nitrification decreases disinfectant (chloramine)
residuals and thus leaves the water system more susceptible to increased microbial
activity and contamination events.
According to the AwaaRF guide, Optimizing Chloramine Treatment, 2nd Edition
(AwwaRF 2004), nitrification is most easily observed by reviewing monitoring results for
reduced disinfectant residual (total chlorine) and elevated nitrite. Nitrification is also
associated with elevated nitrates, higher HPCs, and elevated free ammonia.
Since September of 2004, DCWASA has regularly monitored for nitrate, nitrite,
and free ammonia at TCR and supplemental sites in the system. It should be noted that
DCWASA only measured nitrate if nitrite results for a sample exceeded 0.1 mg/L or if
free ammonia results either exceeded 0.4 mg/L or were below 0.2 mg/L. Our dataset
contains more than 100 free ammonia results each month and nearly that many nitrite
results each month. The sample size for nitrate is smaller, representing an average of 25-
30 samples each month. In addition to these parameters, total chlorine is always recorded
in the field using HACH brand test kits. Exhibits 3.5.6 through 3.5.9 show monthly
maximum, minimum, and average values for total chlorine, nitrite, nitrate, and free
ammonia from September 2004 to December 2005.
IOCCT Review
Final Draft
3-44
March 2007

-------
Exhibit 3.5.6 Total Chlorine Results by Month
8.0
5.0
| Monthly Maximum
- I Monthly Minimum
~ Monthly Average
3.0
2.0
Source: TCR and Supplemental Monitoring Sites, taps and hydrants.
Exhibit 3.5.7 Nitrite as Nitrogen Monitoring Results by Month
0.4
0.35
0.3
O)
E
(n
ra
0.25
0.2
¦= 0.15
0.1
0.05
Max of N
Min of N
a Avg of N
2004
Ul
1 , 1
E °
2005
Source: TCR and supplemental Monitoring sites, tap and hydrant samples.
IOCCT Review
Final Draft
3-45
March 2007

-------
Exhibit 3.5.8 Nitrate as Nitrogen Monitoring Results by Month
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
I I T


CD
CD
-Q
-Q
E
O
CD
O
-*—>
Q_
O
CD

CO

CD
-Q
E
CD
>
O
2004
CD
-Q
E
CD
O
CD
Q
03
03
=5
-Q
CD
M ax of N
Min of N
a Avg of N
o
03
Q_
<
03
O)
=5
<


CD
CD
-Q
-Q
E
O
CD
O
-*—>
Q_
O
CD

V)

2005
CD
-Q
E
CD
>
O
CD
-Q
E
CD
O
CD
Q
Source: TCR and supplemental monitoring sites, taps and hydrants.
Exhibit 3.5.9 Free Ammonia as Nitrogen Monitoring Results by Month
0.8






Max of N
Min of N
a Avg of N


















k
I












k
i



k
k

> *
t i
i
k
*
t i

k
i
k i
i

k
i
k
* .








































September

ro
.2 0.4
1 0.3
E
< 0.2
0)
£
£ 0.1
0
Source: TCR and supplemental monitoring Sites, taps and hydrants.
IOCCT Review
Final Draft
3-46
March 2007

-------
Overall, there does not appear to be a strong relationship among the four
parameters (total chlorine, nitrite, nitrate, and free ammonia levels) since the start of
orthophosphate application in August 2004. While nitrite and total chlorine do not
exhibit a strong pattern in concentration over time, average nitrate levels were generally
greatest in the winter months of both 2004 and 2005. Free ammonia levels remained, on
average, steady over time. DCWASA observed a spike in average free ammonia in
September 2005. Evaluation of average HPC counts did not reveal a related increase in
bacteriological activity during this time frame or a relationship with nitrite or nitrate
levels. To determine if this spike was localized or system-wide, we graphed free
ammonia results by site for September 2005. Exhibit 3.5.10 shows average, maximum,
and minimum free ammonia for each supplemental and TCR site for which data were
available during this month. No patterns appear by site or pressure zone.
Exhibits 3.5.11 and 3.5.12 show nitrite and nitrate concentration by Site ID,
respectively, for both TCR and supplemental monitoring sites. Exhibit 3.5.11 reveals
high nitrite maxima only at a few specific sites; Exhibit 3.5.12 reveals a nitrate spike at
one site (A2H-1 BKJV). The TCR site 3H-5 and the supplemental monitoring sites L-5
BKJV, 1H-1 BKJV, 3H-2 BKJV, and A1H-2 BKJV represent those locations with the
highest nitrite results, all in excess of 0.1 mg/L. Elevated nitrite concentrations at these
five sites—possibly indicative of localized nitrification events—are graphed over time in
Exhibit 3.5.13. It appears that, while DCWASA experienced high nitrite levels for
selected locations in late 2004, nitrite levels for these same sites declined over time. This
is confirmed in Exhibits 3.5.14 through 3.5.18, which track total chlorine and nitrite level
for each site over time. Note that the graphs of individual sites have different scales for
nitrite and chlorine. In almost every graph, higher nitrite values are accompanied by
lower total chlorine results, suggesting that, indeed, nitrification was occurring.
DCWASA personnel reported that at at least one location, they discovered a cross
connection which likely contributed to high nitrite levels.
IOCCT Review
Final Draft
3-47
March 2007

-------
Exhibit 3.5.10 Free Ammonia as Nitrogen Results by Site (September 2005)
TT
A	*
A
A	A
E
E
<
0.1
| Max
| Min
AAvg

Source: TCR and supplemental monitoring sites, taps and hydrant data.
Exhibit 3.5.11 Nitrite as Nitrogen Results by Site (Sept. 2004 - Dec. 2005)
0.4
0.35
o> 0.3
>>>>>>
Supplemental
>>>>>>> CO
IIX X X X X
T" T- T- T- T- CM CM
<<<<<<<
- CM lO CD r-~
III
CM CM
< < <
Source: TCR and Supplemental monitoring sites, taps and hydrants.
IOCCT Review
Final Draft
3-48
March 2007

-------
Exhibit 3.5.12 Nitrate as Nitrogen Results by Site (Sept. 2004 - Dec. 2005)
E 1
| Max
I Min
AAvg
0.4

i
i •
> > >
2 2 2
m cq co
M (O "t
XXX
> > >
2 2*:
CD CO CD
CN (O t—
X I <
>	>	>
2	2	2
m	cq	co
¦t	in	Y
±	±	±
T-	T-	CN
<	<	<
i i :
T- CM C
< < «
Supplemental
Source: TCR and Supplemental Monitoring Sites, taps and hydrants.
Exhibit 3.5.13 Maximum Nitrite Values as Nitrogen for Sites with High Nitrite
Peaks
0 3
0 1H-1BKJV B3H-2BKJV A3H-5 XA1H-2 BKJV • L-5 BKJV
02
0.1

Q. O ^
a)	H
CO	z
2004	2005
Source: TCR and Supplemental Monitoring Sites, taps & hydrants.
IOCCT Review
Final Draft
3-49
March 2007

-------
Nitrite Concentration as Nitrogen (mg/L)
o
o
o
o
cn cn
OJ CD
September
K)
o
o
October
November
December
January
February ~
March
April
May
June
o
o
U1
July
August
September
October
November
December
o
cn
Total Chlorine Concentration (mg/L)
m
o
C
H
$
&
Nitrite Concentration as Nitrogen (mg/L)
N>
O
O
September
October
November
December
January
February
N>
o
o
Ui
March
August
September
October
November
December
m
o
So
N>
Ui
o
co

CO CO
11
CJ1 CJ1
a ~
» a
CD
O
N>
N>
Ui
co
co
bi
m
x
CO
cn
O
3
D
¦o
73
(D
(/>
c_
p+
V)
CO
I
cn
Total Chlorine Concentration (mg/L)

-------
Nitrite Concentration as Nitrogen (mg/L)
o	o	o	o	o	o
o	o	o	o	o	o
o	—>¦	K>	CO	Js.	ui	cn
September
October
o
o
November
December
January
February
March
April
May
June
o
o
cn
July
August
September
October
November
December
o
o
CO
CO
Total Chlorine Concentration (mg/L)
in
o
e
Nitrite Concentration as Nitrogen (mg/L)
m
x
G>
cn
O
3
D
¦o
7J
(D
(/>
c_
p+
(/>
-*1
0
>
1
io
CO
*
c_
<
H
$
&
oooooooo
Q-iMCOJi.OlO)>IOO
N>
O
O
-t^
September
October
November k*
December f
January
February
March
April
May
N>
O
O
ai
June
July
August
September
October
November
December
\
~~
o
3L
O
~~
m
x
w
cn
o>
o
3
D
"0
73
(D
(/>
c_
p+
(/>
00
*
Total Chlorine Concentration (mg/L)

-------
21
i.
a
o
o
n
H
£
ft)
a'
3
in
o
H
u>
i
Ui
to
2
£2
O
tr
K>
O
o
Nitrite Concentration as Nitrogen (mg/L)
oooo oooo
O ¦ ¦ ¦ ¦ o
oooo
O M	O) CO
September
M
O
O
October
November
December
January
February
March
April
N>
o
0
01
May
June
July
August
September
October
November
December
CD 00 M
II ~
OJ	OJ
I	I
I	I
N>	K)
00	00
c_	c_
<	<
£	I
03	3;
O
CD
03
CO
O
CO
CD
CO
Ol	Ol	Ol
Total Chlorine Concentration (mg/L)
m
x
w
cn
oo
o
3
D
¦o
73
(D
(/>
c_
p+
(/>
o>
I
IO
DO
*

-------
4. Conclusions and Recommendations
This chapter provides conclusions and recommendations based on the research
summarized in Chapter 2 and the water quality data reviewed in Chapter 3. As noted in
the relevant chapters, pipe loop data are through March 2006, and water quality data are
through December 2005. Research conducted and data collected since that time could
potentially change the conclusions presented in this chapter.
1)	Orthophosphate has effectively reduced lead levels in D.C. drinking water.
There is a wealth of laboratory and field data that support the success of the
orthophosphate treatment. Since the addition of orthophosphate, lead levels are generally
near or below the AL after 8 hours of stagnation in both the DCWASA and WA pipe
loops. Lead profiling results show significant reductions in dissolved and total lead
following the orthophosphate treatment. Profiling results include some evidence that the
orthophosphate scale is stabilizing over time as the treatment progresses. LCR
compliance monitoring shows significant reductions in total lead levels at customers taps
after only a few months of orthophosphate treatment. No other corrosion inhibitor
(including zinc orthophosphate, blended polyphosphate, pH adjustment, or stannous
chloride) outperformed orthophosphate in the DCWASA or WA pipe loop studies.
2)	Circulation and flow through pipe loops are useful tools in selecting and
optimizing corrosion control treatment.
The circulation loops constructed by DCWASA were invaluable in selecting the
optimal corrosion control treatment for D.C. The overall lead concentration in the pipe
loop studies correlated well with the peak lead concentration from the lead profiles of
actual service lines and internal home plumbing. WA flow through pipe studies
corroborated circulation loop findings, although beginning lead concentrations in the WA
study were somewhat lower.
3)	Partial Lead Service Line Replacement is an effective strategy to reduce lead
in D.C. drinking water.
Field investigations by DCWASA have shown that the potential mechanical
disruption of lead corrosion scales during partial LSL replacements is not a serious threat
if care is taken during the cutting process and the line is vigorously flushed. Lead
profiles conducted before and after partial replacements support DCWASA's field
studies. Laboratory experiments have shown that when a new length of copper tubing is
connected to a well-passivated LSL during partial LSL replacement, the area of galvanic
influence is very small and decreases rapidly. Thus, any potential lead release caused by
galvanic corrosion is minimal.
IOCCT Review
Final Report
4-1
March 2007

-------
4)	Operation of the orthophosphate treatment is progressing well, but needs
improvement in some areas.
LCR compliance monitoring and lead profiling show slightly higher lead levels
than are measured in either pipe loop study. This is expected given that water quality is
more stable under controlled laboratory conditions than in the distribution system.
However, improved lead control may still be achieved through more consistent water
quality in the distribution system. A comparison of observed orthophosphate, pH, free
ammonia, and nitrate/nitrite levels in the distribution system to WQP goals set by EPA
show some deviation. pH in particular is critical to the success of orthophosphate
treatment and should be closely tracked to ensure that it stays within acceptable treatment
limits.
Further investigation into the following areas may be warranted to determine:
•	The cause of intermittent low orthophosphate readings in June and October
2005;
•	The cause of high free ammonia in September 2005; and
•	The cause of consistently elevated free ammonia levels (average monthly free
ammonia in the distribution system was higher than the WQP goal in all
months evaluated except December 2005).
5)	The potential drawbacks of a chlorine burn may outweigh the benefits.
The review of water quality data did not reveal significant nitrification problems
in the D.C. system. In fact, there is limited evidence to suggest that occurrence of
nitrification may have decreased since the orthophosphate application. Both DCWASA
and WA pipe loop data show potential increases in lead leaching resulting in the change
back to chloramines after a chlorine burn (DCWASA Pipe Loop 3 and WA Rack 3),
although measured changes are very small and may not be meaningful. Given some LCR
compliance samples still exceeded 15 ppb through the end of 2005, any decision to
modify finished water quality should be approached with extreme caution. The potential
benefits of a chlorine burn may not outweigh the risk of an AL exceedance.
While bacterial activity in the D.C. distribution system appears to be under
control, nitrification is a potential problem as long as the D.C. system is using
chloramines as a secondary disinfectant. Continued monitoring for nitrification
parameters throughout the distribution system is recommended.
6)	A further reduction in orthophosphate dose may not be advisable at this
time.
Based on discussions with DCWASA and outside experts, WA reduced the target
orthophosphate dose from 3.5 mg/L to 2.5 mg/L in February 2006. Technical experts
agreed that 3.5 mg/L is much higher than typical maintenance doses for orthophosphate
IOCCT Review
Final Report
4-2
March 2007

-------
and that maintaining a dose at this concentration over the long term could potentially
cause problems because of the formation of other solids.
This report does not include a comprehensive review of 2006 data to evaluate
impacts of the change in orthophosphate dose on lead levels in the distribution system.
DCWASA Pipe Loops 1 and 3 and WA Rack 7 all show a very slight increase in lead
concentration at a reduced orthophosphate concentration of 2.5 mg/L. Furthermore, WA
Rack 6 revealed a more substantive lead increase when orthophosphate dose was reduced
to 1.0 mg/L.
It is difficult to predict how slight changes in lead release in the pipe loops will
correlate to changes in lead release in the distribution system. However, based on the
data reviewed in this document, this report recommends that WA and DCWASA
maintain an orthophosphate target dose of 2.5 mg/L until evidence shows that the
orthophosphate scale has stabilized in the distribution system.
7) Orthophosphate treatment has not resulted in a long-term increase in
microbial activity, and may be responsible for a slight decrease.
As predicted by industry experts, D.C. experienced a spike in HPC levels and a
total coliform violation soon after the orthophosphate application began system-wide.
This spike was likely the result of biofilm sloughing as orthophosphate reacted with iron
corrosion scale. Between September 2004 and December 2005, D.C. did not observe
elevated HPC or total coliform levels. Data suggest that occurrence of coliforms, HPC
levels, and incidence of high nitrite levels have all declined through 2005.
It should be noted that modifications to DCWASA's unidirectional flushing
program in 2004 and 2005 may also have contributed to biofilm and nitrification control.
IOCCT Review
Final Report
4-3
March 2007

-------
This page left intentionally blank.
IOCCT Review	4-4	March 2007
Final Report

-------
5 References
AWWA. 2004. Proceedings of Getting the Lead Out: Analysis & Treatment of Elevated
Lead Levels in D.C.'s Drinking Water. Proceedings AWWA Water Quality
Technology Conference. San Antonio, TX.
AWWA. 2005. Managing Change and Unintended Consequences: Lead and Copper
Rule Corrosion Control Treatment.
AwwaRF and DVGW-Technologiezentrum Wasser. 1996. Internal Corrosion of Water
Distribution Systems. 2nd edition. AwwaRF Report 90508. Project #725.
AwwaRF. 2004. Optimizing Chloramine Treatment. 2nd Edition.
Bell, G.E.C. 1995. Observation on the Effect of Grounding on Water Piping,
CORROSION/95. Orlando, FL.
Bell, G.E.C. 1998. Effects of Grounding on Metal Release in Drinking Water. AWWA
Inorganics Contaminants Workshop. San Antonio, TX.
Code of Federal Regulations. 1991. 40 CFR, Part 141.85.
Cottis, R.A., S. Turgoose, and R. Newman. 1999. Corrosion Testing Made Easy:
Electrochemical Impedance and Noise. National Association of Corrosion
Engineers. Houston, TX.
DCWASA. 2005. Research Newsletter.
Duranceau, S.J. Electrical Grounding, Pipe Integrity and Shock Hazard. AwwaRF Report
90702. Project #913.
Duranceau, S.J., D. Townley, and G.E.C. Bell. 2004. Optimizing Corrosion Control in
Distribution Systems. AwwaRF Report 90983. Project #2648.
Economic and Engineering Services. 1990. Lead Control Strategies. AwwaRF Report
90559. Project #406.
Edwards, M. and S.H. Reiber. 1997. A General Framework for Corrosion Control Based
on Utility Experience. AwwaRF Report 90712. Project #910.
Edwards, M., and T. Holm. 2001. Role of Phosphate Inhibitors in Mitigating Lead and
Copper Corrosion. AwwaRF Report 90823. Project #2587.
Edwards, M., S. Jacobs and D. Dodrill. 1999. Desktop Guidance for Mitigating Pb and
Cu Corrosion Byproducts. Journal of American Water Works Association.
91:5:66-77.
IOCCT Review
Final Report
5-1
March 2007

-------
Giani, R., M. Edwards, C. Chung, and J. Wujek. 2004. Use of Lead Profiles to
Determine Source of Action Level Exceedances from Residential Homes in
Washington , D.C. Proceedings AWWA Water Quality Technology Conference
Sunday Workshop. San Antonio, TX.
Giani, R., W. Keefer, and M. Donnelly. 2005. Studying the Effectiveness and Stability of
Orthophosphate on Washington D.C.'s Lead Service Line Scales. Proceedings
AWWA Water Quality Technology Conference. Quebec City, Quebec.
Giani, R., M. Donnelly, and T. Ngantcha. 2005. The Effects of Changing Between
Chloramine and Chlorine Disinfectants on Lead Leaching. Proceedings AWWA
Water Quality Technology Conference. Quebec City, Quebec.
Hecht, P.M., and E. A. Turner. 2004. Washington Aqueduct Desktop & Flow-Through
Pipe Study. Presented at Getting the Lead Out: Analysis & Treatment of Elevated
Lead Levels in D.C.'s Drinking Water. Proceedings AWWA Water Quality
Technology Conference. San Antonio, TX.
Keefer, W., H.M. McDonald, and R. Giani. 2004. Lead Leaching Rates from Fixtures.
Presented at Getting the Lead Out: Analysis & Treatment of Elevated Lead Levels
in D.C.'s Drinking Water. Proceedings AWWA Water Quality Technology
Conference. San Antonio, TX.
Kirmeyer, G.J. et al. 2000. Distribution System Water Quality Changes Following
Implementation of Corrosion Control Strategies. AwwaRF Report 90764. Project
#157.
Kirmeyer, G.J., A.M. Sandvig, G.L. Pierson, and C.H. Neff 1994. Development of a
Pipe Loop Protocol for Lead Control. AwwaRF Report 90650. Project #604.
Kirmeyer, G.J., B.M. Murphy, A. Sandvig, G. Korshin, B. Shaha, M. Fabbricino, and G.
Burlingame. 2004. Post Optimization of Lead and Copper Control Monitoring
Strategies. AwwaRF Report 90996F. Project #2679.
Korshin, G. 2005. Fundamental Mechanisms of Lead Oxidation: Effects of Chlorine,
Chloramine and Natural Organic Matter on Lead Release in Drinking Water.
Proposal to the National Science Foundation.
Korshin, G.V. and J.F. Civardi. 2004. Real Time Lead Leaching Rates During Stagnation
Using Corrosion Potentiostat Methods. Presented at Getting the Lead Out:
Analysis & Treatment of Elevated Lead Levels in D.C.'s Drinking Water.
Proceedings AWWA Water Quality Technology Conference. San Antonio, TX.
Korshin, G.V. and J.F. Ferguson. 1999. Corrosion and Metal Release for Lead
Containing Materials: Influence of NOM. AwwaRF Report 90759. Project #182.
IOCCT Review
Final Report
5-2
March 2007

-------
Lytle, D.A. and M.R. Schock. 2005. The Formation of Pb(IV) Oxides in Chlorinated
Water. Proceedings AWWA Water Quality Technology Conference. Quebec
City, Quebec.
Maas, R.P., J.C. Pitch, and A.M. Smith. 2005. Effects of Fluorides and Chloramines on
Lead Leaching from Leaded-Brass Surfaces. Asheville Environmental Quality
Institute Technical Report #04-137.
Marshall, B., J. Rushing, and M. Edwards. 2003. Confirming the role of aluminum solids
and chlorine in copper pitting corrosion. In Proceedings of AWWA Annual
Conference. Denver CO.
Maynard, B. and B. Mast. 2005. The Application of Raman Spectroscopy to Mineralogy
of Lead Pipe Scales. Project Workshop Presentation AwwaRF 3018 -
Contribution of Service Lines to Lead and Copper Rule Compliance Issues.
Seattle, WA.
Reiber, S. 1991. Galvanic Stimulation of Lead/Tin Solder Sweated Joints. Journal of
American Water Works Association. 83(7).
Reiber, S. and R. Giani. 2005. National Impacts from D.C.'s Lead Experience. Water
Environmental Federation Technical Conference October, 31 2005
Reiber, S., W. Keefer, L. Dufresne, and R. Giani. 2004. Circulation Loop Testing
Provides Rapid Assessment of Corrosion Control Strategies. Proceedings AWWA
Water Quality Technology Conference Sunday Workshop. San Antonio, TX.
Sarin, P., V.L. Snoeyink, J. Bebee, K.K. Jim, M.A. Beckett, W.M. Kriven, J.A. Clement.
2004. Iron Release from Corroded Iron Pipes in Drinking Water Distribution
Systems: Effect of Dissolved Oxygen. Water Research. 38: 1259-1269.
Schock, M.R. and D.A. Lytle. 2005. The Formation of Pb(IV) Oxides in Chlorinated
Water. Journal of American Water Works Association. 97(11): 102-114.
Schock, M.R. et al. 2005. Mode of Occurrence, Treatment, and Monitoring Significance
of Tetravalent Lead. Proceedings AWWA Water Quality Technology Conference.
Quebec City, Quebec.
U.S. EPA. 2004. Action Plan to Reduce the Occurrence of Lead Leaching from Service
Lines, Solder, or Fixtures Into Tap Water In the District of Columbia And
Arlington County and Falls Church, Virginia.
Vasquez, F.A. et al. 2006. Effect of Free Chlorine and Chloramines on Lead Release in a
Distribution System. Journal of American Water Works Association. 98(2): 144-
153.
IOCCT Review
Final Report
5-3
March 2007

-------
Washington Aqueduct and CH2MHill. 2004. Desktop Corrosion Control Study prepared
for US EPA Region III. Philadelphia, PA.
Washington Aqueduct. 2005. Report of Water Analysis.
Wujek, J.J. 2004. Minimizing Peak Lead Concentrations after Partial Lead Service Line
Replacements. Proceedings AWWA Water Quality Technology Conference. San
Antonio, TX.
Wujek, J.J. and H.M. McDonald. 2004. Evaluation of Lead Service Line Replacement
Methodologies to Optimize Peak Lead Concentrations. Presented at Getting the
Lead Out: Analysis & Treatment of Elevated Lead Levels in D.C.'s Drinking
Water. Proceedings AWWA Water Quality Technology Conference. San
Antonio, TX.
IOCCT Review
Final Report
5-4
March 2007

-------
Appendix A IOCCT Designation Letter, USEPA 2004

-------
This page intentionally left blank.

-------
^osr^
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
REGION III
1650 Arch Street
Philadelphia, Pennsylvania 19103-20
PRO^°
w.
By Federal Express
Thomas P. Jacobus
August 3, 2004
General Manager
Washington Aqueduct
5900 MacArthur Blvd., N.W.
Washington, DC 20016-2514
Jerry N. Johnson
General Manager
District of Columbia Water and Sewer Authority
5000 Overlook Ave., SW
Washington, DC 20032
Gentlemen:
The United States Environmental Protection Agency Region III ("EPA") has primacy for
the Public Water System Supervision ("PWSS") Program in the District of Columbia. The
primacy agency is responsible for implementing the PWSS Program and the National Primary
Drinking Water Regulations ("NPDWRs"), including designation of optimal corrosion control
treatment ("OCCT") under the Lead and Copper Rule ("LCR") for public water systems. The
NPDWRs define OCCT at 40 C.F.R. § 141.2 as "the corrosion control treatment that minimizes
lead and copper concentrations at users' taps while insuring that the treatment does not cause the
water system to violate any national primary drinking water regulations." The Preamble to the
LCR states that the effect of corrosion control treatment on the waste water stream also may be
considered in selecting OCCT. 56 Fed. Reg. 26460, 26480 (June 7, 1991).
On July 16, 1997, EPA conditionally designated an OCCT for the drinking water
treatment and distribution system for the District of Columbia and required additional study. In
February 2000, EPA designated the use of pH adjustment as the OCCT for the drinking water
distribution system for the District of Columbia, which required the Washington Aqueduct to
maintain a pH in the finished water between 7.7 and 8.5. On May 17, 2002, EPA revised its
designation of OCCT with respect to the monthly pH goals.
On August 26, 2002, the District of Columbia Water and Sewer Authority ("DCWASA")
submitted a final report to EPA Region III stating that, during the compliance period July 1,
2001 - June 30, 2002, the level of lead in first draw water samples from 53 residences served by
the District of Columbia drinking water distribution system was 75 parts per billion ("ppb") at
the 90th percentile. This monitoring result exceeded the lead action level of 15 ppb at the 90th
percentile. On July 29, 2003, DCWASA reported to EPA Region III that, during the compliance
period January - June 2003, the level of lead in the first draw water samples from 104 residences
was 40 ppb at the 90th percentile. For the July - December 2003 compliance period, DCWASA
Printed on 100% recycled/recyclable paper with 100% post-consumer fiber and process chlorine free.
Customer Service Hotline: 1-800-438-2474

-------
reported that the level of lead in first draw water samples was 63 ppb at the 90th percentile. For
the January -June 2004 compliance period, DCWASA reported that the level of lead in first draw
water samples was 59 ppb at the 90th percentile.
On May 28, 2004, the U.S. Environmental Protection Agency Region III ("EPA")
approved an interim modification of the OCCT for the drinking water distribution system for the
District of Columbia. The interim modification consisted of an application of the corrosion
inhibitor orthophosphate to the 4th High Pressure Zone of the District of Columbia drinking
water distribution system. The 4th High Pressure Zone is hydraulically isolated from the
remainder of the District of Columbia's drinking water distribution system, but is representative
of the entire system in terms of component materials (lead service lines, unlined cast iron pipe,
etc.). The purpose of the proposed partial system application was to assess, prior to any full
system application, operational characteristics and any unanticipated effects. At the time EPA
approved this interim modification, it was expected that, absent any unresolvable problems and
subject to EPA's approval, the system-wide OCCT ultimately would be modified to include
application of orthophosphate to maintain reduced levels of lead in the entire District of
Columbia drinking water distribution system.
This letter modifies EPA's interim designation of the OCCT for the District of Columbia
distribution system. The interim OCCT for the District of Columbia drinking water distribution
system shall consist of the application of the corrosion inhibitor orthophosphate subject to the
conditions and water quality parameters ("WQPs") set forth below. This designation is being
considered an "interim" designation because it applies only to the passivation period. A final
designation for maintenance of corrosion control will be issued once the system is passivated.
The Washington Aqueduct will use an orthophosphate product in the form of phosphoric acid
that meets ANSI/NSF Standard 60: Drinking Water Chemicals — Health Effects. The
Washington Aqueduct will apply an initial passivation dose that will continue until the lead level
in the 90th percentile of tap water samples is equal to or below the 0.015 mg/1 (15 ppb) lead
action level, or until water quality results indicate the need to reduce the dosage earlier. The
initial passivation dose should be designed to achieve a residual of > 3.0 mg/L measured as
orthophosphate in tap samples. Following initial passivation, it is anticipated that the
Washington Aqueduct will apply a maintenance dose sufficient to achieve a residual of
approximately 0.5 - 1.5 mg/L measured as orthophosphate in tap samples, or a dose sufficient to
ensure lead levels remain equal to or below 0.015 mg/1 (15 ppb) at the 90th percentile of tap
samples.
The interim WQPs set herein apply to the initial passivation dose. The LCR
contemplates that the primacy agency will establish final water parameters following passivation
of the system. EPA will review monitoring results and system operation records after
passivation has been reached and will establish final WQPs for the Washington Aqueduct and
DCWASA for maintenance of corrosion control following passivation. EPA anticipates that it
will establish final WQPs that will allow for smaller variations in the parameters than the interim
WQPs for the passivation dose set forth in this letter.
The Washington Aqueduct is a wholesaler of water and has no distribution system of its
own. The Washington Aqueduct sells water to a number of other water systems. DCWASA,
2

-------
Arlington County Public Works and the City of Falls Church are consecutive community water
systems and provide no additional treatment to the water received from the Washington
Aqueduct before they distribute it to their customers. The Washington Navy Yard is a
consecutive community water system that purchases its water from DCWASA. The Anacostia
Annex, the Naval Observatory and the Naval Security Station are consecutive non-transient,
non-community water systems that purchase water from DCWASA. Ronald Reagan National
Airport is a consecutive non-transient, non-community water system which has the capability of
providing additional disinfection to the water it receives from the Washington Aqueduct. Thus,
any treatment, including OCCT, applied by the Washington Aqueduct will affect all of its
customer water systems. The public water systems affected are:
PWS Identification Number	Public Water System
DC0000001
Washington Aqueduct Division, U.S. Army Corps of

Engineers
DC0000002
District of Columbia Water and Sewer Authority

("DCWASA")
DC0000003
Naval Station Washington - Washington Navy Yard
DC0000004
Naval Station Washington - Anacostia Annex
DC0000005
Naval Observatory
DC0000006
Naval Security Station
VA6013010
Arlington County Public Works
VA6013080
Ronald Reagan Washington National Airport
VA6610100
City of Falls Church Public Utilities
Background

Following DCWASA's report that it had exceeded the LCR lead action level in 2002,
EPA recognized the need to conduct additional research into the cause of elevated levels of lead
in the District of Columbia drinking water distribution system. (Arlington County and the City
of Falls Church have not reported elevated lead levels in their drinking water distribution
systems.) EPA contracted with an independent corrosion expert in May 2003 to research the
cause of the increased lead levels. The expert presented a written report to EPA in October
2003. DCWASA developed a research strategy, which it presented to the Washington Aqueduct,
Arlington County, the City of Falls Church and EPA in January 2004. EPA formed the
Technical Expert Working Group ("TEWG") to address the problem of elevated lead levels in
tap water in the District of Columbia in February 2004. The TEWG consists of representatives
from EPA Region III, EPA Headquarters' Office of Ground Water and Drinking Water, EPA's
Office of Research and Development, the Washington Aqueduct, DCWASA, the District of
Columbia Department of Health, Arlington County, Falls Church and the Centers for Disease
Control and Prevention.
The TEWG's Production Treatment Operations Team, led by the Washington Aqueduct
and its contractor, developed a Desktop Study. The Desktop Study considered various treatment
options, including maintaining a constant high pH at the Dalecarlia and McMillan water
treatment plants using either quicklime (current practice) and/or sodium hydroxide (caustic
3

-------
soda), and feeding a corrosion inhibitor, such as orthophosphate, while maintaining a constant
pH throughout the year of about 7.7. The Desktop Study reviewed the various reports and
recommendations previously prepared for the Washington Aqueduct and/or EPA, conducted a
telephone survey about treatment techniques employed by drinking water treatment and
distribution facilities similar to Washington, D.C.'s, performed mathematical modeling of
corrosion abatement strategies, and reviewed water treatment industry accepted corrosion control
practices.
The TEWG and the Washington Aqueduct originally recommended introduction of
orthophosphate as a corrosion inhibitor. The Desktop Study and its recommendations were
reviewed by an Independent Peer Review Panel assembled by EPA's Office of Ground Water
and Drinking Water in Washington, D.C. Based upon one of its members' greater familiarity
with the use of zinc orthophosphate, the Peer Review Panel recommended the use of zinc
orthophosphate. On April 30, 2004, EPA designated use of zinc orthophosphate for partial
system application in the 4th High Pressure Zone.
On May 28, 2004, EPA modified its April 30, 2004 designation, and EPA designated use
of orthophosphate (rather than zinc orthophosphate) for the partial system application of a
corrosion inhibitor in the 4th High Pressure Zone. This modification was based on concerns
raised by Arlington County regarding its wastewater treatment plant's ability to handle the
anticipated added zinc load from any future full system application and on data suggesting that
zinc orthophosphate and orthophosphate are equally effective in achieving corrosion control
endpoints. For an explanation of EPA's considerations, see Letter from Jon M. Capacasa to
Thomas P. Jacobus and Jerry N. Johnson (May 28, 2004).
Orthophosphate (in the form of phosphoric acid) is an approved and commonly used
drinking water additive. Phosphoric acid, one of the three common forms of orthophosphate and
the form proposed for the full system treatment by the Washington Aqueduct, is a proven
corrosion inhibitor that is currently being used by the Washington Suburban Sanitary
Commission for corrosion inhibition on Potomac River water. It also is used in a number of
large distribution systems, including distribution systems in New York, Wisconsin and
elsewhere. See The Cadmus Group, Inc., Investigation of Potential Environmental Impacts due
to the use of Phosphate-based Corrosion Inhibitors in the District of Columbia (July 22, 2004)
("Cadmus Report"). As noted in TEWG's Desktop Study, orthophosphate "has been used for
many years as a reliable, known and safe chemical additive that has been shown to reliably
reduce lead and copper corrosion." See Letter from Jon M. Capacasa to Thomas P. Jacobus and
Jerry N. Johnson (May 28, 2004).
EPA has considered the known studies and data. In addition, EPA has consulted with
members of the TEWG, the Independent Peer Review Panel, other experts attending the recent
Lead and Copper Rule Workshop, and regulators in states that have water distribution systems
using orthophosphate and/or zinc orthophosphate. EPA has concluded that zinc orthophosphate
and orthophosphate are likely to be equally effective in achieving corrosion control end points in
the District of Columbia drinking water distribution system. It should be noted that the proposed
application of orthophosphate will not immediately decrease lead levels in the tap water. It is
expected that lead levels will decrease over the course of implementing the proposed treatment.
4

-------
A measurable reduction of lead levels may take more than six months and possibly more than a
year.
The partial system application to the 4th High Pressure Zone commenced on June 1,
2004. After reviewing the available data the Technical Expert Working Group reached
consensus that there were no water quality monitoring results that would warrant delaying full
system application of orthophosphate as a corrosion inhibitor. Although data from the 4th High
Pressure Zone application have not yet shown a reduction in lead levels, this was expected based
on experts' opinions and TEWG members' experiences elsewhere. Data did show some elevated
numbers of heterotrophic plate count bacteria at several sample sites and elevated color and iron
levels in about one third of samples taken from fire hydrants. These results can reasonably be
expected during the start-up phase of a phosphate-based corrosion inhibitor treatment. There
have been no customer complaints of red water in the 4th High Pressure Zone. The equipment
installed to perform this temporary chemical feed has performed well. In summary, no results
indicated unresolvable problems in connection with application of orthophosphate in the 4th
High Pressure Zone, and no unexpected results from the water quality monitoring were seen.
The TEWG's consensus, from its discussion on July 28, 2004, is that there is no reason to delay
application of a full system treatment.
EPA considers this interim OCCT designation to be part of an ongoing process. Pursuant
to 40 C.F.R. § 141.82(h), "[u]pon its own initiative or in response to a request by a water system
or other interested party, [EPA] may modify its determination of the optimal corrosion control
treatment... where it concludes that such change is necessary to ensure that the system continues
to optimize corrosion control treatment. A revised determination shall be made in writing, set
forth the new treatment requirements, explain the basis for [EPA's] decision and provide an
implementation schedule for completing the treatment modifications." EPA's interim OCCT
designation is informed by its understanding that additional studies are being undertaken. For
example, the TEWG is conducting pipe loop experiments to evaluate optimal treatment dose, pH
and other factors. The Washington Aqueduct also is studying means to optimize pH stability.
Other ongoing research includes: investigation into galvanic corrosion related to water meter
replacement; flow-through pipe loop studies; lead corrosion behavior studies in household
plumbing (lead profiling); pipe scale analysis; a study of lead leaching rates; and pipe loop
studies to compare the relative effectiveness of zinc orthophosphate and orthophosphate in
reducing lead levels. In addition, studies are planned on galvanic corrosion related to partial lead
service line and water meter replacement and potential impacts on corrosion rates from electrical
system grounding to home plumbing systems.
Conditions and Water Quality Parameters
The Washington Aqueduct will use an orthophosphate product that meets ANSI/NSF
Standard 60: Drinking Water Chemicals - Health Effects. Based on the NSF certification, the
application of orthophosphate is not expected to cause adverse human health effects. In
addition, the application of orthophosphate is not expected to have an adverse effect on the Blue
Plains Wastewater Treatment Plant or the Arlington County Water Pollution Control Plant. See
Cadmus Report. Discharges of potable water from the drinking water distribution systems in the
District of Columbia, Arlington County and the City of Falls Church to receiving streams
5

-------
through planned and unplanned events such as line flushing, water main breaks, combined sewer
overflows, lawn watering, etc. are not expected to cause any adverse effects to the receiving
streams.
The application of orthophosphate may cause temporary rust-colored or "red water"
events in the tap water, a potential increase in total coliform bacteria due to breakdown of
biofilm on the pipes, and an increase in calcium (lime) deposits in water mains and residential
plumbing. Total coliform are indicator bacteria and any increase in total coliform bacteria
caused by the application of orthophosphate does not present a human health risk. Information
regarding these possible effects and what to do if there is "red water" was provided by EPA in
two public information sessions conducted on April 27 and 29, 2004, by DCWASA in a public
information session conducted May 24, 2004, and by the TEWG's fact sheet, which is posted on
the District of Columbia Department of Health's website. EPA, DCWASA, the District of
Columbia Department of Health and the Washington Aqueduct will continue with outreach
programs designed to inform consumers of steps that should be taken as a result of the
application of orthophosphate. DCWASA has informed EPA that DCWASA intends to send a
letter to its customers informing of them of the application of orthophosphate and the steps that
should be taken if they experience discolored water. EPA and the TEWG are scheduling
additional public information meetings as well.
DCWASA, with support from Washington Aqueduct contractor flushing crews, will
proceed with a unidirectional water main flushing program as quickly as possible to complete
flushing the entire DCWASA distribution system prior to the onset of freezing weather. During
this and all subsequent water main flushing events, DCWASA shall implement best management
practices (in addition to dechlorination) to minimize discharges associated with water main
flushings to storm sewers and receiving streams. Such best management practices shall include,
but not be limited to, exercising best efforts to avoid conducting line flushings in combined
sewer overflow ("CSO") service areas during or immediately after storm events and identifying
and monitoring relevant CSOs during line flushings to determine whether the line flushings are
associated with any discharges from CSOs.
No later than December 1, 2004, or within ten (10) days of completing a study to analyze
pH control (whichever is sooner), the Washington Aqueduct shall submit to EPA a study
analyzing methods of pH control designed to achieve the WQP goals set forth herein.
Monitoring
Pursuant to 40 C.F.R. § 141.82(f), EPA is required to set WQPs for water supplies
implementing corrosion control treatment. The interim WQPs and WQP goals set forth herein
apply both to water entering the distribution system and to water quality as measured in tap
water samples from the distribution system collected pursuant to 40 C.F.R. § 141.87 and this
letter. The Washington Aqueduct will be responsible for monitoring and achieving the WQPs
for water entering the distribution system. DCWASA will be responsible for monitoring and
achieving the WQPs in the distribution system. The interim WQP for orthophosphate in water
entering the distribution system is set as a range to account for the possibility that the
Washington Aqueduct may need to adjust treatment for a short period of time to respond to
6

-------
temporary conditions in the distribution system (such as red water). The pH values for waters
entering the distribution system are expressed as a range to allow the Washington Aqueduct to
make adjustments to consistently attain WQPs in the distribution system. The interim WQP
goals will serve as targets which both the Washington Aqueduct and DCWASA should strive to
achieve.
Along with the typical parameters required of systems using a phosphate-based corrosion
inhibitor, EPA is requiring that DCWASA monitor for and report supplemental parameters in the
distribution system to help determine whether the application of orthophosphate causes any
unexpected water quality changes. Because the purpose of monitoring for and reporting the
supplemental parameters is to assist EPA, DCWASA and the Washington Aqueduct in
evaluating and fine-tuning operations, the requirement is for monitoring and reporting, and no
numeric values have been assigned to the supplemental parameters.
The Washington Aqueduct shall conduct monitoring for WQPs according to the
requirements in 40 C.F.R. § 141.87. DCWASA shall conduct monitoring for WQPs according
to the requirements in 40 C.F.R. § 141.87, with the following modifications. With respect to
frequency, DCWASA shall monitor for WQPs monthly at all sample locations. DCWASA also
shall monitor all locations selected pursuant to 40 C.F.R. § 141.87 for all parameters set forth
below, including those parameters designated as "monitor and report." DCWASA's compliance
with the numeric interim WQPs established herein shall be assessed based upon monitoring
conducted at the locations selected pursuant to 40 C.F.R. § 141.87.
In addition to monitoring at the locations selected pursuant to 40 C.F.R. § 141.87,
DCWASA also shall monitor at least twenty-five (25) additional or "supplemental" locations to
provide additional information on any changes in the water chemistry during the passivation
period. Monitoring at the supplemental locations shall consist of monitoring and reporting for
all parameters set forth below, including pH, orthophosphate, free ammonia nitrogen and
nitrite/nitrate nitrogen. Although DCWASA must monitor and report parameter values for these
twenty-five supplemental locations to comply with 40 C.F.R. §§ 141.82 & 141.87, compliance
with the numeric interim WQPs for pH, orthophosphate, free ammonia nitrogen and
nitrite/nitrate nitrogen will not be assessed based on the data from these supplemental locations.
Monitoring at the supplemental locations shall be conducted in accordance with a supplemental
monitoring plan described below.
Prior to the full system application of orthophosphate, DCWASA shall develop and
submit to EPA for review a supplemental water quality monitoring plan. This plan shall
identify at least twenty-five (25) additional or "supplemental" sample locations beyond those
required by 40 C.F.R. § 141.87. The additional or "supplemental" sample locations shall be
representative of dead-end and low flow areas of the distribution system confirmed using
DCWASA's calibrated hydraulic model. The supplemental water quality monitoring plan also
shall include monitoring for all parameters listed below at all sampling locations, both those
identified in the supplemental water quality monitoring plan and those identified pursuant to 40
C.F.R. § 141.87.
7

-------
Reporting
The Washington Aqueduct and DCWASA shall report WQP monitoring data as required
by 40 C.F.R. §141.90 with the modifications below. WQP reports are due to EPA within ten
(10) days of the end of each monthly monitoring period. Where the tenth day falls on a weekend
or holiday, reports are due the first business day thereafter.
DCWASA shall also report to EPA data collected under the supplemental WQP
monitoring plan within ten (10) days of the end of each monthly monitoring period. Data for the
parameters identified below as "monitor and report" will be used by EPA, the Washington
Aqueduct, and DCWASA for evaluating and fine-tuning operations.
At this time, EPA is setting interim WQPs and WQP goals for the passivation period. As
stated above, once the distribution system is passivated, EPA will establish more refined WQPs
to be achieved in connection with maintenance of corrosion control. EPA anticipates that the
WQP goals provided herein will form the basis of the more refined WQPs associated with the
maintenance of corrosion control that will be established by EPA following the initial
passivation period.
Interim Water Quality Parameters for the Passivation Period
For water entering the distribution system during passivation period
(These apply to Washington Aqueduct):
pH
Orthophosphate
InterimWOPs
7.8-7.9 ±0.3
1.0-5.0 mg/1*
WOP Goals
7.8 ±0.1
3.0 mg/1*
*dose necessary to reach this residual in tap samples
For water samples from the distribution system during passivation period
(These apply to DCWASA):
pH
Orthophosphate
Interim WQPs
7.7 ±0.3
1.0-5.0 mg/1
WOP Goals
7.7 ±0.1
3.0 mg/1
residual in tap samples
free ammonia nitrogen
nitrate/nitrite nitrogen
0.5 mg/1
0.5 mg/1
0.2 mg/1
<0.1 mg/1
Supplemental Parameters
free chlorine
total chlorine
temperature (°C)
alkalinity
monitor & report
monitor & report
monitor & report
monitor & report
8

-------
Calcium hardness as CaC03
Calcium dissolved hardness
iron
aluminum
total dissolved solids
oxidation-reduction potential
sulfate
color
heterotrophic plate count bacteria
total coliform bacteria and fecal
monitor & report
monitor & report
monitor & report
monitor & report
monitor & report
monitor & report
monitor & report
monitor & report
monitor & report
coliform or E. coli testing of
total coliform positive samples
free ammonia
total ammonia nitrogen
dissolved P04
total P04
monitor & report
monitor & report
monitor & report
monitor & report
monitor & report
Thank you for your efforts to help secure a long term solution to elevated lead levels in
the District of Columbia drinking water distribution system. If you or your staff require
additional information, please contact Rick Rogers, Water Protection Division, EPA Region III
at (215) 814-5711.
cc: Hugh J. Eggborn, Director, Office of Water Programs, Culpepper Field Office, Virginia
Department of Health,
Robert J. Etris, Director of Public Utilities, City of Falls Church, Virginia
Randolph W. Bartlett, Arlington County Department of Public Works
William J. Brown, Ronald Reagan National Airport
Thomas Calhoun, District of Columbia Department of Health
Thomas Lewis, Naval District Washington
Sincerely,
Jon M. Capacasa, Director
Water Protection Division
EPA Region III
9

-------
This page intentionally left blank.

-------
Appendix B Oxidant/Disinfectant Chemistry and Impacts of Lead
Corrosion

-------
This page intentionally left blank.

-------
Oxidant/Disinfectant Chemistry and Impacts on Lead Corrosion
Michael R. Schock
Chemist
Water Supply & Water Resources Division
U. S. Environmental Protection Agency
Cincinnati, OH 45268
Richard Giani
Water Quality Manager
District of Columbia Water & Sewer Authority
Dept of Water Services
Water Quality Division
3900 Donaldson PI, NW
Washington, DC 20016
Background
In response to continued elevated lead levels throughout the District of Columbia's distribution
system, a collaboration was begun with the District of Columbia's Water & Sewer Authority
(WASA) and Water Resources Division of U. S. Environmental Protection Agency's (USEPA)
Office of Research and Development in late winter of 2004 to investigate the causes of the
sudden increases in lead release. Lead levels had been slightly above the Action Level in 1991
through 1994. Although increases in pH and the addition of orthophosphate were investigated in
pilot studies and desktop corrosion studies at that time, the 90th percentile lead levels dropped to
below 0.012 mg/L throughout the mid- to late 1990s. This coincided with a program of
increased flushing and considerably elevated free chlorine dosages (often as high as 4 mg/L) to
provide improved control of biofilms and bacterial regrowth. However, by 2000, concern with
the upcoming increases in stringency of disinfection byproduct regulations led to a decision to
change to chloramination in an attempt to keep trihalomethane (THM) levels below the new
standards. The change was made in November of 2000, and in the next Lead and Copper Rule
(LCR) monitoring round, the 90th percentile lead level surprisingly was found to have jumped to
over 0.07 mg/L. Subsequent LCR 90th percentile results have remained over 0.04 mg/L. The
general history of lead levels is illustrated in Figure 1, keeping in mind that some sampling
periods are represented by some different monitoring sites in the site pool, and there were also
differences in the number of sampling sites required.
Although there was considerable skepticism that chloramination could have been the cause of the
sudden elevation of lead levels, some apparent increases of lead and copper release and attack on
brass had been reported before in several investigations (Larson et al., 1956; James M.
Montgomery Consulting Engineers, 1982; Francis, 1985a, 1985b; James M. Montgomery
Consulting Engineers, 1985; Schock, 1999). In fact, reviews of general aspects of the redox
chemistry of disinfectants and consideration of the different chemical behavior of metals in their
different common valence states (eg. Fe, Mn, Cu, Pb) do suggest that any substantial change in
the oxidation/reduction potential (ORP) could substantially alter the behavior of pipe scales and
1
Auwft <&<-, 20°4

-------
the potential for corrosion byproduct or contaminant release (AWWARF-TZW, 1996; Schock,
2000; Schock & Holm, 2003).
While most water treatment and water chemistry specialists are very familiar with the radically
different solubilities and scale formation properties of ferrous iron versus ferric iron, or cuprous
copper versus cupric copper, there is much less realization of the same potential behavior of lead.
Potential-pH diagrams for the lead system going back many years have a prominent stability
field for the highly insoluble lead dioxide (PbOs) solid (Delahay et al., 1951; Pourbaix et al.,
1966; Pourbaix, 1973; Schock, 1980, 1981; Schock & Wagner, 1985; Schock etal., 1996;
Schock, 1999). Thus, an analogy can be drawn between the Pb(IV)-Pb(II) redox couple and the
Fe(III)-Fe(II) redox couple. The higher valence state forms oxide or oxyhydroxide scale phases
of much lower solubility than those of the lower valence state. In the case of lead, however, the
ORP required for the transformation of Pb(II) to Pb(IV) is much higher than for the ferrous to
ferric iron transformation. Because of typical free chlorine dosages, consumer dislike of
chlorinous tastes and odors, normal water residence times, and the usual pipe wall and bulk
water oxidant demands, such highly oxidizing conditions will not be common amongst public
water systems in the United States.
USEPA analyses of scales from lead pig-tails and service lines in the late 1980's and early
1990's verified that one or both of the common polymorphs of PbC>2 (plattnerite and scrutinyite)
were present in varying degrees on the pipes from several different water systems (Schock et al.,
1996). Thus far, of more than 85 lead pipe specimens obtained from 34 water systems, at least
16 specimens representing 9 systems have either a-Pb02, or p-PbC>2, or both present in clearly
identifiable quantities. More samples may have trace amounts that are hard to positively
confirm. Usually, the PbC>2 exists in the form of patches or a thin surficial layer at the water
boundary. PbC>2 comprises nearly the entire scale material in the pipe samples from WASA and
Cincinnati, and the majority of the scale material in some Oakwood, OH specimens. In
Madison, WI, it formed a rather distinct surficial layer in contact with the water. The presence of
PbC>2 is associated with waters of persistently high ORP. The elevated ORP could be caused by
any of several mechanisms. For example: pristine low-NOM ground waters with little bulk
oxidant demand allowing significant persistence of free chlorine; waters that very effectively
passivate iron and remove its oxidant demand (such as hard waters with high buffering
intensity); waters with low oxidant demand resulting from oxidative treatments such as
greensand filtration (enabling the stability of high ORP); and many other possible scenarios. An
additional cause can be the use of very high dosages of free chlorine to combat biofilm problems
or to overcome corrosivity towards iron and its pipe wall demand.
More pipe analysis investigations by the U. S. Environmental Protection Agency in 2000 and
2001 revealed much more evidence for the importance of tetravalent lead compounds as very
large primary components of lead service line scale material, particularly those in Cincinnati,
Ohio and Madison, Wisconsin (Schock et al., 2001). In the Cincinnati pipe, representing long
periods of relatively high disinfectant concentrations, followed by a combination of granular
activated carbon and free chlorine treatment, essentially the entire scale was composed of the
two Pb02 polymorphs, with only traces of basic lead carbonate. It is not known when the Pb02
scale formed in the Cincinnati pipe, as there were no historical scale analyses during several
earlier treatment schemes. For the distribution system area where the Cincinnati pipe specimens
2

-------
were obtained, it has employed elevated pH (8.5 to 9.2) since the 1980s for corrosion control.
The lead service line specimens from Madison had a thin layer of PbC>2 at the water interface,
with PbCC>3 making up the bulk of the underlying scale. An important amount of tetravalent
lead scale material was also found in pipe from Oakwood, Ohio (another pH 7 high alkalinity
water where greensand filtration for iron removal is employed). Taken in combination, these
observations finally provide a reasonable hypothesis to explain the apparent anomaly observed in
many field studies in which high alkalinity waters did not tend to produce nearly as high levels
of lead release as would be expected from the knowledge of lead solubility chemistry and bench-
scale tests (Dodrill & Edwards, 1994; Dodrill, 1995; Dodrill & Edwards, 1995; Edwards et al.,
1999).
Pipe Scale Analyses Results
Lead service line specimens from residential homes in the District were shipped to the WSWRD,
U. S. Environmental Protection Agency lab in Cincinnati, with their ends sealed to preserve
humidity and moisture. The specimens were cut longitudinally with a band saw having a fine
metal-cutting blade, and were photographed with a stereomicroscope at 6 to 66 X. Figures 2 and
3 illustrate some of the specimens from the WASA system that were analyzed, and how they
compare to similar scales composed of large amounts of tetravalent lead compounds from other
water systems.
~	Scale was removed and analyzed by X-ray powder diffraction using the same procedures as
described previously (Schock et al., 2001). When lead carbonate and hydroxycarbonate solid
phases are present, the positive identification of PbC>2 can be somewhat problematic, because
some of the significant diffraction peaks of p-PbC>2 and a-PbC>2 overlap with some of the peaks
from PbO (litharge) and PbCOs (cerussite). While the d-space accuracy of the carefully-
calibrated diffractometer should be more than sufficient to positively identify the phases of
interest, the naturally-formed solids tend to have lattice distortions and peak broadening from
small crystallite sizes. Both of these factors complicate the positive identification of PbC>2
polymorphs.
Therefore, to confirm the existence of tetravalent lead phases and corroborate the XRD results,
the Pb pipe scales were additionally analyzed by X-ray absorption near edge (XANES) and X-
ray absorption fine structure (XAFS) spectroscopies. For XANES and XAFS studies, a thin
layer of a Pb scale was smeared onto Kapton tape and folded back on itself. Pb (13035 eV) Lm-
XANES and XAFS data were collected at Sector 20-BM (Pacific Northwest Consortium -
Collaborative Access Team (PNC-CAT)) at the Advanced Photon Source at Argonne National
Laboratory, Argonne, IL. The electron storage ring operated at 7 GeV. Three scans were
collected at ambient temperature in fluorescence mode with an Ar-purged Lytle detector. A 0.5
mm premonochromator slit width and a Si(III) double crystal monochromator detuned by 20% to
reject higher-order harmonics was employed. The beam energy was calibrated by assigning the
first inflection of the absorption edge of lead metal foil to 13 035 eV. Reference samples of PbO
(massicot), PbC03 (cerussite), Pb3(COs)2(OH)2 (hydrocerussite), Pb5(P04)3Cl
(chloropyromorphite), Pb(N03)2, P-Pb02 (plattnerite), and Pb304 were commercially obtained
3

-------
for comparison with the XANES and XAFS spectra. The phase identities were confirmed by
XRD. The collected scans for a particular sample were averaged, the data were then normalized,
and the background was removed by spline fitting using WinXAS 2.0 (Ressler, 1998).
Because the scales were very thin and tenaciously adherent to the pipe surfaces, only small
amounts (tens of milligrams) were available for analysis. Subsamples have also been sent by
USEPA to the U. S. Geological Survey laboratory in Denver, for digestion and elemental
analysis to investigate general scale chemistry and to try to help identify some of the other trace
compounds present in the diffraction patterns.
Figures 4 and 5 show examples of the XRD and XANES patterns that confirmed that the WAS A
lead pipes are coated with a thin and uniform layer of a- and p-PbC>2.
Discussion
Conventional lead corrosion control theory, which is normally based on divalent lead chemistry,
would predict solubility behavior approximately as represented in Figure 6. This was done using
the LEADSOL computer program (Schock, 1980; Schock et al., 1996) with some representative
concentrations for chloride, sulfate and total inorganic carbon. Figure 7 illustrates the difference
in trends and order of magnitude of solubility for tetravalent lead as opposed to divalent lead.
Trying to model Pb(IV) solubility is full of difficulties, as there is little relevant data appropriate
to potable water systems. Some of these issues have been addressed previously (Schock et al.,
2001). Due to lack of stability constant data and speciation data, the model followed suggestions
of Pourbaix (1966) as a first approximation. Only PbC>32" and PbC>34" complexes were included.
The accuracy of the thermodynamic data and the proper species to choose for the aqueous model
is highly questionable. However, even if considerably off, it still strongly argues that tetravalent
solubility is remarkably lower than in well-treated systems working with divalent lead scales.
Using the existing tentative aqueous solution model for Pb(IV), tetravalent lead solubility is
predicted to be at its lowest level at a pH even below neutrality. Hence, the solubility
minimization trend with pH is intriguingly opposite that of Pb(II) in the normal pH range for
controlling corrosivity to metallic materials, approximately pH 7 to 10.
No specific information on identified Pb(IV) orthophosphate or carbonate-containing solids has
been uncovered so far. Whether or not orthophosphate interacts with tetravalent lead is also hard
to determine reliably, because of conflicting published interpretations of experimental results
from lead-acid battery performance investigations (Voss, 1988). At very high phosphoric acid
concentrations, some evidence has been found for some stable Pb(IV)-orthophosphate complexes
or poorly-crystalline materials (Amlie & Berger, 1972). Some research suggests that
orthophosphate facilitates reversibility of oxidation and reduction by surface sorption and
modification of the PbC>2 phases, while other research suggests that the orthophosphate could
bind with Pb(II) and reduce the formation of PbSC>4 (Bullock & McClelland, 1977; Bullock,
1979a; Bullock, 1979b, 1980; Voss, 1988). The high acidity and extremely high concentrations
of sulfuric and phosphoric acids present in such systems cannot be directly applied to estimate
the relative impacts at concentrations thousands of times lower in drinking water solutions. One

-------
of the most interesting aspects of the literature, however, is that PbC>2 solids may readily function
as semiconductors (Greninger et al., 1975). Thus, there is likely considerable electrochemical
reversibility and ease of electron transport between the water and the underlying lead metal of
the pipe, making responses of the scale to changes in ORP rather fast and measurable.
The critical role that the concentration and type of disinfection plays in the formation and
solubility of passivating films on lead service line piping is clarified by Figure 8. The top figure
is a simple potential-pH diagram for 1 mg/L free chlorine, showing the speciation of the chlorine
system and the high ORP necessary for free chlorine stability. These fields are considerably
above the thermodynamic water stability boundary. The bottom two graphs show the
comparison of ORP values obtained using different concentrations of monochloramine solution
and free chlorine in recent USEPA laboratory studies by James, et. al. (2004). Note the ORP
produced by monochloramine concentrations are far lower than those produced by the same
concentration of free chlorine. Referring to the potential-pH diagram for lead (Figure 9), it can
be seen that free chlorine at high dosages can produce sufficiently high ORP to form Pb02.
These experimental data are in good agreement with values extracted from the research literature
in papers relating to virus inactivation studies or breakpoint chlorination studies (Schock et al.,
1996).
There are some uncertainties in the Pb02 field boundary because of the imprecision and possible
inaccuracy of the tabulated free energy of formation data, although it is qualitatively consistent
with the analyzed scale material and the solubility behavior in the actual lead piping. The
boundary would shift upward (higher ORP needed) if the Pb02 material is less soluble than
predicted by the published data. However, there is clearly a straightforward mechanism that can
readily explain the sudden rise in dissolved lead release (as well as some particulates) when the
ORP is lowered. In the pH range of normal operation (high 7's to low 8's), divalent lead
solubility is considerably higher than the Action Level, and observed lead levels in targeted
samples representing water in direct contact with the lead service lines for "overnight" standing
periods were in an amazingly similar range (100 to 200 fig/L) to that predicted by the solubility
model. Unfortunately, specific lead service line sample data was not available from the late
1990's when the 90th percentile values were very low and the ORP was very high from the use of
free chlorine. However, given the first-draw relationships to the service line concentrations in
the current sampling (post-chloramination), the lead levels were probably very low during that
time period, consistent with tetravalent lead solubility trends. Note that of 5 second-draw sample
collections in 1997-2000, the 90th percentile for the second-draws were equal or lower than the
90th percentiles for the first draws in 3 cases and only slightly higher in the other two cases.
Since the change to chloramination and the lead scale destabilization, the 90th percentile of the
second draw samples has always been at least approximately 50% higher than the 90th percentile
of the first draw samples, and has generally been 2 to 4 times the Action Level.
Two other lines of investigation further corroborate the operation of the Pb(IV) to Pb(II)
conversion mechanism as being the cause of the sudden increase in lead levels. During April of
2004, the normal springtime seasonal switch to 1 month of free chlorine residual (4 mg/L) was
made. After 3 weeks of going back to free chlorine, several lead profiles were conducted in
residential homes containing lead service lines that previously had lead profiles conducted during
5

-------
chloramination. In all cases, lead levels decreased substantially over that time frame (Figures 10
& 11), showing the reversibility of the lead redox reaction.
Support is also provided by data from the exhumed pipe section used in one of the "stagnation"
loop tests described by Thomas et, al, (2004), shown in Figure 12. After equilibrating with the
chloramines "control" condition, the disinfection was changed to hypochlorite solution (dashed
reference line). The lead concentrations immediately began to drop. After about a month of
operation at approximately 5 mg/L free chlorine dosage, the dosage was approximately doubled
(solid reference line). As was demonstrated in the USEPA laboratory experiments (Figure 13),
the ORP did not change significantly by the additional free chlorine addition. Lead levels
continued to nearly linearly decrease. It is hard to tell from the data if the trend was beginning to
reduce in slope after about 3 weeks of this elevated ORP when approximately 10 mg/L (as PO4)
was added to test its effect (dot/dashed line). Therefore, it is not possible at this time to
unambiguously determine if the apparent stabilization is caused by interference with the
oxidation of existing Pb(II) solids in the scale to Pb02, or not. In comparison to the "control"
experiment with chloramine only (not shown), the lead concentration after 2 months of return to
free chlorine was more than a factor of 15 times lower. Interestingly, the total lead concentration
in the control loop is similar (particularly within modeling uncertainty) or slightly higher than
that predicted by the diagram in Figure 6, and the dissolved lead concentrations in
orthophosphate-dosed chloraminated loops were mostly between approximately 45 and 80 |J.g/L,
again consistent with model trend predictions, though slightly higher.
The plausibility of relatively rapid (months) formation of PbC>2 under drinking water conditions
(DIC = 10 mg/L, free chlorine residual, pH 6.5 - 10) from the addition of lead chloride to water
has been proven in bench experiments at USEPA (Lytle & Schock, 2004). The PbC>2 evolved
from a hydrocerussite or cerussite precursor phase, and when the ORP decreased after the
chlorine residual was reduced or lost, the Pb02 decomposed. The phase transformations were
observed to follow the expected trends, i.e. decreasing dissolved lead concentration during Pb02
formation, and increases back to the carbonate phase equilibrium values after oxidant depletion
and reversion. The induction period for Pb02 formation varied with pH, but was generally only
a few weeks. Decomposition after the loss of sufficient ORP was similarly only a few weeks in
duration.
Future Research Needs
More lead pipe specimens will be collected and analyzed from both the service line removal
program, and also from the laboratory test systems, to get an even better understanding of the
scale transformations taking place.
The field data and the chemical models show that high ORP conditions will mitigate the lead
solubility problem caused by the breakdown of the Pb02 passivating scale. However, if the
maintenance of high ORP conditions through free chlorine are not desired for other reasons,
there is relatively little firm information upon which to base the development of an alternative
treatment strategy that would be as effective against lead release.
6

-------
The viability of different alternate treatment schemes are totally dependent upon the exact
reaction pathway of the chemical reduction/breakdown pathway of the Pb02, and the relative
rates of the breakdown reaction(s) versus the rates of formation reactions for divalent lead
compounds. For example, the breakdown of PbC>2 could follow any of several reaction paths to
release lead into solution, such as (but not confirmed to be)
Pb02 (s) + 4H+ Pb2+ + 4H20	(1.1)
or
Pb02 (s)+2H+ + 2 e' <	> PbO(s) + H2Q	(1.2)
PbO(s)+2H+ <	> Pb2+ + H20	(1.3)
Pb(II) oxide and hydroxide are both extremely soluble at any drinking water pH, so other
precipitation reactions would have to be operative to limit lead levels. Once in solution, the
activity of the free lead ion will be governed by the amount of complexation, primarily by
bicarbonate, carbonate, and hydroxide ions (Hunt & Creasey, 1980; Schock, 1981; Schock &
Gardels, 1983; Schock etal., 1996; Schock, 1999). The free lead ion can then react with
carbonate or orthophosphate in the water to precipitate one of the conventional passivating
solids. For example:
3Pb2+ + 2H20 + 2C02- <	> Pb3(C03)2(0H)2(s)	(1.4)
5Pb2+ +3PO4" +H2Q(	>Pb5 (P04)30H(s)	(1.5)
3Pb2+ +2PO4" <=~ Pb3 (P04 )2 (s)	(1.6)
Presumably, these reaction rates would necessarily be very dependent upon pH and the activities
of the passivating and compilexing ligands, such as bicarbonate and orthophosphate.
Unfortunately, kinetic information for these dissolution and precipitation reactions is almost
completely lacking, making a priori estimates of lead levels and time to achieve them very
unreliable. If the existing scale dissolves faster than the released aqueous lead(II) species can re-
precipitate into a sufficiently insoluble passivating film, then prolonged elevated lead levels will
persist, until all of the prior scale is converted. If the precipitation reactions are as fast or
potentially faster than Pb2+ ions are released from the PbC>2 breakdown, and if the scale material
adheres to the pipe surface, then the lead concentrations in the pipes will stabilize in the usual
timeframes typified by the experiences of many other water systems that have successfully used
pH/carbonate adjustment or orthophosphate passivation.
The current hardness of the water and the use of lime for pH adjustment essentially precludes the
pH/alkalinity/TIC adjustment approach, because the necessary pH (over 9 based on theory and
the experiences of other water systems) could not be achieved without major scaling problems in
the filters and distribution system. Supplemental softening or some kind of carefully balanced
threshold sequestration would be needed. The remaining alternative approach, using
orthophosphate dosing, needs to be investigated to determine empirically what the relative rates

-------
of divalent lead passivation reactions are in comparison to the breakdown of the existing pipe
scale. As noted above, if the reaction of the Pb(II) released from the breakdown of the PbC>2
scale with orthophosphate in the water (equations 1.5 and 1.6) is equal to or faster than the rate
of dissolution and release into the water, then it should be possible to achieve sufficiently low
lead levels in the water relatively quickly. If it is not, there is not a good basis to estimate the
length of time it would take to achieve complete conversion of PbC>2 to the passivating divalent
lead phosphate solids.
Very basic questions pertaining to Pb(IV) chemistry are critical to answer to provide important
information necessary for revisions to lead corrosion control guidance, and to properly evaluate
disinfection alternatives. These fundamental questions include:
•	What are the solubility constants for the a-PbC>2 and P-PbCh polymorphs?
•	What factors govern the formation of one polymorph as opposed to the other?
•	What are the important aqueous complexes of Pb4+ (e.g. PO4, SO4, CI, HCO3, CO3, OH")?
•	What are the stability constants of those complexes?
•	What reaction pathways are taken for formation and breakdown of PbC>2 phases?
•	Are there other important Pb(IV) solid phases for drinking water conditions?
Little is known about the passivation mechanism for exposed soldered joints and brass devices.	,
Whether or not PbC>2 solids can form and be stable on these kinds of surfaces is another
important question, because of its relevance to the origin of lead concentrations caught in 1-liter
first-draw samples.
Several studies to shed light on some of the aspects of tetravalent lead chemistry are currently
underway at WASA and at USEPA. Corrosion control is intimately interrelated with other
finished water quality objectives, for consumer satisfaction and regulatory compliance.
Therefore, these recent discoveries of the importance of Pb(IV) chemistry in some water systems
support the idea that more resources need to be mobilized quickly to gather the information
needed to provide timely guidance for water systems confronted with needing to evaluate
complex and costly major treatment upgrades to meet new regulatory requirements.
Disclaimer
Any opinions expressed in this paper are those of the author(s) and do not, necessarily, reflect the
official position and policies of the EPA. Any mention of products or trade names does not
constitute recommendation for use by the EPA.
Acknowledgements
The authors wish to thank Darren Lytle of U. S. Environmental Protection Agency, and Rachel
Copeland and Cheryl James of the University of Cincinnati for sharing information and
experimental data on redox potentials generated by different oxidants. Kirk Scheckel of U. S.
Environmental Protection Agency assisted with the XANES analyses of the pipe scales. Use of
V.
8

-------
the Advanced Photon Source was supported by the U. S. Department of Energy, Office of
Science, Office of Basic Energy Sciences, under Contract No. W-31-109-Eng-38.
In addition, the authors would like to thank Will Keefer, John Cirvardi and Chris Thomas of
Baker Killam Joint Venture for their assistance and Laura Dufresne from Cadmus Group for
helpful discussions.
Bibliography
Amlie, R.F. & Berger, T.A., 1972. Polarographic Analysis of Lead(IV) Species in Solutions
Containing Sulfuric and Phosphoric Acids. Journal of Electroanalytical Chemistry,
36:427.
AWWARF-TZW, 1996 (Second ed.). Internal Corrosion of Water Distribution Systems.
AWWA Research Foundation/DVGW-TZW, Denver, CO.
Bullock, K.R., 1979a. Effect of Phophoric Acid on the Positive Electrode in the Lead-Acid
Battery II. Constant Potential Corrosion. Studies. Journ. Electrochem. Soc., 126:3:360.
Bullock, K.R., 1979b. Effect of Phosphoric Acid on the Positive Electrode in the Lead-Acid
Battery III. Mechanism. Journal of the Electrochemical Society, 126:11:1848.
Bullock, K.R., 1980. Effect of Anion Activity on Electrochemical Equilibria: Three-Dimensional
Potential-pH Diagram for Pb/H2S04/H20 System. Journal of the Electrochemical
Society, 127:3:662.
Bullock, K.R. & McClelland, D.H., 1977. The Effect of Phosphoric Acid on the Positive
Electrode in the Lead Acid Battery. Journ. Electrochem. Soc., 124:10:1478.
Delahay, P., Pourbaix, M. & Van Rysselberghe, P., 1951. Potential-pH Diagram of Lead and its
Application to the Study of Lead Corrosion and to the Lead Storage Battery. Journal of
the Electrochemical Society, 98:57.
Dodrill, D.M., 1995. Lead and Copper Corrosion Control Based on Utility Experience. Master
of Science in Environmental Engineering thesis, University of Colorado at Boulder.
Dodrill, D.M. & Edwards, M., 1994. A General Framework for Corrosion Control, Proc. AWWA
Water Quality Technology Conference, San Francisco, CA.
Dodrill, D.M. & Edwards, M., 1995. Corrosion Control on the Basis of Utility Experience.
Journal of the American Water Works Association, 87:7:74.
Edwards, M., Jacobs, S. & Dodrill, D., 1999. Desktop Guidance in Mitigation of Pb and Cu
Corrosion By-Products. Journal of the American Water Works Association, 91:5:66.
9

-------
Edwards, M., Giani, R., Wujek, J., Chung, C., 2004, Use of Lead Profiles to Determine Source
of Action Level Exceedances from Residential Homes in Washington, D.C., AWWA
Water Quality Technology Conference, San Antonio, TX
Francis, R., 1985a. Effect of Pollutants on Corrosion of Copper Alloys in Sea Water, Part 1:
Ammonia and Chlorine. British Corrosion Journal, 20:4:167.
Francis, R., 1985b. Effect of Pollutants on Corrosion of Copper Alloys in Sea Water, Part 2:
Sulphide and Chlorine. British Corrosion Journal, 20:4:175.
Greninger, D., Kollonitsch, V. & Kline, C.H., 1975. Lead Chemicals. International Lead Zinc
Research Organization, Inc., New York.
Hunt, D.T.E. & Creasey, J.D., 1980. Calculation of Equilibrium Trace Metal Speciation and
Solubility in Aqueous Systems by a Computer Method, With Particular Reference to
Lead. Technical Report TR-151, Water Research Centre, Medmenham, Bucks.
James, C.N., Copeland, R.C. & Lytle, D.A., 2004. Relationships between Oxidation-Reduction
Potential, Oxidant, and pH in Drinking Water, Proc. AWWA Water Quality Technology
Conference, San Antonio, TX.
James M. Montgomery Consulting Engineers, 1982. Internal Corrosion Mitigation Study Final
Report, Bureau of Water Works, Portland, OR.
James M. Montgomery Consulting Engineers, 1985. Water Treatment Principles & Design. John
Wiley and Sons, Inc., New York.
Larson, T.E., King, R.M.. & Henley, L., 1956. Corrosion of Brass by Chloramine. Journal of the
American Water Works Association, 48:1:84.
Lytle, D.A. & Schock, M.R., 2004. Formation of Pb(IV) Oxides in Drinking Water, manuscript
in preparation.
Pourbaix, M., 1973. Lectures on Electrochemical Corrosion. Plenum Press, New York.
Pourbaix, M., De Zoubov, N., Vanleuenhaghe, C., and Van Rysselberghe, P., 1966. Section 17.5.
Lead. Atlas of Electrochemical Equilibria in Aqueous Solutions Vol. Chapter IV. Section
17.5, pp. 485-492. National Association of Corrosion Engineers, Houston, TX.
Ressler, T., 1998. WinXAS: A program for X-ray absorption spectroscopy data analysis under
MS-Windows. Journal of Synchrotron Radiation, 5:118.
Schock, M.R., 1980. Response of Lead Solubility to Dissolved Carbonate in Drinking Water.
Journal of the American Water Works Association, 72:12:695.
10

-------
Schock, M.R., 1981. Response of Lead Solubility to Dissolved Carbonate in Drinking Water.
Journal of the American Water Works Association, 73:3:36.
Schock, M.R., 1999 (Fifth ed.). Internal Corrosion and Deposition Control. Ch. 17 In: Water
Quality and Treatment: A Handbook of Community Water Supplies, pp. 17.01-17.109.
McGraw-Hill, Inc., New York.
Schock, M.R., 2000. Lead Corrosion: What Research is Needed?, Proc. International
Distribution Research Symposium, Denver, CO.
Schock, M.R. & Gardels, M.C., 1983. Plumbosolvency Reduction by High pH and Low
Carbonate—Solubility Relationships. Journal of the American Water Works Association,
75:2:87.
Schock, M.R., Harmon, S.M., Swertfeger, J., and Lohmann, R., 2001. Tetravalent Lead: A
Hitherto Unrecognized Control of Tap Water Lead Contamination, Proc. AWWA Water
Quality Technology Conference, Nashville, TN.
Schock, M.R. & Holm, T.R., 2003. Are We Monitoring in the Right Places for Inorganics and
Radionuclides? Journal of the New England Water Works Association, 117:2:102.
Schock, M.R. & Wagner, I., 1985. The Corrosion and Solubility of Lead in Drinking Water.
Internal Corrosion of Water Distribution Systems, pp. 213-316. AWWA Research
Foundation/DVGW Forschungsstelle, Denver, CO.
Schock, M.R., Wagner, I. & Oliphant, R., 1996 (Second ed.). The Corrosion and Solubility of
Lead in Drinking Water. Ch. 4 In: Internal Corrosion of Water Distribution Systems, pp.
131-230. AWWA Research Foundation/TZW, Denver, CO.
• Thomas, C., Kim, J., Korshin, G., Civardi, J. & Giani, R., 2004, Evaluation of Lead Leaching
Rates During Stagnation Using Real-Time Corrosion Potential Monitoring and Modeling
Methods, AWWA Water Quality Technology Conference, San Antonio, TX
Voss, E., 1988. Effects of phosphoric acid additions on the behaviour of the lead—acid cell: A
review. Journal of Power Sources, 24:3:171.
11

-------

0.09

0.08

0.07

0.06
_l

0.05
Q.

Ui
0.04
E
0.03


0.02

0.01

0.00
pH Adju&tmenl
Vnryinc; pH
Elevated tree eh orine (2.5 ¦ 4 rngiL) Change to 3 - 4 mg/L chloramloe
RaieesORP: PbO. formation	ORP drops: PbOj breakdown
) Firsl Draw
rs/A Second Draw
Af.lrO!! I
^ ^ ^ ^ ^ ^	ri.	v5-, vi.
'£> \ 'tp % % % % % O % °o °o b
Lr 1? 'J* O* —4 CP 


-------
Cincinnati, OH
Lead Service Line
J 225 Gilder AVE., it 914?-1 2/2002J
1 19 CATALPA RD. (1925? - 1 /24/03J
3030 JUNtETTA AVE 11930? 1 /2d/03t
Oakwood, Ohio
Lead Service Lines (rem. 2002>
Figure 3. Comparison of similar PbO;-r.ch scales in two other water
systems, showing the similarity to the DC-WASA pipe scales.
13

-------
c
fij
QQ-C">-Q5J9> Soutinyitc syn Pb02
DCMONW L1
DCUKKl I 1	00-013-0131 > Hydrwefussife syn - Pb3(C03)?(Ot I)?
nr.l IENW7 L1	00-046-1045> Quartz. syn - SIO?
DCHtFWWI I t	00 02S 044?1, PSaSnetfte - PteQ2
1500
1000
600
0
tr>op
1.0DD
500
0
Two-THeta (tfeg)
I Mi 'l V. . I .
Civr/tfrM'l I i
1 • .• t r _¦>> 1HVI'
i-THETAi5:
Figure 4. X-ray diffraction pat (Cms of lead service line specimen? from the WAS A system.
(Top) Predominance of plattnerite and scrutinyite phases in four samples and similarity ol'
mineralogical compositions in the different locations. Hvdrocernssite and quartz are minor
phases on some specimens. (Bottom) Reference I ines for cerussite and hydrocerossite for two
scale layers (LI nearest water), showing potential peak overlap for PbCO? with Ph()?. and the
minor amount of PbstCOj.fcf'Oin; present. Large peak near 271 is quartz (SiCb).
14
i

-------
Energy (keV)
Pb{IV)
characteristic ,
t
DCM0NW1LI
0CVIQNVV1L2
PhO
Pb,Ot
Pb^CO^^OH),
Pb,(P04)ACI
Figure 5. XANKS spectra example showing uniformity of scale sample from 1083 Monroe St.
NW (two layers scraped) and similarity so reference specUum for fi-PbO...
13

-------
0.250
1 mg PO/L
3 mg PO«/L
0 mg PO^L
0.200
PbC03 + Pb3(C03)2(0H)3
d 0.150
JQ
U)
E 0.100
0.050
Action Level
0.000
6
pH
Figure 6. Solubility diagram for the impact of orthophosphate on Pb(II) species for DIC
mg (VI.. CI" - ">1 mg/l., SO,"" 44 mg/L, 0.005 M ionic strength, and 25°C.
16

-------
1e+Q
^°3 * Pb.fCPj^QH), J
1e-i
PO
1e-2
PO
1e-3
^ 1 e-4
_Q
ft. 1e-5
U)
E 1e-6
1e-7
1e-8
1e-9
//
1e-10
10
11
9
6
7
a
pH
Hgurc 7. Computed solubility diagram for PbO;, using several tabulated Gibbs lice energy oJ"
formation values to compute the solubility constants. Shown also is the relationship to divalent
lead solubility ibr the same water.
17

-------
V)
+*>
o
>
LL1
0
1 T
r g * t »
> i i i i
i
0
Ct ?
HOCI0
OC!

h—-
o*ld.
-

CI



	
-
r 1
-•	1	'	1	*—
—'	1	»	-L-—-*¦

0
6 8
pH
10
12
14
—fir- -fr-	' V
- • r ?--r 1

fcr-2_ —• -
pH 7 P'r^t-n/if "
pH " FJrrr-iVtr T
j-H * FtoHMi i
JlH 1 F.'fJ^fMlt I
pH C\ F.Y"
pf? •? FJx-iiWWtf :
in i * ic :
Ml.A Dosage i:#;wK Irnjf.'l I
1.5
Figure 8. (Trip) Potential-pH diagram for 1 rag L free chlorine, showing the speciation of the
chlorine system and the high ORP necessary tor tree chlorine stability, and the relationship to the
water stability boundary (Bottom) Comparison, of ORP values obtained using different
concentrations of monochloramine solution (left) and free chlorine (right) in USEPA laboratory
studies by James, et. al. (2004) Note ihe consistently much lower ORPs produced by
chiora nxincs.
18

-------
GT
x
w
§
£
o
>
LU
1.50
.00
0.50
0.00
-0.50
1.00
-t .50
¦MF-pH Diagram for Pb - H2D - CO2System
Pb species « 0.015 mg/L; D)C 1S mg C/L
l=0; 25"C
PbOo i's)

0 1
{po;-* 0.37
atmj
Pb
2+

J Vs 1 efrry
I f
X
¦ ' •	4_
\
pH0j^ 3; I'bOg
/ w /
"" tX,
Pb{OH)

:
V l \
'l \
I I 1 1 '
	¦¦¦-.'
4
6 7 8 9
pH
10 11 12 13 14
Figure; 9, Potent ial-pH diagram for the lead system corresponding to DC-W ASA water, showing
how the sequence of treatment changes over the past decade formed and then destabilized the PbO;
passivatmg film. (1) represents the initial conditions of the early 1990's where there were some
lead release problems. The initiation of high free chlorine residuals and flushing in 1994 moved the
system chemistry to approximately point ( 2). The change to chloramines secondary disinfection in
2001 moved the ORP back into approximately the area of (3).
19

-------
Leitl Profile -3-31-CM Cfttorammei
1.5)0 Service
In-Hnujw?
Plumbing
¦n
,11
Main
di Min i
B Leac Filtered 1
Iirn id a 1111 in Hi
n i
;« a -a .7 fl II 13 16 IB ?l 74 Tt
y/+11!
1 iter i
Lead Profile 5-7-04 Chlorine
Id'HuUtrfi
Plumbing I
O Lead pph
O Lead Filtered pptri
Figure 10. Lead profile conducted at the same home during chlorarnine (top) and chlorine. The
x-axis represents the liter number taken from the tap into the main (i.e. 1 - first draw, 2 - 2m:
liter. 27 - 27"' liter. 27+3 represents 3 minutes after the 2"!:: liter was taken, X - water hammer).
Lead profile procedures taken from Edwards, Giani et al 2004)
20

-------
Peak dissolved lead levels in homes during lead profiles
i iPo.ik lead lewis during
r,h5nramlr(attnn
*"'l 'eak lewd levels during chlOi'inialton
Kesidenl 1 Resident 2 Resident 3 Resident <5 Residents RosWnni G
residents
f igure 11 Peak dissolved lead concentrations taken from 6 residential homes duriui-
ehlorarmnation and chlormation.
Reference
electrode
Multichannel
Voltmeter
g Data
Acquisition
Drain
Figure 12. Stagnation tonp setup diagram.

-------
Pipeloop S Initial


t1 o
a Tot CI
• Free CI
s
- K ay «j il d >n iii tr> o> a>
OOOOOOOO o o o
Date
i too
1000
900
800
Plpeicrop 6 Final
>ORP
ui 700
w 600
£ 500
& 400
° 300
200
100
.

"
>
t •
m
Im*
• • * •


.. !•
•

# •* i


• • |
*
i
¦r J
1 » « 1
r—r ^ f—?—* v r*
of 5 § t 5  CM ®
O «- f¦4 N
g 3
0*te
•	pii
•	Pb Chlorine
¦ Filtered Lead Values
Plp-eioop 6 Flncjl
§Sp?S§£2?8gi£
8 S 3 3 3 3 8 8 £! 8 8 8 8
Date
s § s e a §
8 3 S 3 8 8
Figure 13-, Laboratory "stagnation loop"' data for actual removed pieces of service line, testing tbc effects ot
free chlorine and subsequent ortbopbospbatc (10-15 mg/L as P0<) addition on redox potential and lead
release. Tlie magenta reference line indicates the end of the conditioning period of exposure to chlorauiiiistetl
water, before shifting to free chlorine (top graph). The green line represents a boost in free chlorine
cnncent ration, and the blue dotted line represents the introduction of the orthophospbatc dosing.
::

-------
Appendix C Galvanic Corrosion and Grounding Effects Study

-------
This page intentionally left blank.

-------
Final Report
Effects of External Currents and Dissimilar Metal Contact
on Corrosion from Lead Service Lines
Prepared for:
George Rizzo, Work Assignment Manager
U.S. Environmental Protection Agency Region III
1650 Arch Street
Philadelphia, PA 19103-2029
Contract Number 68-C-02-069
Work Assignment Number 47
Prepared by
Dr. Steve Reiber
Formerly of HDR Engineering
and
Laura Dufresne
The Cadmus Group, Inc.
Finalized November 2006

-------

-------
Table of Contents
Executive Summary	1
Background	3
Research Protocol	5
Electrochemical Measurements	5
Galvanic Coupling Experiments	6
Indirectly Coupled Cells	6
Evan's Diagram	8
Cathode/Anode Ratio Effects	9
Directly Coupled Pipe Sections	10
Water Quality and Galvanic Impacts	12
Dielectric Insertion	15
Grounding and Impressed Currents	16
Impressed Current Experiments	16
Observations and Conclusions	20
Passivation	20
Lead Electrochemistry	20
Area of Galvanic Influence	20
Cathodic Effect of Copper Pipe	21
Water Quality and Galvanic Impacts	21
Dielectric Effects	21
Impressed Current Effects	21
References	22
Final Report	i	November, 2006

-------
Table of Figures
Figure 1.	Polarization (Evan's) diagram of coupled lead and copper surfaces	4
Figure 2.	Schematic of a typical pipe rig configuration using indirectly-coupled cells... 7
Figure 3.	Photo of a pipe rig with indirectly-coupled cells in operation	8
Figure 4.	Effect of cathode/anode ratio on LSL surface potential	10
Figure 5. Schematic of a test rig showing the direct coupling of LSL and copper pipe
sections	11
Figure 6. Photo of directly coupled LSL and copper sections used in a portion of
this study	11
Figure 7. Surface potential along the length of coupled LSL and copper service linesl2
Figure 8. Effect of chlorine concentration on galvanic impacts relative to passivated
LSL specimen	14
Figure 9. Comparative effect of free versus combined chlorine on galvanic impacts
relative to passivated LSL specimen	14
Figure 10. Effect of conductivity on galvanic impacts relative to passivated LSL
specimen	15
Figure 11. Effect of inserting a dielectric between the passivated LSL and copper
sections on galvanic impact	15
Figure 12. Schematic illustration of impressed current test rig	17
Figure 13. Impressed current impacts (separate DC and AC tests) on surface
potential of an LSL coupled to copper tubing	19
Final Report
ii
November, 2006

-------
Executive Summary
This study set out to investigate two basic issues and has largely succeeded at resolving
both. The first issue was whether grounding or impressed currents have a significant
and prolonged impact on the electrochemistry and corrosion of lead service lines (LSLs)
in a water distribution system, or, for that matter, on any metal plumbing appurtenance
that may be hydraulically and electrically connected to a household service line used as
an electrical ground. The second principal issue was to characterize the electrical
impacts associated with galvanically-coupled copper and lead service lines to determine
if replacing a portion of a lead pipe with copper piping might cause accelerated lead
release. Both issues relate to the potential for accelerated corrosion on LSLs leading to
the release of metals in drinking water. Both have relevance to the DC WASA corrosion
control program and LSL replacement program.
The specific objectives of the research were to establish under controlled laboratory
conditions the absolute magnitude of the electrical impacts on LSLs associated with both
grounding and galvanic coupling under a variety of pipe geometries and water
chemistries. This research did not intend to explore grounding currents in an existing
home or to replicate actual distribution system conditions where a partial lead service
line replacement (PLSLR) had occurred. Rather, the goal was to demonstrate whether
or not grounding currents or galvanic coupling could generate lead release. If a
meaningful impact could not be demonstrated under conditions designed to exacerbate
lead release, then it would be unlikely that a PLSLR as practiced in the DC WASA
system (where conditions would be much less challenging than in the laboratory) would
produce accelerated metal release. If positive effects were found, the study would then
serve as a foundation for further testing.
All the LSL pipe sections used in these tests had been removed recently from
residences in the DC WASA distribution system.
In brief, this study has shown that grounding and/or impressed currents moving along
LSLs, end eventually leaving the pipe to ground, have no meaningful impact on internal
pipeline corrosion and do not contribute to metals release. Therefore, we believe the
long-debated controversy about whether or not grounding currents generate accelerated
corrosion and metal release can now be considered closed. Also, while the study found
that galvanic impacts can be substantial on unpassivated, newly-exposed lead surfaces,
the magnitude of galvanic impacts on aged and passivated LSL surfaces and on new
copper surfaces is minimal, and, in the long term, likely to be inconsequential. Therefore
there is now a basis for discounting concerns relative to the long-term impacts
associated with PLSLR.
A caveat that must be attached to these finding is that testing was restricted to waters
with low mineral content similar to the water distributed within the DC WASA system;
galvanic impacts in systems having water with a substantially higher mineral content
could be more extensive and possibly more prolonged.
Final Report
1
November, 2006

-------
The significant conclusions to be drawn from this study are as follows:
•	Well-aged DC WASA LSL specimens - including those that have been exposed
to an orthophosphate inhibitor-are exceptionally well passivated and highly
resistant to electrical perturbations of any kind.
•	When a well-passivated LSL is coupled to a new length of copper tubing (as in a
partial LSL replacement) the area of galvanic influence is very limited. The actual
reach of the galvanic current is partially a function of the water quality, but is
likely limited to the first inch of the LSL.
•	A conventional plumbing dielectric junction removes even the minor corrosion
risks associated with galvanic coupling. Any break in electrical continuity
between the copper and LSL lines effectively eliminates the potential for
significant galvanic effect.
•	A chlorine residual (free or combined) does elevate the galvanic effect on the
LSL/copper couple by accelerating the cathodic current exchange process. The
impact overall, however, is largely limited to the galvanic influence on the copper
service line. The overall impact on the LSL surface is nearly imperceptible.
Interestingly, water conductivity has a more important effect on the galvanic
process than chlorine residual.
•	Impressed currents, whether AC or DC, on LSLs or copper service lines
(including grounding type currents), have no impact on the internal corrosion of
the household service lines (or any other plumbing appurtenance for that matter).
There is no acceleration of corrosion associated with the conventional practice of
electrical system grounding to household water systems.
Final Report
2
November, 2006

-------
Background
In theory, it is conceivable that replacing a portion of a lead line with a new copper
service line could create a strong galvanic couple with an initial Cu/Pb electromotive
difference in the 400 - 500 mV range (Reiber, 1991). If a significant portion of the
remaining section of lead service line were shifted in the anodic direction by even a
fraction of this amount, there would be a substantial acceleration of the corrosion rate
and associated metal release rates.
In a similar sense, for well over a decade, there has been substantial conjecture within
the drinking water industry that electrical currents impressed, or, more often, shunted
onto water service lines as a result of grounding practices in individual homes, create a
similar scenario. Supposedly these impressed currents shift the surface potential of the
corroding pipeline surface, generating accelerated corrosion and metal release, and in
some cases producing other corrosion-related problems such as localized pitting. There
are few texts on distribution system corrosion that do not cite impressed currents as a
potential cause of the interior pipeline corrosion (Bell, 1996; AWWARF, 1996). The
suggestion has been made that these currents may be responsible for some of the
abnormally high lead release levels observed in isolated homes. By extension, it could
be assumed that if grounding currents are important, then perhaps a portion of the
randomness associated with observed lead levels may be related to the presence of
different magnitudes of grounding currents.
It is important to note that while the proposed mechanisms of galvanic and/or impressed
current influence are plausible, and that some limited evidence is supportive, it has not
been demonstrated that either grounding currents or galvanic coupling meaningfully
increase LSL corrosion rates. Moreover, partial LSL replacement in the DCWASA
system has not resulted in observed increases in lead release (Wujek, 2004). In fact, the
recent DC WASA experience relative to LSL replacement suggests that in the long term
PLSLR does not exacerbate lead release rates, but rather reduces overall household
drinking water lead concentrations in proportion to the amount of LSL replaced. While
this evidence appears strong, for several years there has been a debate on the potential
galvanic effects associated with replacing a portion of old LSLs with new copper tube, or
for that matter coupling any lead-containing alloy to a dissimilar metal.
Final Report
3
November, 2006

-------
Freely Corroding
Cu Specimen
100
/'Cu
-100 --
x Theoretical Effect of
\ Calvanically Coupled
I Pb and Cu Cells
.£ Couf^led
> -200 --
-300 --
(Coupled
-400 -¦
Freely Corroding
Pb(LSL) Specimen
-500 --
Ai
-600
10s	10"
/ (A/cm2) - Corrosion Current Density
Figure 1. Polarization (Evan's) diagram of coupled lead and copper
surfaces.
Theoretically, the coupling of a new copper surface to a lead surface should produce a
substantial galvanic impact. The Evan's Diagram above illustrates this point. The
diagram presents the observed polarization behavior of uncoupled copper and lead
surfaces, as well as the theoretical polarization behavior of the surfaces that would occur
if they were coupled. The uncoupled surfaces are unsealed and unpassivated, and
hence represent practical worst-case scenarios. Relative to lead, the coupling of the two
surfaces results in a theoretical initial increase in the anodic (lead) exchange current (Ai,
corrosion current density) of approximately two orders of magnitude, which of course
would have a profound impact on corrosion and metal release rates if it were sustained
at this level.
What cannot be discerned from the above representation is how long that accelerated
current exchange can be sustained after the initial coupling. In a practical sense, since
all of the existing LSLs in the DC WASA system are well passivated after many decades
of service, the more important question becomes how passivation of the lead and/or
copper surfaces affects the galvanic current. A large portion of this modest study
focuses on that question.
Final Report
4
November, 2006

-------
Research Protocol
Electrochemical Measurements
At the core of this study was the search for the substantial electrochemical impacts that,
theoretically, should be associated with the galvanic and impressed currents imposed on
the LSLs. The principal measure of these impacts would be a significant shift in the
electrochemical potential of the interior surface of the LSLs away from the freely
corroding surface potential. Surface potential can be directly and accurately measured
using straightforward electrochemical tools (AWWARF, 1996).
This research did not attempt to create laboratory conditions that exactly replicate field
conditions. Instead, the goal was to demonstrate whether or not extremes of grounding
currents or galvanic coupling could affect the LSL electrochemistry. It was also beyond
the scope of this work to define how differences in passivation states of copper tubing, or
quality of plumbing fabrication may influence the respective current impact. Testing was
generally short-term, inexpensive and designed to answer the simple question, "Can
grounding and/or galvanic currents under a worst-case scenario meaningfully contribute
to lead corrosion and metals release?"
The study used a series of electrochemical cells which allowed the mounting of sections
of LSLs under flow conditions and the placement of electrodes capable of quantifying
shifts in surface potential. In most cases the electrode of choice was a calomel
electrode (Hg/Hg2CI2), selected because of its stability and resistance to external
electrical noise. This electrode, coupled with a sensitive potentiostat, can measure
surface potential shifts of a millivolt or less. This is an important analytical factor, since
the shifts in surface potential theoretically resulting from the galvanic and impressed
currents were thought to be hundreds of millivolts or greater.
The surface potential measurement is sensitive, easy to use, and allows speedy
measurements, but its principal advantage is that it is influenced only by the
electrochemistry of the metal surface and the water in contact with that surface. It is not
substantially or directly influenced by the mechanical stability or chemical solubility of the
corrosion scale covering the surface - unless the rapid loss of that scale is changing the
underlying electrochemistry. The surface potential measurement reflects the corrosion
conditions of the underlying metal, which, in this case, is the factor most directly
influenced by application of the galvanic and/or impressed currents in question. At the
same time, the limitation of this electrochemical testing is that it tells us very little about
the stability or changing mineralogic makeup of the corrosion scale. Nonetheless,
relative to the issue of galvanic and impressed currents, electrical perturbation will
precede any long-term change in the nature of the corrosion scales.
The short-term tests used in this study did not readily lend themselves to standard metal
release monitoring used in other aspects of the DC WASA corrosion-control optimization
studies (Giani et al, 2005). While monitoring of metal-release rates was attempted, it
was found that the physical cutting and manipulation of LSL sections generated
frequent, but irregular, particulate release. The individual tests, which generally ran for
periods of only a few days each, did not provide sufficient time to condition and stabilize
the scales on the different test sections.
Final Report
5
November, 2006

-------
Galvanic Coupling Experiments
The original approach to the galvanic coupling research was to utilize polarization cells
mounting individual sections of LSLs and copper tubing. These cells could be
connected in a hydraulic series, with the electrical connections between the individual
cells manipulated at will. Because the pipe specimens of each cell were not in direct
contact, these cells were referred to as indirectly coupled. The importance of the
indirectly coupled cells relates primarily to the ability to control cathode/anode ratios. It
is critical to the appreciation of the galvanic couple concern to understand that it is not
the contact of dissimilar metals, per se, that creates the corrosion risk, but rather the fact
that the cathodic surface (the more electropositive metal), if present in abundance, can
affect a shift in the surface potential of the anodic surface. The greater the
cathode/anode ratio the greater the potential shift, and the greater the area of anodic
impact. Any meaningful shift in the anodic surface in a more positive direction generates
a higher corrosion rate on that surface.
A second approach to galvanic testing utilized longer segments of LSLs and copper pipe
coupled together in a manner similar to an actual PLSLR. Because these pipe
specimens are in direct contact, this type of testing is referred to as directly coupled
pipe specimens. This form of testing yielded more useful results about the nature of the
galvanic couple formed between copper and LSL sections.
Indirectly Coupled Cells.
Recirculation pipe loops with individual cells holding LSL and copper pipe sections were
fabricated with acrylic polarization cells and vinyl connecting tubing. The schematic
presented in Figure 2 shows the arrangement of the individual cells, hydraulics and
electrical connections. In this arrangement the cells are hydraulically connected in
series, and electrically connected via external circuits that can be configured as needed.
The arrangement offers the opportunity to manipulate cathode/anode ratios, measure
the current flow between cells, evaluate metal release and, most importantly, accurately
assess the surface potential of the individual pipe sections, all while controlling flow and
water quality conditions. Figure 3 presents a photo of such a loop in operation.
Final Report
6
November, 2006

-------
Potentiostat

Cathodic
Coupling
(parallel)
20L
Reservoir
Soldered Pigtail
Reference Electrode
Figure 2. Schematic of a typical pipe rig configuration using indirectly-
coupled cells
Final Report
7
November, 2006

-------
Figure 3. Photo of a pipe rig with indirectly-coupled cells in operation.
Evan's Diagram.
The data presented in the Evan's diagram of Figure 1 was prepared using results
obtained from an indirectly-coupled pipe rig of a type similar to that in Figure 2. In that
test a single LSL section was coupled to a single copper pipe section of comparable
internal surface area (20 cm2). Originally, both the LSL and copper surfaces were
abraded and then polished (300 grit wet jeweler's paste) so as to represent truly
unsealed and unpassivated surfaces. This attempt to create an unpassivated LSL
surface generated a surface potential of approximately -400 mV (vs. SCE), which is not
substantially different than the surface potential of a passivated LSL. Discussions with
Michael Schock (US EPA - ORD) suggested that a truly unpassivated lead surface
should be closer to -500 mV (vs. SCE), rather than the -400 mV measured.
Final Report
8
November, 2006

-------
Revisiting the techniques used to remove the LSL corrosion scale, a milling machine
with high speed de-burring tool was used to ream out the interior of the short LSL
sections, exposing bare metal with the assurance that no corrosion scale or passivation
layer remained. Surface potential measurements on these specimens were
approximately -550mV (vs. SCE) when first exposed to water flow. Owing to the
electroactive nature of the pure lead surface, the absolute magnitude of the
measurement quickly began to decrease as the surface began to passivate in the
presence of water.
The rapidly changing electrochemical nature of a bare lead surface makes it difficult to
accurately assess the surface potential of a truly unpassivated specimen. The data
presented in the Evan's diagram of Figure 1 is probably best described as a partially
passivated LSL surface.
Other experiments conducted to define the character of an unpassivated LSL surface
were, at best, only partially successful. The most substantial observation is the speed
with which an unsealed lead surface will begin to passivate. Although we have not tried
to quantify the rate of passivation, we note that an unpassivated LSL section within a
matter of weeks will take on the character of a passivated LSL section having decades
of exposure, and it will do this in a low mineral content water very similar in chemistry to
that distributed by DC WASA. Moreover, it is clear that while orthophosphate corrosion
inhibitors can over a substantial period of time enhance the passivation of LSL surfaces
(based on on-going metal release pilot studies), these surfaces will effectively passivate
absent orthophosphate or other specific corrosion inhibitors (Reiber and Giani, 2005). It
is the nature of the lead surface to quickly form an effective passivation layer, which is
why LSLs are still found in service even after a century of exposure to drinking water
flows.
Cathode/Anode Ratio Effects.
Using the indirectly coupled cell rig, the effect of coupling passivated LSL sections to
unpassivated copper sections quickly became apparent. Figure 4 presents the results of
coupling a single LSL section (20 cm2 internal surface) to multiple copper pipe sections
of equivalent surface area. At most, the electrical perturbation of the passivated LSL was
a few millivolts, regardless of how many copper sections were electrically coupled to it.
A substantially different observation was made when a relatively unpassivated LSL
section (mechanically reamed) was coupled to the same copper pipe sections as above.
The unpassivated LSL surface was substantially influenced by the galvanic coupling -
producing an anodic shift of approximately 100 mV at a cathode/anode ratio of three.
Clearly, galvanic coupling is important on relatively unpassivated surfaces, yet largely
irrelevant to passivated LSL specimens.
While this experiment evaluated the macro effects of galvanic coupling, it is not
unreasonable to extrapolate these macro observations to the micro surface chemistry
associated with leaded brasses. The question has been raised (Korshin, 2005) as to
whether brasses having any lead inclusions will accelerate corrosion of the lead by virtue
of the intrinsic galvanic couple created by a surface with very small lead anodes
surrounded by the much more abundant and more electropositive copper/zinc alloy
(brass). The macro galvanic coupling experiment described above would suggest that
cathode/anode ratio is largely irrelevant to the corrosion of a passivated lead surface.
Final Report
9
November, 2006

-------
Hence, it seems likely that lead corrosion on an aged brass surface is unlikely to be
influenced by the more electropositive alloy surrounding it.
u
un -1004-
-400"
Passivated LSL Section
M Unpassivated LSL Section
Uncoupled
112 13
Cathode/Anode Ratio
Figure 4. Effect of cathode/anode ratio on LSL surface potential.
Directly Coupled Pipe Sections.
Following the test with indirectly coupled pipe sections, the study attempted to quantify
galvanic effects by directly coupling copper tubing sections and DC WASA LSL sections.
New %-inch diameter copper tubing was selected to ensure as high a galvanic driving
force as possible. To ensure electrical coupling, the end of the LSL was grooved to
accept the end of the copper tubing, which mated directly with the LSL. A hydraulic seal
was achieved by mounting both the copper and LSL sections between compressive
headpieces. Holes placed at strategic locations along the copper and LSL sections
allowed for the placement of reference electrodes capable of reading the surface
potential on the pipe opposite to those locations. To simulate operational service, water
was circulated through the pipe sections while the surface potentials were recorded.
Figure 5 presents a schematic of a directly coupled cell. Figure 6 presents a picture of a
small coupled pipe cell used for one portion of this testing.
Final Report
10
November, 2006

-------
The LSL pipe sections used in these tests were recently removed (July, 2005) from
residences in the DC WASA distribution system. As such, they had been exposed for a
period of almost one year to the phosphoric acid corrosion-control mitigation strategy
implemented in the summer of 2004. The passivated LSL sections described in this
section were used as received. An attempt was made to create unpassivated LSL
sections for this testing by polishing the interior of relatively long sections of the LSLs by
forcing a tightly wadded plastic abrasive sheet (Scotch-Brite) back and forth along the
length of the LSL section. This was at least partially successful and did remove a
portion of the very adherent and very protective passivating layer on the aged LSL
sections. Because of the length of the test sections it was not possible to use a
mechanical reaming tool as was done in the indirectly coupled cell testing. Although the
specimen geometry did not allow for a rigorous polishing, the effort did produce a
substantial change in the surface electrochemistry that was evidenced in subsequent
testing. For purposes of this discussion, partially polished LSL sections are referred to
as unpassivated.
Potentiostat
Reference Electrode
Soldered Pigtail
Electrode
Insertion Port
Sample Port
Cu Service Line (3/4 dia.
Pb Service Line
Pump
Figure 5. Schematic of a test rig showing the direct coupling of LSL and
copper pipe sections
Figure 6. Photo of directly coupled LSL and copper sections used in a
portion of this study
Final Report
11
November, 2006

-------
Figure 7 presents a comparison of the surface potentials measured along passivated
and unpassivated LSL sections connected to equivalent lengths of copper tubing in a
directly coupled cell. The comparative results are significant at two levels: first and
foremost, as in the testing in the indirectly coupled eels described earlier, direct coupling
of new copper tubing to well-passivated LSL sections has almost no discernible
electrochemical impact on the LSL. Secondly, coupling to an unpassivated LSL section
shows the converse, and is in fact strongly influenced by the connection to the copper
tubing. This second point underscores the earlier observations that an unsealed and
unpassivated LSL section is highly electroactive, but that once it is passivated it is
remarkably polarization resistant.
Lead Service Line
Freely Corroding Cu Surface Potential^
Freely Corroding Pb Surface Potential
Passivated
V= sample locations
Figure 7. Surface potential along the length of coupled LSL and copper
service lines
Water Quality and Galvanic Impacts.
The directly coupled cell approach was used to assess the impact of important water
chemistry changes including chlorine chemistry and concentration, and the impact of
water conductivity. The baseline water chemistry used in this testing was a simulated DC
WASA water having similar pH, alkalinity, hardness and conductivity profiles.
Figure 8 shows the effect of chlorine concentration on the galvanic impacts on a
passivated LSL section. An increasing free-chlorine residual elevates the galvanic effect
by accelerating the cathodic current exchange process on the copper pipe, however, the
Final Report
12
November, 2006

-------
impact overall is limited to the copper service line, while the impact on the passivated
LSL surface is nearly imperceptible.
Figure 9 shows the comparative effect of equal concentrations of free and combined
chlorine relative to galvanic coupling using a passivated LSL sections. The test shows
no meaningful difference in galvanic impact on either the lead or copper surfaces. (Note:
While the DC WAS A lead solubility issues were the result of redox chemistry impacts
associated with the change from free to combined chlorine, the test used here is capable
of discerning only fundamental changes in surface electrochemistry, and tells us nothing
about solubility of the existing corrosion scales).
Figure 10 shows the impact of increasing conductivity levels on the galvanic impact
relative to a passivated LSL section. Conductivity increases were brought about by the
simple addition of NaCI to the recirculating water in the test rig. Interestingly, water
conductivity has a more important effect on the galvanic process than chlorine residual,
or chlorine type. The area of galvanic influence on the LSL specimen is marginally
expanded as the conductivity of the electrolyte (water) increases, while the area of
influence on the copper service line is substantially expanded. This would appear to be
because the higher conductivity lessens the resistance of the electrolyte circuit (water),
expanding the "reach" of the galvanic current. (Note: DC l/V/AS/A distributes a low
conductivity water (< 100 microSiemens), which, in part, explains the minimal galvanic
impacts observed.)
200-
Lead Service Line

5 mq/L C!
Freely Corroding Cu Surface Potentials,
1 mg/LCl
mg/L C\?
Freely Corroding Pb Surface Potential
UO -400
-500J-r
m
20
10	0	10
Junction Displacement (cm)
20
V= sample locations
Final Report
13
November, 2006

-------
Figure 8. Effect of chlorine concentration on galvanic impacts relative to
passivated LSL specimen.
200-
Lead Service Line
Freely Corroding Cu Surface Potential
Freely Corroding Pb Surface Potential
Ji -400-
10	0	10
Junction Displacement (cm)
V= sample locations
Figure 9. Comparative effect of free versus combined chlorine on
galvanic impacts relative to passivated LSL specimen.
200
u
U1
>
£
c
a;
4-J
o
CL

-------
Figure 10. Effect of conductivity on galvanic impacts relative to
passivated LSL specimen.
Dielectric Insertion.
A dielectric is an insulating device that prevents direct electrical contact between
dissimilar metals, and hence avoids at least some of the problems associated with
galvanic coupling. Although not always used, it is generally considered good plumbing
practice to use a dielectric when different metal plumbing materials are to be connected.
It is standard policy for DC WASA to use dielectric couplers when performing partial LSL
replacements (DC WASA, 2004).
Figure 11 shows the effect of inserting a dielectric coupler between the passivated LSL
and copper sections of the directly coupled test rig. While in the previous water quality
testing it was shown that the galvanic effect of the direct coupling was largely limited to
the copper line and had little effect on the LSL, the insertion of a dielectric removes any
galvanic impact from either surface. This is a particularly important finding, and along
with the general polarization resistance of passivated LSL sections, explains why the
partial LSL replacement program in the DC WASA system has not exacerbated lead
conditions, but rather has helped to reduce household lead levels.
200+
100
Dielectric Insertion (1 in. separation)-
U
1/1
~ -100
5 -200
•*-»
o
CL
g -300
TO
<3? -400
-500
*
Lead Service Line	( |) Copper Service Lit
Uncoupled
ICj>U|
Freely Corroding Cu Surface Potential^

10	0	10
Junction Displacement (in.)
—r—
20
V= sample locations
Figure 11. Effect of inserting a dielectric between the passivated LSL and
copper sections on galvanic impact.
Final Report
15
November, 2006

-------
Grounding and Impressed Currents
Does a current flowing in the pipe wall, and exiting the pipe via an external connection (a
typical electrical-system grounding scenario in many older households), change the
electrical character of the internal pipe surfaces? If it does, a variety of corrosion and
water quality impacts are to be expected, including accelerated metal release. However,
a clear absence of a measurable electrical change on the internal surface would mean
the grounding circuit is irrelevant to the internal corrosion processes.
That is the question that was addressed by this portion of the study which focused on
investigating the interior LSL surfaces relative to simulated grounding currents. The
objective was to demonstrate under controlled conditions whether it was possible to
create a grounding scenario that accelerates internal corrosion, and to extrapolate its
relevance to household plumbing practices.
Although this topic has been previously researched, there is still considerable debate
about the impact of externally imposed grounding currents on the electrochemistry of
domestic plumbing. The bulk of available research has focused on copper tubing -
largely ignoring grounding impacts on LSLs. While some research has suggested an
important internal corrosion role for grounding currents (Bell, 1998), other laboratory
simulations and field tests have discounted them relative to copper release and
associated water quality effects (AWWARF, 1996).
At first glance, it seems intuitive that imposing a (grounding) current on a buried pipe
would change the surface potential of the internal and external surfaces. Certainly, as
the current is dissipated to ground, the surface potential of the external surfaces does
change. However, internally, unless some portion of the grounding current is lost to the
electrolyte (water in the pipe), these surfaces will show no change in surface potential
relative to the water contacting them. In effect, imposing an external current on the
pipeline changes the potential of all surfaces (internal and external) and everything in
contact with the pipe. Internally, however, the surface potential relative to the electrolyte
(which determines corrosion rates) may not change since the electrolyte potential has
also been shifted an amount equal to the internal surface.
Impressed Current Experiments
A flow-through recirculation loop consisting of DC WASA LSL segments, new copper
tubing, water reservoir, flow control and pumping hardware was employed for this
testing. As in the previous galvanic work, the LSL segments were modified to accept
high impedance reference electrodes penetrating the pipe wall at multiple locations
along its length. The electrodes monitor surface potential on the interior of the pipe
relative to the electrolyte, yet allow for pipeline pressurization. Internal surface potential
along the pipeline was monitored, while different current forms, amperages, voltages
and grounding scenarios were applied to the test pipes. Figure 12 presents a schematic
of the test rig.
Quantifying the actual interior surface potential change vis-a-vis the electrolyte (water) of
these pipes was key to assessing grounding current impacts. Any meaningful change in
the corrosion condition of the internal surface could be assessed by measuring any
substantial change in surface potential, which could be monitored with a high degree of
Final Report
16
November, 2006

-------
accuracy (+/- 0.1 mV). Table 1 presents a summary of the basic electrical testing profile
used in this examination.



<5> Q
r
Potentiostat
AC and DC
Current
Source
Current Flow
Reference Electrode
Soldered Pigtail
Electrode
Insertion Port
Sample Port
Headpiece
Pb Service Line
/"Calomel
X Ref.Elec.
Rubber
Stopper
Pipe Wall
mm gap
Figure 12. Schematic illustration of impressed current test rig
Ta
ble 1. Grounding Currnet Testing profile
Impressed grounding
current forms
•	Full wave AC
•	DC
Voltage range
0-120 Volts AC, 0 - 12 Volts DC
Current range
0-20 Amps AC, 0-6 Amps DC
Grounding scenarios
•	Single-point ground
•	Multi-point ground along pipe length
•	Variable resistance reservoir ground: By providing a
current path from the pipe wall through the electrolyte to
ground, it was possible to dissipate some of the applied
external current to the interior surface of the pipe.
Final Report
17
November, 2006

-------
With the test rig in operation, the actual assessment could be conducted quickly,
generally requiring no more than a few hours per test condition. The testing proceeded
from impressing direct currents at minor voltages and amperages to upwards of a 12-
Volt current at up to six amps of current flow. Attempts to measure the impact of the
impressed current on the interior surface potential were made at different locations along
the pipe rig. Different grounding scenarios were tested in conjunction with the
impressed current in an attempt to force the impressed current to flow the full length of
the pipe samples, as well as to force as much of the current to transfer to the water flow
as possible. (Presumably, current transfer from pipe wall to water creates a corrosion
cell). The most rigorous of the grounding protocols involved adding a ground to the
recirculation water reservoir in parallel with the pipe rig ground, thus allowing a direct
current path from the pipe wall to the recirculating water.
Following the DC testing, alternating currents were imposed on the pipe rig using a
standard 120-Volt (breakered at 20 amp) wall-type circuit. Grounding scenarios similar
to the DC testing were employed. A summary of actual results from this testing is
straightforward: impressed currents, whether AC or DC, had no meaningful impact on
the surface potential of the pipe rig regardless of voltage or amperage. Figure 13
summarizes the results of the highest voltage DC and AC tests, demonstrating that
these impressed currents did not meaningfully shift the interior surface potentials of the
test specimens.
Impressing an AC current, however, does create substantial electrical noise, making it
difficult to measure a stable potential. Yet, while the noise effect expanded the range of
variability by about 5 mV, the baseline potential did not shift. There is an
electrochemical argument to be made about the capacitive effect of an AC current
applied across a corrosion scale, yet, if the corrosion potential of the interior surface
does not change vis-a-vis the water, capacitance is irrelevant.
From this testing we conclude the obvious: currents flowing in pipe walls take the path of
least resistance to ground, producing no change relative to the corrosion potential of the
internal surface, whether it be a copper or lead service line.
Final Report
18
November, 2006

-------
y u
Q
>
E
£
3
OO
Q_
~D

QJ
I—
CL
e
u
<
20
u
-100'
DC Current Off
DC Current On (6 V)


Instantaneous Spike



i i i i i i i i i i i i i
3 100 200 300 400 500 600 700 800 900 1000 1100 1200
Time Domain (sec.)
1
AC Current Off
AC Current On (120 V)
>
i -200+
«->
c
O)
f -300+
I
3
"AC Induced (60 Hz) Monitoring Noise
-400+
-500r
0 100 200 300 400 500 600 700 800 900 1000 1100 1200
Time Domain (sec.)
Figure 13. Impressed current impacts (separate DC and AC tests) on
surface potential of an LSL coupled to copper tubing.
Final Report
19
November, 2006

-------
Observations and Conclusions
This study has shown that grounding and/or impressed currents moving along LSLs, and
eventually leaving the pipe to ground, have no meaningful impact on internal pipeline
corrosion and do not likely contribute to metals release. Secondly, while the study found
that galvanic impacts can be substantial on unpassivated lead surfaces, the magnitude
of the impact on aged and passivated LSL surfaces (as well as on copper service lines)
is so minimal as to be inconsequential. The study provides a strong basis for
discounting claims and concerns relative to accelerated metal release associated with
PLSLR. Moreover, we believe that the long-debated controversy about grounding
currents impacts can now be considered closed.
The most significant observations to be drawn from this study are summarized below.
Passivation.
Lead is a highly electroactive metal, and in pure form oxidizes extremely rapidly. An
unsealed lead surface, even under natural environmental conditions, has an exceedingly
high initial corrosion rate. Fortunately, lead also passivates strongly and quickly.
Observations in this study suggests that meaningful passivation on LSLs can be
achieved within a matter of days. Well-aged DC WASA LSL specimens - especially
those that have been exposed to an orthophosphate inhibitor - are exceptionally well
passivated and highly resistant to electrical perturbations.
Lead Electrochemistry.
Passivated LSL specimens are highly polarization resistant - meaning that it takes an
exceptional surface perturbation to affect the underlying corrosion rate. The actual
degree of polarization resistance expressed as a Tafel Value is in excess of 500 - 600
mV per decade of current shift. Overall, this explains, at least in part, why the galvanic
coupling has little apparent effect on passivated lead surfaces.
Area of Galvanic Influence.
When coupled to a new length of copper tubing (as in a partial LSL replacement) the
area of galvanic influence on a well-passivated LSL is likely limited to less than the first
inch of LSL pipe in the immediate vicinity of the coupling. The galvanic area of influence
on an unpassivated LSL specimen is larger, but likely limited to the first few inches of
pipe in the vicinity of the coupling. As the LSL passivates, the area of galvanic influence
decreases rapidly. The period of transition can be as short as a few days under normal
distribution system conditions.
A potential reason why galvanic impacts do not generate a more significant corrosion
response relates to the respective geometries of the anodic and cathodic surfaces of the
pipeline couple. Because sequential pipelines (LSL to copper tubing) are connected at
only a single location, only a small portion of the LSL is polarized by the galvanic current.
And, given the relatively rapid rate at which both copper and lead surfaces passivate, the
duration of the polarization is relatively brief. Hence even the meager galvanic effect, is
short-lived.
Final Report
20
November, 2006

-------
Cathodic Effect of Copper Pipe.
The cathode/anode ratio on a well-passivated LSL surface is unimportant relative to the
galvanic effect. This means that even an exceptionally long length of copper pipe
connected to a partial LSL does not elevate the galvanic effect. (It had been argued that
long lengths of copper service line connected to short LSL sections would exacerbate
the galvanic effect.)
Water Quality and Galvanic Impacts.
A free-chlorine residual does elevate the galvanic effect by accelerating the cathodic
current exchange process. Conversely, chloramine has a lesser galvanic impact than
free chlorine. The impact overall, however, is largely limited to the galvanic influence on
the copper service line. The overall impact on the LSL surface is nearly imperceptible.
Interestingly, water conductivity has a more important effect on the galvanic process
than chlorine residual. The area of galvanic influence on the LSL specimen is marginally
expanded as the conductivity of the electrolyte (water) increases, while the area of
influence on the copper service line is substantially expanded. This is because the
higher conductivity lessens the resistance of the electrolyte circuit (water), expanding the
"reach" of the galvanic current.
DC WASA distributes a low conductivity water (< 100 microSiemens), which, in part,
explains the minimal galvanic impacts observed.
Dielectric Effects.
While galvanic impacts relative to DC WASA PLSLRs are likely minimal, any break in
electrical continuity between the copper and LSL lines effectively eliminates the potential
for a galvanic effect. In short, a conventional plumbing dielectric junction removes even
the minor corrosion risks associated with galvanic coupling.
Impressed Current Effects.
Impressed currents (AC or DC) on LSLs and copper service lines, including grounding
type currents, have no impact whatsoever on the internal corrosion of the household
service lines (or any other plumbing appurtenance for that matter). There is likely no
acceleration of corrosion associated with the conventional practice of electrical system
grounding to household water systems.
Final Report
21
November, 2006

-------
References
Reiber, S., "Galvanic Stimulation of Lead/Tin Solder Sweated Joints," Journal AWWA,
1991
Bell, G.E.C., "Observation on the Effect of Grounding on Water Piping,"
CORROSION/95, Orlando, 1995
Wujek. J, "Minimizing Peak Lead Concentrations After Partial Lead Service Line
Replacement," WQTC, San Antonio, 2004
AWWA Research Foundation, Internal Corrosion of Water Distribution Systems - 2nd
edition, Denver, 1996
Giani, R., Keefer, W., Donnelly, M. 2005. "Studying the effectiveness and stability of
orthophosphate on Washington DC's Lead Service Line Scales," WEFTEC, Washington
DC, 2005.
Reiber. S, Giani, R, "National Implications of the DC WASA Lead Experience,"
WEFTEC, Washington DC, 2005
Korshin, G.. Fundamental Mechanisms of Lead Oxidation: Effects of Chlorine,
Chloramine and Natural Organic Matter on Lead Release in Drinking Water, proposal to
National Science Foundation, 2005
DC WASA, personal communication with Richard Giani, 2004
Bell. G.E.C., "Effects of Grounding on Metal Release in Drinking Water," AWWA
Inorganics Contaminants Workshop, San Antonio, 1998
AWWA Research Foundation, "Electrical Grounding, Pipe Integrity and Shock Hazard,
Denver, 1996
Final Report
22
November, 2006

-------
Appendix D Lead Profile Results

-------
This page intentionally left blank.

-------
D.1 Total and Dissolved Lead Profile for a Residence Without a Lead Service Line, Before
Orhtophosphate Application
180
160
In House Plumbing
Copper Service
Line
7-7-04 (Profile a)
Main
140
120
100
80
60
40
20
0 1 2 3 4 5 7 9 11 13 15 18 21 24 24+3 24+10 X 01(15)07(15)
~ Total Lead ~ Dissolved Lead
D-l

-------
D.2 Total and Dissolved Lead Profiles from Samples Drawn Prior to Initiation of
Orthophosphate Treatment Program
12-8-03 (Profile No. 1)
ln-house Plumbing

LS
L
Main




—


















—
—






-i





-i










— 	 	
m






















0	1	2	4	5	7	9	13	17	21	25 25+3 X
Liter
~ Total Lead ¦Dissolved
12-15-03 (Profile No. 2)
	



—

'1




ln-house Plumbing
LSL





—
Main



—
































~I





¦ i-L n





























0	1	2	3	4	5	6	7	911	13 13+3 X
Liter
~ Total Lead ¦ Dissolved Lead
1-5-04 (Profile No. 3)
180
ln-house Plumbing
LSL
Main
160
140
120
Q. 100
80
60
40
20
0
2
4
5
6
8
10
12
14
16
16+3
X
0
Liter
~ Total Lead B Dissolved Lead |
D-2

-------
1-13-04 (Profile No. 4)
ln-house
Plumbing
LSL
Copper
Replacement
Main



































m. Ilk
	






1	2	4	5	7	9	13	17 21	25	45	X	0
Liter
~ Total Lead IB Dissolved Lead~|
2-9-04 (Profile No. 5)
180
ln-house Plumbing
LSL
Main
160
140
120
Q. 100
80
60
40
20
0
2
4
5
6
7
8
9
13
16
19
19+3
0
X
Liter
~ Total Lead U Dissolved Lead |
2-24-04 (Profile No. 6)
ln-house Plumbing
LSL
Copper
Replacement
Main
































1 2 4 5 6 8 10 12 14 16 16+3 16+5 16+7 16+9 16+11 X 0
Liter
[~Total Lead M Dissolved Lead |
D-3

-------
3-2-04 (Profile No. 7)
180 -I
160 -
140 -
120 -
o. 100 -
Q.
"O
« 80-
—I
60 -
ln-house Plumbing
40
20
o -M
11	13 15 18 21 21+3 21+10 X	0
Lm
~ Total Lead ¦ Dissolved Lead
3-9-04 (Profile No. 8)
ln-house Plumbing
160 -
140 -
120 -
Q. 100 -
Q.
"O
§ 80-
—I
60 -
40 -
II

11 14 17 20 24 29 29+3 29+10 00 X
Liter
~ Total Lead ¦ Dissolved Lead
180 -,
160 -
3-16-04 (Profile No. 9)
2 15 18 21 24 27 31 35 39 39+3 39+10 X 0
Detailed information was not available to
differentiate among in-house plumbing, lead
service lines, and the water main
~ Total Lead ¦ Dissolved Lead
D-4

-------
3-24-04 (Profile No. 10)
180
ln-house
Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
2
3
4
6
8
10
12
14
17
20
23
26
30
0
30+3 30+10
X
Liter
~ Total Lead U Dissolved Lead |
3-24-04 (Profile No. 11)
180
ln-house
Plumhinn
LSL
Main
160
140
120
100
80
60
40
20
0
2
4
6
8
10
12
14
17
20
23
26
29
33
33+3 33+10
0
X
Liter
~ Total Lead U Dissolved Lead |
3-30-04 (Profile No. 12)
180
ln-house Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
2
3
5
7
9
10
13
15
18
21
24
27
0
27+3 27+10
X
Liter
~ Total Lead ¦ Dissolved Lead
D-5

-------
3-31-04 (Profile No. 13)
180 -I
160 -
140 -
120 -
a. 100 -
Q.
¦a
« 80-
ln-house Plumbing
60 -
m
nmmmmmmin
Information was not available for
Dissolved Lead at Liter 9
11 13 15 18
Liter
~ Total Lead ¦ Dissolved Lead
4-1-04 (Profile No. 14)
24 27 27+3 27+10
ln-house
Plumhino
Q. 100
10 12 14 17 20 23 26 30 30+3 30+10 00 X
~ Total Lead ¦ Dissolved Lead
4-5-04 (Profile No. 15)
ln-house Plumbing
Q. 100
1 13 15 18 21 24 27 27+3 27+10 X
~ Total Lead ¦ Dissolved Lead
D-6

-------
4-6-04 (Profile No. 16)
In -house , LSL . Main
Plumbina i i
i i
i i
i i
I I
i i

—
i i
I I
I i
i i
i i
I I


I I
i i

—
i i
n — |-, i n—I "I i
1 r| n-1 M-11 ¦ 1 r"L_ i
ill ill ill ill ill 111 n~i m~i i it—i n~i m~i m~i rn n~i n~i n~i n~i
1 2 3 4 5 7 9 11 13 15 18 21 24 27 0 27+3 27+10 X
Liter
| DTotal Lead U Dissolved Lead~|
4-6-04 (Profile No. 17)
180
In -house
LSL
Main
160
140
120
100
80
60
40
20
0
2
3
4
6
8
10
12
14
16
19
22
25
28
28+3 28+10
X
Liter
~ Total Lead ¦ Dissolved Lead
4-13-04 (Profile No. 18)
180
In -house
Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
2
3
4
6
8
10
12
14
16
19
22
25
28
00
28 + 3 28 + 10 X
Liter
~ Total Lead ¦ Dissolved Lead
D-7

-------
4-26-04 (Profile No. 19)
180 -I
160 -
140 -
120 -
a. 100 -
Q.
¦a
« 80-
ln -house Plumbing
60 -
10 12 14 16 18 21 24 27 27+3 27+10 X
~ Total Lead ¦ Dissolved Lead
4-27-04 (Profile No. 20)
In -house
Plnmhinn
o. 100
12 14 17 20 23 26 29 33 33+3 33+10
180
160 -
140 -
120 -
60 -
40 -
~ Total Lead ¦ Dissolved Lead
4-29-04 (Profile No. 21)
ln-house Plumbing
13 17 21	25 25+3 25+10
Liter
~ Total Lead ¦ Dissolved Lead
D-8

-------
4-30-04 (Profile No. 22)
ln-house , LSL . Main
Plumbinq i i
i i
i i
i i
i i
i i
--
i i
i i
i i
i i
i i
i i
i i
i i

i i
i i
i i
i i
m r~^ i ^ i rb r~^ i—_r_—_r_—_r_—_r_—_r_—_r_—_r_—_r_—_r_—
1 2 3 5 7 9 10 11 13 15 18 21 24 27 27+3 27+10 0 X
Liter
~ Total Lead U Dissolved Lead |
5-3-04 (Profile No. 23)
180
ln-house
Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
2
3
4
6
8
10
12
14
17
20
23
26
30
30+3 30+10
0
X
Liter
~ Total Lead ¦ Dissolved Lead
5-3-04 (Profile No. 24)
180
LSL
Main
ln-house
Plumbing
160
140
120
100
80
60
40
20
0
2
3
4
6
8
10
12
14
17
20
23
26
30
30+3 30+10
0
X
Liter
~ Total Lead ¦ Dissolved Lead
D-9

-------
5-7-04 (Profile No. 25)
180
ln-house
Plumbing
LSL
Main
160
140
120
Q. 100
80
60
40
20
0
2
3
4
5
7
9
13
15
18
21
24
27
00
27+3 27+10
X
Liter
~ Total Lead U Dissolved Lead |
5-18-04 (Profile No. 26)
180
ln-house Plumbing
LSL
Main
160
140
120
Q. 100
80
60
40
20
0
2
4
5
6
7
8
9
13
16
19
19+3 19+10
0
X
Liter
~ Total Lead ¦ Dissolved Lead
6-28-04 (Profile No. 27)
180
ln-house
Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
0
2
3
4
6
8
10
12
14
17
20
23
26
30 30+3 30+10 X 01(15) 06(15)
Liter
~ Total Lead ¦ Dissolved Lead
D-10

-------
7-6-04 (Profile No. 28)
180
ln-house Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
00
2
3
5
7
9
13
15
18
21
24
27
27+3 27+10
X
Liter
~ Total Lead ¦Dissolved Lead
7-16-04 (Profile No. 29)
ln-house Plumbing i
i
LSL i
•
Main


i i
i i
i i
i i
i i
i i
i i


i i
i i


i i
i i


i i
i i


i i
i i


i i
¦ ri- I l
1 m 1 II n~i m i m n~i I I—i n-i n-i» i-^ r~^ i-m


00 1 2 3 4 5 7 9 11 13 15 18 21 24 27 27+3 27+10 X
Liter
~ Total Lead ¦ Dissolved Lead
D-ll

-------
D.3 Total and Dissolved Lead Profiles from Samples Drawn After Initiation of
Orthophosphate Treatment Program
11-30-04 (Profile No. 30)
80
ln-house Plumbing
60
40
20
00
80
60
40
20
0
2
3
5
6
8
10
12
17
20
23
26
0
26+3 26+10
Liter
~Total Lead ¦Dissolved Lead
12-6-04 (Profile No. 31)
180
Plumbing
160
140
120
100
80
60
40
20
0
1
2
3
4
5
7
9
11
12
13
15
17
20
23
26
00
26+3 26+10
X
Liter
~ Total Lead ¦ Dissolved Lead
1-6-05 (Profile No. 32)
ln-house
Plumbing
LSL
Main





















n_ ri—.
n fL ¦ fl_ n_
n
1 2 3 4 6 8 10 12 14 17 20 23 26 30 0 30+3 30+10 X
Liter
~ Total Lead IB Dissolved Lead~|
D-12

-------
1-25-05 (Profile No. 33)
180
ln-house Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
2
3
4
6
8
10
12
14
16
18
21
24
27
27+3 27+10
0
X
Liter
~ Total Lead ¦Dissolved Lead
2-22-05 (Profile No. 34)
1
ln-house Plumbing i
i
i
LSL i
i
Main


i i
i i
i i
i i
i i
i i
i i
i i
i i
n

i i
i i


i i
i i


i i
i i


i i
i_I i i
1 L n— —, i i n


0 1 2 3 5 7 9 11 12 13 15 17 20 23 26 26+3 26+10 X
Liter
|~DTotal Lead U Dissolved Lead |
3-30-05 (Profile No. 35)
180
ln-house Plumbing
Main
160
140
120
100
80
60
40
20
0
00
2
3
5
7
9
12
13
15
18
21
24
27
27+3 27+10
X
Liter
~ Total Lead B Dissolved Lead |
D-13

-------
4-29-05 (Profile No. 36)
180
ln-house
Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
0
2
4
6
8
9
10
12
14
15
17
20
23
26
26+3 26+10
X
Liter
~ Total Lead U Dissolved Lead |
5-16-05 (Profile No. 37)
180
ln-house Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
0
2
4
5
6
7
9
10
13
15
17
20
23
23+3 23+10
X
Liter
~ Total Lead ¦ Dissolved Lead
6-1-05 (Profile No. 38)
180
ln-house
Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
00
2
3
5
6
7
9
12
13
15
17
20
23
23+3 23+10
X
Liter
~ Total Lead ¦ Dissolved Lead
D-14

-------
6-7-05 (Profile No. 39)
180
ln-house Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
0
2
3
4
6
8
9
10
13
15
18
21
24
24+3 24+10
X
Liter
~ Total Lead U Dissolved Lead |
7-25-05 (Profile No. 40)
.00
ln-house
Plumbing
LSL
Main
.00
.00
.00
.00
.00
.00
.00
.00
.00
o
2
3
4
5
7
8
9
13
15
17
20
23
23+3 23+10 X
Liter
~ Total Lead ¦ Dissolved Lead
9-28-05 (Profile No. 41)
180
ln-house
Plumbing
LSL
Main
160
140
120
100
80
60
40
20
0
00
2
4
5
6
7
9
10
13
15
17
20
23
23+3 23+10
X
Liter
~ Total Lead ¦ Dissolved Lead
D-15

-------
10-5-05 (Profile No. 42)
ln-house
Plumbing
140 -
120 -
100 -
11 13 15 17 20 23 23+3 23+10 X
~ Total Lead ¦ Dissolved Lead
11-29-05 (Profile No. 43)
ln-house
Plumbing
120 -
100 -
80 -
0 1 2 3 4 5
7 9 11 13 15 17 20 23 23+3 23+10 X
Liter
~ Total Lead ¦ Dissolved Lead
12-12-05 (Profile No. 44)
ln-house
Plumbing
140 -
120 -
60 -
40 -
0 1 2 3 4 5
11 13 15 17 20 23 23+3 23+10 X
~ Total Lead ¦ Dissolved Lead
D-16

-------
1-27-06 (Profile No. 45)
ln-house Plumbing
20+3 20+10 X
~ Total Lead ¦ Dissolved Lead
D-17

-------
This page intentionally left blank.

-------
Appendix E Technical Memorandum from Dr. Anne Camper,
July 30, 2004

-------
This page intentionally left blank.

-------
TECHNICAL MEMORANDUM
TO:
The Cadmus Group
FROM:
Anne K. Camper
DATE:
July 30, 2004
RE:
Evaluation of Washington, DC Distribution System; Orthophosphate

Addition Effects on Microbial Water Quality
In accordance with my agreement with The Cadmus Group, and based upon the information I
have received from them and from telephone conversations with the Technical Expert Working
Group, I have prepared this technical memo to address specific questions posed by Cadmus.
Questions included:
1.	Considering WASA's distribution system infrastructure (e.g., extent and age of cast iron
pipes) and water quality, do you believe that orthophosphate addition will help control biofilm
growth? More specifically, how will the addition of orthophosphate affect TCR compliance?
Do you have any other recommendations for controlling biofilm growth and improving TCR
compliance?
2.	What detrimental effects could there be in a distribution system such as WASA's from
maintaining an orthophosphate residual of 3 mg/L?
3.	Can the Aqueduct reduce the disinfectant residual concentration in the DC distribution system
without adversely impacting TCR compliance?
4.	WASA practices unidirectional flushing in their distribution system. They routinely flush the
entire system approximately every two years. How important is unidirectional flushing for the
DC distribution system? In your opinion, how often should WASA flush their entire system
(e.g., once a year, once every two years, once every five years)?
Before these questions are addressed, some general comments are necessary. Each distribution
system is unique. The nature and extent of biofilms in distribution systems is influenced by the
interaction of many factors including pipe materials, water quality (including disinfectant and
organic matter) etc. Therefore, extrapolating the performance of a corrosion inhibitor,
disinfectant, or other treatment change on the response of biofilms from one system to another
should be done with caution. It is also important to note that there can be a difference between
the control of biofilms and the suspended cells that are measured during routine monitoring. The
tendency is to assume that a reduction in HPC counts also means a reduction in biofilms. It is
only possible to infer some connection between suspended cell counts and biofilm cell numbers
in non-disinfected systems. Even then, detachment of biofilm cells may not be constant even
under steady flow conditions. This is further complicated by flow reversals, surges in velocity,
etc. If a disinfectant residual is present, the suspended cells may be reduced in number but the
biofilm organisms are unaffected. When evaluating the impact of a treatment change on biofilm
or suspended HPC and coliform counts, historical trends prior to the change must be taken into
consideration. Separation of variables in full-scale systems is nearly impossible; a decrease in
1

-------
HPC counts after the change may not be related to the new practice but to some other variable
like decreasing water temperatures. This can be partially ascertained by evaluating historical
data at that location.
QUESTION 1. DISTRIBUTION SYSTEM INFRASTRUCTURE AND ORTHOPHOSPHATE
Distribution system materials
A review of the provided information indicates that the majority of the WAS A system is unlined
iron or steel, with a large portion being cast iron (Section 6, Distribution Piping Improvements,
CDM Report). In general, the literature and practice has shown that biofilms on iron surfaces are
more problematic than those on other types of surfaces (LeChevallier et al. 1998, LeChevallier
1997, Delanoue et al. 1997, Neden et al. 1992, Niquette et al. 2000, Kerr et al. 1999).
LeChevallier et al. (1996) showed that coliform occurrences correlated directly with the number
of miles of unlined cast iron pipe in the distribution system. Another study showed that the
biofilm densities on iron were significantly higher than those on plastic-based materials such as
PVC with densities on cement intermediate to plastic-based materials and iron (Niquette et al.
2000). Neden et al. (1992) reported that cast iron pipes had the highest counts, while PVC was
the lowest. Camper (1996) demonstrated in laboratory studies that mild steel had 10-fold higher
HPC and coliforms than polycarbonate surfaces in reactors operated in an identical fashion. In
another laboratory study, Kerr et al. (1999) found the highest counts and species diversity on cast
iron when compared to medium-density polyethylene and uplasticized PVC. In another
laboratory study where iron, PVC, cement and epoxy were compared, biofilm levels on iron
were always higher than the other materials with the exception of when the systems were fed
biologically treated water (Camper et al. 2003). It was also found that increases in DOC led to
general increases in biofilm and effluent HPC, and this effect was most pronounced for reactors
that contained iron. Parallel field testing results indicated that either iron had the highest
regrowth, or the type of material had no influence on the number of bacteria present.
Involvement of organic matter
One of the difficulties in interpreting the importance of iron surfaces on regrowth is the
interaction of iron oxides with natural organic matter. As will be explained below, these iron
surfaces have an affinity for natural organic matter, but the potential for immobilized organic
matter to act as a carbon and energy source for biofilm organisms has been largely ignored. The
industry generally attempts to correlate fractions of the organic matter (AOC and BDOC)
obtained from bulk water fractions with growth determined in suspended systems to what can be
expected in the distribution system. However, in many distribution systems, AOC and/or BDOC
levels have not been correlated with regrowth. There are several reasons for this inconsistency:
(1) an inadequate number of AOC or BDOC measurements to truly represent the level of organic
carbon available for growth, (2) a significant interaction of other factors (disinfectant,
distribution system materials, etc) that govern microbial growth, and/or (3) the presence of
organic carbon promoting biological growth not measured by these tests. In fact, Najm et al.
(2000) report that utilities should not rely on AOC or BDOC levels alone to assess the potential
for regrowth; other factors should be considered as well. The third possibility, that there are
components of the organic material recalcitrant to degradation in the bulk fluid, is discussed here
in the context of interaction with iron oxide surfaces and corrosion control.
Part of the recalcitrant components of total organic carbon in water is humic and fulvic acids.
Until recently, there has been only one reference suggesting that biofilm bacteria are capable of
using humic materials (Volk et al. 1997). In our laboratory it was demonstrated that soil derived
humic substances fed at a level of 1.4 mg/L, resulted in a biofilm that was at least one log higher
than the same water before the addition of humic substances (Burr et al. 2004). In earlier work,
2

-------
humic substances supported nearly the same number of biofilm and suspended cells in a
simulated drinking water distribution system as more readily available carbon sources (amino
acids and carbohydrates)at the same concentration of carbon (Butterfield et al. 2000b, Ellis et al.
2000). Across the reactor, nearly 80% of the humic substances were removed. This may be
caused by the large amount of substrate bound to the biofilm (8.3 -11 jug C/cm2), which would
lead to growth rates independent of the bulk fluid concentration. This adsorption may influence
the ability for microorganisms to degrade humics. This phenomenon is most widely studied on
iron oxides (Parfitt et al. 1977, Tipping 1981, Tipping et al. 1981, Davis 1982, Gu et al. 1996).
Gu et al. (1995) showed that different fractions of NOM are adsorbed by iron oxide with
different affinities and capacities.
Implementation of corrosion control on distribution system biofilms
As will be described in more detail below, implementation of corrosion control has been seen to
decrease biofilm cell numbers and suspended cell counts in systems containing corroding iron
when all other parameters are held constant. It is unknown if the effect relates to improvement
of disinfectant efficacy by reducing the amount of available iron to react with the disinfectant, to
the reduced availability of the adsorbed organic matter, or other causes. It is likely that there is a
combination effect that is dependent on the type of organic matter, the surface, and the
disinfectant.
To see if corrosion control may improve disinfectant efficacy, work has been done with reactors
exposed to free chlorine or monochloramine with and without the presence of corrosion control
chemicals (orthophosphate, polyphosphate, pH adjustment). In ductile iron annular reactors,
there was a substantial demand for both disinfectants. Influent concentrations of 3.25 mg/L in
both cases resulted in barely measurable residuals. When compared to the control reactor (no
disinfectant), the chlorinated reactor had higher numbers of culturable cells in the biofilm. The
counts from the monochloramine system were lower than that of the parallel control system, and
lower than when chlorine was used. Overall, monochloramine appeared to be a slightly better
disinfectant under these conditions and the addition of a corrosion control chemical improved
biofilm control, especially in the chloraminated reactor, with orthophosphate being most
effective (Abernathy and Camper, 1998a, b). These results support field observations that
monochloramine may be more effective at controlling biofilms grown on corroding ferrous metal
distribution system materials, and the combination of monochloramine and corrosion control for
improving microbial water quality has been demonstrated (Schreppel et al. 1997). Since
corroding surfaces are often considered to problems in full-scale distribution systems, the use of
monochloramine as a secondary disinfectant may be indicated when the primary material in the
network is unlined cast or ductile iron pipes, and the use of corrosion control schemes may also
improve biofilm control. However, even this observation must be extrapolated with care. As
shown by Batte et al. (2000), phosphate added to a reactor system containing non-corroding
material (polycarbonate) seemed to protect the biofilms from both chlorine and monochloramine
(both at an applied dose of 0.5 mg/L) when compared to systems with the disinfectants but
lacking the phosphate.
Because of the suspected interaction of humic substances with iron oxides, a project was
specifically designed to determine if the presence of iron oxides enhanced biofilm growth in the
presence of humic substances with phosphate and pH adjustment for corrosion control
(Butterfield et al. 2002a). For this work, the biofilms were grown in small columns containing
glass beads, glass beads covered with a synthetically created iron oxide, or crushed corrosion
products taken from a cast iron drinking water distribution pipe. The surfaces were initially
exposed to humic substances (termed "loading") with the exception of a control. After initial
exposure, all but one column received influent humics at a concentration of ca. 3 mg/L. For
3

-------
glass beads, the humics addition was pH adjusted to 7.3, 8 or 9. For corrosion products, only pH
7.3 and 8 were used. For both corrosion products and glass beads, a set of columns at pH 7.3
received chlorine at an effluent residual of 0.15 - 0.2 mg/L, another was chlorinated plus
phosphate added. A control with humics and phosphate at pH 7.3 was also run. The data for cell
counts in the effluent and on the beads were subjected to a weighted hierarchical process called
the simple multiattribute rating technique was used to determine which columns resulted in the
best performance, which was selected to be the least number of bacteria in the biofilm and
effluent. The addition of humic substances was a major factor in biofilm formation. The
absence of added humics was the second most important factor following chlorine. The
interaction between iron oxides and humics also was high for the corrosion products, suggesting
that the interaction of humics with a reactive iron surface is favorable for biofilm formation. The
corrosion products removed from the distribution system had a much higher adsorption affinity
for the humics. They also retained sufficient humics during the initial loading that even if no
other humics were added to the column influent, far more biofilm than the equivalent iron oxide
bead system was supported. Of the corrosion control schemes tested, pH adjustment had little
positive effect (in fact, the pH 9 system had higher counts than the column held at pH 7.3).
Phosphate addition with chlorine was slightly more effective than chlorine alone when the
corrosion products were used.
In a full-scale system (Southern California Water Company-Southwest District) with over half of
the pipeline being unlined iron and steel, polyphosphate blend was added to a portion of the
distribution system in a pilot scale test to determine if this would reduce disinfectant (free
chlorine) demand by decreasing corrosion. The polyphosphate blend was utilized because the
deposits formed should be softer than those produced by orthophosphate, thus making removal
by flushing easier to accomplish. A secondary anticipated benefit was reduction in HPC. For
approximately the first six months after phosphate addition, the HPC counts increased from 600 -
800 CFU/mL up to 1600 CFU/mL. The counts then gradually declined over a two year period to
an average below 100 CFU/mL. During the first six months, the chlorine residual steadily
increased and then remained constant. This study suggested that it may take three years to
achieve an overall improvement in water quality after implementation of corrosion control and
that flushing should be used to help remove any loosened deposits (Cohen et al. 2003).
A report issued by the American Water Works Association Research Foundation (Kirmeyer et al.
2000) provides information on changes in water quality from five extensively monitored systems
and case studies from eight utilities that implemented corrosion control. Several of these utilities
added orthophosphate, polyphosphate, or both with and without pH adjustment. Specifically,
Detroit utilized orthophosphate at 3 mg/L phosphate, Philadelphia increased their phosphate
dose, and Hartford, Connecticut gradually added blended phosphate to attain a level of 1 mg/L
phosphate. For the case histories, it was found that Springfield MA, Portland ME, Cedar
Rapids, IA and Charleston SC implemented phosphate addition. Evanston, IL added a
phosphate/polyphosphate blend. A general conclusion was that phosphate corrosion control had
minimal secondary impacts (red water, elevated microbial counts, etc.) provided that adequate
residuals were maintained and the pH was held at 7.3 - 7.8. Systems most at risk for adverse
effects included those with large amounts of unlined iron pipe, highly varying distribution water
quality due to many changes in practice or unbuffered water with pH swings, and those systems
that implemented large changes in water quality. This report specifically indicates that for
phosphate, doses should be increased incrementally at values of approximately 0.2 - 0.5 mg/L as
phosphate. They also recommend flushing to help control any negative impacts.
The data from the 4th High monitoring program (Table 1) are somewhat ambiguous on the effect
of phosphate addition on HPC counts. At some locations there seems to be elevated counts
4

-------
while at others the numbers remain low. A comparison was made to data collected prior to
phosphate addition at the 4-H 4 site in April and May; all of these counts are below 84 CFU/ml.
Data from August of 2003 to present at this site have counts as high as 4400 (7/14/04) but the
count the next day is only 33 CFU/ml. This illustrates the inherent variability in HPC counts.
With the information provided, it can only be said that there does not appear to be a dramatic
impact of phosphate addition on HPC counts in the 4th High system.
Based on the results listed above and other sources, it is probable that phosphate addition will
have a beneficial effect on regrowth in the WASA system in the long term. Although there are
some circumstances where phosphate addition could stimulate microbial growth (seen in the
Scandinavian countries) the WASA water is likely to be carbon limited rather than phosphate
limited. Better biofilm control could be from improved disinfectant efficacy or in modification
of the surface chemistry so that the organisms are not as likely to grow. However, the effect may
take some time to occur, and in the short term, there may be elevated counts. In the remainder of
the distribution system that has not been flushed, increased bacterial counts are probable. This
may be seen with coliforms as well; phosphate will be added at a time when there has been a
general upward trend in coliform positives. In addition, late summer is associated with warm
water conditions and is the season when violations are most likely to occur. There could be
sporadic coliform releases from the pipe surfaces until the new surface equilibrium is reached.
Unfortunately, the disinfectant concentration present is not likely to kill these organisms,
especially if they are released in clumps or if they associated with particulate matter released
from the pipe surface from the softening of the scales. Ideally, it would have been beneficial to
have the entire system flushed before the addition of phosphate to minimize the potential for
increased microbial counts.
Other recommendations for controlling biofilm growth and improving TCR compliance
For overall water quality improvement, including reducing disinfection by products, reducing
microbial growth, and potentially improving taste and odor, reduction in natural organic matter is
a high priority. There is a great deal of published literature and practical experience that has
shown this approach to be beneficial in improving microbial water quality during distribution.
As a case in point, the study on pipe material conducted in our laboratory (Camper et al. 2003)
showed the same level of biofilm on iron, PVC, cement and epoxy when exposed to biologically
treated water in the absence of disinfectant. When monochloramine or free chlorine were added
to maintain a residual of 0.2 mg/L, bacterial levels on the iron surfaces became elevated and the
suspended counts also increased; the disinfectants could not control the increased amounts of
biofilm growth and detachment. However, the importance of the disinfectant in controlling
bacterial growth, even when organic matter is reduced, is apparent in the literature. The results
from an American Water Works Association Research Foundation funded field project (Najm et
al. 2000) suggests that the suspended cell counts are more heavily influenced by chlorine
residuals than the level of BDOC or AOC. Waters with very low concentrations of AOC/BDOC
still required the presence of residual to keep bacterial numbers within acceptable levels. Along
with the information presented earlier, it is apparent that there is an interaction between many
water quality factors that contributes to microbial numbers in distribution systems, with the type
and quantity of natural organic matter and disinfectants being key players.
Another recommendation is the removal/relining of the iron pipes. The reported interactions of
natural organic matter, disinfectants, and iron pipes suggest that the iron surfaces are a major
factor in regrowth.
5

-------
QUESTION 2. WHAT DETRIMENTAL EFFECTS COULD THERE BE IN A
DISTRIBUTION SYSTEM SUCH AS WASA'S FROM MAINTAINING AN
ORTHOPHOSPHATE RESIDUAL OF 3 MG/L?
These comments are restricted to the potential impact on microbial growth/biofilms. There is no
information available in the literature on the effect of dose of phosphate on regrowth, provided
that phosphate is not the limiting nutrient for bacterial growth. As stated above, it is unlikely
that the WASA water is phosphate limited (more likely organic carbon limited). The only
potential effect will be the rate at which the surfaces are impacted by the corrosion inhibitor and
the time required to reach equilibrium. There may be initial sloughing of the deposits, causing
higher bacterial counts, followed by a gradual decline in numbers. The slope of this line may be
impacted by the dose of phosphate, but there are no data to support this. Experience has shown,
however, that the goal should be to reach equilibrium and not to change practices just because
deleterious effects are seen initially; these may be short-term and overcome as the system adjusts
to the new water quality (Kirmeyer et al. 2000).
QUESTION 3. CAN THE AQUEDUCT REDUCE THE DISINFECTANT RESIDUAL
CONCENTRATION IN THE DC DISTRIBUTION SYSTEM WITHOUT ADVERSELY
IMPACTING TCR COMPLIANCE?
If corrosion control has a positive impact on microbial counts with the HPC numbers declining
with time and a downward trend in the detection of total coliforms, it is possible that the
chloramine dose can be reduced. It would be best to determine if the corrosion control scheme
was producing desirable results prior to reducing the disinfectant, however, to ensure that control
can be maintained. In addition, it is important to ascertain that there are measurable residuals in
the dead ends and low flow sections of the distribution system. Again, if corrosion control is
adequate, it may be feasible to deliver a residual to these locations even if the concentration
leaving the plant is decreased.
QUESTION 4. WASA PRACTICES UNIDIRECTIONAL FLUSHING IN THEIR
DISTRIBUTION SYSTEM. THEY ROUTINELY FLUSH APPROXIMATELY EVERY TWO
YEARS. HOW IMPORTANT IS UNIDIRECTIONAL FLUSHING FOR THE DC
DISTRIBUTION SYSTEM? HOW OFTEN SHOULD THEY FLUSH?
Flushing the distribution system has several benefits. Good unidirectional flushing will remove
loose deposits that contain organisms and materials that cause color, may scour the pipe surface,
moves disinfectant residuals into areas that may have levels that are too low, decreases water
age, helps control nitrification, and in the case of the corrosion inhibitor, will ensure that it
reaches all points of the distribution system. In dead ends and low flow areas, flushing, looping,
and bleed-off may ensure that the potential benefits from corrosion control access these problem
sites. Flushing of the DC system, especially in light of the presence of a predominance of
unlined iron pipes, is critical for maintaining water quality. This practice will have the most
benefit in areas where flows are typically low enough that loose deposits accumulate. It would
have been beneficial to have the DC system flushed prior to phosphate addition to minimize any
adverse impacts on scale release, microbial counts, and color. In lieu of this, however, spot
flushing should be used to help remediate sites if adverse effects are seen.
For microbial control, flushing can remove loose deposits that contain bacteria (Barbeau et al.
1999, Gauthier et al. 1997). Typical flushing velocities will do little to remove the tightly
adherent deposits such as tubercles. These deposits also contain organisms that will be protected
from shear, disinfection, and predation. There may be some effect on biofilm bacteria on the
6

-------
surfaces of the deposits, although the effects are probably limited. McMath et al. (1997) tested a
system where biofilms had developed under low flow conditions and then incrementally
increased the flow and collected the bacteria that were detached from the surface. Bacteria were
released from the surface for over 48 hrs, which far exceeds the time a part of the distribution
system is flushed. Donlan and Pipes (1988) showed a linear relationship between flow velocity
and HPC density on coupons placed in a full scale system. It is possible, though, that this effect
was due to greater transport of disinfectant to the surface rather than increase in shear at higher
flow velocities. After flushing, the biofilm can and will regrow within a time scale much shorter
than the flushing interval. Actual recovery times will depend on water quality, the amount of
biofilm remaining after flushing, and the surface upon which the organisms are attached. In
Zurich, Switzerland, the water is unchlorinated and the entire distribution system is flushed twice
a year to keep microbial counts low (Klein and Forster 1998). Regardless, flushing has been
used by many utilities to help improve microbial water quality, and it should be considered as
part of the program in the DC system.
It is difficult to determine the frequency at which the flushing should take place. Some
information suggests that flushing should initially take place frequently to remove loose deposits
that have accumulated over the years (Friedman et al. 2003). Once the system is relatively clean,
it may be possible to increase the time period between flushing. The interval may be determined
by other water quality parameters, including color, ability to maintain a disinfectant residual,
potential for nitrification, etc. Other important parameters are cost, ability to dechlorinate and
discard the flushed water, water availability during drought conditions, personnel, etc.
REFERNECES
Abernathy, C. and A.K. Camper. 1998a. Interactions between Pipe Materials, Organics,
Corrosion Inhibitors, and Disinfectants on Distribution Biofilms. Volume 1. Bench and Pilot-
scale Studies. National Water Research Institute, Fountain Valley, CA
Abernathy, C. and A.K. Camper. 1998b. The effect of phosphorus based corrosion inhibitors
and low disinfectant residuals on distribution biofilms. Proc. Water Qual. Technol. Conf.
AWWA, Denver, CO
Barbeau, B., K. Julienne, V. Gauthier, R. Millette and M. Prevost. 1999. Dead-end flushing of a
distribution system: Short and long term impacts on water qualtiy. Proc. Water Qual. Technol.
Conf. AWWA, Denver, CO
Batte, M., B. Koudjounou, P. Laurent, J. Coallier and M. Prevost. 2000. Does phosphate-based
anticorrosion treatment enhance biofilm resistance to disinfection? Proc. Water Qual. Technol.
Conf. AWWA, Denver, CO
Burr, M., L. Goodrum, J. Buls and A.K. Camper. 2004. Effect of UV light on humics and
biofilm formation in a model drinking water system. Submitted. Water Research.
Butterfield, P.W., A.K. Camper, J.A. Biederman, and A.M. Bargmeyer. 2002a. Minimizing
biofilm in the presence of iron oxides and humic substances. Water Research. 36:3893-3910.
Butterfield, P.W., A.K. Camper, B.D. Ellis and W.L. Jones. 2002b. Chlorination of model
distribution system biofilm: Implications for growth and organic carbon removal. Water
Research . 36:4391-4405.
7

-------
Camper, A.K., 1996. Factors Limiting Microbial Growth in the Distribution System:
Laboratory and Pilot-Scale Experiments. AWWA Research Foundation, Denver, CO
Camper, A.K., K. Brastrup, A. Sandvig, J. Clement, C. Spencer and A.J. Capuzzi. 2003. Impact
of distribution system materials on bacterial regrowth. J. Amer. Water Works Assoc. 95(7): 107-
121.
Cohen, Y.K., C.G. Abernathy and C.P. Hill. 2003. Corrosion control aids residuals. Opflow.
29(1). American Water Works Association, Denver, CO
Davis, J. A.. 1982. Adsorption of natural dissolved organic matter at the oxide / water interface.
Geochimica et Cosmochimica Acta 46:2381-2393.
Delanoue, A., D.M. Holt, C.A. Woodward, S.M. McMath and S.E. Smith. 1997. Effect of pipe
materials on biofilm growth and deposit formation in water distribution systems. Proc. Water
Qual. Technol. Conf. AWWA, Denver, CO
Donlan, R.M. and W.O. Pipes. 1988. Selected drinking water characteristics and attached
microbial population density. J. Amer. Waterworks Assoc. 80(1): 70-76.
Ellis, B., P. Butterfield, W.L. Jones, G.A. McFeters and A.K. Camper. 2000. Effects of carbon
source, carbon concentration, and chlorination on growth related parameters of heterotrophic
biofilm bacteria. Microbial Ecology 38:330-347.
Friedman, M.J., K. Martel, A. Hill, D. Holt, S. Smith, T. Ta, C. Sherwin, D. Hildebrand, P.
Pommerenk, Z. Hiedi and A. Camper. 2003. Establishing Site Specific Flushing Velocities.
AWWA Research Foundation, Denver, CO
Gauthier, V. C. Rosin, L. Mathieu, J.M. Portal, J.C. Block, P. Chaix and D. Gatel. 1997.
Characterization of the loose deposits in drinking water distribution systems. Proc. Water Qual.
Technol. Conf. AWWA, Denver, CO
Gu, B., J. Schmitt, Z. Chen, L. Liang and J.F. McCarthy. 1995. Adsorption and desorption of
different organic matter fractions on iron oxide. Geochimica et Cosmochimica Acta 59:219-229.
Gu, B., T.L. Mehlhorn, L. Liang and J.F. McCarthy. 1996. Competitive adsorption,
displacement, and transport of organic matter on iron oxide: I. Competitive adsorption.
Geochimica et Cosmochimica Acta 60:1943-1950.
Kerr, C.J., K.S. Osborn, G.D. Robson, P.S. Handley. 1999. The relationship between pipe
material and biofilm formation in a laboratory model system. J. Appl. Microbiol. 85:29S-38S.
Kirmeyer, G.J., G. Pierson, J. Clement, A. Sandvig, V. Snoeyink, W. Kriven and A.K. Camper.
2000. Distribution System Water Quality Changes Following Corrosion Control Strategies.
AWWA Research Foundation, Denver, CO
Klein, H, and R. Forster. 1998. Network operation without safety chlorination. Zurich Water
Supply 16(3/4): 165-174.
LeChevallier, M.W. 1997. Disinfection of biofims in drinking water systems. Proc. Water Qual.
Technol. Conf. AWWA, Denver, CO
8

-------
LeChevallier, M.W., C.D. Lowry, R.G. Lee, and D.L. Gibbon. 1993. Examining the
relationship between iron corrosion and the disinfection of biofilm bacteria. J. Amer. Water
Works Assoc. 85(7): 111-123.
LeChevallier, M.W., C.D. Norton, A.K. Camper, P. Morin, B. Ellis, W. Jones, A. Rompre, M.
Prevost, J. Coallier, P. Servais, D. Holt, A. Delanoue, J. Colbourne. 1998. Microbial Impact of
Biological Filtration. Amer. Water Works Assoc. Research Foundation, Denver, CO
LeChevallier, M.W., N.J. Welch, and D.B. Smith. 1996. Full-scale studies of factors related to
coliform regrowth in drinking water. Appl. Environ. Microbiol. 62:2201-2211.
McMath, S.M., A. Delanoue, D.M. Jolt, S. Maier and C.A. Woodward. 1997. Clumps shed
from pipe walls in water distribution systems. Proc. Water Qual. Technol. Conf. AWWA,
Denver, CO
Najm, I.N., L. Boulos, M. LeChevallier, C. Norton, C. Volk, A. Randall, I. Escobar, L. Kiene,
and C. Campos. 2000. Case Studies of the Impacts of Treatment Changes on Biostability in Full
Scale Distribution Systems. AWWA Research Foundation, Denver, CO
Neden, D.G., R.J. Jones, J.R. Smith, G.J. Kirmeyer and G.W. Foust. 1992. Comparing
chlorination and chloramination for controlling bacterial regrowth. J. Amer. Water Works
Assoc. 84:80-88.
Niquette, P., P. Servais and R.Savoir. 2000. Impacts of pipe materials on densities of fixed
bacterial biomass in a drinking water distribution system. Water Resources 34(6): 1952-1958.
Parfitt, R.L., A.R. Fraser and V.C. Farmer. 1977. Adsorption on hydrous oxides III. Fulvic acid
and humic acid on goethite, gibbsite and imogolite. J. Soil Science 28:289-296.
Schreppel, C.K., D.W. Fredericksen and A. A. Geiss. 1997. The positive effects of corrosion
control on lead levels and biofilms. Proc. Water Qual. Technol. Conf. Amer. Water Works
Assoc., Denver, CO.
Tipping, E. 1981. The adsorption of aquatic humic substances by iron oxides. Geochimica et
Cosmochimica Acta 45:191-199.
Tipping, E., C. Woof and D. Cooke. 1981. Iron oxide from a seasonally anoxic lake. Geochimica
et Cosmochimica Acta 45:1411-1419.
Volk, C.A., C.B. Volk and L.A. Kaplan. 1997. Chemical composition of biodegradable
dissolved organic matter in stream water. Limnol. Oceanogr. 42:39-44
9

-------
This page intentionally left blank.

-------