xvEPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington, DC 20460
EPA/625/R-93/001
May 1993
Seminar Publication
Control of Lead and
Copper in Drinking Water
Water
Service
Main
Service
Line
* Bathroom 1
Faucet
Water Meter
' Soldered
Joints
Kitchen
Faucet
/Household!
i Plumbing ,
s«
i^':
Jj|.«*--^;;
Rt"
^Possible)
Sources
,of Lead>
-------
EPA/625/R-93-001
May 1993
Seminar Publication:
Control of Lead and Copper in
Drinking Water
Office of Research and Development
Washington, DC 20460
> Printed on Recycled Paper
-------
Notice
This document has been reviewed in accordance with the U.S. Environmental Protection Agency's peer and
administrative review policies and approved for publication. The mention of trade names, commercial products,
or services does not convey, and should not be interpreted as conveying, official EPA approval, endorsement, or
recommendation.
-------
Acknowledgments
This document is based in part on presentations made at the September 1991 U.S. Environmental Protection
Agency/American Water Works Association National Workshop on Lead and Copper in Drinking Water. The
agenda and names of presenters from the Workshop are listed in Appendix A. Appreciation is expressed to those
individuals for making presentations and providing written materials for the Workshop and for this document,
and to EPA Region 5 and Jon DeBoer of the American Water Works Association for their work in arranging the
Workshop. Appreciation is expressed to the following individuals for contributing their work to this publication:
Chapter Four
Chapter One 1.1
1.2
Chapter Two
Chapter Three 3.1
3.2
3.3
3.4
3.5
4.1
4.2
4.3
4.3
4.4
5.1
5.2
5.3
5.4
5.5
5.5
5.6
5.7
Chapter Five
Jeff Cohen, U.S. Environmental Protection Agency
William Parrish, Maryland Department of the Environment
Vernon Snoeyink, University of Illinois
William Richards, Roy F. Weston, Inc.
Jack Dice, Denver, Colorado Water Department
Douglas Neden, Greater Vancouver Regional District
Jack DeMarco, Cincinnati Water Works
Thomas Bailey, Durham, North Carolina Department of Water Resources
Michelle Frey and Leland Harms, Black & Veatch, Inc.
Chester Neff, Kent Smothers, Mark Brooks, and Mark Warnock, Illinois State Water
Survey
Anne Sandvig and Boris Prokop, Economic and Engineering Services, Inc.
Michael Schock, U.S. Environmental Protection Agency
Steve Reiber, University of North Carolina at Charlotte
Michael Schock, U.S. Environmental Protection Agency
Richard Moser, American Water Works Service Company, Inc.
Jonathan Clement, Black & Veatch, Inc.
Albert Ilges, Champlain, Vermont Water District
John Allen, Chippewa Falls, Wisconsin
William Barry, Ayres Associates
Thomas Sorg, Michael Schock, and Darren Lytle, U.S. Environmental Protection
Agency
Rita Gergely, Iowa Department of Public Health
Appreciation is expressed to the following individuals for providing guidance, review, and/or helpful sug-
gestions for this document: Ronnie Levin, Michael Schock, and Thomas Sorg, all of the U.S. Environmental
Protection Agency.
Dr. James E. Smith, Jr. of EPA's Center for Environmental Research Information managed the preparation
of this document, with assistance from Richard Scharp. Jennifer Helmick and other staff members of Eastern
Research Group, Inc. provided overall editing and document preparation. Jonathan A. Clement of Black & Veatch,
Inc. served as technical editor.
111
-------
Contents
Introduction xiii
Chapter 1 Regulatory Issues 1
1.1 EPA's New National Primary Drinking Water Regulation for Lead and Copper 1
1.1.1 Introduction 1
1.1.2 Tap Water Monitoring for Lead and Copper 1
1.1.3 Monitoring for Water Quality Parameters 1
1.1.4 Corrosion Control Optimization 2
1.1.5 Source Water Treatment 2
1.1.6 Public Education 2
1.1.7 Lead Service Line Replacement 5
1.1.8 Regulatory Schedule 5
1.1.9 Impacts of the Lead and Copper Rule 5
1.2 A Smaller State's Perspective 5
Chapter 2 Corrosion Characteristics of Materials 9
2.1 The Corrosion Cell 9
2.2 Uniform Corrosion and Pitting 10
2.3 Passivation 10
2.4 Galvanic Corrosion 10
2.5 Corrosion Rate vs. Metal Uptake 12
Chapter 3 Monitoring Design and Implementation 13
3.1 Characterizing the System: Baseline Monitoring 13
3.1.1 Introduction 13
3.1.2 Characterizing the Water System 13
3.1.3 The Materials Survey 14
3.1.4 Information Sources 14
3.1.5 Conclusions 15
3.2 Selection of an Analytical Laboratory 15
3.2.1 Introduction 15
3.2.2 The Regulation 15
3.2.3 Decision Time 15
3.2.4 Selection Criteria 17
3.2.5 Conclusions 17
3.3 "At the Tap" Monitoring 17
3.3.1 Materials Surveys and Site Selections 17
3.3.2 Sample Collection 18
3.3.3 Other Water Quality Parameters 18
3.3.4 Case Study One—Greater Vancouver Water District Experience 18
3.4 Monitoring Program Design Using Utility Employees and Customers 18
3.4.1 Introduction 18
3.4.2 Case Study Two—The Cincinnati Water Works System 19
3.5 Integrating Water Testing and Occupancy Certification 22
3.5.1 Case Study Three—Durham, North Carolina 22
Chapter 4 Corrosion Control Assessment 27
4.1 Basics of a Corrosion Control Study 27
4.1.1 Regulatory Requirements 27
-------
4.1.2 Study Components 27
4.1.3 Desktop Evaluations 29
4.1.4 Corrosion Study Organization 36
4.1.5 Demonstration Testing 36
4.1.6 Data Handling and Analysis 40
4.1.7 Secondary Testing Programs 40
4.1.8 Quality Assurance/ Quality Control Programs 40
4.1.9 Example of Selecting Optimal Treatment 41
4.1.10 Example of a Flow-Through Demonstration Testing Program 41
4.2 Design Considerations and Procedures for Coupon Tests 43
4.2.1 Summary of Method 43
4.2.2 Basic Corrosion Measurement Considerations 44
4.2.3 Purchasing and Preparation of Corrosion Specimens 44
4.2.4 Duration Guidelines for Corrosion Studies 45
4.2.5 Processing of Corroded Specimens 47
4.2.6 Chemical Cleaning Procedures 47
4.2.7 Evaluation of Localized Corrosion 48
4.2.8 The Corrosion Rate Calculation 48
4.2.9 Interpretation of the Corrosion Data 48
4.2.10 Summary 50
4.3 Design Considerations for Pipe Loop Testing 50
4.3.1 Introduction 50
4.3.2 Pipe Loop Design and Construction Considerations 50
4.3.3 Pipe Loop Operational Considerations 51
4.3.4 Characteristics of Pipe Loop Data 52
4.3.5 Data Evaluation Considerations 52
4.4 Electrochemical Methodologies for Corrosion Measurement in the Distribution System 52
4.4.1 Polarization Techniques 53
4.4.2 Electrical Resistance and Electrochemical Noise 54
4.4.3 Summary 54
4.5 References 54
Chapter 5 Control Strategies 57
5.1 Overview of Control Strategies for Lead in Drinking Water 57
5.1.1 Chemical Treatment Strategies 57
5.1.2 Selection Criteria 63
5.1.3 Treatment Chemicals 64
5.1.4 Summary 65
5.2 Secondary Effects and Conflicts with Lead Corrosion Control Strategies 66
5.2.1 Carbonate Passivation 66
5.2.2 Corrosion Inhibitors 67
5.2.3 Materials 68
5.2.4 Conclusions 68
5.3 Full-Scale Performance Testing of Sodium Silicate to Control the Corrosion of Lead, Copper,
and Iron: York, Maine 68
5.3.1 Introduction 68
5.3.2 Findings 69
5.3.3 Recommendations 69
5.3.4 Methodology 70
5.3.5 Results and Discussion 71
5.4 Assessing Zinc Orthophosphate vs. pH Adjustment: Champlain, Vermont 74
5.4.1 Introduction 74
5.4.2 Materials and Methods 75
5.4.3 Results 79
5.4.4 Discussion 80
5.4.5 Conclusions 83
5.4.6 Recommendations 83
5.5 Reducing Corrosion Products in Municipal Water Supplies: Chippewa Falls, Wisconsin 83
5.5.1 Background 83
5.5.2 Water System 84
5.5.3 Regulations 84
VI
-------
5.5.4 Hot Water Flushing 84
5.5.5 Aging of Service Pipe 86
5.5.6 Chemical Stabilization 86
5.5.7 Administrative Order 86
5.5.8 Referendum 86
5.5.9 Legal Action .-. 87
5.5.10 Pilot Study 87
5.5.11 Goals of the Pilot Study 87
5.5.12 Implementation of the Pilot Study 89
5.5.13 Decision to Treat 89
5.5.14 Implementation of Central Treatment 89
5.5.15 Facilities Constructed 90
5.5.16 Monitoring 90
5.5.17 Sampling Protocol 91
5.5.18 Feed Rates 93
5.5.19 Operation and Maintenance Costs 93
5.6 Evaluating a Chemical Treatment Program to Reduce Lead in a Building: A Case Study 94
5.7 Iowa's Lead in Schools' Drinking Water Program: More Than Just a Monitoring Program.... 97
5.7.1 Introduction 97
5.7.2 Requirements of the LCCA 97
5.7.3 More Than a Monitoring Program 98
5.7.4 Implementation in Iowa: Monitoring Results 98
5.7.5 Implementation in Iowa: Technical Assistance Program 98
5.7.6 Test Results from Iowa's Program 98
5.7.7 Example of a Solution: Finding a Solution for New Hampton High School 99
5.7.8 General Observations from Investigations 100
5.8 References 100
Appendix A Workshop Agenda: EPA/AWWA National Workshop on Control of Lead and Copper in
Drinking Water 103
Appendix B Units and Conversions 105
vn
-------
Figures
Number Page
1-1 Tap water monitoring (lead and copper) 2
1-2 Monitoring for water quality parameters 2
1-3 Implementation pathways for large public water systems 3
1-4. Implementation pathway for medium-sized and small public water systems 4
1-5 Regulatory schedules for large, medium-sized, and small systems 6
2-2 Typical anodic (a) and cathodic (b) reactions 9
2-1 A diagram of the corrosion cell 9
2-4 Corrosion rate as a function of time 11
2-5 Scale composition on the surface of iron pipe 11
2-3 The oxygen concentration cell (a) and pitting and tuberculation for iron pipe (b) 11
2-6 A micrograph of a cross-section of brass (xlOO) 12
3-1 Tier One sampling site requirements 14
3-3 Tier Three sampling site requirements 14
3-2 Tier Two sampling site requirements 14
3-4 Copper levels from the Greater Vancouver Water District monitoring program 18
3-5 Lead levels from the Greater Vancouver Water District monitoring program 18
3-6 Cincinnati Water Works: Schematic of treatment system for the Ohio River supply (a)
and lime softening treatment system for the ground water supply (b) 19
3-7 Percent of samples failing lead, copper, coliform, and standard plate count tests 23
3-8 Percent of samples failed for lead (a) and copper test (b) 23
3-9 Percent samples failed and passed for copper, coliform, and standard plate count tests 24
3-10 Number of samples exceeding 50 vs. 15 jig/L 25
4-1 Logic diagram for evaluating alternative corrosion control approaches 30
4-2 Suggested corrosion control approaches based on water quality characteristics 32
4-3 Conceptual layout of flow-through testing schemes 37
4-4 Reduction in metal concentrations (a) and coupon weight-loss (b) by alternative
treatments 41
4-5 Corrosion specimen data form 46
4-6 Planned-interval pipe insert exposure during EPA/ISWS corrosion study 49
4-7 Corrosion of galvanized steel specimens at Site 302 49
4-8 Corrosion of copper specimens at Site 307 49
4-9 Effect of water corrosivity on galvanized steel 50
5-1 Stagnation lead levels 58
5-2 Alkalinity/DIC relationships 58
5-4 Lead speciation for 25°C, I = 0.01, DIG = 50 mg/L 60
5-3 Lead speciation for 25°C, ionic strength (I) = 0.01, dissolved inorganic carbonate
(DIG) = 3 mg/L 60
5-5 Lead solubility (I = 0.01, 25°C) 60
5-6 Variation in lead solubility (ph 7.0) as a function of orthophosphate dosage for
different alkalinities 61
5-7 Variation in lead solubility (ph 7.5) as a function of orthophosphate dosage for
different alkalinities 61
5-8 Variation in lead solubility (ph 8.0) as a function of orthophosphate dosage for
different alkalinities 61
5-9 Zinc solubility (I = 0.01) 62
5-10 Zinc solubility (pH 7.5,1 = 0.01, 25°C) 62
5-11 Path of lead response to treatment changes 65
viii
-------
Number Page
5-12a Distribution of HOC1 and OC1" in water as a function of pH 66
5-12b Effects of pH and oxidant dosage on the formation of TOX and THMs (CHC13)
at 20°C in distilled water solutions of 5 mg humic acid/L 67
5-13 Corrosion rate vs. pH, 114-hour laboratory test with aerated tap water 67
5-14 Tricalcium phosphate saturation 68
5-15 Map of the York Water District distribution system 71
5-16 Temperature of the filtration plant finished water (a) and monthly water production (b) 72
5-17 Average monthly silica dosages and raw water silica concentrations 72
5-18 Average pH (a) and alkalinity (b) from the distribution sampling events 72
5-19 Silica concentrations from selected sites within the distribution system (a) and in
first- and second-draw samples (b) 73
5-20 Average lead concentrations in the first-draw samples (a) and the number of samples
exceeding specified concentrations in first-draw samples (b) 73
5-22 Average copper concentrations in the first-draw samples 74
5-21 Average iron concentrations in the first- and second-draw samples 74
5-23 Coupon studies on corrosion rates in four cell units 76
5-24 Comparison of municipal and regional water treatment using the same source waters
(Lake Champlain, Vermont) 77
5-25 Schematic of Champlain Water District water treatment process 78
5-26 Champlain Water District laboratory coupon procedure change (06/01/89) 78
5-27 Well locations, Chippewa Falls, Wisconsin 85
5-28 Pilot test area, Chippewa Falls, Wisconsin 88
5-29 pH, copper, and lead at the 461 A Street copper services during pilot study 89
5-30 pH, iron, and zinc at the 467 Chippewa Street galvanized service 89
5-31 Lead levels in samples collected at 1301 Waldheim Road 91
5-32 Copper levels in samples collected at 1301 Waldheim Road 91
5-33 pH, lead, and copper at 1301 Waldheim Road 92
5-34 pH, lead, and copper at 43-45 Stump Lake Road 92
5-35 Annual operation and maintenance costs for the chemical feed system 93
5-36 Cost of caustic soda per anhydrous ton 93
5-37 Lead (a) and zinc (b) concentrations in samples collected sequentially (Room 3329) 95
5-38 Lead (a) and zinc (b) concentrations in samples collected sequentially (Room 1618) 95
5-39 Water usage study—lead concentrations over time in Room G402 96
5-40 Water usage study—lead concentrations over time in Room 3325 96
5-41 Water usage study—average lead concentrations from the ground floor 96
5-42 Water usage study—average lead concentrations from the third floor 97
5-43 Water usage study—number of samples with less than 50 pg/1 and 15 jig/1
lead from the ground floor 97
5-44 Water usage study—number of samples with less than 50 jo.g/1 and 15 (ig/1
lead from the third floor 97
IX
-------
Tables
Number Page
2-1 Examples of Galvanic Corrosion 10
3-1 Analytical Methods 16
3-2 Water Main Materials 19
3-3 Joint Materials 20
3-4 Service Branch Materials 20
3-5 Lead Levels in First-Draw Samples as Part of Employee Monitoring Program 20
3-6 Outline for Implementation 21
3-7 Lead and Copper List of Monitoring Questions for Ohio EPA 21
3-8 Lead Concentrations in Samples Collected as Part of Durham Lead Survey 22
4-1 Recommended Corrosion Control Study Components for Large PWSs Based
on Lead Levels 28
4-2 Source Water Treatment Guidelines 29
4-3 Schedule of Drinking Water Regulatory Activity: 1990-2000 34
4-4a Constraints Worksheet for pH/Alkalinity or Calcium Adjustment Treatment Alternatives 35
4-4b Constraints Worksheet for Inhibitor Treatment Alternatives 35
4-5 Organization of the Major Components in Corrosion Control Studies 36
4-6 Pipe Volumes by Tubing Length and Diameter 39
4-7 Corrosion Control Treatment Performance Ranking Matrix 42
4-8 Final Corrosion Control Treatment Selection Matrix 42
4-9 Lead Concentrations from Pipe Loop Testing 42
4-10 Skewness Coefficients for Lead Data 43
4-11 Calculated Student's t Values 43
4-12 Suppliers of Corrosion Specimens and Pipe Loops 44
4-13 Typical Coupon and Pipe Loop Costs 45
4-14 Typical Corrosion Rates for Pipe Inserts in Illinois Waters 45
4-15 Density of Selected Metals 48
4-16 Significance of Coupon Weight Loss Measurements 49
5-1 Langelier Index (LI) vs. Calcium Carbonate Precipitation Potential (CCPP) 59
5-2 Average Finished Water Quality Summary 71
5-3 Corrosion Rate Reductions of Laboratory Steel Coupons 84 days exposure time 79
5-4 Corrosion Rate Reductions of Laboratory Lead Coupons 84 days exposure time 79
5-5 Corrosion Rate Reductions for the Distribution System Steel Coupons 79
5-6 Corrosion Rate Reductions for the Distribution System Lead Coupons 79
5-7 Lead Concentrations in Sequential Samples 80
5-8 Average Lead Concentrations at Consumer Taps Cone. \ig/L (number of samples) 82
5-9 Hot Water (140°F) Hushing Results (lead in ng/L) 86
5-10 Construction Costs 90
5-11 Lead Levels in the Samples Collected at 1301 Waldheim Road (|Jg/L) 91
5-12 Lead and Copper Levels in the Samples Collected at 1301 Waldheim Road 91
5-13 Lead and Copper Levels in Samples .Collected at 47 Stump Lake Road 91
5-14 Lead and Copper Levels in Samples Collected at 1100 West River Street 91
5-15 Chemical Feed Rates of Caustic Soda 93
5-16 Annual Operation and Maintenance Costs 93
5-17 Caustic Soda Costs 93
5-18 Water Quality Characteristics 94
5-19 Lead Levels in Samples of Flushed and Static Water from Various Locations 94
5-20 Percentage of Lead in Solder Samples 94
-------
Number Page
5-21 Lead Baseline Data Collected at the Ground Floor and at the Third Floor 96
5-22 Summary of Lead Levels Found by Institutions 99
5-23 Number of Facilities Reporting Lead Levels Above 20 |J.g/L and the Highest
Lead Levels Recorded from those Facilities 99
XI
-------
Introduction
On September 23, 24, and 25, 1991, the U.S. Environ-
mental Protection Agency (EPA) and the American Water
Works Association (AWWA) held a national workshop on con-
trol of lead and copper in drinking water. The objectives of the
workshop were to help participants:
• Become familiar with EPA's new national primary drinking
water regulation for lead and copper and its anticipated im-
pact on utilities.
• Learn investigative requirements and control strategies.
• Learn methodologies for implementation.
• Share field experience.
• Become familiar with laboratory and pilot testing proce-
dures.
The workshop speakers included individuals from EPA,
states, industry, academia, AWWA and the AWWA Research
Foundation, consulting firms, and utilities. More than 300 par-
ticipants heard presentations on regulatory issues, corrosion
characteristics of materials, monitoring design and implemen-
tation, and control strategies (see Appendix A, Workshop
Agenda). Two breakout sessions addressed design considera-
tions and procedures for pipe loop and coupon testing. This
publication is based in part on the information presented at the
1991 national workshop, updated and supplemented with addi-
tional material.
How To Use This Document
Chapter One of this publication discusses regulatory is-
sues, presenting both an overview of the new federal require-
ments and a state perspective on implementing these
requirements. Chapter Two presents information about the cor-
rosion characteristics of materials. Chapter Three discusses
the design and implementation of a corrosion monitoring pro-
gram. Topics include baseline monitoring, selecting an analyti-
cal laboratory, monitoring at the customer's tap, designing a
monitoring program using utility employees and customers, and
integrating water testing and occupancy certification. Chapter
Four focuses on corrosion control assessment, including cou-
pon tests, pipe loop tests, and electrochemical methodologies
for corrosion measurement. Section 4.1, Basics of a Corrosion
Control Study, presents recommendations for states and utilities
for performing and evaluating corrosion control studies. Fi-
nally, corrosion control strategies are addressed in Chapter
Five, which includes an overview of control strategies as well
as secondary effects. Throughout, the document presents the
experience of utilities in monitoring, assessment, and control
strategies.
xui
-------
Chapter 1
Regulatory Issues
On June 7,1991, EPA promulgated maximum contaminant
level goals (MCLGs) and national primary drinking water regu-
lations (NPDWRs) for lead and copper in drinking water. The
MCLG for lead is zero, and the MCLG for copper is 1.3 mil-
ligrams per liter (mg/L). EPA promulgated an NPDWR for lead
and copper consisting of a treatment technique requirement that
includes corrosion control, source water treatment, lead service
line replacement, and public education.
This chapter presents an overview of EPA's new NPDWR
for lead and copper, hi addition, it discusses implementation of
the rule at the state level, from the perspective of the state of
Maryland.
1.1 EPA's New National Primary Drinking Water
Regulation for Lead and Copper
1,1.1 Introduction
EPA's final rule for lead and copper in drinking water (see
Federal Register, June 7, 1991, 56 FR 26460) is part of a
federal effort to reduce lead exposure from all sources. The rule
was one of the most controversial regulations ever proposed by
the Agency, receiving more than 3,000 comments. The final
lead rule was developed through a cooperative effort by EPA's
regulatory staff, the American Water Works Association
(AWWA), and others. This process resulted in a regulation
based on the practical realities faced by water utilities as well
as the need to protect human health.
Although drinking water generally does not contain high
concentrations of lead, it can be a source of lead to which a
large number of people are exposed. In addition, recent scien-
tific evidence shows that children, in particular, suffer adverse
health effects from lower levels of lead exposure than pre-
viously thought harmful. The potential health effects of lead in
children can include impaired mental development, reduced IQ,
shortened attention span, diminished hearing, lowered birth
weight, and altered heme synthesis and vitamin D metabolism.
In adults, the health effects include increased blood pressure.
Because of these health effects, EPA has set an "action
level" for lead in drinking water of 15 Hg/L, measured at the
90th percentile (e.g., if there are 100 samples, no more than 10
may exceed the action level). In contrast, the maximum con-
taminant level (MCL) for lead (which was promulgated as an
interim drinking water regulation in 1975 and was effective
until December 7, 1992) was 50 |0.g/L in samples obtained at
the point of entry into the distribution system. The action level
is not an enforceable standard, but it triggers corrosion control
treatment EPA also has set an action level for copper of 1.3
mg/L (also at the 90th percentile).
The final rule for lead and copper applies to community
and nontransient, noncommunity systems. The rule includes
requirements for tap water monitoring (lead and copper, water
quality parameters), corrosion control optimization, source
water treatment, public education, and replacement of lead
service lines.
1.1.2 Tap Water Monitoring for Lead and Copper
The dates by which tap water monitoring for lead and
copper must begin are shown in Figure l-l(a); the number of
sites required for additional monitoring are shown in Figure
l-l(b). If a system complies with the action levels, the state
may reduce the monitoring requirements, as shown in Figure
l-l(c).
Targeted high-risk homes include those homes with lead
solder installed after 1982, lead pipes, and lead service lines. A
tiered approach, worked out between the system and the state,
should be used to select the sample sites.
The samples should be first flush, 6-hour standing time, 1
liter. The system can furnish bottles to residents and train those
residents to collect the samples. Acid preservative need not be
added until the water sample reaches the laboratory. Acidifica-
tion may be done up to 14 days after the sample is collected.
1.1.3 Monitoring for Water Quality Parameters
Systems serving more than 50,000 people, as well as small
and medium-sized systems that exceed action levels, must moni-
tor for water quality parameters to identify optimal treatment
and to determine compliance. These parameters include pH,
alkalinity, calcium, conductivity, orthophosphate (if used in
treatment), silica (if used in treatment), and water temperature.
Figure l-2(a) shows the number of samples required for
initial monitoring of water quality parameters. Reduced moni-
toring, as shown in Figure l-2(b), may be authorized by the
state. The sampling site locations for water quality parameters
may be different from those for lead and copper, but they should
-------
(a) Start Dates for Monitoring
System Size (Population)
Start Dates
Large Systems (more than 50,000)
Medium-Sized Systems (3,301 to 50,000)
Small Systems (3,300 or fewer)
January
1992
July 1992
July 1993
(b) Initial Monitoring (Samples collected every 6
months.)
System Size (Population)
Number of Sampling Sites
More than 100,000
10,001 to 100,000
3,301 to 10,000
501 to 3,300
101 to 500
100 or fewer
100
60
40
20
10
5
(c) Reduced Monitoring
System Size (Population)
Number of Sampling Sites
More than 100,000
10,001 to 100,000
3,301 to 10,000
501 to 3,300
101 to 500
100 or fewer
50
30
20
10
5
5
Figure 1-1. Tap water monitoring (lead and copper).
be representative taps (e.g., they may be the same as those for
coliform monitoring). In addition, one sample must be collected
at every point of entry to the distribution system.
1.1.4 Corrosion Control Optimization
Optimal corrosion control treatment (required as shown in
Figures 1-3 and 1-4) minimizes lead and copper in drinking
water at the tap while ensuring that the system does not violate
the NPDWRs. The system must identify constraints for differ-
ent treatments and fully document its treatment recommenda-
tion. Elements of corrosion control optimization are:
• Laboratory study. A laboratory study is used to evaluate
alternative treatments (e.g., pH and alkalinity adjustment,
calcium adjustment, and use of corrosion inhibitors).
• Recommendation to the state.
• Treatment installation. After the system makes a recommen-
dation, the state approves the recommendation or designates
an alternative. The system has 24 months to install the treat-
ment and 12 months for follow-up monitoring.
• Follow-up monitoring. Different tests are allowed (e.g., pipe
loops, coupons, partial systems, and analyses based on
analogous systems).
• State-specified operating parameters. These parameters (e.g.,
pH, alkalinity, calcium, orthophosphate, and silica) become
compliance measures.
• Compliance with specified parameters.
(a) Initial Monitoring (Two samples every 6 months.)
System Size (Population) Number of Tap Sampling Sites
More than 100,000
10,001 to 100,000
3,301 to 10,000
501 to 3,300
101 to 500
100 or fewer
25
10
3
2
1
1
(b) Reduced Monitoring
System Size (Population) Number of Tap Sampling Sites
More than 100,000
10,001 to 100,000
3,301 to 10,000
501 to 3,300
101 to 500
100 or fewer
10
7
3
2
1
1
Figure 1-2. Monitoring for water quality parameters.
l.Jf.5 Source Water Treatment
If the tap action level is exceeded, it becomes necessary to
monitor for lead and copper in the source water. If the source
water contains lead or copper, the water system must investi-
gate treatment alternatives. The system makes a recommenda-
tion, and the state either approves the recommendation or
designates an alternative course of action. Treatment alterna-
tives for source water include ion exchange, reverse osmosis,
lime softening, and coagulation/filtration. The system has 24
months to install the treatment system and 12 months for fol-
low-up monitoring. The state designates the maximum permis-
sible lead and copper concentrations for finished water entering
the distribution system.
1.1.6 Public Education
EPA has developed a package of public education materials
that the system must use if action levels are exceeded. This
package provides the minimum materials for public education
as specified in the rule: an introduction, information about
health effects and sources of lead, and steps that can be taken
at home to reduce lead levels in water. The system may add
information to this package.
Public education program delivery must begin within 60
days after the lead action level is exceeded. Program delivery
-------
SAMPLE PLAN AND MATERIAL SURVEY
INITIAL MONITORING
• Pb-Cu Tap
• WQP1 Distribution System
• WQP Entry Points
Meets ALs2
90% TL - POE < POL3
90% TL - POE5 POL3
Existing
Treatment
Optimal
Corrosion Control
Study
Recommend Optimal
Treatment
Corrosion Control Study
State Approves
Existing
Corrosion Control
Treatment=Optimal
Install Treatment
Follow-Up Monitoring
State Specifies
WQP Ranges
Routine Monitoring
Exceeds WQPs Ranges
Recommend Optimal
Treatment
State Approves
Install Treatment
Follow-Up Monitoring
Public Education
andLSLRP"
State Specifies
WQP Ranges
Meets WQPs Ranges
—
Meeh
Routine Monitoring
i WQPs Ranges
Exceeds WQPs Ranges
Reduced
Monitoring
Exceeds ALs
W
Public Education
and LSLRP
1WQP = Water quality parameter
2AL = Action level
*rhe 90th percentile tap water level (TL) minus the highest source water concentration (Point of Entry) Is < or >
the practical quantitation level (PQL) of 5 |ig/L.
T.SLRP = Lead service line replacement
Figure 1-3. Implementation pathways for large public water systems.
Source: U.S. EPA, Lead and Copper Rule Guidance Manual, Volume 2 (1992).
-------
SAMPLE PLAN AND MATERIAL SURVEY
Initial Monitoring
• Pb/Cu Tap
Meets ALs'
Exceeds ALs1
Existing Treatment Is Optimal
J_
90%TL-POEaPQL3 I
Initial Monitoring
• Pb/Cu Entry Points
•WQP7 Entry Points
• WQP Distribution System
90%TL-POE
-------
must include bill staffers, pamphlets to selected groups (such
as pediatricians), notices to major newspapers, and public serv-
ice announcements (PSAs) to local radio and television sta-
tions. These must be delivered every 12 months for as long as
the lead action level is exceeded (with the exception of the PSA,
which must be delivered every 6 months).
LI. 7 Lead Service Line Replacement
If corrosion control and source water treatment do not
work for systems containing lead service lines, and the system
continues to exceed lead action levels, the lead service lines
must be replaced. The rule requires that 7 percent be replaced
each year (over a 15-year period); only those lines that are
under system control, however, must be replaced. Control, as
defined by state statutes, municipal ordinances, or public serv-
ice contracts, is the authority to set standards for construction,
repair, or maintenance; the authority to replace, repair, or main-
tain; or ownership. No replacement is required for an individual
line if the lead concentration in all service line samples from
that line is less than or equal to 15 (ig/L. Monitoring methods
include (1) tapping into the water line, (2) measuring tempera-
ture changes, and (3) determining flush volume between the
end of the line and the tap.
1.1.8 Regulatory Schedule
The regulatory schedules for large, medium-sized, and
small systems are shown in Figure 1-5; the steps that water
systems must take are shown in Figures 1-3 and 1-4. Deadlines
are set for the initial monitoring period, the completion of stud-
ies, state approval, treatment installation, and follow-up moni-
toring.
1.1.9 Impacts of the Lead and Copper Rule
The total capital costs are estimated to be between $2.9
and $7.6 billion; operation and maintenance costs, $240 million
per year; and total annualized costs, between $500 and $790
million. Corrosion control treatment required by the rule is
estimated to cost $1 per household per year for large systems
and $2 to $20 per household per year for smaller systems. Tap
water monitoring will be required for 79,000 community and
nontransient, noncommunity water systems. Monitoring costs
are estimated to be $40 million per year nationwide. Total an-
nualized costs for lead service line replacement are estimated
to be between $80 and $370 million. State implementation costs
are estimated to be $40 million per year.
1.2 A Smaller State's Perspective
This section presents the progress and plans made by the
state of Maryland in preparing for implementation of the lead
and copper rule. It discusses the results of an assessment pre-
pared for upper-level management in Maryland's Department
of the Environment Water Supply Program. This assessment
examines monitoring, treatment, lead service line replacement,
compliance and enforcement, training, and resources.
In Maryland and throughout the United States, EPA's
NPDWRs likely will have a substantial impact on small system
compliance. Training will be essential for educating the indus-
tries, especially the small systems that lack technical staff and
resources.
Maryland has a population of 4.8 million, 80 percent (3.2
million) of whom are served by public water supplies. Approxi-
mately 8 percent of those are served by Washington's Urban
Sanitary Commission and the Baltimore Metropolitan Water
Supply System. Approximately 530 community water supply
systems and 520 nontransient, noncommunity water supply
systems, primarily schools and day care centers, serve the re-
maining 92 percent. About 980 systems serve a population of
fewer than 3,300. The 50 to 60 medium-sized and large systems
are not expected to have problems implementing the rule. The
small systems, however, almost certainly will have problems,
primarily with monitoring and costs, and this is where the
state's resources will be used most.
The assessment looked at the impact that monitoring will
have on water supplies in the state. A survey determined that
44 state-certified laboratories were available for lead and cop-
per analyses; about half of those are out-of-state laboratories.
A question therefore arises concerning whether adequate capac-
ity for the analyses will be available, especially hi the final
phase of the rule, when the systems that serve fewer than 3,300
people are required to monitor. Therefore, the state has made
plans to spread out the workload over time.
The laboratory survey identified representative costs for
lead and copper monitoring, with an upper limit of about $65
per sample. This monitoring included collection by the utility,
transport to the laboratory, analysis, and recording of the re-
sults. Analyses for water quality parameters would at least
double that figure per sample. For larger systems, the total cost
estimated for lead and copper monitoring was between $8,000
and $13,000 per year; for small systems, between $650 and
$1,300 per year. Sample costs, including those for water quality
parameters, increase these costs to approximately $1,300 to
$2,600 per year. Many smaller utilities will incur significant
costs in conducting corrosion control studies. At the time of the
assessment, the estimated cost was between $10,000 and
$15,000, although a study in the District of Columbia's system
cost more than $300,000. It is hoped that most systems will not
incur such expenses, and some of the earlier studies provide
valuable lessons in controlling costs.
Treatment costs probably will not have a significant impact
on larger utilities because many of these utilities have been
practicing corrosion control for many years. For the small sys-
tems, treatment costs probably will include purchasing feeders,
chemical storage tanks, and other related work. Costs for these
systems are approximately $5,000. About 70 percent of the
systems in the state will require treatment for lead and copper,
with an estimated total of $3 million statewide for capital costs.
Operation and maintenance (O&M) costs are another fac-
tor to consider. O&M costs for corrosion control treatment are
approximately $1,500 per year. This cost is not large for a
system serving a population of 3,300, but it is a major expense
for small systems of approximately 15 connections.
-------
Regulatory Schedule for Large Systems
(a)
Initial
Monitoring
!
2 .9
Conduct Studies
1
3 '94
|1
W|
'9
Install Treatment
1
5 '96 '?
Follow-up
Monitoring
i
r '9
Maintain
Optimal
Treatment
8
'9S
Regulatory Schedule for Medium-Sized Systems
Without Studies
Initial
Monitoring
2Sf
&3&
cn
-------
the District of Columbia, and Pennsylvania have spent consid-
erable energy and funds over the past few years trying to clean
up the Chesapeake Bay. Part of that effort is included under the
nutrient control strategy to reduce nitrogen and phosphorus by
40 percent. One initiative implemented to reach that goal is a
ban on phosphate detergents. As a result of that ban, phosphorus
levels in rural wastewaters going to plants in Maryland have
dropped significantly. In addition, controls on most of the
wastewater plants that discharge directly to the bay are very
stringent, with a NPDES phosphorus limit of 0.3 mg/L and
stringent toxicity standards for zinc.
For planning purposes, it was estimated that about 30 per-
cent of Maryland's systems would not comply with the moni-
toring requirement after the first round. Approximately
two-thirds of the systems not hi compliance would require treat-
ment or optimization of existing treatment. It was estimated
that the remaining one-third, or about 230, would require some
kind of enforcement action to ensure compliance with the regu-
lation.
Current enforcement procedures have two levels. One level
involves issuing a public notice of violation and, in many cases,
providing technical assistance to identify the nature of the com-
pliance problem and to reach compliance. This level of enforce-
ment is usually 80 to 90 percent effective in getting systems
back into compliance, but it is very resource-intensive. For
those systems that do not comply after the first step, the stand-
ard process is to issue an administrative order. An estimated
110 administrative orders might have to be issued for the 230
systems. A significant number of these will be referred to the
attorney general's office or to the court. Court action can result
if the system does not comply with the order. These situations
frequently end up in some kind of civil action or appeal, which
becomes a long and convoluted process and is very resource
intensive. This year, a proposal will be submitted to the Mary-
land legislature for a bill that would give the state the authority
to levy administrative penalties against noncompliant systems.
In so doing, the state could avoid bringing these situations to
court. This proposal is not a panacea, but it is one of a number
of tools available to bring systems into compliance. One posi-
tive aspect is the creation of a fund that will use fines for
research, technical assistance, and training.
Another key to implementing the lead and copper rule is
an effective training program for water suppliers, engineers,
and state agencies. Training is especially critical to meeting
monitoring and recording requirements, selecting the optimum
treatment, and safely operating and maintaining treatment sys-
tems. In addition, chemical dosage control can be critical in
controlling corrosion in systems, and proper operation of facili-
ties will be very important.
Operator certification might be affected by the rule. In
Maryland, a system that provides simple chlorination is a Class
1 facility, but a system providing any other treatment such as
corrosion control would increase its classification level to Class
2. Training programs would enable operators to upgrade their
certification. In addition, training in analytical methods for
monitoring lead, copper, and water quality parameters should
be provided. Water quality parameters such as pH and alkalinity
should be monitored at least once a day, especially hi smaller
systems, and more frequently in larger systems. (EPA requires
such monitoring only every 2 weeks.)
Maryland is developing some interesting approaches to
training. The state training center initially was developed under
the Safe Drinking Water Act. This center was set up to provide
water and wastewater training for operators and managers. The
training is conducted by engineers, scientists, and operating
specialists. The center receives local, state, and some federal
funds, and the training is provided at a network of community
colleges across the state. The center offers 25 different courses
per year that are developed with assistance from the Water
Supply Program. Last fall, the center conducted training on the
total colifonn rule in 12 different locations, reaching 250 peo-
ple. The state anticipates that it will need to train as many as
2,000 people regarding the lead and copper rule.
In addition, a number of agencies are setting up state train-
ing coalitions. The agencies involved include EPA, AWWA, the
Association of State Drinking Water Administrators, the Na-
tional Environmental Training Association, the National Rural
Water Association, and the Rural Community Assistance Pro-
grams. These agencies are working together at the national level
to encourage state drinking water programs to coordinate and
direct the available training resources. Maryland is one of two
states where this approach will be tested At the first meeting,
these agencies will identify the types of training needed; then
they will develop a specific plan to perform the training. Train-
ing for the lead and copper rule will receive priority.
The staff of the Water Supply Program and eight large
water utilities are expected to review the rules and require-
ments, the progress each utility has made, and the problems
they have identified. Ideally, issues raised and lessons learned
by larger systems will be applied to smaller systems. All state-
certified laboratories will meet to discuss the rule and to ana-
lyze the required samples.
State staff training is critical. The lead and copper rule
places substantial responsibility on state agency personnel who
review treatment plans, identify optimum corrosion control, and
deal with many other issues. State staffs frequently are short of
personnel and expertise, and training for those who are avail-
able is necessary to counter the deficit.
Finally, the matter of state resources is probably the most
critical issue facing state programs. States will have to develop
some sources of funding, such as fees, operating permits, and
taxes. Maryland requires an additional $1.3 million to imple-
ment all of the federal regulations through the radionuclide rule
and $0.3 million to fund the six and a half positions needed to
implement the lead and copper rule. A proposal will be submit-
ted to the legislature for a water-use tax assessed from water
suppliers ($1 per year per household), state property taxes, or
income taxes. (This also could benefit private well protection
programs.)
-------
For additional information on how to comply with the
technical requirements of the lead and copper rule
see the Lead and Copper Rule Guidance Manual,
Volume 1 (Monitoring) and Volume 2 (Corrosion
Control Treatment) developed by the U.S. Environ-
mental Protection Agency. The manual can be or-
dered from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
(800) 533-NTIS
Ask for PB92112101 (Volume 1) and PB93101533
(Volume 2).
-------
Chapter 2
Corrosion Characteristics of Materials
This chapter presents an overview of the corrosion charac-
teristics of materials. It describes the corrosion cell, uniform
corrosion and pitting, passivation, galvanic corrosion, and cor-
rosion rate vs. metal uptake. Corrosion occurs only when there
is a corrosion cell, consisting of an anode, a cathode, metal to
conduct electrons from the anode to the cathode, and a con-
ducting solution that transports excess positive or negative ions
produced during corrosion. Some corrosion products deterio-
rate water quality, and others react with chemicals in solution
to produce scales on the corroding surface that significantly
reduce the rate of corrosion. Changes in water quality that cause
dissolution might cause periodic high concentrations of corro-
sion products in solution.
In addition, corrosion rates sometimes increase when dis-
similar metals are connected. Examples are copper pipe joined
with solder and brass fittings in contact with galvanized pipe.
A thorough understanding of corrosion-related reactions
will enable water systems to make scientifically valid judg-
ments in order to minimize corrosion problems.
2.1 The Corrosion Cell
Corrosion essentially consists of four components: an an-
ode, a cathode, a conducting solution, and a conducting metal.
The anode is the point at which corrosion takes place and
electrons are released (Figure 2-1). The released electrons travel
through the conducting metal to the cathode. The cathode can
be referred to as an electron acceptor. Once the cathode has
(?) Conducting Solution
A"
3) Cathode
received these electrons, ions move from the cathode through
the conducting solution back to the anode. If the anode is a
different type of metal from the cathode, such as in the case of
pipe and an attached fitting, a gasket between the fitting and
the pipe will prevent the flow of electrons and stop the corro-
sion current Equally important is the conducting solution—in
this case, water. At the anode, positive ions are produced and
at the cathode, negative ions are produced. A flow of positive
ions toward the cathode and a flow of negative ions toward the
anode must exist to maintain corrosion. If water is eliminated
from this cell, corrosion stops because metal ions and anions
can no longer be conducted.
Examples of typical anodic reactions (Figure 2-2) are ele-
mental copper converting to cupric ions (Cu+2), lead converting
to lead ions (Pb+2), and iron converting to ferrous ions (Fe+2).
Once the ionic form of the metal is released into the solution,
it can undergo secondary reactions (Figure 2-2). Under the
appropriate conditions, ferrous iron (Fe+2) can precipitate to
Typical Anodic Reactions (a)
Primary
Pb -» Pb2* + 2e
Secondary
2Fe2+ + 1/2Oz + 4OI-T -» 2FeOOH(s)
3Pb2* + 2OHT++2CO3 2- -» Pb3(OH)2(CO3)2(s)
Typical Cathodlc Reactions (b)
Primary
2e + 2H* -» H2
(2) Conducting Metal
Secondary
OH~ + HCOs ~ -> CO3 2~ + H2O
CO3 2" + Ca2+
Figure 2-1. A diagram of the corrosion cell.
Figure 2-2. Typical anodic (a) and cathodic (b) reactions.
-------
form FeCO3. When iron scales on pipes are analyzed, FeCO3
commonly is found, indicating a deposit or a corrosion scale.
Fe+2 can be oxidized to the ferric ion (Fe+3), which then can
form precipitates such as FeOOH. During these reactions, hy-
droxide ions are consumed; as a result, the pH drops in the
region where these reactions take place. Secondary reactions
with lead can result in the formation of lead precipitate, which
includes hydroxide ions (OH") and carbonate ions (CO3-2)
(lead hydroxycarbonate). The species that deposit, the manner
in which they deposit, and the amount of the deposit are very
important; they affect subsequent corrosion reactions. It is im-
portant to know whether they attach to the pipe, forming an
adhering layer, or whether they become a paniculate and do not
adhere to the pipe.
The factors that determine whether an adhering scale or
particulate is formed are not well understood. Oxygen plays an
important role in cathodic reactions because it accepts elec-
trons. The corrosion rate would be reduced by eliminating oxy-
gen, but oxygen is also a component essential to scaling, which
helps reduce the corrosion rate. When the pH drops below 4.5,
hydrogen ions can accept electrons, but this is unlikely to occur
in most water systems. Disinfectants such as chlorine also can
serve as electron acceptors, sustaining the corrosion reaction.
When oxygen accepts electrons, it also reacts with hydrogen to
form hydroxide ions. These hydroxide ions can convert bicar-
bonate (HCO3-) to carbonate (CO3-2). Calcium carbonate and
ferrous carbonate men can be formed in the presence of car-
bonates. At the point where these reactions take place (e.g., on
the pipe wall), localized high pH can occur, causing metal
carbonates to precipitate. This pH might be significantly differ-
ent than the pH of the water away from the pipe surface (bulk
solution). Since the pH is localized, the high pH values will not
be detected by collecting samples from the bulk solution.
2.2 Uniform Corrosion and Pitting
For a single metal to corrode, there must be an anode and
a cathode and a difference in the electrical potential between
them. The difference in potential must come either from within
the material, perhaps from a difference in the crystalline struc-
ture, in the way the atoms are put together to make the metal,
or in the concentration of the electron acceptor. For corrosion
to occur uniformly, the anode and cathode must be moving
rapidly across the surface of the pipe. Pitting corrosion results
if the anode is fixed, causing metal loss at one point.
A local differential in oxygen concentration can support
corrosion of a metal (Figure 2-3a). Low dissolved-oxygen con-
ditions can prevail under sludge or a suspended solid that has
attached to the surface of the pipe. In the area surrounding the
attached particle, higher concentrations of dissolved .oxygen
will exist. Corrosion taking place at the anode, underneath the
particle, will produce electrons that will be transmitted to the
surrounding area. Corrosion does not occur in the region with
high dissolved oxygen because it functions as the cathode.
Pitting and tuberculation (Figure 2-3b) are particular prob-
lems with iron. The point at which corrosion takes place be-
comes fixed for an extended period of time, resulting hi pitting
corrosion. The electrons produced from this reaction are con-
sumed by the surrounding dissolved oxygen. Ferric ions pro-
duced will react with hydroxide ions or oxygen to form pre-
cipitates that attach to the pipe. As these precipitates attach
themselves to the pipe, a porous tubercle is formed. The result
of this corrosion is the formation of pits and tubercles, giving
the pipe a rough surface.
2.3 Passivation
Passivation involves the development of a layer of material
resistant to corrosion on the surface of the metal. Initially, the
corrosion rate of a fresh bare metal is relatively rapid, but over
time the corrosion rate slows because of the accumulation of
deposits (Figure 2-4). The corrosion rate of lead/tin solder can
be reduced by 90 percent in a period of 2 weeks. The purpose
of changing water chemistry through chemical additions is to
promote the formation of these deposits or scales.
Earlier literature suggested that corrosiveness could be re-
duced by an eggshell-thin layer of calcium carbonate; if it is
not possible to form such a layer, the water is corrosive. The
occurrence of eggshell-thin layers of calcium carbonate in pip-
ing systems is very rare. Data from Hanover, Germany, indicate
that several distinct layers of scale exist, which form over a
period of time (Figure 2-5). The outermost layer of scale, the
layer in contact with water, consists of a mixture of Fe"*"3, Mn*4
(the oxidized forms of iron and manganese), and some calcium
carbonate. Thus, this outer layer consists of a mixture of dif-
ferent compounds and elements. Residing underneath this layer
is a dense, shell-like layer of Fe+3. Beneath these two layers,
the conditions are more reduced (low dissolved-oxygen con-
centration) and the iron is in the Fe+2 or iron solid [Fe(s)] state.
This layer is particularly dense, and inhibits the passing of
different constituents of the corrosion reaction. The presence of
this dense layer might, in part, explain the relatively low cor-
rosion rate associated with iron pipe. In contrast to this dense
film is the formation elsewhere in the pipe of loosely packed,
localized scale produced by microbial action. Iron bacteria de-
rive energy by converting Fe+2 to Fe+3. The iron, released from
the water in the Fe"1"3 state, is arranged in a nonordered, porous
array. As a result, constituents necessary for sustaining corro-
sion can pass through this type of scale. Equivalent data for
copper and lead pipe are lacking, but the possibility of similar
reactions must be recognized.
2.4 Galvanic Corrosion
In galvanic corrosion, two different kinds of metals are hi
contact with each other: the anode, with a higher potential, and
the cathode, with a lower potential (Table 2-1). In this respect,
potential is a measure of a metal's capacity to give electrons:
Table 2-1. Examples of Galvanic Corrosion
Anode (corrosion) Cathode
Galvanized (Zn)
Lead/tin solder
Lead
Zinc
Copper
Copper
Brass
Cast iron
10
-------
Sludge.
High DO
(a) Oxygen Concentration Cell
High DO
Water-
Porous Tubercle
(Fe ppts)
OH'
(b) Pitting and Tuberculation
Figure 2-3. The oxygen concentration cell (a) and pitting and tuberculation for iron pipe (b).
CORROSION
RATE
Water
t
Scale
I
Fe (III), MnOz, CaCOs
Shell-Like Layer
Original Pipe Surface
Graphite, Fe (II), Fe (III)
Cast Iron Pipe
Figure 2-4. Corrosion rate as a function of time.
Figure 2-5. Scale composition on the surface of iron pipe.
11
-------
the higher the potential, the higher the tendency to lose elec-
trons. In a situation where galvanized pipe is attached to copper
pipe, the difference in the potential of these metals causes the
galvanized pipe to serve as anode and the copper to serve as
cathode. Lead has the potential to undergo galvanic corrosion
when it is in contact with brass. Brass fittings can contain up
to 8 percent lead (by weight). Brass is a mixture of copper and
zinc with lead added to make the brass more machinable. The
lead contained in brass is not spread uniformly but in pockets
along the grain boundaries (Figure 2-6). In these circumstances,
the lead (anode) corrodes, discharging its electrons to the adja-
cent brass. This can be one way in which lead is corroded from
brass. Another way in which lead held within the brass can be
corroded is by the action of dezincification of brass. As brass
51%Cu/46%Zn
3% Pb
Figure 2-6. A micrograph of a cross-section of brass (x100).
dezincifies, the underlying pockets of lead can be exposed to
water passing by.
2.5 Corrosion Rate vs. Metal Uptake
As pipe material corrodes, metal will be lost at the anode.
This metal can pass into the bulk solution, leading to water
quality problems. It is possible, however, to find corrosion
without any noticeable impact on water quality, because the
metal released is retained as scale at another point on the sur-
face of the pipe. Scale often is formed by the combination of
iron and oxygen. As the oxygen combines with the iron, the
dissolved oxygen in the vicinity of the pipe is depleted. Reduc-
ing conditions occur with the low dissolved oxygen and, by
accepting electrons, the Fe+3 is converted to Fe"". Low dis-
solved oxygen in water distribution systems can be caused, for
example, by dead ends or heterotrophic bacteria that consume
oxygen. In the absence of dissolved oxygen, another element,
hi this case iron, serves as the electron acceptor. After the Fe+3
is converted to Fe+2, Fe+2 is free to migrate into the bulk
solution because of the low dissolved-oxygen concentration.
Once hi the bulk solution, away from the surface of the pipe,
the dissolved-oxygen concentration is higher, and the Fe+^ is
converted back to Fe+3. Suspended hi the bulk solution, the
Fe+3 reacts with hydroxide ions to form ferric hydroxide, which
causes red water problems.
The reactions described above also can apply to the corro-
sion of copper. As with iron pipe, it is possible for oxygen to
be depleted on the surface of copper pipe and for Cu+2 to be
converted to Cu+1 or Cu(s) by accepting electrons. Studies have
shown that the copper concentration in water sitting motionless
in contact with copper pipe increases and then decreases. The
decrease has been attributed to the formation of a layer of
cuprous oxide, which prevents loss of copper.
12
-------
Chapter 3
Monitoring Design and Implementation
EPA's lead and copper rule contains requirements for tap
water monitoring for lead and copper, monitoring for water
quality parameters, source water monitoring for lead and cop-
per, and studies for evaluating corrosion control treatment.
This chapter describes the part of the rule that pertains to
analysis, reviews the analytical procedures, and discusses the
criteria for selecting a laboratory. It discusses baseline moni-
toring to characterize the system, drawing on the experience of
the District of Columbia. It also presents three case studies
illustrating issues related to sampling and analysis:
• "At the tap" monitoring: Greater Vancouver Water District.
• Monitoring program design using utility employees and cus-
tomers: Cincinnati Water Works.
• Integrating water testing and occupancy certification: Dur-
ham, North Carolina.
3.1 Characterizing the System: Baseline Monitoring
3.1,1 Introduction
The recent lead and copper regulations set requirements
for monitoring lead levels at high-risk residences connected to
community water systems. This baseline monitoring will be
used to determine regulatory compliance for the water system
and also can be used to evaluate the effectiveness of any cor-
rosion control treatment required by the regulations. The meth-
odology for developing the sampling pool is specified as part
of the rule. The appropriate selection of monitoring locations
will be extremely important in helping both water utilities and
regulatory agencies meet the baseline monitoring requirements.
To assist in the sampling pool selection, the regulation also
requires a characterization of the water system by a materials
survey. This characterization benefits utilities beyond providing
the basis for the selection of sampling locations. By identifying
pipe materials within the traditional water distribution system,
in the customers' service lines, and as much as possible, in the
customers' indoor plumbing, the utility can better understand
its own system and the extent of lead materials in the system.
Knowledge of lead and copper materials and their location
helps the utility optimize corrosion control. In addition, if a
water utility falls under the requirement to replace lead service
lines, the materials survey used to characterize its system
should be invaluable in locating these lead service lines for
replacement.
The baseline monitoring program for compliance with the
lead rule is unique, with a completely different philosophy from
the standard practice of trying to obtain samples representative
of the water distribution system. The monitoring program at-
tempts to identify the level of lead exposure for individuals who
drink water when the lead level is likely to be highest (first
draw for water standing in interior plumbing or services for 6
to 10 hours) in residences where the risk for lead sources is
high. In essence, the monitoring program seeks to obtain rep-
resentative samples from a nonrepresentative portion of the
system, the high-risk homes with lead service lines and other
lead sources the location of which probably is not well known.
For many systems, this is not an easy task.
3.1.2 Characterizing the Water System
The lead and copper rule identifies the method for charac-
terizing the water system as a materials survey. The rule states
that the level of effort put into this survey needs to be only
what is necessary to select the sampling pool from which base-
line monitoring is required, as long as the highest category of
sampling pool is achieved. Several categories exist, based on
the availability for sampling of single-family residences that
have interior plumbing with lead solder installed after 1982 or
lead service lines. The highest category of sampling pool con-
sists entirely of single-family homes and is made up of equal
numbers of homes with lead service lines and post-1982 lead
solder. This category is designated as Tier One—Category A
(Figure 3-1).
If a water system cannot obtain 50 percent of its sampling
sites from lead service lines and 50 percent from homes with
post-1982 lead solder, it should try to meet the next highest
category, which consists entirely of single-family residences
with an unequal mix of lead service line sites and post-1982
lead solder sites. Tier One—Category B (Figure 3-1) must stay
as close to the 50 percent/50 percent mix as possible. If a
system can get enough sampling locations from single-family
homes, but only with either all lead service lines or all post-
1982 solder (but not both) the sampling pool is Tier One—Cate-
gory C (Figure 3-1).
A sampling pool becomes Tier Two only if it needs to
include multiple-family residences to obtain enough lead ser-
13
-------
Tier One - Category A
All Single Family Residential (SFR)
Tier One - Category B
Ml Single Family Residential (SFR)
Tier One - Category C
All Single Family Residential (SFR)
Figure 3-1. Tier One sampling site requirements.
vice lines and post-1982 solder locations to meet monitoring
requirements (Figure 3-2). Presumably, the preference for the
even mix that differentiated classes A and B in Tier One would
also apply to Tier Two, but all Tier Two sampling pools are
classified as Category D. Finally, any system that cannot meet
Tier Two is allowed to include homes with lead solder from
before 1982 and will be classified Tier Three—Category E
(Figure 3-3). It is presumed that every system in the United
States will fall into one of these categories.
Tier Two - Category D
MFR/BLD: Multi-Family Residential and Public/Private Buildings
Figure 3-2. Tier Two sampling site requirements.
3.1.3 The Materials Survey
Any water system that cannot obtain a sampling pool of
Tier One—Category A will have to document why it cannot do
so, and will have to document that its sampling pool category
is as high as possible. The materials survey can be used to
Tier Three - Category E
LS-: Plumbing with Lead Solder Before 1983
Figure 3-3. Tier Three sampling site requirements.
provide this documentation. The survey should attempt to de-
termine the location and material of water mains, service lines,
service line connections, and interior plumbing throughout the
utility's distribution system. These should be categorized by
building type to help with sampling site selection. For large
systems, the survey should have been completed before final-
izing site selection and beginning the baseline monitoring, and
should have been submitted along with the first monitoring
results. Systems of medium size should have completed the
materials survey by June 1992. Small systems should complete
the survey by June 1993.
If it is available, additional information can be included in
the survey such as estimates of the age of lead solder in interior
plumbing and a breakdown of portions or numbers of lead
service lines under the control of the water system. (The rule
defines "control" as any one of the following: ownership of
service lines; authority to replace, repair, or maintain service
lines; or authority to set standards for construction, repair, or
maintenance of service lines.) Based on plumbing practices
within a utility's customer service area, the water system might
want to characterize plumbing with solder used before and after
implementation of the lead-solder ban. Or it might prefer to
abide by the rule's assumption that 1982 is a reasonable date
demarcating the use of lead and nonleaded solder.
3.1.4 Information Sources
If a water system needs to go beyond the minimum effort
for a materials survey (meaning that it cannot meet the Tier
One—Category A sampling pool criteria and must document
why), the effort that it must put into the survey could be con-
siderable. For the characterization of materials within the dis-
tribution system, most of the data should be available from the
water utility's own records. These could include permit or tap
files, distribution maps or drawings, maintenance records, me-
ter records, information from senior and retired staff, contract
documents and dates, and water quality data.
The need to determine materials used on customer property
presents a greater challenge. Except for information on cus-
tomer service lines, probably little or no information on resi-
dential materials is available from the water system's own
recordkeeping systems. Thus, numerous external information
sources will need to be researched. These might include plumb-
ing codes, building/plumbing permits, water quality data, dates
of construction, interviews with plumbers and/or building in-
spectors, and community surveys.
14
-------
In larger systems, several information sources usually will
be available from a number of different agencies that can con-
tribute to the materials survey. Many of these sources are com-
puterized data bases. The development of a master data base
can be very useful in accumulating, compiling, and analyzing
these data. Frequently such a data base also will provide addi-
tional advantages and applications unrelated to the materials
survey.
3.1.5 Conclusions
To develop a sampling pool for baseline monitoring, and
to document the type of sampling pool within the various sam-
pling categories, the lead and copper rule requires that water
utilities characterize their distribution system and their custom-
ers' plumbing systems by conducting a materials survey. The
survey need be only of sufficient effort to select a sampling
pool for monitoring. If the water system cannot meet the highest
sampling category (50 percent lead service lines, 50 percent
interior plumbing with post-1982 solder in single-family
homes), then it must extend its materials survey effort to docu-
ment why it could not meet this category and mat its sampling
pool is at the highest category possible. The water system then
would carry out a search of its own and other agency records
to determine the materials of construction of its own and of
customers' piping and plumbing. The schedule for completion
would be to conduct the materials survey prior to the required
dates for carrying out the baseline monitoring sampling.
3.2 Selection of an Analytical Laboratory
3.2.7 Introduction
The part of the lead and copper rule that regulates the
analysis of the parameters contained in the rule is a very small
part of a complex regulation. The results obtained from these
analyses could play a very large role, however, in the ways in
which a utility must respond to the regulation. Selecting a labo-
ratory to conduct the analysis, therefore, becomes very impor-
tant.
The regulated concentration of lead, for example, is being
reduced from 50 \ig/L to 15 (ig/L. It is well known in the
analytical community that the smaller the concentration of an
element to be measured, the larger the chance of missing the
true value. So, as the regulated maximum contaminant level
becomes smaller and more difficult to analyze accurately, it
becomes more important that laboratories provide accurate
compliance monitoring data. Many utilities' responses to por-
tions of this rule will depend on the analytical results obtained
from the monitoring.
3.2.2 The Regulation
Section 141.89 of the lead and copper rule addresses the
methods required by the regulations and presents a list of meth-
ods that may be used to analyze the requisite parameters. There
are three methods and five references to the methods listed for
lead (Table 3-1). The methods are the Atomic Absorption (AA)
Furnace, Inductively Coupled Plasma Mass Spectrometry
(ICPMS), and AA Platform Furnace. There are three EPA ref-
erences and one each for American Society of Testing and
Materials (ASTM) and Standard Methods. For copper, five
methods are listed along with 10 references to the methods.
Methods for copper are the AA Furnace, AA Direct Aspiration,
ICP, ICPMS, and AA Platform Furnace. Five EPA references
are listed along with two ASTM references and three Standard
Methods.
The next paragraphs in the regulation address laboratory
certification. Paragraph 1 states that "analyses under this sec-
tion shall only be conducted by laboratories that have been
certified by EPA or the State." It further says that, to obtain
certification, these laboratories must have analyzed perform-
ance samples containing lead and copper and must meet the
quantitative acceptance limits.
A major portion of this rule is the regulation requiring
monitoring for water quality parameters. Large utilities and the
small and medium-sized utilities that exceed the action levels
must monitor for pH, conductivity, calcium, alkalinity, ortho-
phosphate, silica, and temperature. The regulation specifies the
approved methodology for the analyses for these parameters,
but it does not require laboratory certification for the analyses
and reporting of these data.
3.2.3 Decision Time
A utility must make some important decisions about data
analysis: should it become certified and conduct all the analyses
in-house; should it go to an outside laboratory for lead and
copper testing only and conduct the analyses of water quality
parameters in house; or should it go to an outside laboratory
for all analyses, including field sampling and field analyses for
the water quality parameters?
At the Water Quality Technology Conference in 1988, a
paper was presented on "A Utilities' Perspective of Laboratory
Certification." It was reported that most large utilities preferred
to be certified, for both chemical and microbiological analyses,
and preferred to conduct their own analyses. Most medium-
sized utilities are certified for bacteriological analyses only and
most cannot afford the personnel and equipment needed to be
certified for chemistry parameters. The report also noted that
small utilities depend on outside laboratories to provide their
compliance monitoring data. Since 1988, more medium-sized
utilities, specifically those on the upper end of the "population-
served" scale, have been considering in-house capabilities.
Numerous scenarios show that a medium-sized utility
could enter the analytical field. For example, two utilities in
Colorado, which fit the picture of midsized utilities, are both
certified for bacteriological parameters and want to obtain cer-
tification for organic analyses. They currently have qualified
personnel in charge of their laboratory operations and are pro-
ducing (unofficial) in-house data. Both have adequate space in
which to expand. Both need to purchase atomic absorption
instrumentation, which costs between $30,000 and $60,000, a
high price if dedicated to analyzing lead and copper only. A
utility probably should consider obtaining an instrument that
15
-------
Table 3-1. Analytical Methods
Contaminant
Lead
Copper
PH
Conductivity
Calcium
Alkalinity
Methodology
Atomic absorption; furnace technique
Inductively coupled plasma; mass spectrometry
Atomic absorption; platform furnace technique
Atomic absorption; furnace technique
Atomic absorption; direct aspiration
Inductively coupled plasma
Inductively coupled plasma; mass spectrometry
Atomic absorption, platform furnace
Electrometric
Conductance
EDTA titrimetric
Atomic absorption, direct aspiration
Inductively coupled plasma
Titrimetric
Reference
EPA
239.2
200.8
200.9
220.2
220.1
200.7
200.8
200.9
150.1
150.2
120.1
215.2
215.1
200.7
310.1
(Method Number)
ASTM
D3559-85D
D-1688-90C
D-1688-90A
3120
D1293-84B
D1125-82B
D511-88A
D511-88B
3120
D1067-88B
AWWA
Standard
Method USGS
(SM) Procedure
3113
3113
3111-B
4500-H
2510
3500-Ca-D
3111-B
2320
Orthophosphate,
unfiltered, no digestion
or hydrolysis
Silica
Temperature
Electrometric titration 1-030-85
Colorimetric, automated, ascorbic acid 365.1 4500-P-F
Colorimetric, ascorbic acid, two reagent 365.3 4500-P-F
Colorimetric, ascorbic acid, two reagent 365.2 D515-88A
Colorimetric, phosphomolybdate; 1-1601-85
automated-segmented flow; 1-2601-85
automated discrete 1-2598-85
Ion chromatography 300.0 D4327-88 4110
Colorimetric, molybdate blue; 1-1700-85
automated-segmented flow 1-270-85
Colorimetric 370.1 D859-88
Molybdosilicate 4500-Si-D
Heteropoly blue 4500-Si-E
Automated method for molybdate-reactive silica 4500-Si-F
Inductively coupled plasma 200.7 3120
Thermometric 2550
can, at a minimum, analyze all the inorganic metal MCLs, and
that is capable of both flame and furnace procedures.
Personnel, space, and major instrumentation purchase are
the three main factors to be considered when establishing an
analytical laboratory. Capital outlay and annual O&M costs will
be the major stumbling blocks to obtaining management ap-
proval. On the positive side are the utility's ability to be flexible
and control monitoring and analytical programs, including en-
suring that data and reports are produced in a timely manner.
Over the long term, in-house laboratory capability will pay for
itself. There is no way of telling how the lead and copper rule
and other rules will affect analytical capacity nationally, but it
is strongly recommended that utilities take a comprehensive
look at establishing in-house capability.
Small systems serving a population of fewer than 3,300,
including many nontransient, noncommunity systems, depend
on commercial or outside laboratories to conduct analyses for
lead and copper. Some of these, however, could and should
consider conducting their own analyses for the required water
quality parameters if they exceed the lead and copper limits.
Laboratory certification will not be required for those parame-
ters to be reportable, but specific analytical procedures are re-
quired. EPA has specified the electrometric method as the
approved method for testing pH and the conductance method
as the approved method for testing conductivity (formerly
known as specific conductance) (Table 3-1). Portable field in-
struments are available on the market for both of these analyses.
Titrimetric methods are specified for alkalinity measurement.
A well-trained technician could conduct these analyses and
provide valid, accurate data. In some instances, therefore, con-
ducting these analysis in house will be beneficial.
On the other hand, many small and medium-sized systems
will choose not to enter the analytical laboratory business and
will select outside assistance. The following section presents
16
-------
criteria to help systems select an appropriate analytical labora-
tory.
sample locations, must be developed between the commercial
entity and the utility.
3.2.4 Selection Criteria
The first consideration is whether to choose a state- or
EPA-certified laboratory. Most state health departments and all
regional EPA offices can provide a list of certified laboratories.
It is important to make sure that the laboratory is certified to
analyze all of the parameters desired, especially for lead and
copper.
The instrumentation and methodology must be investi-
gated when a system is selecting a laboratory. Familiarity with
the rule is important because the rule contains information on
the methods that the laboratory must use to analyze the parame-
ters. The laboratories should be asked what instruments and
which of the three approved methods for lead and the five
approved methods for copper they will use to analyze the sam-
ples. For example, an ICP method is approved for copper; an
ICPMS, however, is required for lead. If a laboratory has only
ICP capabilities, it cannot provide valid lead data.
A discussion should be held with the laboratory manager
about detection limits. The various methods have various sen-
sitivities; ICPMS, for example, is more sensitive than AA Fur-
nace. If necessary, a laboratory can provide lower detection
limits than is its usual practice, but such testing might cost more
than usual. Conversely, acceptable detection limits can be re-
ported through the use of less sensitive instrumentation than is
the laboratory's norm and can be less expensive.
Along with discussing detection limits, an agreement
should be reached about the procedures that the laboratory will
use in reporting the quality assurance/quality control (QA/QC)
data. The system must have documentation that QA/QC proce-
dures were carried out and that the sample data are verifiable.
Analysis time is important and must be guaranteed. Section
141.91 of the lead and copper rule reporting requirements states
that utilities must report data to the primacy agencies within 10
days of the end of the monitoring period. Sampling must be
timed so that the analysis can be conducted and a report pre-
pared within the required time frame. Most commercial labo-
ratories can improve analysis time at additional cost.
Prices and costs must be checked, compared, and verified.
Some laboratories have minimum costs, for example $50 for a
single parameter; as already mentioned, lower detection limit
reporting and quicker turnaround times can increase the costs.
Supply and demand will probably play an important role in
future analytical costs.
If necessary, a small utility might desire to contract with
an outside laboratory to conduct the analyses for the water
quality parameters. If this is the case, then further investigation
and discussion must be conducted with the laboratory. Since
certification is not required, laboratory personnel qualifications
must be ascertained. The type of field equipment and method-
ology to be used must be verified A monitoring plan, including
3.2.5 Conclusions
Only time will tell if all of the rules being promulgated,
including the lead and copper rule, will make it difficult to
produce reliable compliance data. For now, utilities can only
do their best under the prevailing conditions to comply with
the rules and submit timely reports to their primacy agencies.
It might be prudent for many utilities to attempt to conduct
in-house analyses. If, however, outside laboratory services are
required, a utility should be selective in hiring this service.
Criteria to be considered include certification status, instrumen-
tation available and methods to be used, the laboratory's
QA/QC program and reportable detection limits, turnaround
time, and costs for analyses.
Concerns have been raised about whether adequate analyti-
cal services will be available to meet the requirements of all
the rules. Supply and demand usually dictate availability and
cost, however, and experience indicates that sufficient services
will be available for utilities to exercise their selection exper-
tise.
3.3 "At the Tap" Monitoring
Another requirement of EPA's lead and copper rule is "at
the tap" monitoring at high-risk locations, which are homes
with newer lead solder, lead pipes, or lead service lines.
3.3.1 Materials Surveys and Site Selections
A materials survey is required to establish areas with high-
risk sites. Information on expected plumbing materials can be
obtained from area plumbing codes, building department files
for plumbing age and materials, utility records for age and
materials of service lines, and water quality data for potentially
corrosive water. The materials survey will identify potential
high-risk sites with site-selection priority as required by the
EPA regulations (lead service lines and lead solder plumbing
for single-family homes). Highest priority site selections will
be single-family home subdivisions constructed after 1982 and
before 1987 and older areas where lead service lines connect
the street water main to the house.
When the highest risk priority areas have been chosen and
sufficient sample sites selected, questionnaires should be sent
to each potential site owner to confirm information such as the
age of plumbing, type of plumbing and joints, type of service
line, plumbing modification, and fixture staining, and to obtain
an agreement to allow sampling. An adequate number of suit-
able sites should be obtained to provide an excess (10 to 20
percent) of the statutory number to allow for attrition through
homeowners' moving, plumbing changes, lack of interest in
further sampling, and incorrect sampling.
17
-------
3.3.2 Sample Collection
The most practical sampling of the first draw of overnight
standing water is performed by the resident rather than a utility
employee. To ensure proper sampling, clear and simple instruc-
tions must be provided to residents who take their own samples.
It must be made clear that it is better to report poor sampling
procedures and repeat them if necessary than to provide a sam-
ple that is not properly taken (e.g., water flushed during the
minimum 6-hour standing period). Lead service lines might
require special sampling procedures as outlined in the EPA
regulations. Monitoring frequency is dependent on system
population and is outlined in the regulations.
3.3.3 Other Water Quality Parameters
All large water systems (serving more than 50,000 people)
and smaller systems exceeding the copper and lead action levels
must carry out monitoring for other water quality parameters,
including pH, alkalinity, calcium, conductivity and temperature,
as well as orthophosphate or silica if such inhibitors are used.
These samples must be taken from the distribution system and
from each water source entering the distribution system (they
can be taken from coliform sampling sites). It might be easiest
to take these samples from home water taps when the copper
and lead monitoring samples are taken.
copper concentrations in the newer home samples (Figures 3-4
and 3-5).
Home Tap Samples, 1-L
Tukey Box Plots, Cu
Cumg/L 2
1
E3
New (22) Old (35)
UBC Study, 1990
Figure 3-4. Copper levels from the Greater Vancouver Water District
monitoring program.
3.3.4 Case Study One—Greater Vancouver Water
District Experience
The Greater Vancouver Water District (GVWD) whole-
sales water to 1.5 million people through 17 municipalities. The
water supply comes from three lake impoundments sited in the
mountains north of the city. The lake watersheds are closed to
the public and are imfiltered sources with chlorination as the
only treatment process. The sources provide very soft, low-pH
water with corrosive characteristics.
The GVWD has undertaken an intensive water quality im-
provement investigation in recent years. Initiatives included pri-
mary disinfection, secondary disinfection, and corrosion control.
As part of the corrosion control investigation, a number of pro-
grams monitoring metals corrosion arid leaching were under-
taken. Plumbing water samples were tested in schools, homes,
apartments, office buildings, and hotel rooms as well as in simu-
lated plumbing systems in a corrosion control pilot plant.
In a 1988 monitoring program, 36 homes in the GVWD
service area were tested for lead and copper. First-draw 1-L sam-
ples were taken; 21 percent exceeded a lead concentration of 20
mg/L, and 52 percent had copper levels exceeding 1.3 mg/L.
A monitoring program of 60 single-family homes and 72
apartment suites was carried out in 1990. It was found that 46
and 50 percent of first-draw 1-L samples in apartments and
homes, respectively, exceeded 1.3 mg/L of copper; 32 and 35
percent of samples in apartments and homes, respectively, ex-
ceeded 15 u.g/L of lead. Tukey box plots of lead and copper
levels in newer (less than about 10 years) and older single-fam-
ily homes in the same study clearly showed higher lead and
Home Tap Samples, 1-L
Tukey Box Plots, Pb
Pbng/L
ou
40
on
oU
20
10
n
—
—
—
_ —
New (22) Old (35)
UBC Study, 1990
Figure 3-5. Lead levels from the Greater Vancouver Water District
monitoring program.
3.4 Monitoring Program Design Using Utility
Employees and Customers
3.4.1 Introduction
During the regulatory activities carried out in the past few
years regarding corrosion by-products (CBPs), the most con-
troversial issues, from the utility viewpoint, have centered
around first-draw samples at the customer's tap. The require-
ment of ensuring compliance with action levels or providing
18
-------
optimal treatment related to CBPs presents a major challenge
to utility managers. One of the major reasons for concern is
that in most instances the major CBP, lead, occurs beyond the
point where utilities have direct control regarding the materials
used, methods of construction, and factors required to collect
appropriate samples that describe the problem. Unlike micro-
biological problems generally caused by inappropriate treat-
ment or distribution cross connections, the major cause of CBPs
lies in the building owners' piping and plumbing fixtures. Utili-
ties do have control, however, over the corrosivity of the water
that comes in contact with the homeowner's plumbing. Realiz-
ing their important role, many utilities collected data in advance
of the January 1992 date for implementing the monitoring
regulations.
(a)
_^
Primary
Sedimentation
(1984)
1
'».
Reservoirs
T° ,
System
3.4.2 Case Study Two—The Cincinnati Water Works
System
The Cincinnati Water Works (CWW) consists of a surface
and ground water treatment plant to provide water to a common
distribution system. The surface water treatment plant (Figure
3-6a) processes water from the Ohio River water by coagula-
tion, settling, and rapid sand filtration. Alum, polymers, and
sometimes ferric sulphate are used for solids removal. Chlorine
is used for disinfection, and fluoride is added for prevention of
tooth decay. The raw water pH of about 7.5 is raised to a
finished water pH of about 8.5 by lime addition. About 88
percent of the distributed water is produced by this surface
water treatment plant, which is located in the southeastern part
of the system. The remaining 12 percent of the distributed water
is ground water processed by a lime softening treatment plant
located at the northwest portion of the distribution system (Fig-
ure 3-6b). Raw water is pumped from 10 wells located along
the bank of the Great Miami River. The conventional lime
softening treatment facilities include primary and secondary
basins and dual media filters. Chlorine is added for disinfection,
and sodium hexametaphosphate is used as a sequestering agent.
Fluoride also is added under state mandate. The raw water pH
of about 7.5 is raised to about 9.5 in the finished water.
(b)
10 Wells
To
Distribution •<-
System
Wastewater Recovery
Figure 3-6. Cincinnati Water Works: Schematic of treatment system
for the Ohio River supply (a) and lime softening treatment
system for the ground water supply (b).
The piping network under CWW's direct control (Table
3-2) comprises a variety of distribution system materials. Iron
pipe constitutes the majority of the water mains, because iron
pipe was used during the largest part of the expansion, with the
most popular size installed being 6 and 8 inches.' In 1975,
ductile iron pipe was installed to replace cast iron pipe.
Prestressed concrete and steel pipe are used for larger diameter
pipe installations; steel pipe is used where special conditions
warrant the added expense. About 8 miles of asbestos-cement
pipe are still in use. Small copper mains were put into service
as a means of minimizing stagnant water quality concerns at
dead-end locations. Prior to 1947, all pipe was unlined. Ce-
ment-lined grey and ductile iron pipe have prevailed as the
largest part of the system since that time.
Tables 3-3 and 3-4 list representative joint and service
branch materials used in the distribution system. Lead and
leadite joints were discontinued for new main use in the late
Table 3-2. Water Main Materials
Type
Grey Iron
Ductile Iron
Concrete
Steel
Transite
Copper
Period of
Major Use
1850-1975
1975-present
1956-present
1920-1953
1940-1952
1975-1985
Miles in
Use
2203
288
104
15
8
4
Predominant Size
6", 8", (10"-60")
8", 12", (16")
24", 36", 48", (54")
(36"), 42", 48"
6", 8"
2"
'English units (inches and miles) are used in this publication to facilitate its use by
the intended audience. Appendix B contains a table for conversion to metric units.
1950s. Rubber gaskets, both mechanical and compression
joints, have been used in new main construction since 1958.
Table 3-4 shows that lead service branches have not been in-
stalled in Cincinnati since 1927. CWW records indicate that
about 31,000 lead service branches are still in active service
among the 212,000 customer taps in the system.
19
-------
Table 3-3. Joint Materials
Water Mains
• Period of Major Use
Lead Joints
Leadite Joints
Rubber Gaskets
1860-1958
1931-1958
1958-present
Table 3-4. Service Branch Materials
Types
Period of Major Use
Lead
Brass
Copper
1837-1927
1923-1927
1927-present
3.4.2.1 The Awakening
A coupon study was performed to evaluate the corrosivity
of the finished water. The past state and federal MCL of 50
M-g/L for lead never posed any problems, primarily because of
the required sampling methods. Water was sampled from the
water distribution system, rather than from water standing in
residential plumbing. In September 1985, EPA performed a
short-term monitoring study of employee homes in the Greater
Cincinnati area. Of the 81 homes monitored, 50 were supplied
from the CWW distribution system. First-draw 125-mL and
1,000-mL samples were collected and analyzed for eight met-
als: lead, copper, cadmium, chromium, zinc, iron, sodium, and
calcium. Thirty-eight of the 50 samples did not show detectable
lead levels in either sample (Table 3-5). Only two of the 50
Table 3-5. Lead Levels in First-Draw Samples as Part of Employee
Monitoring Program
Sample
1-38
39
40
41
42
43
44
45
46
47
48
49
50
Percentile
2-76
78
80
82
84
86
88
90
92
94
96
98
100
1 -Liter Sample
(H9/L)
BDL'
BDL
BDL
BDL
BDL
BDL
7
8
16
21
34
72
94
'Below detection limit.
residences had standing sample results that exceeded the 50
\igfL regulation for flowing water (94 \iglL and 72 Hg/L). Five
of the 50 samples exceeded the current 15 \ig/L action level.
These nontargeted locations, selected at random, would have
had a 90th percentile concentration of 8 Hg/L. The concentra-
tion corresponding to the 92nd percentile was 16 Jig/L. Al-
though CWW was concerned about the few sporadic high lead
levels, there was no sense of urgency in addressing these
"worst-case" results because they were well below the regula-
tions in effect at that time.
Nevertheless, in 1985, CWW started to conduct monitoring
studies to determine the extent of lead contamination. Sampling
taps were installed at a residence served with a lead pipe and
containing plumbing with lead solder. The results from this
initial study indicated that the plumbing presented a larger
problem than the lead service branch, with temperature effects
especially evident The results of this initial study prompted
further investigation into the problems with lead.
In June 1987, a pipe loop was constructed with 50/50 lead
solder to determine the length of time required for lead levels
to stabilize. Lead concentrations in the first-draw samples taken
through mid-1991 typically exceeded 15 p.g/L. The lead levels
in samples collected from service lines and water mains seldom
contained lead but still exceeded 15 |j.g/L concentrations on
occasion.
A one-time sampling of 25 drinking water locations was
conducted within the various CWW facilities: detectable lead
concentrations were discovered at 12 of the sites, and 3 loca-
tions contained lead levels greater than 15 jig/L. Another survey
performed by the Cincinnati Health Department found that 86
of 656 samples from electric water coolers at various sources
had lead concentrations at or above 15 |J.g/L.
The 2-year CWW monitoring program of about a dozen
employee homes resulted in data on first-draw and service line
standing water. None of the locations tested consistently had
lead levels in excess of 15 |J.g/L. This was true even of the three
locations with lead service lines, both the first-draw and service
branch samples. CWW also began a 1-year monitoring program
of a home with a lead service line. Lead concentrations were
consistently detected in the first liter and in samples collected
during each 1-minute interval for 5 minutes after the first-draw
sample. Concentrations appeared to follow seasonal water tem-
perature variations. A number of lead service branches might
be added to the study in the area being monitored.
Other random samplings of routine bacterial sample loca-
tions and storage tanks showed sporadic lead levels. Most re-
cently, a program has been developed to collect and analyze tap
water samples before and after replacement of city-owned por-
tions of leaking lead service branches. It appears that replacing
a portion of a lead service branch will improve the quality of
water at the consumer's tap. No efforts have been made to
control the standing time before sampling. A more structured
study might be attempted at a later time.
The results of the studies have demonstrated that elevated
concentrations of lead are. present in water that has remained
motionless while in contact with residential plumbing and dis-
tribution piping. Thus, it is important for CWW to initiate the
structured monitoring required by U.S. EPA and the Ohio EPA
for compliance with the lead and copper rule. Table 3-6 is a
simple outline of the plan for implementation. CWW's review
of draft rules and the final rule resulted in a series of questions
from the utility (Table 3-7). The Ohio EPA answered these
questions as of September 1991 and CWW proceeded with its
overall plan. The first phase of work for the plan consisted of
establishing representative Tier 1 locations, efficiently solicit-
20
-------
Table 3-6. Outline for Implementation
1 • Review Final Rule for Monitoring Implementation Plan
2. Obtain Federal and State Answers to Questions, Pose Own
Answers Based on Rule Review
3. Establish Representative Sample Locations for Tier 1
4. Prepare Monitoring Plan Packet for Ohio EPA Approval
5. Solicit Volunteers from Questionnaire
6. Screen Volunteers and Resolicit as Needed
7. Train Dispatchers, Valvemen, and Homeowners
8. Perform Materials Evaluation
9. Begin Monitoring After Ohio EPA Approval Is Received
10. Collect All Locations Within One Month and Repeat in Six
Months
Table 3-7. Lead and Copper List of Monitoring Questions for Ohio EPA
1. Are the Bolton & California plants two different systems?
2. How do we determine the number of people served by each
plant?
3. Is an estimate of population served, based on pumpage
acceptable? i.e.,
Bolton service
area population 96,360 based on 12% total CWW pumpage
WTP service
area population 706,640 based on 88% total CWW pumpage
Total CWW service
area population 803,000
4. What are OEPAAJSEPA criteria for selection of targeted sites?
5. Will OEPA allow CWW employees to sample their own
residences?
6. Are commercial sites considered single-family structures?
7. Can monitoring be spread over 6 months or must it be done all
at the same time?
If at 6 months frequency, must the repeat samples be precisely
6 months apart?
8. How should the materials survey be conducted?
9. What is the purpose of each of the two samples at each of the
25 sites and the distribution system entry points?
10. Do WQ parameter samples have to be collected at official
bacteriological sites?
11. Will OEPA accept homeowner sampling?
12. How can one guarantee 6-hour static time prior to first draw
sampling?
13. Do we need to survey for water-using appliances or leaky
plumbing?
14. How do we guarantee that the solder in the 1962 and newer
sampling sites is 50/50?
15. Will sites need to be approved prior to sampling?
ing and screening volunteers from questionnaires, and training
samplers.
Since CWW has a separate Public Water Supply Identifi-
cation (PWSID) number for each of its water treatment plants,
the total number of sites would be 160. The U.S. EPA appears
to consider such situations to be one distribution system, but
the Ohio EPA considers the CWW to be two separate distribu-
tion systems. CWW shows a fair distribution of copper
branches, but the lead service branches only occur in clusters
in the northern half of the distribution system and hardly at all
in the northwestern part of the system, which is supplied by the
Bolton plant. If there are two distribution systems, how does
CWW delineate the two? Are there any additional requirements
in the mixing zone (wherever it might be on a given day)?
These questions had to be resolved to the best of CWW's
ability.
CWW determined that it would need to collect 100 sam-
ples from its California Ohio River treatment plant service area
and 60 samples from its Bolton service area. These numbers
were based on the population served by each system. Each of
the systems includes lead service lines and copper service
branches that were installed after 1983. Therefore, CWW as-
sumed that Tier 1 sampling was required and that half of the
samples in each system had to be lead and the other half fairly
new copper installations with 50/50 solder. Obtaining a repre-
sentative sampling of lead services in the Bolton system would
be possible only in a cluster area. The other requirement of pipe
installed after 1982, however, would easily yield a group of
sites scattered evenly throughout the distribution system. Cur-
rent and former CWW employees, as well as employees of U.S.
EPA's Drinking Water Research Division and the Ohio EPA
who reside in the area served by the distribution system, would
be asked to perform sampling, as long as representative sam-
pling could be achieved. This approach could provide the most
credible set of samples possible, given the knowledge base of
these potential sample-location homeowners. Private citizens
who wished to participate would not be excluded if they could
meet the requirements. Obviously, locations with automatic ice-
makers, humidifiers, and leaks would not provide adequate
samples without precautions. A questionnaire identified poten-
tial problem areas for follow-up discussions. All of the first
6-month monitoring was planned for January or February 1992
and samples would be analyzed the following month, thus es-
tablishing the cold weather conditions; a repeat 6 months later
would establish the warm weather conditions. Also, this method
would provide a finished program in time to evaluate any nec-
essary follow-up prior to January 1993.
CWW put its lead and copper program together using ex-
perts from each of the pertinent divisions involved with water
distribution. The Water Quality and Research Division has re-
sponsibility for adding the proper chemicals and ensuring op-
timum treatment and the distribution of quality water to the
customer. The Distribution Division has responsibility for en-
suring that water pipes are properly selected and laid to deliver
potable water with proper pressure. These representatives know
precisely where the water from each plant goes and the location
of various types of mains and service branches. The Commer-
cial Division determined when various materials were installed
and provided target lists for representative sampling. The En-
gineering Division has design and contracting responsibility for
pipe installed in the system. This team provided a CWW re-
sponse to a monitoring program that appears to satisfy the intent
of the federal law. From the studies that CWW has conducted,
it is apparent that lead can be present when water is allowed to
stand in contact with plumbing and piping materials for ex-
tended periods of time. The challenge now is to understand the
magnitude of the problem in CWW's distribution system and
implement a program that will minimize the presence of harm-
21
-------
fill corrosion by-products while keeping other health-threaten-
ing constituents under control.
3.5 Integrating Water Testing and Occupancy
Certification
One way to ensure high-quality water at the consumer's
tap is to integrate water testing with occupancy certification for
facilities. The experience of Durham, North Carolina, demon-
strates the usefulness of such a program in ensuring that drink-
ing water meets standards for lead as well as other parameters.
3.5.1 Case Study Three—Durham, North Carolina
The City of Durham, North Carolina, has a new facility
water testing program that has improved the quality of water
at the customer's tap. The water is tested for standing and
running lead, standing and running copper, and heterotrophic
plate count (HPC) and coliform bacteria. The water at the fa-
cility must meet minimum standards before an occupancy per-
mit is allowed.
3.5.1.1 Background for Developing the Program
In 1985, a survey was conducted in Durham to determine
the presence of elevated lead levels. Sampling was conducted
at 582 buildings and elevated lead levels were discovered, es-
pecially in samples collected from new facilities. Lead levels
in excess of 15,000 \ig/L were observed in a few unoccupied
new homes in which water had been standing in the line for an
undetermined period of time (Table 3-8). Sixty-two of the 582
Table 3-8. Lead Concentrations in Samples Collected as Part of Dur-
ham Lead Survey
Lead Concentration,
H9/L
Location
Sample Date Standing Running
#34 Clearwater Place 09/16/85 17,000 20
#34 Clearwater Place 09/18/85 76 10
(1st Resample)
#34 Clearwater Place 09/23/85 10 <10
(2nd Resample)
3414 Shady Creek Dr. 09/16/85 11,000 20
3414 Shady Creek Dr. 09/18/85 950 10
(1st Resample)
3414 Shady Creek Dr. 09/23/85 20 <10
(2nd Resample)
samples exceeded 50 (Xg/L (the city lead limit prior to August
1991). All 62 sites in violation were less than 2 years old. Even
new facilities soldered with 95-5, tin/antimony solder were
found to be in violation of the 50 |ig/L lead standard (due to
lead impurity hi the solder and lead in fixtures).
Of the 62 locations that exceeded 50 |J.g/L lead in the
standing water sample in 1985, 58 were resampled in January
1988. No standing sample exceeded 50 jig/L lead and only 2
exceeded 20 |ig/L of lead. Only 8 standing samples exceeded
5 ng/L and no running sample exceeded 5 (ig/L.
To further demonstrate that Durham's lead problem existed
primarily with new facilities, approximately 100 new facilities
were sampled in cooperation with the Inspections Department.
Using the standards finally adopted into the program, more than
30 percent of the facilities failed one of the three parameters
tested Gead, copper, and bacteria). In addition to lead, bacteria
and copper were found to be major contaminants of these new
facilities during this survey.
3.5.1.2 Implementation of the Program
Since it was demonstrated that Durham had a problem with
lead, copper, and bacteriological contamination hi new facul-
ties, the new facility water testing program was developed and
presented to the City Council for approval. The City Council
approved the program effective July 1,1987. The program was
initiated in June 1988. The implementation of this program was
slow because of the coordination needed among various city
departments. The city water distribution system also serves
areas of Durham County beyond the city limits. Therefore, both
city and county Plumbing Inspections Divisions had to be in-
volved for sample collection. The Engineering Department,
which controls die distribution and collection system, was in-
volved whenever flushing of the distribution system was
needed to improve water quality. In addition, sampling and
testing procedures had to be established and local organizations
representing real estate agents, building contractors, and plumb-
ing contractors had to be notified about the new procedures. As
a result of making these contacts, the implementation went
relatively smoothly. There were some problems with real estate
agents and owners, especially when "closing" deadline dates
were being postponed by test failure. But with extra efforts by
all parties involved, most of these problems were resolved.
3.5.1.3 Sampling and Analysis
A 100-mL sample is taken for standing and running lead
and copper. The standing sample is taken after a minimum of
8 hours standing time. The running metal and bacteriological
samples are taken after running the water for at least 2 minutes.
The samples are collected by the plumbing inspectors during
final inspection. If resamples are required, they are taken by
Water Resources personnel.
The standards established by the city are:
1. Lead: standing and running 15 fig/L (August 1, 1991)
2. Copper: standing and running 1.3 mg/L(August 1,1991)
3. Heterotrophic Plate Count Bacteria: 100 colonies/mL
4. Coliform Bacteria: 0 colonies/100 mL
If these standards are not met, the occupancy permit is
withheld.
This testing program requires access to a free-flowing out-
let for sample collection with no question of access authority.
There have been no problems with authority to gain access to
free-flowing outlets with other programs, such as cross-connec-
tion control, and Durham definitely has control over facilities
that have not been approved for connection to the water system.
Once facilities, especially private homes, are occupied, it is
22
-------
much more diffinilt to arrangp, fnr sampling Standing samples bacteria tasting, and S?fl, nr 15 percent, failed the, standard.plate
are almost impossible to obtain.
The city maintains a certified laboratory equipped with
three atomic absorption (AA) spectrophotometers; two are
equipped with a graphite furnace, used for lead analysis. Al-
though it is possible to analyze lead at a lower level, 5 |ig/L
has been established as the analytical detection limit. Copper
is analyzed with standard flame AA.
For the standard plate count bacteriological test, a 24-hour
incubation period is used instead of the 48 hours usually used by
the city. This shorter time is used to expedite the testing procedure.
The coliform analysis is by membrane filter procedure.
Since the occupancy permit is withheld until the water
meets the water quality criteria, it is important to complete the
analyses as soon as possible. Samples are received from the
Inspections Department at about 5:00 p.m. Bacteriological
analyses are Initiated immediately, and lead and copper are
analyzed within 24 hours. This system produces final results in
less than 24 hours.
If the facility fails any of the parameters, the company that
requested the test, usually the builder, is notified and requested
to flush the system thoroughly. Each outlet, hot and cold, is to
be flushed for a minimum of 30 minutes at maximum velocity.
The facility is resampled 24 hours after flushing is completed.
The resample is analyzed on the day it is collected. If the
sample fails the resample, further investigation is made to de-
termine the reason for failure before reflushing and resampling.
The program is partially financed by a $40 fee collected
with the meter fee. There is a $30 resample fee for the first
resample. Any additional resamples are without charge.
3.5.1.4 Results of the Program
This program has improved the water quality at residential
plumbing taps. From June 1988 through December 1990,4,826
facilities were sampled and tested (Figure 3-7). Some 1,521, or
27 percent, failed one or more of the three tested parameters.
Without this program, the occupants of 1,500 facilities would
have consumed water that failed to meet the city standards. If
five people in each facility consumed the water, then 3.8 per-
cent of the 130,000 people served by the water system would
have consumed water that failed to meet water standards. In
the same period, 297 facilities, or 5 percent, failed coliform
New Facilities Sampled
Percent Failed
CuSUn4ng(7.3%^ ^Cu Rumlng (2.1%)
fb Running (O.I%K
PbStmBng(4.1%)-
ToUIPI««(H.»%)
Conform (Six)
PMMd(M.O%)
count for bacteria. Also, 234, or 4 percent, failed the standing
lead standard of 50 Hg/L. Even 47, or 1 percent, would have
failed the running lead test. Without the program, the occupants
of 412 facilities would have consumed water with a copper
content in excess of 1 mg/L (the city copper limit prior to
August 1991). Some copper levels were found in excess of 100
mg/L.
Failures to meet the standing lead standard decreased from
15 percent, when the program started, to less than 1 percent in
December 1990. The failures because of standing copper have
decreased from 11 percent to 5 percent from June 1988 to
December 1990. These results probably are because of better
workmanship by the builders (Figure 3-8).
1880
Jun S«p 0*c
0 STANDING
"RUNNING
Sap OK
0 STAN DING
» RUNNING
Figure 3-7. Percent of samples failing lead, copper, coliform, and
standard plate count tests.
Figure 3-8. Percent of samples failed for lead (a) and copper test (b).
Figure 3-9 illustrates that bacteria and copper contamina-
tion are also water quality problems for new facilities. It is
important to correct for these parameters as well as for lead.
No new facility testing program is responsible for meeting
any water standard. The testing program only indicates whether
23
-------
The practice of grounding electrical systems, both AC and
(a) Percent Copper Standing
Failed and Passed
Cu Standing (7.3%)
(b) Percent Copper Running
Failed and Passed
Cu Running (2.1%)
F>»Md{«.OV
(c) Percent Conform
Failed and Passed
PMMd (M.0%)
(d) Percent Standard Plate Count
Failed and Passed
TOUI Plate (14U%)
PMMd(W.O%)
P«>Md(M.O%)
Figure 3-9. Percent samples failed and passed for copper, coliform,
and standard plate count tests.
standards are being met. Action to correct the water problem
must accompany the testing program. No system with corrosive
water such as Durham's should expect to meet a lead or copper
limit without a correctly applied corrosion control program. The
City of Durham has used zinc-orthophosphate for corrosion
control since 1976. This phosphate-based compound was tested
extensively from 1974 through 1976 and found to be very
effective in drinking water for controlling both iron and copper
corrosion. Although lead corrosion was not tested, the com-
pound also has been proven effective for lead control in other
systems. Some products tested, such as metaphosphates, a form
of polyphosphate, increased copper corrosion and thus possibly
lead corrosion.
A flushing program also is essential to ensure high-quality
water within the distribution system. This program must in-
clude flushing all water lines on a regular basis. All new facili-
ties must be adequately flushed. No water line should be
constructed without a way to flush the line. Hydrants or blow-
offs must be installed on the ends of all lines. A sampling
program is of limited value without mechanisms in place for
correcting potential problems.
A cross-connection control program is essential if water
quality standards are to be met at the residential taps. No new
facility sampling program is adequate without a cross-connec-
tion control program.
A new facility sampling program is valuable in policing
the illegal use of lead solder, but an education and training
program to forestall the use of lead in water systems is probably
even more effective. A new facility sampling program also will
identify high levels of copper. Good workmanship and proper
use of solder flux will help prevent high levels of copper in
drinking water.
DC, to the water system should cease. Although it was not
identified as a problem in Durham, electrical grounding has
been implicated in causing copper corrosion. If lead corrosion
is controlled by isolation of the copper cell in a dissimilar metal
cell consisting of lead and copper, then electrical grounding
also could cause lead corrosion.
3.5.1.5 Acceptance of the Program
Reaction to and acceptance of the program was varied. The
Inspections Department's initial reactions were all negative.
Such comments as the following were common:
• There is no way we can handle the extra workload.
• The program is just too much trouble. Our people are not
trained to collect samples.
• The delay in occupancy will make the program unworkable.
• The public will never stand for the delay.
Although the Water Resources Department accepted this
program, the overtime requirements caused difficulty. As a re-
sult, the requirements have been reduced by allowing the labo-
ratory staff to work "flex time."
Appreciation and support for the program, as well as com-
plaints, have been received from builders and contractors. A
frequent complaint is "no one else is doing it." Many have
complained about the extra expense. "I can't close and will miss
the sale" is probably the most frequent comment in opposition
to the program from the builders and contractors. A few cases
of illegal use of lead solder were discovered. Of course, build-
ers objected to having to re-plumb the facility. Individuals are
now aware that inspections are being performed, and therefore
little if any lead solder presently is being used. When problems
that are not directly attributable to the builder—usually bacteria
in Durham's water system—cause the facility to fail, extensive
complaints result. On the other hand, many builders have rec-
ognized the value of the program for assuring their customers
that the water meets quality standards, especially in regard to
lead.
Most opposition from consumers has been because of a
delayed move. A few commercial establishments have had to
delay opening after widely advertising an opening date. Al-
though some individuals concluded that the problem must be
with the water system, many have expressed support for the
program. Lead in drinking water has received significant atten-
tion in the Durham area. It has helped to have facts to share
with the public and to have a positive program to deal with the
lead contamination problem.
The State Division of Health Service Water Supply Branch,
which has primacy in North Carolina, has been complimentary
but noncommittal about the new facility water testing program.
Other water systems representatives have commented that the
Durham testing program is making matters difficult for them.
An investigator from the University of North Carolina who
24
-------
a survey hy actual 1p.ad analysis of water throughout
North Carolina has stated that Durham has the least serious
problem with lead in the entire state. Of the 120 homes with
copper plumbing tested, only 4, or 3.3 percent, exceeded a
first-draw lead level of 15 |ig/L. The highest lead level found
was only 31 |lg/L, probably as a result of the new facility
sampling program and the corrosion control program.
3.5.1.6 Effect of Lowering the Lead Standard
The lead limit has been lowered to 15 ng/L in drinking
water. Although the lower limit results in a larger number of
failures, additional flushing by the builder still meets the lower
lead standard (Figure 3-10).
3.5.1.7 Summary
The new facility sampling program has resulted in im-
proved water quality at the consumer's tap. The program has
proven to be economical and without jurisdictional problems
concerning the purveyor's authority on private property. The
new facility sampling program is recommended to all water
purveyors.
o FAILED 50 |ig/L
• FAILED IS |»g/L
Figure 3-10. Number of samples exceeding 50 vs. 15 (ig/L.
25
-------
Chapter 4
Corrosion Control Assessment
Large public water systems (PWSs) will be required to
conduct corrosion control studies, and medium-sized and small
systems might need to conduct studies if required by the state.
This chapter provides guidance on how utilities should conduct
corrosion control studies to meet the requirements of the lead
and copper rule.
Several methodologies, including coupon tests, electro-
chemical testing devices, and pipe loops, can be used to assess
the effectiveness of various corrosion control strategies. Cou-
pon tests are based on weight loss measurements. Electro-
chemical measurements use devices that sense the flow of
electrons, providing a direct measurement of corrosion. Of par-
ticular interest to utilities are pipe loop systems that simulate
residential plumbing systems and whose key measurements
consist of metal levels. This chapter provides an overview of
each of these corrosion control assessment methodologies.
4.1 Basics of a Corrosion Control Study
The lead and copper rule requires corrosion control studies
to be performed by large PWSs and those small and medium-
sized PWSs required to do so by the state because they exceed
the lead or copper action level (AL). The lead and copper rule
defines certain conditions that must be met by these studies,
but it does not specify (1) the investigative components neces-
sary to accomplish the study, (2) the testing protocols to be
used, (3) the procedures for evaluating data, or (4) the basis for
identifying "optimal" corrosion control treatment This section
discusses these issues and provides recommendations for states
and utilities for performing and evaluating corrosion control
studies. It also presents examples of corrosion control studies
to illustrate alternative approaches and rationales used in the
design, implementation, and interpretation of findings gener-
ated by these studies.
4.1.1 Regulatory Requirements
The lead and copper rule (141.82(c), 56 FR 26550) speci-
fies six conditions that must be met when performing a corro-
sion control study:
(1) Evaluate the effectiveness of each of the following
treatments and, if appropriate, any combinations of
these approaches:
(a) pH/alkalinity adjustment (carbonate system passi-
vation)
(b) Calcium hardness adjustment (calcium carbonate
precipitation)
(c) Phosphate- or silicate-based inhibitors (phosphate
or silicate passivation)
(2) Protocols should include the use of pipe rig/loop tests,
metal coupon tests, partial-system tests (full-scale), o'r
analyses based on documented analogous treatments
with other systems of similar size, water chemistry, and
distribution system configuration.
(3) Analytes are to include the following water quality
parameters in the course of testing: lead, copper, pH,
alkalinity, calcium, conductivity, water temperature,
and orthophosphate or silicate when an inhibitor con-
taining the respective compound is used.
(4) Constraints (chemical or physical) that can limit the
application of a particular treatment option are to be
identified and the existence of one of the following
conditions should be documented:
(a) A particular corrosion control treatment has ad-
versely affected other water treatment processes
when used by another PWS with comparable water
quality characteristics.
(b) From the experience of the PWS, a particular cor-
rosion control treatment has been demonstrated to
be ineffective and/or to adversely affect other
water treatment processes.
(5) Secondary impacts due to the effect of corrosion con-
trol treatment on other water treatment processes are to
be evaluated.
(6) Recommendation of the optimal corrosion control
treatment, as identified by the PWS based on an analy-
sis of the data generated, is to be provided to the state
with supporting documentation and rationale.
While these elements present important pieces of a corro-
sion control study, they do not clearly delineate how to organize
and execute a study.
4.1.2 Study Components
Three major elements are available to PWSs in defining
optimal treatment through a corrosion control study:
27
-------
Desktop evaluations to Determine the range of treatment nr rhrcy, trc,atmf;nt options) for additional evaluation via dem-
altematives.
• Demonstration testing to define the performance of alter-
native corrosion control treatment approaches.
• Source water evaluations to assess whether removal of lead
and copper is necessary through the treatment facilities prior
to distribution.
The full scope of corrosion studies will vary from system
to system, and the methods and procedures used to reach a
recommendation necessarily will reflect this level of site speci-
ficity. During the state review of these studies, the following
criteria can provide a framework for evaluating PWS findings
and recommendations for optimal treatment:
• Reasonableness of the study design and findings.
• Technical integrity of the data handling and analysis proce-
dures.
• Best professional judgment of the state regarding the deci-
sion-making criteria used by the PWS in determining the
recommended optimal corrosion control treatment.
The following sections describe the scope of the testing
and evaluations that PWSs might be required to perform.
4.1.2.1 Scope of Corrosion Control Testing Activities
By requiring all systems conducting studies to evaluate
specific treatment alternatives, EPA did not intend for all PWSs
to construct pipe rigs or conduct bench-scale tests to accom-
modate any and all treatment options. It is anticipated that
desktop evaluations will be used as a preliminary step in the
study. Alternatives are to be screened on the basis of the avail-
able findings from: (1) other corrosion control studies for sys-
tems with comparable water quality, (2) theoretical and applied
research efforts, and (3) the potential adverse impacts associ-
ated with treatment modifications. As a result of this desktop
evaluation, primary alternatives are to be selected (at most, two
onstration testing.
Beyond the desktop evaluation, the specific components,
or steps, included in performing corrosion control studies de-
pend in part on the extent of testing required. EPA believes that,
in certain cases, the results of the desktop evaluation would
suffice in the selection of optimal treatment and additional test-
ing would not be required.
Small and medium-sized systems must recommend opti-
mal corrosion control treatment to the state within 6 months of
exceeding an AL. EPA envisioned the use of a desktop evalu-
ation to be a sufficient level of effort for these systems to
identify optimal treatment. The state retains the discretion to
require additional testing should the supporting documentation
and rationale provide insufficient justification.
Some large PWSs might not need to perform demonstra-
tion testing to identify optimal treatment. Table 4-1 presents a
recommended matrix of the degree of testing to be performed
by large PWSs based on the results of initial monitoring for
lead. The rule classifies the existing treatment of large PWSs
as optimized for corrosion control only when the difference
between the 90th percentile tap water lead level (Pb-TAP) and
the highest source water lead concentration (point of entry [Pb-
POE]) is less than the practical quantitation level (PQL) of 5
|0.g/L for each 6-month period of the initial monitoring program.
If this condition is met, then no study or testing is required and
the monitoring results for copper are irrelevant. It is recom-
mended, however, that states give some consideration to the
presence of copper in tap samples when determining whether
the treatment in place is optimized.
Large PWSs not experiencing problems with lead corro-
sion might find elevated levels of copper for which corrosion
control treatment would be warranted. The recommended level
of effort for corrosion control studies by large PWSs based on
copper is as follows:
Table 4-1. Recommended Corrosion Control Study Components for Large PWSs Based on Lead Levels
Source Water (POE) Lead Level,
Tap Lead Level as the
90th Percentile, ug/L
Pb-POE < PQL
PQL < Pb-POE < 10
Pb-POE > 10
Pb-TAP < PQL
PQL < Pb-TAP <10
10 < Pb-TAP <1 5
None required
None required
Desktop evaluation
None required
If (Pb-POE - Pb-TAP) <
No corrosion control testing
Pb-TAP >15
Desktop evaluation and
demonstration testing
PQL, then none; otherwise,
desktop evaluation
Desktop evaluation and
demonstration testing
Source water treatment recommended or
required
If (Pb-POE - Pb-TAP) < PQL, then only
source water treatment required.
Otherwise, desktop evaluation and
demonstration testing and source water
treatment recommended or required.
28
-------
Pripppr AT.
and snnrre water mpper is low (Cu- • Ion exchange
POE < 0.2 mg/L): Desktop Evaluation + Corrosion Testing.
• Copper AL exceeded and source water copper is high (Cu-
POE > 0.2 mg/L): Desktop Evaluation + Corrosion Testing
+ Source Water Treatment.
• Cu-TAP (90th percentile) contribution is > 0.5 mg/L: Desk-
top Evaluation + Corrosion Testing.
• Cu-TAP (90th percentile) contribution is < 0.5 mg/L: Desk-
top Evaluation.
4.1.2.2 Evaluating Source Water Contribution
PWSs are required to monitor lead and copper at the points
of entry (Pb/Cu-POE) only if either AL is exceeded on the basis
of first-flush tap samples. Some systems might choose to moni-
tor the source water contribution of these metals simultaneously
with first-flush tap sampling to determine whether the existing
treatment is optimal with regard to corrosion control (90% Pb
- Pb-POE < PQL). Otherwise, this monitoring must be com-
pleted within 6 months of exceeding the lead or copper AL.
All systems must submit source water treatment recom-
mendations to the state within 6 months of exceeding an AL.
While the lead and copper rule is silent with respect to the levels
of lead or copper that mandate treatment, Table 4-2 provides a
guideline for source water treatment needs. If the source water
is contributing more than the AL for either lead or copper, then
source water treatment is required. In cases where a significant
amount of lead or copper is present, treatment is recommended
to reduce the overall lead or copper exposure and to assist
PWSs in meeting the ALs in future monitoring events. Table
4-2 also shows that source water treatment is optional when
moderate levels of metals are found and is unnecessary when
very low levels of either lead or copper are present.
Table 4-2. Source Water Treatment Guidelines
Point of Entry Monitoring Results
Source Water Treatment
Guidelines
Lead, ug/L
Copper, mg/L
Not Necessary
Optional
Recommended
Required
S5
5-10
10-15
> 15
S0.2
0.2-0.8
0.8-1.3
>1.3
In cases where systems find elevated levels of lead or
copper, the sources of supply (raw water) should be monitored
prior to treatment and at various stages within the existing
treatment facility (if currently treating the supply) to determine
the source of the metals. This monitoring also will help the
system assess the performance of the existing treatment in re-
moving lead and copper.
Several types of treatment might be appropriate for re-
moval of lead and copper. EPA specified the following tech-
niques in the lead and copper rule:
• Reverse osmosis
• Lime softening
• Coagulation/filtration
If a PWS currently is providing conventional treatment
(whether alum or ferric coagulation, iron/manganese removal,
or lime softening), modifying these processes might produce
the desired results. If treatment is not available, package treat-
ment units for any of the above technologies can be installed
at individual wellheads (especially when the elevated metals
are contributed by a small number of individual wells) or at a
centralized treatment location. In the case of elevated copper,
eliminating copper sulfate applications might reduce the back-
ground level of copper for some surface water facilities.
States must respond to the recommendations for source
water treatment within 6 months. If required, PWSs have 24
months to install source water treatment once that treatment is
approved by the state. For large PWSs, the installation of source
water treatment could precede corrosion control treatment by
as much as 18 months. Followup monitoring for Pb/Cu-POE
and first-flush lead and copper tap samples will occur simulta-
neously, however, after corrosion control treatment has been
installed.
4.1.3 Desktop Evaluations
The logic diagram shown in Figure 4-1 presents the proc-
ess involved in performing desktop evaluations for selecting
alternative treatments for further investigation or the optimal
treatment for systems not required to perform demonstration
testing. This procedure allows systems to eliminate any treat-
ment approaches that are not feasible and then to determine the
water quality conditions defining the best corrosion control
treatment approaches. Among the remaining alternatives, the
system should select the optimal treatment on the basis of the
following criteria:
• Corrosion control performance based on either the reduc-
tions in metal solubility or the likelihood of forming a pro-
tective scale.
• The feasibility of implementing the treatment alternative on
the basis of the constraints identified.
• The reliability of the alternative in terms of operational
consistency and continuous corrosion control protection.
• The estimated costs associated with implementing the alter-
native treatments.
The first step is to describe the existing conditions of the
PWS in terms of its water quality parameters and the theoretical
estimation of lead and copper solubility as well as the potential
for calcium carbonate precipitation. Changes in water quality
conditions for alternative treatments should be compared to the
29
-------
Stepl
Step 2
Step 3
Step 4
StepS
Step6
/
DEFINE EXISTING CONDITIONS:
PH Lead Solubility
Alkalinity copper Solubility
Calcium corrosion Indllces
Inhibitor
I
Monitor Pb/Cu-POE and
Determine Source Water
Treatment Needs
|
DEFINE CONSTRAINTS:
• Other Water Quality Goals
• Distribution System Behavior
• Wastewater Considerations
I
/Identify Corrosion ,
Control Priorities /
i
/ Eliminate Unsuitable Approaches /
/ Based on Results of Steps 1-4 /
i
/ Evaluate Viable Alternative /
Approaches /
,,
1
1 - A
Carbonate Passivation Inhibitor Passivation ^'pr^Ston*'6
444
Define Alternative Treatment Define Alternative Treatment Goals for Calcium
r!^fc^a^I^Zr Goals for pH, Inhibitor Type Carbonate Precipitation ~ I
Goals tor pH and Alkalinity and Dose Potential (CCPP) 1
4 1 1 I
Find Lead and Copper Find Lead and Copper ..kS^JZlSAffteve 1
Solubility tor Each Alternative Solubility tor Each Alternative Alkalmity^Cateiumto Achieve §
1 1 !
Calculate Reductions In
SoluhiHty: Existing -Aft
Existing X1C
Calculate Reductions In Evaluate Feasibility T
mv Solubility: Existing- Alt of Resultant Water — «
W% Existing X100% Quality Goals
; <- i
Step?
( Reject ^^_
\. Alternatives T*
, r
EVALUATE EACH ALTERNATIVE
BASED ON:
• Performance
• Feasibility
• Reliability
• Cost
_W^ Select OptimaA
\^ Treatment J
Figure 4-1. Logic diagram for evaluating alternative corrosion control approaches.
30
-------
gristing rnnHitinng fn Hp.termine relative pp.rfnrmanrf. and po- should he compiled; (1) water quality data, (2) evidence of
tential to reduce corrosion.
Each PWS operates within certain constraints—such as
conflicting water quality goals, existing coatings in distribution
system piping, multiple sources of supply of varying water
quality, and wastewater permit limits on metals or nutrient lev-
els—that can be improved or compromised by corrosion control
treatment. The PWS should identify and document any con-
straint that could affect the feasibility of implementing an al-
ternative treatment. This information will be important in the
selection of those treatment options that are viable alternatives
for the PWS to consider further.
Based on the water quality characteristics of the supply and
site-specific constraints, the PWS can eliminate corrosion con-
trol treatment approaches that would be infeasible to implement
successfully. The remaining options should be evaluated on the
basis of each PWS's corrosion control treatment priorities. For
example, a system that experiences lead levels greater than the
AL in first-flush tap samples should set lead control as its
primary goal. A second system that finds low lead levels, but
has elevated copper levels in first-flush tap samples, should set
copper as the primary objective of corrosion control treatment.
In the latter case, however, optimal treatment should not worsen
lead corrosion behavior, and the control of lead can be consid-
ered a constraint on the decision-making process for selecting
optimal treatment for copper control.
Each of the corrosion control treatment approaches that are
viable options should be evaluated to determine the water qual-
ity characteristics that describe optimal treatment within each
option. For the passivation methods (pH/alkalinity adjustment
and corrosion inhibitors), alternative treatments are evaluated
by comparing their ability to reduce the solubility of each tar-
geted metal (lead and/or copper). The calcium carbonate pre-
cipitation method is evaluated by comparing the ability of
alternative treatments to produce sufficient potential for scale-
forming conditions to exist hi the distribution system. The "rule
of thumb" guidelines presented hi Appendix A of EPA's Lead
and Copper Rule Guidance Manual, Volume 2 (see Chapter
One for ordering information) can be used to rank the alterna-
tives within this treatment approach.
The final selection of optimal treatment will rest on the
four factors discussed above: performance, feasibility, reliabil-
ity, and costs. Direct comparison of corrosion control perform-
ance for alternative treatment approaches might not be possible.
Professional judgment and experience will be necessary to pro-
vide a basis for ranking alternatives.
The following sections provide more detailed descriptions
of the steps involved in performing a desktop evaluation of
alternative treatments and developing final recommendations
for optimal treatment.
4.1.3.1 Documenting Historical Evidence
The first step of the desktop evaluation is to identify and
document any existing information pertinent to the evaluation
of corrosion control for the system. Four categories of data
corrosion activity, (3) results of corrosion studies performed by
other PWSs as reported in the literature, and (4) results from
prior corrosion studies or testing performed by the PWS. The
most pertinent information is the results of any prior corrosion
control testing performed by the system. Beyond the direct
testing results, the PWS should conduct a comprehensive re-
view of the other sources of information.
Water Quality Data. The PWS should compile and analyze
current and historical water quality data. The key parameters
of interest include pH, alkalinity, hardness, total dissolved sol-
ids or conductivity, temperature, dissolved oxygen, and metals
(e.g., aluminum, manganese, iron, lead, and copper). These
basic water quality parameters only represent those most com-
monly required. The system should consider site-specific re-
quirements when selecting water quality parameters for review.
The data collected should pertain to raw and finished water
conditions as well as to the water quality in the distribution
system, if available. Additionally, the results of the initial moni-
toring program should be considered when available.
Understanding the treatment processes at a PWS facility
and their effects on water quality is an important aspect of
interpreting the water quality data and evaluating the appropri-
ateness of alternative corrosion control treatment techniques
(1). Figure 4-2 illustrates the relationship between water quality
and alternative corrosion control treatment approaches. In many
cases, site-specific water quality conditions will reduce the fea-
sibility of an alternative treatment approach. For example, it
would be reasonable to eliminate the calcium carbonate pre-
cipitation option as a viable treatment approach for PWSs ex-
hibiting low pH, alkalinity, and hardness in the treated water.
Conversely, a PWS exhibiting high pH conditions with moder-
ate to high alkalinity and calcium content might concentrate its
efforts on calcium carbonate precipitation, for the following
reasons:
• Although high pH conditions might be optimal for lead con-
trol, these water quality conditions are very aggressive to-
wards iron corrosion and most likely would cause severe
degradation in distribution system water quality if calcium
carbonate precipitation is not pursued.
• High dosages of corrosion inhibitors might be necessary to
maintain an effective residual throughout the distribution
system due to the presence of calcium. Also, some inhibitors
can cause existing corrosion by-products to be released in
the distribution system, resulting in water quality degrada-
tion (2).
Figure 4-2 is intended to provide general guidelines on
water quality conditions vs. alternative treatment approaches;
it is not intended to serve as the sole basis for selection or
elimination of the available alternatives. Furthermore, a PWS
must use caution any time a corrosion control approach requires
a severe modification of the existing water quality entering the
distribution system. Disruptions and upset of existing corrosion
by-products will affect the overall performance of any corro-
sion control treatment approach.
31
-------
LOW pH
<7.5
Alkalinity
(mg/L CaCO3)
Moderate
(50-150)
Calcium
(mg/LCaCO3)
Moderate pH
7.5-9.0*
Alkalinity
(mg/L CaCO3)
Calcium
(mg/L CaCO3)
* Phosphate Inhibitor only appropriate
for pH conditions less than 8.
High pH
>9
Moderate
(50-150)
Calcium
(mg/L CaC03)
Low
(<50)
Low Moderate High
Alkalinity (<50) (50-150) (>150)
(mg/L CaCO3)
Figure 4-2. Suggested corrosion control approaches based on water quality characteristics.
|= Calcium Carbonate Precipitation
|= Carbonate Passivation
]= Phosphate Inhibitor
1= Silicate Inhibitor
32
-------
Corrosion Activity. The PWS should identify and analyze
Example:
existing records indicating corrosion activity within the distri-
bution and home plumbing systems to obtain information about
the nature and extent of corrosion activity anticipated within
the service area. Evidence of corrosion activity can be obtained
by: (1) reviewing customer complaint records for dirty water
or metallic taste and odor events, (2) performing an informal
survey of area plumbers regarding the frequency and nature of
plumbing repairs (especially, for example, hot water heater re-
placements), (3) reviewing records citing the inspection of dis-
tribution system mains and service lines when they are being
replaced or repaired, and (4) water quality monitoring for met-
als or other corrosion by-products within the distribution sys-
tem or home plumbing environments.
Several factors should be considered in evaluating the use-
fulness of this information: (1) the frequency of data collection,
(2) the number of coupons, if used, and their locations within
the distribution system, (3) the analytical methods and their
respective detection limits, (4) the temporal and spatial consis-
tency of the data, and (5) the reliability of the incidence reports.
The results of the initial monitoring program required by the
lead and copper rule, if available, should be included in this
pool of information.
This information can be used to set priorities among the
corrosion control program elements by identifying the key ma-
terials for protection and to assess the general effectiveness of
the existing treatment approach.
Review of the Literature. The PWS should review the avail-
able literature to ascertain the findings of similar systems when
performing corrosion control testing and the theoretical basis
for alternative corrosion control approaches.
Several water suppliers in the United States have per-
formed corrosion control studies and published the results
(3,4,5,6). Each study has site-specific goals and objectives, as
well as water treatment and quality conditions, relevant to the
testing protocols. The experiences of these systems provide a
useful resource to other PWSs investigating corrosion control
in terms of study design and execution, data handling and in-
terpretation, and recommended treatment given the goals and
constraints acting on the system. EPA's Lead and Copper Rule
Guidance Manual, Volume 2 contains a summary of the avail-
able literature on corrosion control studies.
Prior Experience and Studies. Corrosion control treatment
is not a new concern for water suppliers, and many systems
have performed studies in the past to assist in the design and
implementation of corrosion control treatment. These past ex-
periences and studies should be revisited by PWSs to incorpo-
rate their findings and results in the present evaluation of
corrosion control for lead and copper. In some cases, the prior
testing targeted lead and copper control. These findings would
be directly applicable to the corrosion control study objectives
for the lead and copper rule. Therefore, additional testing might
not be necessary to formulate recommendations for optimal
corrosion control treatment (if not already considered to be in
place).
The Town of Allywad, a small PWS operating a ground water
well, found lead levels above the action level during initial
monitoring. To prepare recommendations for optimal treat-
ment, the PWS operator began collecting information about
the condition of distribution system materials and the experi-
ences of nearby towns and communities. From previous pipe
replacement activities, the PWS operator had noticed a thin,
buff-colored deposit on the walls of the distribution system
piping. Since the ground water source is well buffered, with
an average pH of 7.4, alkalinity of 160 mg CaCOa/L, and
calcium hardness of 140 mg CaCOs/L, this deposit was as-
sumed to be calcium carbonate.
A nearby township with wells located in the same aquifer as
Allywad had installed orthophosphate inhibitor feed facilities
for corrosion control. The township's experience was not al-
together positive. It had a significant number of turbid and
dirty water complaints after the addition of the orthophos-
phate. The township gave up the use of the corrosion Inhibitor
to restore the aesthetic quality of the delivered water supply.
After learning of these experiences, the Town of Allywad
decided to eliminate the use of orthophosphates from its list
of alternative corrosion control treatment approaches.
4.1.3.2 Identifying Constraints
The lead and copper rule provides two conditions by which
a water system may identify constraints that limit or prohibit
the use of corrosion control treatments: (1) the treatment has
been shown to adversely impact other water treatment proc-
esses and cause a violation of a National Primary Drinking
Water Regulation, or (2) the treatment has been shown to be
otherwise ineffective for the PWS.
PWSs should evaluate the impact of alternative corrosion
control treatment options on compliance with existing federal
and state drinking water standards, and with regulations antici-
pated to be finalized within the timeframe for corrosion control
installation by small and medium-sized PWSs. Table 4-3 pre-
sents the schedule for regulatory actions during the next decade
in conjunction with the compliance timeline for medium-sized
and small system implementation steps for the lead and copper
rule. The key regulatory actions that small and medium-sized
PWSs should fully evaluate to select optimal corrosion control
treatment are discussed below.
• Under the Surface and Ground Water Treatment Rules
(SWTR/GWTR), PWSs will be required to meet disinfec-
tion performance criteria. These criteria are pH-dependent
for free chlorine, where less effective disinfection results
under higher pH conditions.
• The Total Coliform Rule (TCR) requires all PWSs to meet
minimum occurrence standards for total and fecal coliforms
in distribution system samples. Some PWSs have noted in-
creases in microbiological growth within the distribution
system after installing corrosion control treatment. In most
cases, however, corrosion control treatment has been found
to have little or no effect on heterotrophic plate counts.
33
-------
Table 4-3. Schedule of Drinking Water Regulatory Activity: 1990-2000
Adverse impacts on the service community, including: (1)
Regulatory Action
Phase 1 Volatile
Organic Chemicals
Phase II Synthetic
Organic Chemicals
and Inorganic
Chemicals
Phase V Synthetic
Organic Chemicals
and Inorganic
Chemicals
Phase lib Arsenic
Surface Water
Treatment Rule
Total Coliform Rule
Radionuclides Rule
Ground Water
Disinfection Rule
Disinfectants/Disinfection
By-Products
Lead and Copper Rule
Proposal
Date
11/85
05/89
07/90
11/92
11/87
11/87
07/91
06/93
06/93
08/88
Final Date
07/87
01/91-07/91
03/92
01/95
06/89
06/89
04/93
06/95
06/95
05/91
Effective
Date
01/89-01/91
07/92-01/93
09/93
07/96
07/93
01/91
10/94
01/97
01/97
07/91-01/99*
commercial users' water quality criteria, (2) health-care fa-
cility water quality criteria, and (3) wastewater operations
(permit requirements for discharges and solids handling pro-
grams).
The particular conditions that define the constraints for
each system will be site-specific. The PWS should investigate
these conditions thoroughly as part of the desktop evaluation
aspect of the corrosion study. Small and medium-sized systems
that exceed the ALs but are not required to perform testing
should consider each of these items when selecting the optimal
treatment for recommendation to the state. Large PWSs re-
quired to perform only a desktop evaluation must present rig-
orous documentation of any constraints to support the
recommended treatment approach for the system. For any PWS
performing corrosion testing, the availability of information
regarding system constraints will assist in limiting the optional
treatment approaches that must be evaluated through the testing
program.
r o
'Dates reflect effective date of the lead and copper rule through small
PWS installation of optimal treatment after the system exceeds ALs
during first round of initial monitoring and is required to perform a cor-
rosion study.
• The Disinfectants/Disinfection By-Products Rule
(D/DBPR), currently under development, will be finalized
when PWSs are installing corrosion control treatment as a
result of the lead and copper rule. Adjusting pH conditions
can affect the level of certain DBPs, especially total triha-
lomethanes (TTHMs) and total haloacetic acids (THAAs).
These two contaminant groups are likely to be included in
the future DBPR, and they exhibit opposite relationships to
pH adjustment; that is, TTHM formation increases with in-
creasing pH, and THAA formation increases with decreasing
pH. An additional consideration is the point of pH adjust-
ment within treatment plants, since lower pH conditions
favor increased removal of DBP precursors during coagula-
tion by alum. Compliance with the DBPR could be compro-
mised by increasing the pH of coagulation as part of the
corrosion control treatment approach, because this might
reduce the efficiency of conventional treatment in removing
precursor material.
Additional constraints that PWSs should consider beyond
those required by the rule include:
be identified and considered in the selection of treatment ap-
proaches either for additional testing or as the recommended
treatment process. Worksheets are provided in Tables 4-4(a) and
4-4(b) for each of the three treatment alternatives (pH/alkalinity
adjustment, calcium adjustment, and corrosion inhibitors) to
assist PWSs in evaluating the constraints on their systems.
Example:
After exceeding the lead AL during initial monitoring, the City
of Dannyport began investigating alternative corrosion control
treatment measures to provide the state with recommendations
for optimal treatment. The city had concerns about the me-
dium-sized surface water facility's compliance with the
SWTR and selection of optimal treatment for corrosion con-
trol. The existing treatment provided by Dannyport is conven-
tional coagulation/flocculation with rapid sand filtration.
Under the SWTR, at least 0.5 logs of inactivation of Giardia
and 2.0 logs of virus inactivation were required. For the
Giardia requirements, the plant's performance is adequate to
meet the C*t required, i.e., the C*tact:C*treq is 1.2 at present.
Virus inactivation performance is satisfactory and is not af-
fected by pH changes. Giardia inactivation performance,
however, is a function of pH, and at the higher pH levels under
consideration for corrosion control, the resulting C*tact:C*treq
ratios are 0.99 and 0.83, respectively. Neither case would
provide adequate disinfection performance.
• Compatibility of a treatment approach with multiple sources
of supply.
• Compatibility of a treatment approach for consecutive sys-
tems.
• Reliability features for the particular treatment approach,
including: (1) process control, (2) operational redundancy
requirements, and (3) chemical supply integrity and avail-
ability.
An additional concern is continued compliance with the total
trihalomethane (TTHM) standard. Currently, an average of 60
ug/L TTHM is found in the distribution system with seasonal
peaks of nearly 100 u/L TTHM. Increasing the pH of the
finished water supply could only increase the probability of
Dannyport exceeding the future TTHM standard, which is
expected to be finalized at the same time that the city initiates
corrosion control treatment.
Given the above regulatory concerns, the City of Dannyport
determined that pH adjustment would not be a feasible option.
34
-------
Table 4-4(a). Constraints Worksheet for pH/Alkalinlty or Calcium Adjustment Treatment Alternatives
AUjuatiny pH/u/kalii lity and/or caiaiuiii fui corrosion control typically consists of augmenting their levels to generate favorable conditions
for lead and copper passivation or calcium carbonate precipitation.
A. National Primary Drinking Water Regulations Constraints
Rule Constraint
Surface Water Treatment Rule Reduces inactivation effectiveness of free chlorine.
Potential for interference with dissolved ozone measurements.
Might increase turbidity from post-filtration precipitation of lime, aluminum, iron, or manganese.
Ground Water Disinfection Reduces motivation effectiveness of free chlorine.
Potential for interference with dissolved ozone measurements.
Disinfection By-Products Higher THM concentrations from chlorination.
Reduced effectiveness of some coagulants for precursor removal.
Coliform Rule Potential for higher total plate counts, confluent growth, or presence of total coliforms when chlorination is
practiced.
Radionuclides In-plant adjustments can affect removal of radioactive particles if precipitation techniques are used for
coagulation or softening.
Removal of radionuclides during softening might be linked to the degree of softening. Modifying softening
practices to achieve corrosion control could interfere with removals.
8. Functional Constraints
Increased potential for post-filter precipitation can give undesirable levels of aluminum, iron, or manganese.
Process optimization is essential. Additional controls, chemical feed equipment, and operator attention might be required.
Multiple entry points will require pH/alkalinity adjustment at each entry location. Differing water qualities from multiple sources will require
adjusting chemical doses to match the source.
The use of sodium-based chemicals for alkalinity or pH adjustments should be evaluated with regard to the total sodium levels acceptable in the
finished water.
Users with specific water quality needs, such as health care facilities, should be advised of any changes in treatment.
Excessive calcium carbonate precipitation can produce "white water" problems in portions of the distribution system.
It might be difficult to produce an acceptable coating of calcium carbonate on interior piping for large distribution systems. High calcium
carbonate precipitation potential (CCPP) levels eventually might lead to reduced hydraulic capacities in transmission lines near the treatment
facility, while low CCPP values might not provide adequate corrosion protection in the extremities of the distribution system.
Table 4-4(b). Constraints Worksheet for Inhibitor Treatment Alternative
Corrosion inhibitors can cause passivation of lead and copper by the interaction of the inhibitor and metal components of the piping
system.
A. National Primary Drinking Water Regulations Constraints
Rule Constraint
Surface Water Treatment The application of phosphate-based inhibitors to systems with existing corrosion by-products can result in the
Rule depletion of disinfectant residuals within the distribution system. Additionally, under certain conditions phosphate-
based inhibitors can stimulate biological growths which can result in high heterotrophic plate counts.
Ground Water Disinfection Same as above.
Disinfection By-products No apparent effects.
Coliform Rule If corrosion by-products are released after the application of inhibitors, coliforms might be detected more
frequently and confluent growth is more likely.
Radionuclides No apparent effects.
B. Functional Constraints
Potential post-filtration precipitation of aluminum.
Consumer complaints regarding red water, dirty water, color, and sediment might result from the action of the inhibitor on existing corrosion
by-products within the distribution system.
Multiple entry points will require multiple chemical feed systems.
The use of sodium-based inhibitors should be evaluated with regard to the total sodium levels acceptable in the finished water.
The use of zinc orthophosphate might present problems for wastewater facilities with zinc or phosphorus limits in their NPDES permits.
Users with specific water quality needs, such as health care facilities, should be advised of any treatment changes.
35
-------
4.1.4 Corrosion Study Organization
PWSs not required to perform testing otherwise would select
The suggested framework for the performance of a corro-
sion study is shown in Table 4-5, presenting a logical sequence
of steps organized to satisfy the requirements and recommen-
dations described in this section. For completing steps 1
through 3, a logic diagram was presented in Figure 4-1, refer-
ring to desktop evaluations. The result of the desktop evaluation
for those systems performing corrosion control studies is the
selection of alternative treatments to be evaluated in the dem-
onstration testing step of the study. (Small and medium-sized
Table 4-5. Organization of the Major Components in Corrosion Control
Studies
Step 1. DOCUMENT HISTORICAL EVIDENCE
• Review PWS water quality and distribution system charac-
teristics
• Review PWS evidence of corrosion activity
• Identify prior corrosion control experiences and studies per-
formed by PWS
• Identify prior corrosion control experiences and studies per-
formed by other PWSs with similar characteristics
Step 2. IDENTIFY CONSTRAINTS
• Interferences with other water treatment processes
• Compatibility of multiple sources of supply
• Compatibility for consecutive PWSs
* Reliability features for particular treatment approach, including
(1) process control, (2) operational redundancy requirements,
and (3) chemical supply integrity and availability
• Adverse impacts on the community: commercial users, waste-
water operations, health-care facilities
Step3. DECISION
For any PWSs NOT Required to Perform Testing to Evaluate
Alternative Treatments:
• Formulate decision criteria
• Select primary treatment alternatives.
• Go to Step 5
For any PWS required to perform testing to evaluate
alternative treatments:
• Formulate minimum feasibility criteria for alternative treatments
• Select the alternative treatments to be included in the testing
program
• Establish overall decision criteria for selection of optimal corro-
sion control treatment
Step 4. ASSESS CORROSION CONTROL PERFORMANCE BY
TESTING
• Develop testing protocols and procedures
• Perform testing program and collect data
• Analyze data generating corrosion control performance results
• Rank performance results by priority of corrosion control pro-
gram goals
Step 5. PRELIMINARY COST ESTIMATES AND FACILITY
MODIFICATIONS
• Prepare preliminary facility design
• Prepare preliminary cost estimate
Step 6. DECISION:
• Based on the decision criteria established at the outset, formu-
late recommended corrosion control treatment and submit to the
state
the recommended treatment on the basis of the desktop evalu-
ation as shown in Figure 4-1.)
When the alternative treatments have been selected for
evaluation, a testing program is formulated and implemented.
This includes such steps as:
• Developing testing protocols, procedures, and frequency for
data collection and evaluation.
• Analyzing the resultant data to generate performance meas-
urements.
• Determining the performance ranking of the alternative treat-
ment approaches on the basis of corrosion control, secondary
treatment impacts, and process operations and control.
The PWS should prepare preliminary design and cost es-
timates for the alternative treatment approaches selected from
the desktop evaluation. Although cost is not directly a factor in
assigning optimal treatment, instances will occur where com-
parable treatment performance is observed among two or more
treatment approaches. Holding all else constant, cost might be
the deciding factor in selecting optimal treatment. Additionally,
preliminary design will be required for the state review process.
The PWS can base the final recommendation of optimal
corrosion control treatment on the results of a decision criteria
matrix and the ranking of the alternative processes. The system
must fully document and present to the state the rationale for
the selection.
4.1.5 Demonstration Testing
A PWS can use a variety of approaches and mechanisms
to evaluate corrosion control treatment through demonstration
testing. Although flexibility exists for the actual design of a
testing program, all such endeavors should clearly define and
document the following elements of the study:
• Measures of corrosion activity, such as weight loss, metal
leaching, corrosion rates, and surface condition.
• Sampling program design, including sampling frequency,
locations, volume, parameters, and analytical methods.
• Materials used to simulate the targeted piping environment,
such as lead, copper, iron, lead soldered joints, and brass.
• Protocols for material exposure, specifically, flow-through or
static environments under predetermined operating conditions.
• Data handling and analysis techniques, including statistical
testing and identifiable approaches to the interpretation of
the findings.
• Secondary testing requirements to determine the potential
impacts of alternative corrosion control treatment on exist-
ing PWS operations and on compliance with other drinking
water standards.
36
-------
• Quality assurance/quality control program elements for 4.1.5.1 Flow-Through Testing Protocol
each asnect of the testing nroffram — ... . .
each aspect of the testing program
The remainder of this section discusses each aspect of
corrosion control testing program design as identified above in
general terms. Each PWS, however, is responsible for the de-
sign and execution of a testing program that meets its own
overall goals and objectives.
The premise underlying corrosion control testing is that
alternative treatment approaches should be evaluated in terms
of their relative reductions (or increases) in corrosion activity
for specific materials of concern. Quite often, testing efforts are
used to predict the behavior of various treatment components.
In this respect, corrosion studies differ. It is NOT the intended
purpose of these studies to either: (1) predict the levels of lead
or copper in first-flush tap samples from targeted consumers'
homes or (2) predict the actual reductions in corrosion activity
within the distribution or home plumbing systems. Instead, the
purpose of corrosion control testing is to demonstrate the rela-
tive performance of alternative treatment approaches.
The use of flow-through testing methods to evaluate cor-
rosion control performance is preferred, because these methods
more accurately simulate the home plumbing environment,
where the majority of lead and copper corrosion originates. The
protocols and methods described below are suggestions that
PWSs undertaking flow-through testing can consider in the
design and execution of their demonstration study.
Flow-through testing refers to continuous or cycled flow-
ing conditions through a testing apparatus where die solution
is not recirculated. Typically, flow-through testing is used to
describe pipe rig operations where pipe loops or coupon/insert
apparatus are attached to a central pipe which distributes the
test water to one or more corrosion testing units. Figure 4-3
illustrates conceptually a flow-through pipe rig.1
For a more detailed description of standardized pipe rig construction and im-
plementation, see the American Water Works Association Research Founda-
tion (AWWARF), Lead Control Strategies (Denver, CO: AWWARF, 1990) or
P. Temkar et al., Treatment Evaluation for Reducing Lead Dissolution from
Plumbing Systems Using CERL Pipe Loop System (Champaign, IL: U.S. Army
Construction Engineering Research Laboratory, 1989).
SOURCE
CORROSION CONTROL
TREATMENT
CORROSION ACTIVITY
TESTING RIGS
Row Equalization Chemical Treatment
Basin Basin
1 Chemical Feed
/—>
Row Equalization Chemical Treatment
Basin Basin
Control Rig
Treatment
Alternative 1
M>-
o
LQ
Treatment
Alternative 2
LEGEND
M - Coupon Row-Through Cell ® = Flow Measuring Device
O] = Pipe Loop, Typically Tubing ® = Water Quality Monitoring Location
Row Discharge Point and Monitoring Locations
Figure 4-3. Conceptual layout of flow-through testing schemes.
37
-------
Row-through testing methods provide the following ad-
vantages for determining corrosion control treatment:
• Evaluation of a limited number of alternative treatment ap-
proaches with more rigor than static tests provide.
• Refinement of the chemical feed and water quality condi-
tions that best describe the selected corrosion control treat-
ment option.
• Improved simulation of the real-world conditions present in
the distribution system that the selected corrosion control
treatment will need to address.
PWSs and others conducting such studies should consider
the following general recommendations regarding the design
and implementation of a flow-through testing program:
• Duration of testing should be 9 to 12 months to capture
seasonal effects. The longer the testing period, the more
confidence a PWS can have in distinguishing treatment per-
formance.
• A standardized sampling program should be established be-
fore initiating the testing period to enhance the analysis of
results.
• Alternative locations for siting the testing apparatus should
be considered: (1) laboratory or water treatment plant, (2)
remote within the distribution system, or (3) distribution
system in situ apparatus. Sites experiencing significant
amounts of vibrations or humidity should be avoided. These
conditions can interfere with the performance of the testing
apparatus.
• The test material surfaces should be evaluated at the conclu-
sion of each test run for each material in order to assess the
corrosion behavior of the treatment alternative more com-
pletely.
• When first-flush samples are being collected, the samples
should be drawn slowly so as not to induce high-velocity
events within the test apparatus.
• For each sample withdrawn, water quality parameters and
inhibitor residuals (if appropriate) should be analyzed in
addition to the metal content of the sample.
• To the extent practical, the test conditions evaluated should
simulate the chemical feed application points and finished
water quality conditions expected during full-scale opera-
tions.
An important feature of this testing method is the in-line
corrosion control treatment that must be performed to generate
the test solutions. This treatment requires some pretreatment
appurtenances, such as chemical feed pumps, constant head
tanks, flow meters, and water quality sampling stations. In
some cases, the operation and control of the corrosion control
treatment component of the test rig can be as complicated as
the pipe rig itself, if not more so. The PWS should pay careful
attention to the feasibility of creating a "continuous" supply of
treated water prior to any final testing decisions.
PWSs might be able to use the flow-through testing system
on a long-term basis to assist in understanding the corrosion
response of the distribution system on the full-scale level. In
many cases, relationships between the flow-through testing sys-
tem and the metal levels found in first-flush tap samples can
be developed in terms of trends in responses to treatment con-
ditions. Calibration of the flow-through testing system to first-
flush tap samples would be required for this use, necessitating
concurrent flow-through testing and first-flush sampling activ-
ity beyond the initial monitoring period. Continued use of the
flow-through testing systems could provide PWSs with an ad-
ditional mechanism to determine the potential effects of treat-
ment changes on the full-scale level.
4.1.5.2 Testing Program Elements
The design and operation of a flow-through testing pro-
gram requires special consideration of several study compo-
nents. These components are briefly discussed below.
Pipe Rig Operation and Fabrication. The required flow
rate through a pipe rig depends on the number of connections
it is supplying. Typically, between 0.5 and 2 gallons per minute
(gpm) of flow through a pipe loop is adequate. If a pipe rig
consists of two or three loops, then at least 1.5 to 6 gpm of
flow is required. Operating a rig at much higher flow rates
could compromise its feasibility, depending on the complexity
of the pretreatment component. For example, a system feeding
soda ash for alkalinity and pH modification at an average rate
of 20 mg/L and operating the testing rig for 16 hours of con-
tinuous flow with 8 hours of standing time each day would
require 29 gallons of stock solution (20 mg/mL) for a 6 gpm
pipe rig. Daily stock solution requirements beyond 30 gallons
become difficult to handle, especially when extremely concen-
trated solutions are used.
Additional attention must be given to the limitations of the
pretreatment component when a slurry chemical feed condition
exists, such as lime. Stock solution strengths of hydrated lime
become problematic when solutions more concentrated than 10
mg/mL are used, depending on the pump head and tubing sizes
used. (The use of quick lime for testing rigs is not very practical
because of the large amount of impurities and the inability to
properly slake the lime.) These solutions also require continu-
ous, rigorous mixing during application to ensure a consistent
suspension of the slurry solids.
When a system uses a corrosion inhibitor, typically requir-
ing much lower dosages and therefore much lower feed rates,
the pretreatment step is less limiting on the design and opera-
tion of the pipe rig system. Systems exploring corrosion inhibi-
tors might have more flexibility in terms of the number of loops
and/or coupon/insert apparatus that a single pipe rig can accom-
modate.
The pipe loops attached to the rig should be of sufficient
length to permit a 1-L sample to be collected without introduc-
tion of water from the central pipe. Table 4-6 presents the
38
-------
Table 4-6. Pipe Volumes by Tubing Length and Diameter
Pipe Volume Table
(Volumes Listed in Liters)
Pipe Diameter (in.)
Pipe
Length
(ft)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
25
30
35
40
60
3/8
.03
.06
.09
.11
.14
.17
.20
.23
.26
.28
.31
.34
.37
.40
.43
.46
.49
.51
.54
.57
.71
.86
1.00
1.14
1.43
1/2
.04
.09
.14
.18
.23
.27
.32
.36
.41
.45
.50
.55
.59
.64
.68
.73
.78
.82
.86
.91
1.14
1.36
1.59
1.82
2.27
5/8
.07
.14
.21
.27
.34
.41
.48
.55
.62
.69
.75
.82
.89
.96
1.03
1.10
1.16
1.23
1.30
1.37
1.71
2.06
2.40
2.74
3.43
3/4
.09
.19
29
.38
.48
.57
.67
.76
.86
.95
1.05
1.14
124
1.33
1.43
1.52
1.62
1.71
1.81
1.90
2.38
2.85
3.33
3.80
4.76
1
.16
.32
.49
.65
.81
.97
1.14
1.30
1.46
1.62
1.78
1.95
2.11
2.27
2.43
2.60
2.76
2.92
3.08
3.24
4.06
4.87
5.68
6.49
8.11
1 1/4
.25
.50
.74
.99
1.24
1.48
1.73
1.98
2.22
2.47
2.72
2.96
3.21
3.46
3.71
3.95
4.20
4.45
4.70
4.94
6.18
7.41
8.65
9.88
12.36
Notes:
1. Volumes can be added together for pipe lengths not listed.
2. Liters can be converted to gallons by dividing by 3.785.
volume of water contained in various lengths of piping by
interior diameter dimension. Standard plumbing materials
should be used for the pipe loop tubing. All materials used for
each rig should be obtained from the same lot of piping. For
example, if copper piping loops are to be used in three different
pipe rigs, evaluating three different treatments, then all of the
copper used in each rig should be purchased at the same time
from the same lot. This minimizes variability in the testing
results due to differences in materials.
For copper loops with lead-tin soldered joints, fabrication
of all of the loops should be done by the same person and at
the same time (do not fabricate one set of loops and then wait
several weeks or months before fabricating the next set). In
addition, the solder should come from the same spool. After
soldering, the piping should be flushed prior to starting the
testing program to remove any excess debris.
Test Monitoring Programs. The sampling program for test-
ing rigs should include: (1) the metals being investigated, (2)
water quality parameters defining the treatment process, (3)
chemical feed rates and stock solution strengths, (4) water flow
rate through each testing apparatus, and (5) sample identifica-
tion criteria such as test run, date, analyst, time of sampling,
sample handling steps, and location of sample.
Prior to initiating the testing program, the system should
define the frequency of monitoring for specific parameters and
the method of sample collection. For example, first-flush sam-
ples can be collected every 2 weeks over a 12-month period for
metals and for water quality parameters representative of tap
samples. Daily water quality parameter sampling and notation
of the appropriate chemical feed and flow rate measurements
can be performed when operating the pipe rig, even though tap
samples are not collected, to document the water quality con-
ditions to which the test loops are exposed during the study.
4.1.5.3 Static Testing Protocols
Static tests can be performed to ascertain the corrosion
behavior of alternative treatments toward different piping ma-
terials. Static testing by definition refers to "no flow-through"
conditions or batch testing (for example, the jar testing many
PWSs perform to evaluate coagulant dosages represents a batch
testing protocol). The most common form of static testing is
immersion testing, where a pipe material, typically a flat cou-
pon, is immersed in a test solution for a specified period of
time. The corrosion behavior then can be described by weight
loss, metal leaching, or electrochemical measurement tech-
niques. Other static testing methods include: (1) using a pipe
segment of the desired material, filling it with test water and
measuring the metal pickup obtained at the conclusion of a
specified holding time, and (2) recirculation testing, where a
reservoir of test water is circulated through pipe segments or
pipe inserts over a period of time. (Although water is flowing
through the piping segments, the same "batch" of water is being
recirculated during the holding time; in this sense, it represents
a static test.)
The general methods described above are not exhaustive.
Testing design will be a function of the overall goals and ob-
jectives of the testing program.
In many cases, static tests can be used to evaluate more
quickly the numerous alternative treatments that might be ap-
propriate for a PWS. This procedure would allow a PWS to
narrow the treatment approaches to a more limited number for
additional testing, if required. Since flow-through testing pro-
grams tend to be more complex and costly, eliminating inap-
propriate treatment alternatives prior to performing
flow-through testing is advantageous. To the extent that static
testing can provide such capabilities, it should be included in
the comprehensive testing program.
For many systems, however, static testing can be sufficient
to identify optimal corrosion control treatment. Small and me-
dium-sized PWSs required to perform corrosion studies should
consider static testing programs to verify the appropriate treat-
ment process. Large PWSs also should consider using static
tests for developing recommendations on optimal treatment
when only a limited number of treatment alternatives are avail-
able and flow-through testing is difficult to perform adequately.
39
-------
Drawbacks associated with using the static testing proto- and datft), (2) sampling event (control vs, test apparatus, loca-
cols as the basis for selecting optimal corrosion control treat-
ment include the following:
• Static testing conditions do not represent the conditions to
which piping systems are subject during normal operations.
Household plumbing environments experience on-and-off
cycles of flow, and the distribution system piping network
experiences continuous flow-through conditions.
• The variability found in testing results might confound a
PWS's ability to differentiate treatment performance among
the alternatives tested. Replicate testing and measurements
are important components of the testing design, providing
additional precision and accuracy assessment capability.
• Comparability of the test results with full-scale performance
is uncertain based on existing information. It might be useful
for PWSs to place coupons or pipe inserts within the service
area and at water treatment plant effluent lines during the
testing program. This would provide a basis of comparison
between the static tests (control conditions only) and the
full-scale system.
4.1.6 Data Handling and Analysis
Data needs are an important consideration in (he design of
the testing program (7,8). Analytical procedures should be de-
fined clearly prior to developing the testing program. These
procedures should: (1) describe the behavior of the testing data,
and (2) generate performance rankings for the alternative treat-
ments. The most useful approach to statistically evaluating cor-
rosion control data involves the application of nonparametric
statistics.
Underlying all statistical measures are certain fundamental
assumptions regarding the "true" behavior of the data or its
universe. The most commonly applied statistical tests (such as
the student's t test, chi-square distribution, difference of means,
and analysis of variance) are preconditioned to describing uni-
verses that exhibit a normal distribution of their values. Corro-
sion control testing data, however, tend to be non-normal, and
therefore conventional statistical measures would not describe
the behavior of the data accurately, or would not reliably gen-
erate results that could be used to rank alternative treatments.
Nonparametric analyses accommodate non-normal conditions
and can be applied to develop relative performance measures
for numerous treatments.
The nonparametric tests of importance are: (1) the Wil-
coxon test, or U-test, which can compare the results of two
conditions to determine whether they behave similarly (i.e., no
difference in corrosion performance can be ascertained) or
whether they behave differently (i.e., one treatment method
produces better corrosion protection), and (2) the Kruskal-Wal-
lis test, or H-test, which is the more general case and can
evaluate more than two test conditions.
The information to be collected for each testing run in-
cludes descriptions of: (1) test conditions (run number, treat-
ment dosages of applied chemicals, water quality parameters,
tion of sampling point, time, and type of material), and (3)
analytical results (water quality parameters such as pH, tem-
perature, alkalinity, hardness, inhibitor residual, disinfectant
residual, lead, copper, iron, etc., and/or coupon weight condi-
tions).
Data base management capabilities for microcomputer ap-
plications are satisfactory for evaluating most corrosion study
data. The use of spreadsheets or data base management soft-
ware in conjunction with statistical analysis programs is essen-
tial when large amounts of data are collected.
4.1.7 Secondary Testing Programs
Secondary testing programs are vital to the overall study
design because this corollary information will be incorporated
into the selection process for defining optimal treatment. A
major area of concern for secondary treatment is how the alter-
native corrosion control treatment can be installed successfully
and operated to meet future state-mandated operating condi-
tions that define compliance with the lead and copper rule.
When pH, alkalinity, or calcium adjustment are components of
a treatment alternative, the stability of these parameters be-
tween the point of adjustment and finished water entry to the
distribution system should be ascertained. The likelihood of
inhibitors and key water quality parameters remaining within
acceptable limits in the distribution system also should be in-
vestigated.
The PWS must achieve compliance with existing and fu-
ture drinking water standards after the installation of corrosion
control treatment. Testing to evaluate these conditions should
be included in the design of the corrosion control study. Of
particular concern might be changes in: (1) the levels and types
of disinfection by-products that might occur, (2) the occurrence
of positive total coliform events, including those induced
through increases in the presence of heterotrophic plate count
bacteria, or (3) disinfectant residual concentrations.
4.1.8 Quality Assurance/ Quality Control Programs
Critical to the interpretation of the data and findings is
ensuring that proper quality assurance and quality control
(QA/QC) procedures were followed during the testing program.
A well-designed QA/QC program permits the investigator to
describe more accurately the variability introduced into the data
by the response of testing materials to the corrosion control
treatment processes being evaluated. Elements to be included
in a QA/QC program include:
• Sufficient sampling frequency for water quality parameters
during the period of time when water is flowing. Sufficient
sampling frequency is necessary to adequately describe the
test conditions to which the materials were subject between
first-draw samples. For example, if standing samples are
collected each week, then at least daily sampling for water
quality parameters should be performed for the treated water
supplied to the pipe rig.
40
-------
• Split samples for metal analyses, especially when metal test
kits are used. It is recommended that at least 5 percent of
the samples collected be split samples.
• Preparation of sample blanks and spikes by someone other
than the chemical analyst to verify routine measurements. A
sample blank and spike should be performed during each
testing period for metals.
• Proper calibration of all analytical instruments at the begin-
ning of each testing period. Chemical feed and flow rate
meters should be fully calibrated prior to the initiation of
testing and checked periodically during the testing pro-
gram.
• Sample handling procedures that follow those required in
the lead and copper rule for metals and water quality pa-
rameters. Special care should be given to the cleaning pro-
cedures used for metals analysis containers to minimize
cross-contamination of sampling events.
Each testing program will have specific QA/QC require-
ments. The PWS should delineate these elements at the begin-
ning to prevent the collection of data that cannot be adequately
verified.
4.1.9 Example of Selecting Optimal Treatment
A large PWS performed a desktop evaluation of its sys-
tem and identified two alternative treatments for further
study by corrosion testing. Flow-through testing was per-
formed using pipe rigs with: (1) iron tubing and copper tub-
ing with lead solder, and (2) copper, lead, and iron coupon
flow-through cells. Figures 4-4a and 4-4b present the results
of the corrosion testing in terms of the percent reductions in
metal solubility for standing samples and average weight
loss for treatment alternatives A and B as compared to the
existing treatment results.
The first step in developing the final treatment selection
decision matrix is defining the performance ranking of each
treatment evaluated. The score for the best treatment option
used in this analysis is 7, for the second, 4, and for the worst
option, 0. Given the priorities of the PWS, the weighting factors
to each metal were 0.45, 0.40, and 0.15 for lead, copper, and
iron, respectively. Because of the increased importance of con-
trolling lead and copper solubility, the measurement weighting
factors were 0.7 and 0.3 for solubility and weight loss results,
respectively, for lead and copper. For iron, however, the meas-
urement weighting factor was 0.3 and 0.7 for solubility and
weight loss results, respectively, because of more concerns
about maintenance and repair of iron piping.
Table 4-7 presents the corrosion control performance ma-
trix with the appropriate weighting factors shown. The resulting
score indicates that treatment A provided the best corrosion
control protection, while treatment B provided the second best,
and the existing treatment provided the worst performance.
These results are used in the final treatment selection matrix.
THT-A TRT-B
F==?|-e°PP« E&gg -l«d |^ -Iron
Reduction In Metal Concentrations by Alternative Treatments
Reduction In Coupon Weight-Lou by Alternative Treatments
Figure 4-4. Reduction in metal concentrations (a) and coupon weight-
loss (b) by alternative treatments.
Table 4-8 presents the final treatment selection matrix for
the PWS. Because a desktop evaluation was performed prior to
the selection of treatments A and B for further testing, it was
determined that all treatment options were equally feasible,
eliminating this parameter from the decision matrix. By far, the
most important consideration for identifying optimal treatment
in this case is treatment performance, shown by setting its
weighting factor at 0.75. The reliability and cost weighting
factors were set at 0.20 and 0.05, respectively. The reliability
of the treatment options is considered more important than the
costs, because compliance eventually will be determined by the
ability of the PWS to consistently produce finished water that
meets its optimal treatment objectives.
Based on the results of the final treatment selection deci-
sion matrix, Treatment A would be recommended as optimal
corrosion control treatment.
4.1.10 Example of a Flow-Through Demonstration
Testing Program
Utility A exceeded the action level for lead during its first
6-month period of diagnostic monitoring and initiated a corro-
sion control study. The utility treats water from a surface supply
41
-------
Table 4-7. Corrosion Control Treatment Performance Ranking Matrix
Performance Criteria
Metal Solubility
Weight-Loss
Treatment Alterna-
tive
Copper Lead Iron Copper Lead Iron
Weighting Factors
Treatment A
Treatment B
Existing
Interim Performance
Treatment A
Treatment B
Existing
Measurement
Technique
Weighting
Factors
0.40
4
7
0
Scores
1.6
2.8
0.0
0.7
0.45
7
4
0
3.2
1.8
0.0
0.7
0.15
5.5
5.5
0
0.8
0.8
0.0
0.3
0.40
7
4
0
2.8
1.6
0.0
0.3
0.45
7
0
4
3.2
0.0
1.8
0.3
0.15
4
7
0
0.6
1.1
0.0
0.7
Measurement Scores
Treatment A
Treatment B
Existing
Total Score
Treatment A
Treatment B
Existing
1.1
2.0
0.0
5.8
4.7
0.5
2.2
1.3
0.0
0.2
0.2
0.0
0.8
0.5
0.0
0.9
0.0
0.5
0.4
0.7
0.0
Table 4-8. Final Corrosion Control Treatment Selection Matrix
Corrosion
Treatment Control
Alternative Performance
Weighting Factors
Treatment A
Treatment B
Existing
0.75
7
4
0
Treatment
Reliability
0.15
7
0
4
Estimated
Costs
0.1
0
4
7
Total
1
6.3
3.4
1.3
to provide treated water with the following general charac-
teristics:
pH = 7.8
Total hardness =
85 mg/L as CaCOs
SC-4 = 40 mg/L
Ca = 52 mg/L as CaCOs Total solids = 275 mg/L Cl = 5 mg/L
Total alkalinity = Na = 10 mg/L
60 mg/L as CaCOs
As illustrated in Figure 4-2, several avenues for treatment
exist. After conducting a desktop study and visiting with some
other utilities using similar water sources, the utility decided to
use pipe loops to further define optimal corrosion control treat-
ment.
Three identical pipe loops were constructed of copper pipe
with lead-tin soldered connections. Loop 1 represented a con-
trol loop without treatment. Loop 2 used finished water treated
with lime addition. Loop 3 used Finished plant water with the
addition of a phosphate inhibitor. The target pH for Loop 2 was
8.3. The alkalinity and final hardness were allowed to fluctuate
to satisfy the final pH goal. Loop 3 water was pretreated by the
addition of a proprietary phosphate inhibitor at a dose calcu-
lated to yield 1 mg/L as P04.
The three loops were run for a period of 35 weeks after
which they appeared to have stabilized somewhat and testing
was terminated. Water was pumped through the loops for 16
hours followed by an 8-hour standing period. Standing water
samples were collected for lead analysis once per week for the
35-week period. Data from the tests are given in Table 4-9.
Unless preconditioned for an extended period, new piping
materials are likely to yield higher metals concentrations than
actual household plumbing systems. Results from testing pro-
Table 4-9. Lead Concentrations from Pipe Loop Testing
Week
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Loop 1
Pb, ng/L
62
78
125
110
175
205
190
162
78
112
95
132
126
103
115
138
92
100
118
107
68
82
97
112
85
78
60
92
75
87
63
72
68
80
91
Loop 2
Pb. ng/L
130
100
80
95
110
135
108
92
79
85
90
76
79
108
87
72
68
52
97
75
48
72
103
96
72
80
52
58
45
53
60
55
52
48
57
Loop 3
Pb, ug/L
78
102
115
109
126
102
98
75
82
70
68
65
81
73
65
68
72
38
55
62
50
68
76
72
75
80
62
54
58
45
52
68
30
51
42
42
-------
grams, therefore, are used to select treatment techniques; final where the numerator represents th(
difference between
action levels after installation of full-scale treatment can only
be estimated. In the testing program discussed here, finished
water from the treatment facility was pumped continuously
through all three loops for 4 weeks to partially acclimate the
pipe rig before the initiation of the weekly sampling program.
Parametric statistics were used to compare the two treat-
ments with the control. Recognizing that water quality data
frequently is skewed, the data were investigated for skewness
(as the moment coefficient of skewness approaches zero, the
data approach a more normal distribution). If the distribution
is normal, or can be made more normal by a transformation,
the statistical techniques based on a normal distribution are
appropriate; otherwise, they are only approximations and the
use of nonparametric statistics might be more appropriate. As
indicated in the example, calculating the skewness coefficient,
y, showed that a logarithmic transformation gave smaller
skewness coefficients, so the data were evaluated in the log
normal mode.
The skewness coefficient is defined as:
where:
-2
.-3
= individual samples, i = 1 to n
Table 4-10 gives the calculated skewness coefficients for
the lead data in Table 4-9 for both normal distributed samples
and the log normal mode. The smaller coefficients for the log
normal distribution were used as indicators that the data would
adapt more appropriately to parametric statistics using a loga-
rithmic transformation.
Table 4-10. Skewness Coefficients for Lead Data
Skewness Coefficient
Mode of Distribution
Loop 1
Loop 2
Loop 3
Normal
Log Normal
1.21
0.53
0.47
-0.04
0.60
-0.32
The student's t statistic was used to compare paired data
among the three loops. These results are presented in Table
4-11. The student's t can be defined as:
paired sample data and the denominator represents the standard
deviation appropriate to the difference between the sample
means. These values then are compared to standard statistical
tables to determine if any statistical difference in treatments
exists.
Table 4-11. Calculated Student's t Values
Comparison
t
Loop 1 and Loop 2
Loop 1 and Loop 3
Loop 2 and Loop 3
5.46*"
6.98"*
2.87**
Notes: All test data transformed to logarithmic values
"Highly significant difference at the 0.01 level
"'Extremely significant difference at the 0.001 level
Results from the testing program indicate that either treat-
ment would be beneficial when considering the entire 35 weeks
of data. Any statistical evaluation of data must be tempered
with good judgment, however, and reviewing the data seems to
indicate fewer fluctuations in all the data during the final weeks
of testing. This is a reasonable result, because one would expect
the pipes to become more acclimated as the testing program
proceeded. Using a data set from week 25 on, the data were
examined once again. These results showed that there was still
a significant difference when each treatment was compared to
the control, but there was no apparent statistical difference
between treatments. Thus, the utility needs to examine other
factors such as initial cost, operating costs, and operating phi-
losophy before deciding which treatment to implement for full-
scale treatment.
4.2 Design Considerations and Procedures for
Coupon Tests
4.2.1 Summary of Method
This section presents guidelines for monitoring the cor-
rosivity of water by coupon weight loss methods. The informa-
tion is based on the ASTM Standard Test Method D 2688-90,
Corrosivity of Water in the Absence of Heat Transfer (Weight
Loss Methods) (9). Two types of corrosion specimens are de-
scribed by the ASTM method: flat, rectangular coupons and
cylindrical pipe inserts. The cylindrical Illinois State Water Sur-
vey (ISWS) and Construction Engineering Research Labora-
tory (CERL) pipe inserts were developed by the ISWS and were
adopted later as an ASTM standard. Coupons have had a long
history of use in industrial and research applications. Both cou-
pons and pipe inserts are used routinely in the ISWS laboratory
to measure corrosion rates in potable and industrial water sys-
tems. Several publications by ASTM, the National Association
of Corrosion Engineers (NACE), and others are available that
provide additional insight on the application and use of corro-
sion specimens to measure corrosion rates by the weight loss
method (10,11,12,13,14).
The weight loss method simply measures the mass of a
metal coupon that has been transformed by corrosion into sol-
43
-------
uble and insoluble corrosion by-products. A clean metal sped- signed to sitrmlRtP th<*. environment nn the; surface, of a pipe,.
men is weighed and exposed to water for a specified time. The
specimen is removed from the water and is cleaned chemically
or mechanically to remove all deposits above the underlying
metal. The specimen then is reweighed and the weight loss is
converted into the desired corrosion terms. Although this prem-
ise is simple, many factors must be considered to obtain reliable
data when using corrosion specimens in field studies. Some of
these factors are addressed in this section, but the section fo-
cuses primarily on the laboratory procedures employed to pre-
pare, process, and evaluate both coupon and pipe insert types
of corrosion specimens.
The procedures require a laboratory setting. An analytical
balance, microscope, fume hood, oven, desiccator, hot plate,
and other common laboratory equipment are used to process
corrosion specimens. The capability to store, handle, and dis-
pose of chemicals safely is an essential requirement of the
procedure. Staff members responsible for carrying out the pro-
cedures should have technical training and laboratory experi-
ence. In the absence of an in-house laboratory, independent
laboratories and consultants might be needed to perform the
corrosion studies and conduct corrosion rate measurements.
4.2.2 Basic Corrosion Measurement Considerations
The objectives for determining corrosion rates in potable
water systems should be well defined. Typical objectives in-
clude determining the water corrosivity, life of materials, and
treatment effects. Corrosion specimens have been used in labo-
ratory, pilot-scale, and field experiments to meet these objec-
tives. Laboratory studies generally are used to evaluate the
factors that influence the corrosion of metals under closely
controlled conditions. The laboratory studies also are useful for
accelerated testing and screening.
Corrosion studies conducted in the laboratory, however, do
not represent actual service conditions, and pilot-scale and full-
scale field studies should be used to complement the laboratory
data. In field studies, the corrosion specimens encounter the
actual environmental conditions of the system and conse-
quently reflect the variability in corrosion because of water
chemistry, temperature, and flow. Since the surface of a corro-
sion specimen is bare metal, it is not representative of a material
in equilibrium with the system. The effects of chemical treat-
ment and water quality on corroded materials will not be re-
produced by the clean specimens. Corrosion specimens
nevertheless are effective tools for studying or monitoring cor-
rosion, as long as proper procedures are applied and results are
interpreted correctly.
Early in the design of a corrosion study, a decision must
be made about the type of corrosion specimen to use, the metal
alloy or alloys to be tested, and the quantity of specimens
needed to complete the study. Both the coupon and the pipe
insert type of corrosion specimen have been widely used in
potable water systems. The coupon type is the least expensive
and the most readily available in a variety of alloys. It also
requires less preparation than pipe inserts. The pipe insert type
of specimen, however, offers some characteristics that might
offset the advantages offered by coupons. Pipe inserts are de-
wall. This environment is usually more representative of the
corrosion that occurs in the plumbing system than the disturbed
flow environment surrounding coupon installations. Pipe in-
serts, which are produced from genuine pipe available from
local plumbing suppliers, help ensure that the exposed surface
and material of a specimen are representative of real piping
systems. The inserts also have three to four times the exposed
surface area of coupons, which produces more weight loss and
sensitivity to surface attack.
Corrosion specimens can be purchased directly from sup-
pliers or they can be prepared in house when adequate labora-
tory and machine shop support is present The choice depends
on the capabilities, expertise, and desire of staff to assume
complete control of a corrosion study. The cost associated with
purchasing specimens is relatively low when compared with
the total cost of a study. There can be a vast difference, how-
ever, in the cost of specimens, depending on the alloy and type
of specimen required. The metals of most concern in public
water supplies are cast iron, steel, galvanized steel, copper,
brass, lead, and solder. Coupons of these materials are readily
available from various suppliers. A list of suppliers of coupons
and representative costs is shown in Tables 4-12 and 4-13.
Suppliers of the ASTM pipe insert specimens are not readily
available, and only one has been licensed to distribute the
CERL pipe loop and corrosion test assemblies. The ISWS labo-
ratory has always constructed the pipe inserts in house or in
local machine shops. The ASTM D 2688-90 method provides
the specifications needed to prepare the inserts and test assem-
blies in house.
Table 4-12. Suppliers of Corrosion Specimens and Pipe Loops (ASTM
Method D 2688-90)
Flat Rectangular Coupons, Coupon Holders, and Pipe Loop
Assemblies:
INSS, Inc.
2082 Mtehelson Drive, Suite 100, Irvine, California 92715,
714-250-3033
Metal Samples Company
Route 1, Box 152, Munford, Alabama 36268, 205-358-4202
METASPEC Company
P.O. Box 22707, San Antonio, Texas 78227-0707, 512-923-5999
Cylindrical Pipe Inserts, Supporting Assemblies, and Pipe Loops:
Evans Machine Company (licensed USA-CERL Pipe Loop
supplier)
410 Summit Avenue, Perth Amboy, New Jersey 08861,
908-442-1144
4.2.3 Purchasing and Preparation of Corrosion
Specimens
Whether corrosion specimens are purchased or prepared in
house, the specimens must be machined from a metal of known
composition and made of a material equivalent to the piping
material to be studied. A mill report should be requested when
coupons are purchased to certify the alloy number, composition,
and other metallurgical information. The pipe used for the fab-
44
-------
Table 4-13. Typical Coupon and Plpp I nnp
to protect all corrosion specimens when they are removed ftoiu
Rat, Rectangular Specimens (1/16" x 1/2" x 3"), preweighed:
Price range, July 91
($ per coupon)
Material
Mild Steel, C1010
Copper, CDA110
Copper, CDA 122 DHP
Zinc, 99.9% pure
Lead, 99.9% pure
Lead/Tin Solder, 60/40
1.60 to 2.65
1.80
3.15 to 4.10
1.90 to 6.00
2.001010.00
Pipe Plug Assemblies, 3/4" plug, 3" nylon stem:
Price range $7.25 to $10.00 per unit
Pipe Loop, PVC, 3/4" Sen. 80, for four (4) plug assemblies:
Price range $71.00 to $140.00 per unit
Cylindrical, Pipe Specimens, USA-CERL type, 3/4' x 4",
preweighed:
Estimated Price $20 to $35 per specimen, depending on
material (does not include lab fees). A complete CERL Pipe
Loop, with meter, pump, etc., can cost $1,200 to $2,000.
rication of pipe inserts should be inspected for metallurgical
defects, physical damage, and surface films. All pipe that does
not meet high quality standards should be rejected. A sufficient
number of specimens should be purchased, or sufficient mate-
rial should be in stock, to provide an ample number of identical
specimens to meet the demand of current and anticipated cor-
rosion studies.
Coupons of the same alloy must be identical, if possible.
Each must be machined and treated in the same manner. They
must have the same size, shape, and surface finish. Steel, cop-
per, and galvanized zinc coupons are impacted with glass beads
for a final finish, whereas lead and solders are scoured with a
fine abrasive powder. During the finishing process, extreme
care must be taken to prevent contamination from being carried
over to the coupon surface by other metals. The surface of pipe
inserts is inspected to ensure that they are metallurgical^ sound
and free of mill scale, and defective inserts are discarded.
A distinctive identification number is assigned to each cor-
rosion specimen. This number should be stamped prominently
on the surface of the specimen and also should identify the
metal alloy used to produce the specimen. All specimens must
be degreased and scoured with a fine abrasive to remove lubri-
cants and debris from machining operations. After degreasing,
the specimens must be handled with gloves or plastic-coated
tongs to prevent further contamination.
The clean, dry specimens are weighed on an analytical
balance to the nearest 0.1 mg. The weight of each specimen is
recorded along with its identification number on a customized
report form (Figure 4-5). The report becomes part of a perma-
nent file for documenting future weight loss and evaluation data
concerning the specimen. Specimens are stored in a desiccator
or similar noncorrosive atmosphere until needed. Steel speci-
mens are especially susceptible to corrosion during handling
and storage and should be kept in envelopes impregnated with
a vapor-phase inhibitor. Sealed plastic envelopes should be used
the moisture-free environment of a laboratory.
Corrosion specimens can be installed in a test loop de-
signed to investigate specific corrosion problems, or they can
be inserted into a conventional plumbing system for routine
monitoring. The specimens must be electrically insulated from
any associated piping during exposure to water to eliminate
galvanic and stray current influences. Coupon-type specimens
are attached to a rod threaded into a pipe plug. The threaded
rod, pipe plug, and associated nuts, screws, and washers are
constructed from PVC, nylon, phenolic, or other nonconducting
materials. The pipe plug and mounted coupon are inserted into
a pipe tee with the coupon protruding into the flowing water.
Coupons also can be inserted in the reverse direction to check
for a flow effect on the corrosion results. Pipe inserts are in-
stalled in a holder assembly consisting of standard PVC pipe
unions, nipples, and fittings. Multiple pipe inserts can be in-
serted into a single PVC assembly if separated by PVC spacers.
The complete test assembly containing the pipe inserts can be
installed in a standard pipe loop or can become an integral part
of a building plumbing system.
All relevant information concerning the installation of cor-
rosion coupons is recorded on the report form assigned to each
specimen: date of installation, site location, water supply, ori-
entation of specimens, and similar details. This report is filed
for future reference until the specimens are removed and re-
turned to the laboratory for processing.
4.2.4 Duration Guidelines for Corrosion Studies
The optimum length of time that specimens need to be
exposed to obtain reliable corrosion rates depends on the sur-
face area exposed, the metal corrodibility, and the water cor-
rosivity. The physical size of specimens is limited by the
weighing constraints of the analytical balance, although the
exposed surface area is designed to maximize weight loss dur-
ing installation. A significant weight loss must be obtained to
assess the corrosion resistance of pipe materials accurately or
to evaluate the effect of water treatment. Because the corrod-
ibility of piping materials is very low by design, long-term
corrosion studies are needed in public water supplies to obtain
the needed weight loss. Table 4-14 lists the typical corrosion
rates for plumbing materials exposed to a variety of Illinois
water supplies.
Table 4-14. Typical Corrosion Rates for Pipe Inserts in Illinois Waters
Pipe Material
Corrosion Rates
(range in mpy)
Copper
Galvanized Steel
Mild Steel
0.05-0.60
0.10-2.00
0.50-10.00
As a general guideline, the ISWS has found that corrosion
specimens require at least 6 months exposure for meaningful
corrosion rates, but under some conditions, these specimens can
require up to 24 months exposure (15). Copper and galvanized
steel pipe inserts are installed for 12 months for routine moni-
toring purposes. The duration for a corrosion study is a variable
45
-------
CORROSION SPECIMEN DATA FORM
Coupon Identification:
Number Metal.
Type Dimensions
Surface Area (sq in) Surface Finish.
Source Date Prepared.
Coupon Weight Loss Data:
Original Weight (prior to installation), gram
Final Weight (after exposure and cleaning), gram
Weight Loss (due to corrosion), gram
Installation Information:
Location
Description
Date Coupon Installed (mm/dd/yy)
Date Coupon Removed (mm/dd/yy)
Exposure Time, days
Visual Examination:
General Appearance
Pitting none isolated general
size & shape
maximum pit depth, inch
Corrosion Rate Results:
Penetration, mils/year mpy =
mm/year mmpy =
Weight loss, mg/dm /day mdd =
Report additional comments or calculations on the back of this report
Signature(s) Date Reported
Figure 4-5. Corrosion specimen data form.
46
-------
that should be determined for each water source. Corrosion
specimens must be installed and removed at regular intervals
to determine the effect of time on corrosion. The accuracy and
quality of the corrosion data are improved significantly by the
use of replicate specimens.
Wachter and Treseder's (16) planned-interval test proce-
dure is recommended for corrosion studies designed to develop
treatment strategies in public water supplies. This is an excel-
lent procedure for evaluating the effect of time on the corrosion
of metals and also for monitoring changes in corrosivity of the
water during a corrosion study.
4.2.5 Processing of Corroded Specimens
After a specified interval, the coupon holders and/or pipe
insert test assemblies are removed from the piping system to
terminate exposure of the specimen(s). Each specimen is sepa-
rated carefully from the corrosion test assembly. Specimens are
air-dried immediately and are kept in a 105°C oven, desiccator,
or similar low-humidity atmosphere until processed in the labo-
ratory. The appearance and condition of each specimen should
be evaluated visually. Any degradation in appearance of the
specimen is recorded on the report form: i.e., localized attack;
physical damage; and color, porosity, and abundance of surface
deposits. When required, the appearance of specimens can be
documented with color photographs for future reference. Speci-
mens then are grouped and processed in sets of like metal
alloys.
Any coating applied to the specimen to confine corrosion
damage to a defined surface area is removed. Pipe inserts usu-
ally are painted to limit corrosion to the internal surface of the
specimen. Paint must be removed carefully to prevent solvents
and water from contacting the corroded surface area and de-
grading the oxide or mineral deposits on the specimen. The
specimens are rinsed with water and acetone before being re-
dried in a 105°C oven. In the ISWS, the specimens are removed
from the oven, allowed to cool, and weighed with the deposi-
tion products intact. Although some deposit might be lost in
handling, the difference in specimen weight before and after
chemical cleaning is an indication of the mass of corrosion and
mineral deposits occurring hi the system. The specimens then
are cleaned by chemical and mechanical procedures to remove
all surface deposits above the base metal.
4.2.6 Chemical Cleaning Procedures
Bulky deposits are removed from the corrosion specimens
prior to chemical cleaning to minimize the time that specimens
are exposed to the aggressive chemical solutions. A plastic
spatula or similar tool is used to scrape the deposits off the
specimens without damaging the underlying metal. These sur-
face deposits often are saved for chemical analyses and evalu-
ation by x-ray diffraction to identify the mineral components.
An ideal cleaning procedure will remove all the corrosion
products and mineral deposition from the surface of a coupon
without any loss of base metal. Some base metal is lost by all
cleaning procedures, however, and this weight loss must be
determined by the use of uncorroded specimens (blanks). Rep-
licate blanks are subjected to the same cleaning procedure and
solutions that are used to process corroded specimens. To obtain
the net weight loss because of corrosion, the mean weight loss
of the replicate blanks resulting from the cleaning process is
deducted from the gross weight loss of specimens. This net
weight loss is used to calculate the corrosion rate.
Various chemical cleaning procedures are cited in the lit-
erature for each type of alloy (10,13,17). The ISWS laboratory
has found that the ASTM D 2688-90 cleaning procedures are
quick, simple, and efficient for processing large numbers of
specimens. An ultrasonic cleaning bath also is used to improve
the chemical cleaning efficiency for removing adherent depos-
its. The cleaning procedures used by the ISWS for copper, zinc,
iron, and lead alloys (both pipe inserts and coupons) are sum-
marized in Sections 4.2.6.1 through 4.2.6.4. The procedures are
based on the ASTM 2688 method, but other acceptable cleaning
procedures are documented in the literature (10,13,17). Note
that the procedure for cleaning lead specimens has been modi-
fied because the weight loss of lead blanks was high using the
ASTM cleaning solution. The cleaning procedures for lead and
lead-solder require further study.
Chemical cleaning solutions employ acids, alkalis, and sol-
vents that can be hazardous to personnel. The handling, use,
and disposal of chemical solutions should comply with current
laboratory safety regulations. Cleaning procedures should be
carried out in a fume hood and personnel should wear protec-
tive clothing and goggles.
4.2.6.1 Iron and Steel Specimens
Specimens are immersed in freshly prepared hydrochloric
acid (10 percent HC1) for 5 minutes at ambient temperature.
Alternately, scour, brush, and acid clean to remove stubborn
deposits. Specimens should not be immersed in cleaning solu-
tion for more than 30 minutes. Specimens are rinsed thoroughly
in order with tap water, deionized water, and a dilute passivat-
ing solution. The specimens are placed immediately in a 105°C
oven to dry for 1 hour. They are removed from die oven, al-
lowed to cool, and are reweighed to the nearest 0.1 mg on an
analytical balance. This final weight is recorded on the report
form to complete the data needed for the corrosion rate calcu-
lation.
4.2.6.2 Copper and Copper Alloys
The copper specimens are immersed in hydrochloric acid
(10 percent HC1) for 1 to 2 minutes at ambient temperature.
The specimens are rinsed thoroughly with tap water, deionized
water, and acetone. They are allowed to dry for 5 to 10 minutes
in a fume hood to remove acetone and are stored in a desiccator
for 24 hours before weighing in the same manner as steel
specimens. Acid solutions used to clean copper specimens must
not be used to clean other metals.
4.2.6.3 Zinc and Galvanized Steel
Zinc and zinc-coated specimens are immersed hi sulfamic
acid solution (10 percent) for 5 minutes at ambient temperature.
Beakers containing the cleaning solution are placed in an ultra-
sonic bath to improve cleaning efficiency. Specimens are alter-
47
-------
nately scoured, brushed, and acid cleaned until the surface de-
posits are removed. The specimens then are rinsed thoroughly
with tap water, deionized water, and acetone. They are imme-
diately placed in a 105°C oven to dry for 1 hour before weigh-
ing.
4.2.6.4 Lead and Lead Solder
Lead specimens are immersed in a 1 percent acetic acid
solution for 2 minutes and held at a temperature of 60 to 70°C.
The specimens are brushed very lightly and rinsed thoroughly
with deionized water and dry acetone. They are placed in 105°C
oven to dry for 1 hour before weighing. Lead specimens must
be handled very carefully to minimize unintentional metal loss.
4,2.7 Evaluation of Localized Corrosion
After the corrosion specimens have been cleaned and
weighed for the final time, the surface of each specimen is
examined for evidence of localized corrosion. A low-powered
microscope (5x to 50x) is a useful tool for examining coupons
and pipe inserts. The degree of attack, pit shape, pit density
(pits/sq in.), and pit depth (mils) are routinely recorded. Byars
& Gallop (18) published photos and terminology that are an
excellent guideline for describing the attack on coupons. Pipe
inserts are evaluated in the same manner as coupons but need
to be split lengthwise to permit visual and instrumental inspec-
tion of the specimen. A dial depth gauge is employed to meas-
ure the pit depth. The visual appearance and pitting
measurements are recorded on the specimen data form.
The pitting data can be equated with the results from other
studies by calculating the Pitting Rate Equivalent (PRE), which
is expressed as mils penetration per year (mpy) and is calcu-
lated by the following equation:
PRE (mpy) =
365 *d
t
where
d = maximum pit depth, thousandths of an inch
t = specimen exposure time, days
The Pitting Factor (PF), which also is used for this purpose,
is the ratio of deepest metal penetration by a single pit to the
average metal penetration as determined by the weight loss
measurement. A value of 1 represents uniform corrosion with
no pitting, whereas higher values indicate an increased pitting
tendency.
4.2.8 The Corrosion Rate Calculation
The corrosion rate is calculated from the recorded net
weight loss of a specimen and is reported in terms of average
surface penetration per specified time interval. This implies that
the metallic corrosion is linear with time, which is seldom true
with potable water. In most instances, the corrosion rate de-
creases with time as oxide films develop or as minerals deposit
on the metal. Reporting of corrosion rates also implies that the
weight loss is due to uniform corrosion and not to pitting,
dealloying, crevice corrosion, or other forms of localized cor-
rosion. This might or might not be true. It is important, there-
fore, that the specimens be inspected carefully before and after
cleaning to identify the presence of localized corrosion.
Corrosion data can be calculated and expressed as weight
loss per unit area per unit time or the equivalent rate of pene-
tration. The generally accepted units are grams per square meter
per day (g/m2/d) and millimeters penetration per year (mmpy).
The ISWS traditionally has used mils per year (mpy) for re-
porting corrosion rates.
The Corrosion Rate (CR) is calculated by the following
equation:
CR(mpy) =
W*F
A*T*D
where
W = weight loss of coupon during exposure, grams
A = exposed surface area of coupon, square inches
T = time coupon was exposed to water, days
D = density of metal coupon, grams per cubic centimeter
(from Table 4-15 and various handbooks)
F = factor for converting units of measurement into mpy, use
22,250 for units listed above
To convert units from mpy to mmpy, multiply the mpy
value by 0.0254.
Table 4-15. Density of Selected Metals
Metal
Density, g/cm
Brass, Red
Carbon Steel
Copper
Galvanized Steel or Zinc
Grey Cast Iron
Lead
Solder, 50Pb/50Sn
Stainless Steel, 316
8.75
7.86
8.94
7.13
7.20
11.33
9.32
7.98
Source: NACE Corrosion Engineer's Reference Book
4.2.9 Interpretation of the Corrosion Data
A planned-interval corrosion study conducted by the ISWS
(15) serves as an example of the interpretation of data. Copper
and galvanized steel corrosion specimens were installed for
various intervals over a 2-year period in different water sup-
plies. The corrosion specimens, identified by the letters A
through G, were exposed for the time span shown in Figure
4-6. Changes in water corrosivity and metal corrodibility were
evaluated for each water supply by examining the relationships
in the weight loss data using the Wachter and Treseder tech-
nique. Multiple relationships may be drawn from the differ-
ences between the various combinations of specimen weight
loss measurements. Table 4-16 summarizes some of these rela-
tionships and their significance.
48
-------
0 100200300400500600700800
DURATION OF CORROSION STUDY (Dayi)
A-G « Corrosion apadrmn*
Figure 4-6. Planned-interval pipe insert exposure during EPA/ISWS
corrosion study.
Table 4-16. Significance of Coupon Weight Loss Measurements
Coupon Weight Loss Results Significance
A = G or B = F
G < A or F < B
A < G or B < F
D-C = G or D-B = F
D-C < G or D-B < F
G < D-C or F < D-B
No change in water corroslvity
Decreased corrosivity
Increased corrosivity
No change in metal corrodibility
Decreased corrodibility
Increased corrodibility
A change in corrosivity and its effect on the weight loss of
galvanized steel specimens were observed at Site 302 during
the aforementioned study. Figure 4-7 shows the weight loss
data for the specimens. The change in corrosivity occurred
because the water utility was not satisfied with its corrosion
control program. The difference was most obvious between
specimens A and G. Specimen A was installed during the period
when a phosphate/zinc product (0.7 mg/L) was being applied.
At approximately the same time, specimen A was removed, the
phosphate/zinc treatment was discontinued, and caustic soda
(19 mg/L) was applied for the remainder of the study to main-
1500
1000
A —— A Tfnw from Start of Study
• • Urn* from End of Study
0 100 200 300 400 500 600 700 BOO
EXPOSURE TIME (Day»)
Figure 4-7. Corrosion of galvanized steel specimens at Site 302.
tain a pH of 8.5. This treatment strategy proved to be effective
in controlling a red water problem in the distribution system,
and the corrosion data indicate it was effective in reducing the
corrosion of galvanized pipe. The weight losses of specimens
F and E also were significantly less than the corresponding
weight losses of specimens B and C, which provide additional
confirmation that the water corrosivity changed during the
study.
The corrosivity of another water supply was found to be
relatively consistent throughout the same corrosion study. The
weight losses of copper specimens at Site 307 were nearly
linear (see Figure 4-8). There was no significant difference in
600
100
* A Time from Start of Study
• •••-• Tlm« from End of Study
0 100 200 300 400 500 600 700 800
EXPOSURE TIME (Days)
Figure 4-8. Corrosion of copper specimens at Site 307.
the weight losses of specimens A and G (or B and F), although
the specimens were exposed at different time spans during the
study. The lower weight loss found for specimen E might be
because of processing or metallurgical factors, since the other
six specimens provided consistent results. The use of replicate
specimens is recommended to reduce the uncertainties of a
single weight loss measurement.
In the previous example, the corrodibility of copper was
examined by the weight loss relationships outlined in Table
4-16. The weight loss of specimen G was greater than the
calculated difference between specimens D and C. The weight
loss of specimen F also was greater than the calculated D-B
value. Both comparisons indicate that the corrodibility of cop-
per decreased during the study. This is an anticipated result
since the corrosion or mineral deposition that occurs gradually
on the surface of coupons will tend to protect the underlying
metal and reduce the apparent corrosion rate.
Coupon test results can be used to compare the corrosivity
of various water sources or differences in the corrosivity within
a distribution system. Figure 4-9 illustrates the effect of time
on the corrosion rate for galvanized steel pipe specimens. The
corrosion rates at the two sites located in water supply A both
declined with time, although the corrosion rates were signifi-
cantly different. Site 302 was located at the water treatment
plant, whereas Site 304 was situated at a remote location in the
distribution system. Site 302 is the example cited previously as
49
-------
0.8
I
0.2
Witor Supply A
•*
Water Supply A
Witer Supply B
.•SITE 304
-» SITE 306
12 18
EXPOSURE TIME (Months)
24
Figure 4-9. Effect of water corrosivity on galvanized steel.
experiencing a change in corrosivity because of a change in
water treatment. The data underscore the importance of con-
ducting corrosion studies at different locations when evaluating
the effects of water treatment in a distribution system. The data
likely reflect the difference in the chemical equilibria of the two
sites.
Figure 4-9 also illustrates the corrosivity of two water
supplies and how they compare. Water supply B is a lime-sof-
tened ground water source using the calcium carbonate satura-
tion indices for corrosion control. Water supply A is a clarified
and filtered surface water source that tried both a zinc/phos-
phate and a caustic soda treatment to control corrosion. The
corrosion of galvanized steel attained equilibrium and a very
low corrosion rate within 6 months in water supply B. A much
longer interval was required in water supply A to reach equi-
librium, which will be at a higher corrosion rate than water
supply B.
4.2.10 Summary
The preceding examples demonstrate various techniques
for interpretation of the corrosion data obtained from coupon
weight loss measurements. Corrosion specimens can be very
effective tools for accessing the corrosivity of water or the
corrodibility of metals. Because many factors influence the
corrosion of plumbing materials in potable water, the variability
due to the preparation and processing of specimens must be
minimized. Pipe insert and/or coupon type specimens can pro-
duce valuable data for evaluating the effects of water treatment
on plumbing materials. The advantages and limitations of the
coupon weight loss procedure, however, must be considered
carefully in designing a corrosion study to meet this objective.
4.3 Design Considerations for Pipe Loop Testing
4.3.1 Introduction
Traditionally, corrosion pilot plants have consisted of
either static bench-scale immersion tests or flow-through loops
containing metal coupons for weight loss evaluations. Flow-
through pipe loops designed to evaluate the leaching charac-
teristics of a particular metal have not been as prevalent as these
other methods for evaluating corrosion in a system. With the
recent lead and copper rule comes a need to evaluate corrosion
control with respect to the leaching of these metals. Since
weight loss and metals concentrations have not been correlated
sufficiently, a pilot system that assesses the leaching potential
of the system might be most appropriate when evaluating op-
timum corrosion control treatment. Pipe loops can be used in
corrosion optimization studies in the following ways:
• To compare the impacts of various water qualities on metal
levels.
• To compare the ability of various treatments to reduce metals
levels.
• To evaluate the side effects of various treatments.
The following section provides a brief discussion of the
design, construction, operation, and data evaluation issues re-
lated to conducting pipe loop evaluations for determining op-
timal corrosion control.
4.3.2 Pipe Loop Design and Construction
Considerations
Several operating conditions should be considered when
designing a pipe loop system, including operating pressure,
flow, velocity, pipe diameter, length of loops, total through-put
volume, and on-off cycling. These factors should be controlled
to reduce the amount of variability in the test results. The
American Water Works Association Research Foundation (AW-
WARF) pipe rack model was designed by the Illinois State
Water Survey as part of the Lead Control Strategies manual
(4). The model was designed to enable the metal leaching char-
acteristics of various pipe materials to be evaluated under
equivalent operating conditions. (The Lead Control Strategies
manual contains a complete description of the pipe loop model.)
AWWARF is in the process of revising that initial protocol. The
following discussion lists several key issues that should be
incorporated in the design of the AWWARF pipe loop model
or other models designed to evaluate the leaching potential of
a particular system.
The lead source to be evaluated in the pipe loop should
reflect the sources of lead in the system. If lead service lines
are a major source of lead in a system, lead pipe of a similar
diameter should be incorporated into the pipe loop; a similar
process should be followed for soldered copper pipe. Brass
faucets also have been shown to contribute significantly to lead
levels measured in a standing 1-L sample from the tap. Incor-
porating brass faucets into the pipe loop system might be con-
sidered; the vast number of existing faucets and their varying
lead content, however, make the choice problematic. PVC
should be used for the remainder of the pipe loop to prevent
metals contamination from non-test loop piping sections. Tef-
lon® tape, rather than pipe-joint sealing compounds, should be
used to connect the PVC sections in the manifold.
50
-------
The variability of lead levels measured both at the tap and
from controlled pipe loop studies indicates that multiple loops
of the same material exposed to the same water quality condi-
tions should be incorporated. The AWWARF pipe rack design
includes three replicate loops of the same material. In a corro-
sion optimization study, several of these pipe racks with repli-
cate test loops must be run side-by-side to obtain a comparison
of the various treated water qualities.
Another factor in design of the pipe loop system is the
"location of the apparatus. The facility where these loops are
constructed must have adequate space, heat, power, water sup-
ply, and wastewater drain to accommodate several pipe racks
with replicate test loops of the materials of interest. It also is
important to recognize water quality changes when siting the
apparatus. The quality of water leaving the treatment plant
might be significantly different from that of the water that
reaches the residential units. Construction of the pipe rack can
be accomplished either by in-house or contract staff. The qual-
ity of workmanship should be similar to local plumbing con-
tractors. With soldered copper loops, the amount of solder used
for each loop should be recorded.
4.3.3 Pipe Loop Operational Considerations
4.3.3.1 Startup Issues
Prior to initiating pipe loop operations, several data collec-
tion and operations issues should be considered. Decisions must
be made about:
• What metals levels to evaluate
• The need for corrosion rate information
• Corrosion mechanism evaluations
• Collection of auxiliary water quality parameters
Metals from corrosion reactions might be present as dis-
solved aqueous species, minute colloidal or freshly precipitated
particles that are suspended in the water, or as fragments of
corrosion by-product films that have been eroded or removed
from the pipe by water flow. Knowledge of the form of the
metal and the relative fraction that is dissolved might be im-
portant for developing the optimum treatment. For lead or cop-
per corrosion control, the solubility must be decreased, and the
passivating film must adhere to the pipe.
From a regulatory standpoint, all of the metal is assumed
to be bioavailable, so differentiation between dissolved and
other forms is not necessarily critical for a pipe loop experi-
mental study. Modeling and predictions of metal solubility,
however, are based on establishment of equilibrium with the
dissolved species. For comparison to modeling predictions,
therefore, some type of isolation of the dissolved metal fraction
is necessary. This isolation can be done by complicated analyti-
cal techniques such as anodic stripping voltammetry or by ul-
trafiltration. More often, however, simple membrane filtration
is used, and the cutoff for what size particle is considered
"dissolved" is set at some level, such as 0.4 |0m. Since filtering
of samples removes paniculate metals, it reduces the variability,
allowing improved comparisons between loops.
Filtration appears to be simple, but it can create additional
problems if not done carefully. All materials in contact with the
water should be of plastic or Teflon® and the membranes them-
selves should be of polycarbonate. Filters should be rinsed with
sample water to satiate adsorption sites, with that volume of
water being discarded before the sample is collected. The fil-
tration apparatus should be cleaned thoroughly and acid-rinsed
between filtrations of different samples. In-line filtration during
sample collection generally is more desirable than vacuum fil-
tration.
Another decision to be made when planning a pipe loop
study is whether the corrosion rates are important Clearly, from
a materials performance perspective, lower corrosion rates are
desirable. Lead corrosion rates are relatively low compared to
other metals. Although these low corrosion rates produce trace
quantities of lead solution, they generally are significant
enough to produce health problems. Often, coupons or inserts
that are sections of pipe are used for weight loss determinations.
Soft materials, such as lead, might be extremely difficult to
process accurately for this purpose. Galvanically stimulated
corrosion processes, such as solder joint corrosion, are not eas-
ily amenable to evaluation by weight loss, although some pro-
cedures to do this have been developed (see Section 4.4).
Various electrochemical instruments provide instantaneous (or
nearly instantaneous) readings of corrosion rate. The data fre-
quently can be captured by computers for integration and plot-
ting.
Finally, the water quality characteristics of the source water
both before and after it is treated should be identified, as should
the accuracy of the chemical feed system. Depending on the
treatment, the chemical feed system can be monitored by evalu-
ating pH, alkalinity, concentration of inhibitor, calcium, and
disinfectant residual. Thorough records of the mechanical op-
erating conditions also should be maintained. These conditions
include flow, pressure, and through-put volume. It is advisable
to record the quantity of chemicals used so that comparison can
be made with measured quantities.
4.3.3.2 Sampling Issues
Prior to initiating normal sampling, it might be advisable
to characterize the usable sample volume in each loop, since
mixing fresh influent water with the stagnant water in the loops
might dilute the metals concentrations. This characterization
can be accomplished by collecting a series of small volume
samples (25 to 50 mL) from the loop after a designated standing
time and measuring each sample for metals levels. A "profile"
of lead levels can be obtained, and the usable volume in the
loop can be determined. The sampling protocol for standing
samples in the pipe loop study then can be organized to fit the
total usable volume contained in each loop.
Utilities or researchers with a more scientific orientation
might be interested in trying to understand the mechanisms of
corrosion and inhibitor performance to better predict treatment
goals. The sampling and analysis program must be configured
51
-------
to include special procedures for sampling and chemical analy-
gcs when this information is desired. Measurements uf Jeple-
tion of dissolved oxygen or chlorine, reduction in inhibitor
concentration, increases in pH caused by corrosion reaction,
and changes in metal speciation (i.e., Cu+VCu"1"2, Fe+3) are
examples of analyses that could give insight into corrosion and
treatment mechanisms.
Obtaining accurate information on initiation of stagnation
conditions is important in determining the rate and extent of
metal leaching from the plumbing materials. Whenever the
water source for the pipe rack system is inconsistent, the back-
ground samples should be taken during the flow period imme-
diately before the standing time and as close to the shutoff time
as is practical. Valuable information also is obtained if a sample
is taken of the metal or metals of interest from the loop sam-
pling tap during the flowing period. This shows how much
metal is picked up when water travels through the pipe.
Several special precautions must be taken when certain
sensitive analyses are to be performed on samples in the cor-
rosion study. For example, samples collected for pH analysis
should be taken in closed containers with no air space. The pH
will change, sometimes radically, if the samples come into con-
tact with air. The amount of die change will depend on the
characteristic of the individual water and its buffer intensity.
Waters particularly susceptible to pH drift are low-alkalinity
water with a pH greater than 8, moderate-alkalinity waters of
high pH, and very high-alkalinity waters of low pH where
oxygen and carbon dioxide degassing occur. During analysis,
pH measurements should be made directly on water in the
original container with minimal air contact. Using 25- to 40-mL
glass sample vials with caps having conical polyethylene inserts
has been found to be quite useful. A rubber stopper around the
electrode will enable die samples to be protected from the air.
Sealed containers such as these might enable the preservation
of pH for hours or days. The stability of pH, however, must be
determined for each water supply to be tested. It is possible that
the sensitivity to pH change from atmospheric contact will
differ among waters representing one treatment or another.
Similar to analyses of pH, direct analyses of dissolved
inorganic carbonate (DIG) require that samples be taken with
little disturbance and sealed in bottles with no air space. Dis-
solved oxygen also requires special collection precautions. Spe-
cial preservation requirements for other analytes should be
determined by consulting the laboratory and standard analytical
procedure references for the test of interest.
When metal speciation analysis such as sample filtration
is to be performed, the samples must not be acidified until after
the separation has been done.
Finally, when determining the frequency of sampling and
length of study period, the following interrelated factors should
be considered:
• The analytical precision of the results
• The natural variability of the parameters to be measured
• The length of the test
• The availability of staff and laboratory resources
The length of time to operate a pipe loop system to obtain
stable data for making comparisons will depend on the material
used and the influent water chemistry to each loop.
4.3.4 Characteristics of Pipe Loop Data
Leaching data collected from actual pipe loop studies dis-
plays an intrinsic variability in lead and copper levels. This
variability limits the certainty with which extrapolation of re-
sults from the pipe loop to distribution system standing samples
can be made. The use of new materials and the operating con-
ditions with which these materials are exposed in a pipe loop
system can create film characteristics that rarely represent con-
ditions in the field. These issues place significant constraints
on estimating "optimum" corrosion treatment.
4.3.5 Data Evaluation Considerations
Statistical evaluation of data has commonly been per-
formed using parametric statistics, with which there is wide-
spread familiarity and for which software is readily available.
With parametric statistics, there is a strong assumption that the
probability density function of the data is normal or bell-shaped
(nonskewed). For data distributions that are non-normal
(skewed) or for which there are so few data points that the
distribution cannot be determined, nonparametric statistical
techniques are more statistically efficient. These techniques are
not as widely recognized as parametric or normal distribution
techniques, and until recently, few software packages incorpo-
rated these methods. The simple comparison between paramet-
ric and nonparametric statistics is as follows:
Parametric Nonparametric
Interested in levels Interested in location
Use mean and standard deviation Use median and percentile
T-test Ranked Sign Test
Lead levels from both pipe loop data and standing samples
collected at the tap typically display non-normal (skewed) dis-
tributions that, in many cases, appear to be log-normal. In ad-
dition, pipe loop studies can provide a limited number of data
points with which to evaluate treatments because of the length
of time needed to obtain relatively stable metals levels. Either
of these circumstances suggests that nonparametric techniques
should be used when evaluating the data from a pipe loop study.
4.4 Electrochemical Methodologies for Corrosion
Measurement in the Distribution System
Electrochemical corrosion assessment techniques have
been sufficiently developed to provide a useful tool for corro-
sion control optimization programs. Electrochemical corrosion
assessment is a direct assessment methodology, the type em-
phasized in the lead and copper rule. Calcium carbonate-based
saturation indices, in contrast, are indirect assessment method-
ologies and are inadequate to characterize corrosion processes.
52
-------
The ObviOUS reason for Using electrochemical techniques
shifting potential r»f fop rnrmcinn
the:
is that they provide a nearly instantaneous measure of the un-
derlying corrosion process. They provide a snapshot of what is
taking place on the surface of the corrosion specimen at a
particular time. They are not a cumulative measure, so they are
particularly useful for many process control operations or
screening programs. Once the instrumentation is in place, per-
forming an electrochemical screening program is a rather inex-
pensive process.
Electrochemical techniques can be used with a variety of
different analyses. Assessing corrosion rate could certainly be
high on any system operator's list, but these techniques also
can be used for analyzing passivity phenomena, coating effec-
tiveness, pitting susceptibility, galvanic interactions, and inhibi-
tor evaluations. Electrochemical techniques, however, cannot
measure directly the underlying corrosion current between the
oxidated and the reductant couple, because both the oxidation
and the reduction are taking place on the same sample, perhaps
within a few micrometers or microns of each other. The reduc-
tant couple cannot be isolated; hence, it is not possible to meas-
ure the corrosion current directly. The assessment of the actual
surface potential of the corrosion specimen also is not adequate
to define the corrosion process.
The goal hi performing an electrochemical evaluation of a
particular specimen is to obtain an Evans diagram or a Koppel
plot. In effect, a current is applied to a specimen, a piece of
metal that is presumably at equilibrium. The current is applied
in both the anodic and cathodic direction. The piece of metal
is perturbed by this current, and the offset hi potential, brought
about by those respective anodic and cathodic currents, is meas-
ured. The Koppel plot, which can be developed from these
measurements, graphically demonstrates that the intersection of
the anodic Koppel slopes and the cathodic Koppel slopes
should yield the underlying corrosion current.
Coupon testing is the definitive measure of the actual cor-
rosion rate. Other analytical techniques, including electro-
chemical methodologies, must be compared to some reference,
and a suggested reference is coupon testing using either the flat
coupon technique and the Illinois State Water Survey (ISWS)
technique, or modifications to the ISWS technique using actual
pipe. Making an assessment of a corrosion rate based on a
single specimen or even two or three specimens simply is not
adequate. The variation hi corrosion, most importantly on
steels, is so large that multiple coupon exposures are required.
This definitive technique for corrosion assessment requires 3
to 6 months to carry out.
4.4.1 Polarization Techniques
A potentiostat is required for electrochemical testing. This
device is capable of measuring the surface potential of the
corrosion specimen to an accuracy in the millivolt range and
simultaneously controlling the impressed current applied to a
specimen in the micro range. The cost for a potentiostat is
several thousand dollars. The polarization cell is the device
used to hold the corrosion specimen, consisting of three differ-
ent elements: the test specimen, reference electrode, and
counter electrode. The reference electrode is used to compare
reference electrode has a stable electrochemical potential, all
changes are measured relative to that potential.
Some initial work has been performed at the University of
Washington and subsequently at the University of North Caro-
lina at Charlotte to develop polarization cells that are specific
for the distribution network. The goal of this work was to
develop a simple system that uses plumbing materials as the
actual electrode surface and that has a cell geometry that can
reflect the hydrodynamics of pipe flow. The polarization cell
developed consists of two pieces of plastic that hold a test
specimen between them. The auxiliary electrode penetrates the
test specimen axially, and the reference electrode is directly
above. It is relatively inexpensive to produce, because all that
is required is some plastic machining, a reference electrode, and
auxiliary electrodes. Potentiostats and corrosion monitoring
equipment essentially can be compressed into a single circuit
board, or a single board that fits inside a laptop computer. The
cost to purchase them as a package varies between $8,000 and
$15,000. No unusual laboratory facilities are required. Most
laboratories probably are doing part of their own electrochemi-
cal analyses already. The metal specimens can be fabricated
easily or purchased from one of several different companies
that produce fabricated metal specimens.
Rather than performing a full-blown potentiodynamics
scan of a surface, a linear polarization can be performed. Linear
polarization relies on only a single offset, which shifts surface
potential of the corrosion specimen by 10 or 20 millibles; a
reading of the impressed current is produced at that point. If
Koppel slopes are known for a corroding specimen, then an
interpretation of the corrosion rates can be made from a single
measurement. There are two-electrode and three-electrode vari-
ations on that theme, both of which have worked fairly well.
Linear polarization instrumentation has the advantages of
speed, simplicity, relative low cost, and relative accuracy. In
this context, relative accuracy means that an assessment can be
made to determine whether there are increases or decreases in
the corrosion rate; however, it has low absolute accuracy. A
potentiodynamic scan must be performed if a highly accurate
electrochemical assessment is needed hi terms of absolute ter-
minology.
There is a developing polarization technique that is referred
to as alternating current (AC) impedance. The AC impedance
technique is just now beginning to be applied in distribution
networks. The technique actually has been available for about
a decade and has been used extensively in other industries.
AC impedance does not use direct current; it applies an
alternating current' to the corrosion specimen. The alternating
current and the subsequent perturbation of the corrosion speci-
men produce a variety of information, including the corrosion
rate. An electrical model of the corrosion surface is constructed,
taking into account the different resistances on that surface.
Most importantly, there is very little distortion of the surface
chemistry when AC impedance techniques are used. This is
important in water distribution networks because the other po-
larization techniques often apply or create a potential shift of
53
-------
50 or 100 millibles. This is a substantial change in the nature
of the corrosion surface and alters the underlying electrochem-
istry of the surface. AC impedance gets around this problem by
applying an alternating current that has an offset no greater,
generally, than 5 millibles. The disadvantages of the system of
AC impedance is that it is still an emerging technology, at least
for water distribution systems, and requires fairly extensive
instrumentation. The instrumentation is being reduced to a sin-
gle computer control system, however, and prices are becoming
more affordable. Today an AC impedance system can be pur-
chased for around $20,000 to $25,000.
Electrochemical techniques are best restricted to copper
and its alloys, including brass and bronze. Electrochemical
techniques also work well on lead and lead-tin solders and
tin-antimony and tin-silver solders. The absolute accuracy can
be low unless a rather involved potentiodynamic scan is per-
formed. The data interpretation used to be difficult, but the
reduction is automated by the statistical software available with
the package units. Improved electrochemical techniques are
pending. AC impedance soon will be available for water distri-
bution systems, and widespread application will improve as-
sessment technology overall for the distribution system.
4.4,2 Electrical Resistance and Electrochemical Noise
In addition to polarization techniques, other forms of elec-
trochemical methodologies include electrical resistance and
electrochemical noise. Electrical resistance measures the
change in the resistance of an element that is exposed to the
flow screen. As that element corrodes, its cross-sectional area
changes; hence, its resistivity changes, and that resistivity is
related to a corrosion rate. Advantages of electrical resistance
measurement are that it is very simple, well suited to on-line
measurements, and somewhat sensitive to long-term changes.
A disadvantage is that it is limited to rapidly corroding systems.
Although many systems might believe that they have a serious
corrosion control problem, most water distribution networks
have minimal corrosion rates, at least compared to other indus-
tries, such as the petrochemical industry, where many of these
techniques were developed. The electrochemical resistance in-
strumentation is relatively straightforward.
The electrochemical noise technique uses sensitive infor-
mation to measure the static electricity generated on the corro-
sion surface. On a molecular level, it is possible to interpret a
corrosion rate by the rate of individual molecular events occur-
ring on the surface.
4.4.3 Summary
Electrochemical techniques should be used only when sup-
plemented by other investigative techniques, specifically, the
definitive corrosion assessment techniques of gravimetric cou-
pons and weight loss techniques. Because of their speed, elec-
trochemical techniques can be ideal for screening programs.
Correlation of the electrochemical results with field results can
yield fast and realistic predictive procedures, but only after
electrochemical techniques have been properly calibrated.
Electrochemical measures generally are limited to uni-
formly corroding surfaces. Many attempts have been made to
perform these measurements on pitting surfaces such as galva-
nized steel. The success with such surfaces has been much more
limited than with the uniformly corroding surfaces of copper,
copper alloys, and other materials. Linear polarization might be
a particularly appropriate technique for on-line continuous
measures of finished water corrosivity, specifically for screen-
ing programs and process control. Potentiodynamic techniques
will remain necessary for measurement of absolute corrosion
rates.
4.5 References
1. Wysock, B.M. et al. (1991). A Study of the Effect of Mu-
nicipal Ion Exchange Softening on the Corrosion of Lead,
Copper and Iron in Water Systems. Proc. Annual Conf.
American Water Works Association, Denver, CO.
2. AWWA Research Foundation (1987). Deterioration of
Water Quality in Distribution Systems. American Water
Works Association, Denver, CO.
3. AWWA Research Foundation (1990a). Chemistry of Cor-
rosion Inhibitors in Potable Water. American Water Works
Association, Denver, CO.
4. AWWA Research Foundation (1990b). Lead Control
Strategies. AWWA Research Foundation and American
Water Works Association, Denver, CO.
5. Holm, T.R. and S.H. Smothers. (1990). Characterizing the
Lead-Complexing Properties of Polyphosphate Water
Treatment Products by Cometing-Ligand Spectro-
photometry Using 4-(2-Pyridylazo)Resorcinol. Intern. J.
Environ. Anal. Chem. 41:71.
6. Dges, A. (1991). Control of Lead and Copper in Drinking
Water Champlain Water District Presentation Outline.
Trans. EPA/AWWA National Workshop on Corrosion Con-
trol. American Water Works Association, Denver, CO.
7. Rohlf, F.J. and R.R. Sokal (1981). Biometry: The Princi-
ples and Practice of Statistics in Biological Research, 2nd
Edition. W.H. Freeman and Co., New York, NY.
8. Schock, M.R. (1990). Causes of Temporal Variability of
Lead in Domestic Plumbing Systems. Environmental
Monitoring and Assessment. 15:59.
9. ASTM (1991a). Standard Test Methods for Corrosivity of
Water in the Absence of Heat Transfer (Weight Loss Meth-
ods). American Society of Testing and Materials, Designa-
tion D 2688. Annual Book of ASTM Standards, Vol. 11.01.
Philadelphia, PA.
10. ASTM (1991b). Standard Recommended Practice for Pre-
paring, Cleaning, and Evaluating Corrosion Test Speci-
mens. American Society of Testing and Materials,
Designation Gl. Annual Book of ASTM Standards, Vol.
3.01. Philadelphia, PA.
54
-------
11 NACE (1976). Laboratory Corrosion Testing of Metals for 15. Ncff, Schock, and Mardgn (1987). Relationships Between
the Process Industries. National Association of Corrosion
Engineers, Standard TM0169. Houston, TX.
12. ASTM (199 Ic). Standard Recommended Practice for Ex-
amination and Evaluation of Pitting Corrosion. American
Society of Testing and Materials, Designation G46. Annual
Book of ASTM Standards, Vol. 3.01. Philadelphia, PA.
13.-Ailor, W.H., ed. (1971). Handbook of Corrosion Testing
and Evaluation. The Electrochemical Society and John
Wiley & Sons, New York, NY.
14. Haynes and Baboian, eds. (1985). Laboratory Corrosion
Tests and Standards. American Society of Testing and Ma-
terials, Standard Technical Presentation 866. Philadelphia,
PA.
Water Quality and Corrosion of Plumbing Materials in
Buildings, Vol. I: Galvanized Steel and Copper Plumbing
Systems. EPA/600/S2-87/036.
16. Wachter, A. and R.S. Treseder (1947). Corrosion Testing
Evaluation of Metals for Process Equipment. Chemical
Engineering Progress 43:316-326.
17. Fontana, M.G. and N.D. Greene (1967). Corrosion Engi-
neering. McGraw-Hill Book Company, New York, NY.
18. Byars, H.G. and B.R. Gallop (1975). An Approach to the
Reporting and Evaluation of Corrosion Coupon Results.
Materials Performance, pp. 9-16 (Nov.).
55
-------
Chapter 5
Control Strategies
This chapter provides an overview of control strategies for
lead and copper in drinking water. Control strategies can consist
of any combination of materials selection, materials removal,
point-of-use devices, and chemical water treatment. Chemical
water treatment programs consist of either manipulating the
general water chemistry (such as pH, hardness, and inorganic
carbonate) or-adding a chemical or chemicals (silicates, ortho-
phosphates, or blended phosphates) to the water to produce a
less corrosive water quality.
Usually, chemical treatment employs one of two strategies:
the formation of a coating on the pipe that slows the corrosion
of the underlying pipe or the formation of a relatively insoluble
"passivating" film with the pipe metal itself. Frequently, both
approaches must be used simultaneously. The treatment pro-
gram must consider the nature of the lead or copper contami-
nation source, the initial water chemistry, and the chemistry of
the treatment chemical when dissolved in the water.
This chapter also examines the secondary effects of con-
trolling corrosion through carbonate stability or the use of in-
hibitors. For example, adjusting the pH can benefit oxidation
and coagulation, but it also can hinder a utility in its effort to
comply with the Surface Water Treatment Rule (SWTR) and
can enhance the formation of disinfection by-products. Inhibi-
tor usage can promote algal growth and might create waste
discharge problems due to zinc. Use of nonleaded plumbing
materials can introduce other undesirable contaminants, such
as antimony, into the water.
Finally, this chapter presents five case studies in corrosion
control:
• Sodium silicate for the simultaneous control of lead-, cop-
per-, and iron-based corrosion: York, Maine.
• Assessing zinc orthophosphate vs. pH adjustment: Cham-
plain, Vermont.
• Reducing corrosion products in municipal water supplies:
Chippewa Falls, Wisconsin.
• Evaluating chemical treatment to reduce lead in a building.
• Iowa's lead in schools' drinking water program.
5.1 Overview of Control Strategies for Lead in
Drinking Water
Control strategies for lead in drinking water generally fall
into three categories: physical, point-of-use, and chemical treat-
ment control. Physical control is the removal of lead-containing
materials or the limiting of lead content in materials. Point-of-
use (POU) control is the use of devices attached to water taps
or in lines near water outlets. These devices include filter units,
ion-exchangers, reverse-osmosis units, or adsorber cartridges.
POU control is effective only when the source of lead is located
prior to the device. Many POU devices have terminal brass
faucets or soldered joints and therefore are not effective for lead
removal. In some cases, the devices reintroduce the problem in
a more aggressive (corrosive) water. Chemical treatment means
either that the water has been treated as it comes from the plant,
or that chemical treatment has been used in a building. This
chapter looks only at chemical treatment strategies and the
major treatment chemicals used to apply them to distribution
systems.
5.1.1 Chemical Treatment Strategies
Two modes of effective chemical treatment can be used to
limit lead contamination. The use ofsurficial coatings seals the
surface from interaction with water to prevent either migration
of solubilizing agents into lead-containing materials or migra-
tion of lead out of materials. Alternatively, the creation of pas-
sivating films relies on altering the chemical properties of the
water to form relatively insoluble compounds with lead from
the plumbing material to render the lead relatively immobile.
5.1.1.1 Surficial Coatings
Three categories of surficial coatings can be created, either
naturally or by central chemical water treatment. These are
natural diffusion barriers, calcium carbonate deposition, and
silicate addition.
Natural Diffusion Barriers. Natural diffusion barriers can
consist of a variety of insoluble materials that coat the pipe
surface by means of precipitation reactions within the distribu-
tion system, caused by some chemical imbalance in the source
water or after treatment processes. These solids may be alumi-
num hydroxides or silicates, coming from residual aluminum
present from coagulation. Solids also include magnesium am-
monium phosphate, magnesium silicate, or manganese dioxide,
57
-------
which result from other aspects of water treatment. Natural
diffusion barriers also can be made up of adsorbed organic
material, ferric oxyhydroxides, or a combination of these ma-
terials in colloidal form, which adhere to the interior pipe walls.
The iron can either occur naturally in the water or from the
corrosion of kon mains in the distribution system.
A "stagnation curve" describes diffusion of lead from the
corroding or dissolving surface into the water contained in pipe
when the water is not flowing (2,8,11). Figure 5-1 shows two
ideal stagnation curves computed for a water with an alkalinity
of 30 mg/L calcium carbonate (CaCO3) and a pH of 8 at a
temperature of 25°C. The diffusion barriers function by chang-
ing the shape of the stagnation curve, making the slope of the
initial limb of the curve much shallower. Because of these
diffusion barriers, it takes much longer for the water to attain
equilibrium levels. The standing time represented in most sam-
pling programs is insufficient to allow equilibrium to be at-
tained. Therefore, the amount of lead often tends to be lower
than would even be predicted by the unadjusted stagnation
curve equations.
hydroxide ion generation originating in the corrosion. The de-
gree to which that is the case is a function of the buffering
intensity of the water. Further, the distribution of corrosion cells
is not uniform across the surface of the pipe, so localized spots
of precipitation might exist.
It is important to understand that pH, alkalinity, and dis-
solved inorganic carbonate (DIG) are interrelated (2,11,13).
That is, as you change pH, you change alkalinity, and vice
versa. The variables actually needed to define the conditions of
a water system are pH and DIG, which are linearly related. In
other words, for any given pH, total alkalinity represents a
unique concentration of DIG. Similarly, for the same DIG, the
corresponding total alkalinity changes with pH. Figure 5-2 il-
lustrates the relationship for a hypothetical situation of a total
alkalinity of 25 mg/L CaCO3 at an ionic strength of 0.005 and
a temperature of 25°C. At pH 10, this represents a DIG of 3.4
mg/L carbon, while at pH 6 the same alkalinity is generated by
a DIG of 18.5 mg/L carbon.
Figure 5-1. Stagnation lead levels.
Calcium Carbonate Saturation Considerations. Attempt-
ing to coat pipes with CaCO3 to seal them from corrosion is
historically the most common approach used by utilities. Un-
fortunately, little evidence exists to show that it works. An
underlying (incorrect) assumption in the water treatment field
states that corrosion rates and metal release into the water is
somehow proportional to the amount of oversaturation or un-
dersaturation with CaCO3 in the water. In some cases, particu-
larly in lime-soda softened waters of high pH, uniform thick
CaCO3 films are observed on pipes. These coatings also can
have a small component of silica and lead corrosion product
solid. In the absence of true CaCO3 supersaturation, no chemi-
cal link exists between CaCO3, as measured by a variety of
indices, and corrosivity towards lead.
To actually form CaCO3 barriers, several optimum water
conditioning issues must be satisfied. The first issue is that,
when a water is sampled for analysis, the water conditions do
not necessarily represent the conditions at the surface of the
pipe where corrosion actually occurs. At the pipe surface, the
pH is somewhat higher than in the bulk solution, resulting from
200
180 J M°C.I-M05
PH
6.0 —
5 140 {. T.O--
8.0""
».0—
180. .
< 120' •
N+-
/ ;.^-'
x.<"V'
:•*>"
5 loot 10-0~'
e
i Mt
.X<1<^
10
20 30
mg C/L INORGANIC CARBON
Figure 5-2. Alkalinity/DIC relationships.
An important implication of these relationships is that wa-
ters of low pH and low alkalinity might not necessarily also
have low DIG. Thus, pH adjustment for lead and copper control
might be adequate, without additional carbonate supplementa-
tion through the addition of sodium carbonate or bicarbonate
chemicals. The central question to be determined before treat-
ment is whether a given water has enough DIG to provide
adequate buffer intensity at the targeted pH after adjustment
To form a protective CaC03 pipe coating, a water must
have sufficient available mass of calcium and carbonate species
for precipitation. Enough calcium and carbonate ion must be
delivered to the surface .to create the necessary bulk of a good
coating. It follows that good coatings are likely to be found
only in relatively hard waters, in appropriate total alkalinity and
pH ranges.
A second issue to be considered is what physical or opera-
tional steps must be taken to achieve optimum water conditions
for lead control. A key is to quantify achievable conditions in
the most reliable manner, traditionally by some CaCO3 corro-
sion index or empirical test.
58
-------
Saturation Index DelerMnallon. Historically, the must Tabl» 5.1. Langftlier IndPY (I I) vs Halr-ium OarhnnatR Prnnipitatinn
commonly used index is the Langelier Index (1,2). Basically,
the Langelier Index is an estimate of the thermodynamic driv-
ing force for either precipitation or dissolution of calcium car-
bonate (2). Generally, three forms of the Langelier Index are
found in the literature: approximation, quadratic, and "satura-
tion index" forms. The Langelier Index is defined by the simple
relationship:
Potential (CCPP)
Same Langelier Index
where pHsat represents the theoretical pH at saturation equilib-
rium with calcium carbonate (calcite form), and pHact is the
actual measured pH of the water. Most reported values for a
Langelier Index have been computed using one of the numerous
simplified expressions. Many approximation forms make com-
promises in assuming temperature, ionic strength, and the ab-
sence of significant side reactions with calcium carbonate,
calcium bicarbonate, magnesium carbonate, calcium sulfate,
and other soluble ion pairs. Frequently, these assumptions are
based on numerically outdated or erroneous values for CaCOs
solubility constants (1,2).
The quadratic form (1,2,3) is more precise. It avoids prob-
lems in some mathematical configurations of the Langelier In-
dex where there is a sign change at high pH and the positive
index then represents undersaturated conditions.
The "Saturation Index" (sometimes called a "Disequili-
brium Index") approach has its origins mainly in the geochemi-
cal literature (2). It is a generalized formulation that compares
ion activity products to thermodynamic equilibrium solubility
constants for the given water chemical conditions. It allows for
correction for ion pairs and complexes in a general way, and is
particularly amenable to calculation on personal computers us-
ing a variety of chemical modeling programs that are widely
available (1). The saturation index expression for CaCO3 is
shown in the equation below. The curved braces { } represent
the activity of the ions in solution.
CaC03(s) =
f{Ca21{C032-}]
'I ^ J
The major problem with the Langelier Index and the Satu-
ration Index for estimating the potential for developing surficial
coatings is that they do not clearly quantify the mass available
for precipitation. To overcome this problem, the calcium car-
bonate precipitation potential (CCPP) (1,2,3) was developed. It
is mathematically more complicated than the Langelier Index
and Saturation Index, but with the widespread availability of
programmable calculators and computers, this is not a signifi-
cant problem.
Table 5-1 presents the advantage of CCPP over the Lan-
gelier Index. Comparing the example of a relatively soft water
at high pH with a hard water at low pH shows that the Langelier
Index is the same for both waters. The CCPP is much higher
Parameter
Temperature (°C)
Alkalinity (mg/L)
Calcium (mg/L)
TDS (mg/L)
pH (units)
LI (units)
CCPP (mg/L)
Soft Water
(high pH)
15
25
17
75
6.90
0.10
0.40
Hard Water
(low pH)
15
350
130
750
7.03
0.10
15
for the hard water with the higher alkalinity. In this example,
a factor of almost 40 times more CaCO3 is predicted to be
precipitated from the hard water than from the soft water.
Like the Langelier Index, the CCPP also assumes for-
mation of pure CaCO3 (calcite form) and no kinetic barriers
to deposition. A problem arises in cases where some cations,
such as magnesium, copper, or zinc, might inhibit the for-
mation of well-ordered calcite (CaCO3). Similarly, certain
anions, such as ortho- or polyphosphates, might inhibit the
formation of calcite. These "natural inhibitors" reduce coat-
ings through interaction with growth of crystal nuclei, pos-
sibly by creating distortions of the crystal lattice formation.
Other likely inhibitive mechanisms are by complexation or
sequestration of calcium.
None of the published forms of the Langelier Index or
CCPP can take into account these inhibitory factors, particu-
larly the presence of polyphosphates. Therefore, hi systems
containing polyphosphate either for corrosion control or for
prevention of unwanted calcium carbonate deposition, calcula-
tion of any of the widely published indices of calcium carbonate
saturation or precipitation is invalid.
The third control method is the old empirical test, com-
monly referred to as the "marble" test (1). This test can be done
in several different ways. Empirical tests such as the marble
test are the only valid ways to assess calcium carbonate disso-
lution or solubility potential in the presence of polyphosphates
or some other inhibiting ions.
Silicate Addition. Another chemical approach to creating
surficial coatings is silicate addition. Silicate species, when
present in sufficient concentration under the appropriate water
chemistry conditions, can adsorb to pipe surfaces to create a
film. Sometimes, the silicate operates in conjunction with other
metals present in the water, forming colloidal species that can
adhere to pipe surfaces. Silicate can also react slowly with
existing carbonate, basic carbonate, or oxyhydroxide corrosion
products, either to form fewer soluble reaction products on the
pipe surface or to bind existing corrosion products into more
uniform surface deposits. In this mode of action, the silicate
might act more like a grout or cementing agent.
59
-------
100.0
A second way silicate might function to reduce corrosion
is to buffer against hydroxide production at high pH, because
it can produce increased buffering intensity for the typical soft
water. The role of silicate in augmenting buffer intensity has
been described by Snoeyink and Jenkins (13). Hydroxide ion
production is a normal by-product of most metal corrosion
oxidation reactions in potable waters, so that limiting its pro-
duction would tend to stifle corrosion.
Relatively little quantitative information exists to predict
the effect of silicate addition. It is clear, however, that its ef-
fectiveness depends on pH, silicate concentration, and hard-
ness. Silicate treatment chemicals are added in polymeric form
as a highly viscous, basic chemical. In water, soluble
monomeric silica acts as a diprotic acid, having a negatively
charged species at high pH. Silicate reaction can be relatively
slow. Silicate addition also might need the presence of existing
corrosion by-product films to work. This becomes a complicat-
ing issue when evaluating corrosion treatments in pipe loop
systems or through coupon tests. Because experimental systems
are ordinarily made with new materials, silicates might not give
the same results in actual distribution system use relative to
other experimentally tested treatments that react more readily
with fresh metal surfaces.
5.1.1.2 Passivating Film Formation
Conceptual Approach. The driving concept in the forma-
tion of passivating films is to adjust water quality to form the
most thermodynamically stable phase possible. Even in pure
water with only carbonate species, such as HCO3- and CO32-
ions, lead chemistry is very complicated. Lead forms soluble
complexes such as PbCO3° and PtyCC^l-, in addition to hy-
droxide complexes, depending on pH and carbonate concentra-
tion. Figures 5-3 and 5-4 show lead species distribution in
equilibrium with normal and basic lead carbonate for waters
with relatively low and high DIG concentrations. The figures
show that in neutral to slightly basic systems, as in most treated
drinking water systems, lead exists in complexed forms as
PbCO3° and Pb(OH)2°. The ultimate significance of these com-
plexes can be illustrated by looking at a three-dimensional solu-
100.0
90.0
80.0
70.0
!
! 80.0
t
I
i 50.0
i 40.0
30.0
20.0
10.0
0.0 6o
\
"""a «.---.
..... * ......
7.0
9.0
10.0
11.9
PH
Figure 5-3. Lead speciation for 25°C, Ionic strength (I) = 0.01, dis-
solved inorganic carbonate (DIG) = 3 mg/L.
90.0
10.0
11.0
Figure 5-4. Lead speciation for 25°C, I = 0.01, DIG = 50 mg/L
bility diagram for lead (Figure 5-5). The figure shows that, at
very low lead concentrations, lead is very soluble. However,
the addition of small amounts of carbonate drastically reduces
lead solubility, particularly above pH 8.5 to 9. Further increases
in carbonate levels, however, cause resolubilization of lead be-
cause of the formation of PbHCO3+, PbCO3° , and Pb(CO3)22-
complexes.
0.17
log mg PDA.
pH
mgC«COjA.AHt
Figure 5-5. Lead solubility (I = 0.01, 25°C).
Lead also can form very insoluble orthophosphate com-
pounds, particularly Pb5(PO4)3OH(s) and Pb3(PO4)2(s)
(2,10,11). These orthophosphate solids are less soluble than
Pb3(CO3)2(OH)2(s) (basic lead carbonate, hydrocerussite) or
PbCO3(s) below a pH of about 8. Figures 5-6 to 5-8 show
solubility diagrams for the addition of 0 to 5 mg/L PO4 of
orthophosphate to waters of different total alkalinities, at pHs
of 7.0, 7.5, and 8.0. For these figures, all alkalinity is assumed
to be contributed by a carbonate species or a hydroxide ion,
through the relationship:
TALK = 2[CO3 2~] + [HCO3 "] + [OH~] - [H+]
60
-------
n
thp. hrarlrpts [] represent concentrations in mol/L
These figures are explained further in the section on orthophos-
phate addition.
Control of lead by solubility considerations follows one of
four approaches.
pH Adjustment. For many waters, merely adjusting the pH
is adequate. This adjustment might succeed by decreasing equi-
librium lead solubility to an acceptable level or by reducing the
diffusion rate of lead into solution so that lead levels are low-
ered under the usage patterns of most consumers. The use of
pH adjustment also might be adequate to reduce the lead leach-
ing from soldered joints or brass materials to acceptable levels.
pH/Alkalinity/DIC Adjustment. In some waters, both pH
and DIG need to be adjusted. One reason DIG adjustment is
useful is to decrease lead solubility in conjunction with pH. The
other equally important reason is to provide enough carbonate
concentration to give the water a higher buffering intensity to
help maintain desired pH throughout the distribution system.
Many treatments fail because the pH of water in the distribution
system drops substantially below the pH at which the water left
the plant, rendering conditions unsuitable for the formation of
passivating films on the pipe.
Orthophosphate Addition. Orthophosphate addition has
been shown to be extremely effective in 10 years of application
in Great Britain and Scotland (4,5,12). Published literature from
the United States is more ambiguous, but the utilities reporting
poor results in reducing lead levels almost always use improper
control conditions. The utilities usually are operating in an
incorrect pH range or at an insufficient Orthophosphate dosage
to maintain an adequate level for keeping lead solubility low
in all parts of the distribution system.
Several important factors govern the effectiveness of Or-
thophosphate addition. Effectiveness strongly depends on pH,
DIG, and ortbophosphate dosage; it probably also is influenced
by temperature, but this factor has not been quantified precisely.
Figures 5-6 through 5-8 show, for example, that for treatment
at pH 7.5, the lead level is reduced significantly by the addition
of the first 0.5 to 1.0 mg/L of Orthophosphate as PO4. Additional
dosage has relatively less effect, particularly above approxi-
mately 3 mg/L PO4.
The optimum pH for solubility reduction by Orthophos-
phate also depends on the background DIC/alkaHnity of the
water (2,10,11). Figures 5-6, 5-7, and 5-8 also show that for
higher alkalinities, the level of lead achievable (in terms of
equilibrium solubility) is not as low. For waters with high al-
kalinity, however, Orthophosphate dosage provides much
greater reduction in lead concentration than is possible with pH
and alkalinity adjustment alone.
The dosages of Orthophosphate possible might be limited
by the calcium hardness of the water. Depending on the pH,
hardness, and Orthophosphate dosage, a solid such as octacal-
cium phosphate or other Orthophosphate solid can form and
consume phosphate, creating turbidity in the water (10). It is
important to note that precise and accurate predictions of this
100
10
200 mg M CiCoyL
100 ma «i CiCoj/L
Hmg •• CaCaj/L
1134
ORTHOPHOSPHATE DOSAGE • mg P04/L
Figure 5-6. Variation in lead solubility (ph 7.0) as a function of ortho-
phosphate dosage for different alkalinities.
1,000
1214
OrfMOPWIPHATE DOIAOC - IK| PO|/L
Figure 5-7. Variation in lead solubility (ph 7.5) as a function of ortho-
phosphate dosage for different alkalinities.
1,000
200 «g n ClCoj/L
_ 100 «ig II ClCoj/l
1234
ORTHOPHOSPHATE DOSAOE - mg PO«/L
Figure 5-8. Variation in lead solubility (ph 8.0) as a function of ortho-
phosphate dosage for different alkalinities.
61
-------
limitation have not been established. In many cases,
I I I I I I I I
.\--\—1—J—J—J—J—4-
I I I I I I I I
I I I I I I I I
.!....!....»...:> I....I....I I,
1.0 1.0 9.0 4.0
•I NUL OHTHOniOinUTC
might be more likely to precipitate first before orthophosphate
solids.
A very important, but little recognized, limitation to ortho-
phosphate addition is the interaction of zinc with pH, DIG, and
orthophosphate. Most commercial orthophosphate treatment
chemicals contain zinc in some proportion. Contrary to many
manufacturers' literature and assertions, the orthophosphate ef-
fectively reacts with lead in plumbing materials and does not
function by depositing a zinc orthophosphate coating (2,10,11).
The solubility of zinc depends on both pH and DIG, as is shown
in Figure 5-9. Zinc can precipitate as basic zinc carbonate
Zn5OH6(CO3)2(s), thus causing turbid water, if the DIG or pH
is too high to maintain its solubility.
100.00
60.00
20.00
10.00
5.00
2.00
1.00
0.50
0.20
0.10
SmgC/L
MmgUL
i j i...™!.. i i-—
'"' '~
T.O
pH it 25°C
Figure 5-9. Zinc solubility (1=0.01).
Similarly, given certain pH, DIG, and orthophosphate con-
centration combinations, the precipitation of actual zinc ortho-
phosphate (e.g., a-hopeite, Zn3(PO4)2»4H2O) could take place
(Figure 5-10). This also would cause turbid water, and it would
reduce the concentration of orthophosphate available to react
with the lead elsewhere in the distribution system. For example,
if a zinc orthophosphate formulation were used that had a 1:1
ratio of Zn:PO4, the maximum dosage of orthophosphate that
could be achieved without danger of zinc orthophosphate pre-
cipitation at pH 7.5 would be approximately 1.6 mg/L for a
DIG of 80 mg/L C, or approximately 1.4 mg/L for a DIG of 20
mg/L carbon. The limits would be different at different pHs.
One further issue with orthophosphate addition is the ne-
cessity for zinc in the formulation. For the control of lead pipe
corrosion, it is unlikely to be useful (10). For brass or soldered
joint corrosion control, there is a scarcity of real data. What
data exist suggest that zinc might be helpful (2,6,10,11), pos-
sibly by providing a counter to dezincification in brass by the
addition of the zinc in the water. In the case of brass, the
deposition of zinc orthophosphate solid might also be advanta-
geous. Arguments also are made, with similarly little published
data, that the zinc is somehow useful in providing more effec-
1mgC/L
mmgcn.
MmgCA.
MmgOl
Figure 5-10. Zinc solubility (pH 7.5,1=0.01, 25°C).
live control of iron corrosion, possibly by forming a mixed
Zn-Fe-Ca phosphate film.
Clearly, much more research needs to be done in this area.
Current experiments at EPA show that orthophosphate alone
can be effective in slowing lead leaching from brass, given
correct pH and sufficient orthophosphate dosage. The consid-
erations of zinc orthophosphate solubility discussed above
show that, if zinc is not necessary in the formulations or if a
much lower concentration of zinc than phosphate is useful, then
high PO^Zn ratios would be advantageous for dosing.
Blended Orthophosphate Addition. The remaining viable
approach to formation of passivating films is addition of
"blended" phosphates. These chemicals are mixtures of ortho-
phosphate (often 40 percent) with some combination of linear
polyphosphates. The blended phosphates are used to provide
the necessary sequestration or crystal growth-poisoning prop-
erties for such problems as "red water," CaCO3 precipitation,
or manganese precipitation, without having excess polyphos-
phate to solubilize lead and copper. The orthophosphate com-
ponent is present to form a passivating lead orthophosphate film
on the pipe, as is the case with direct orthophosphate addition
alone. From the standpoint of lead and copper control, the use
of blended phosphates is a "balancing act" between the solu-
bility enhancing properties of the polyphosphate with the solu-
bility decreasing (for lead) properties of the orthophosphate.
As with orthophosphate addition, the effect of blended
phosphate addition will depend on at least the combination of
pH, DIG, and chemical dosage. Temperature is also almost sure
to play an important role by affecting the solubility of the
passivating solid, the aqueous speciation of the metal, and the
aqueous speciation of the phosphate species. The effectiveness
also will depend on the ratio of polyphosphates to orthophos-
phates in the chemical, although what that dependence is cannot
be readily predicted at present. The effect also will depend on
the specific identity of the polyphosphate components and their
speciation under the water quality condition in the distribution
system. Polyphosphates have an intrinsic ability to complex and
62
-------
solubilize lead (and copper) and different
for ralrinm limit lead leaching. Further, assume that because the utility is
magnesium, manganese, ferrous iron, ferric iron, and other sub-
stances (7).
Colloidal and Paniculate Metal Forms. Another important
factor in the formation of passivating films is the possible ex-
istence of colloidal and undissolved metal forms. This problem
manifests itself in several ways. If treatment chemicals form an
insoluble colloid with lead and that colloid does not adhere to
the pipe wall, erratic lead levels can be observed in water
samples and treatment will not produce substantial improve-
ment in lead levels at the tap.
The chemical treatment also might not be effective in pre-
venting the physical creation of particles of lead, such as from
solder or brass, in turbulent water conditions.
5.1.2 Selection Criteria
Several factors need to be taken into account when decid-
ing what strategy to pursue for the control of lead and copper.
5.1.2.1 Mix of Materials in the Distribution System
Distribution systems are not homogeneous. They are made
up of a variety of materials, such as lead pipes; soldered joints;
brass, copper, or galvanized pipe; iron mains; asbestos-cement
pipe; or cement mortar-lined mains.
The source of lead and copper in the water passing through
the system usually is found at the end of the distribution system,
in domestic and commercial plumbing installations. However,
even though the regulatory target is the control of lead and
copper, the utility must devise a control method that is compat-
ible with all of its distribution system materials. Water chem-
istry conditions that effectively control lead and copper
corrosion might not be optimum for controlling cast iron cor-
rosion, for instance, and could even cause an increase in cor-
rosion rates (9).
5.1.2.2 Initial Water Quality
Initial water quality not only dictates the success of a par-
ticular control strategy but also governs the efficiency of em-
ploying a particular strategy. For instance, it would not be
cost-effective to use a CaCO3 saturation control strategy when
the source water has a very low hardness and pH. Similarly,
employing pH adjustment to achieve a good pH for lead con-
trol—approximately 9.0—would be very difficult in a hard
water. As discussed previously, the critical initial water quality
factors that should be considered during the control method
selection process are, at minimum, pH, alkalinity/DIC, hard-
ness, and CaCO3 saturation. Depending on the exact initial
chemical characteristics of a water supply, additional factors
also might be of considerable importance in defining treatment
options and their limitations.
Another critical water quality concern is whether a shift in
treatment strategy could result in the destabilization of existing
corrosion films and a significant increase in exposure to lead
or copper for some time. As an example, consider a utility
currently employing pH adjustment to approximately pH 9.0 to
concerned about disinfection effectiveness and trihalomethane
formation potential, it would prefer to switch to the use of
orthophosphate dosage to enable operation at a considerably
lower pH. Such a pH change could jeopardize the integrity of
the lead films on the surface of the pipe, potentially resulting
in increased lead levels.
5.1.2.3 Source Water Problems
Source water problems can cause severe conflicts when a
particular strategy is used to control copper and lead. Examples
of source water problems include the presence of iron, manga-
nese, volatile organic compounds, humic or fulvic substances,
and high trihalomethane formation potential.
A utility has to judge to what extent it should attempt to
solve the source water conflicts by physical means, or whether
to rely solely on chemical treatment to provide an effective
general treatment.
Chlorine dosage can adversely affect lead and copper con-
trol because chlorine is frequently added as an acidic gas. Con-
sequently, pH in a poorly buffered water is decreased, requiring
additional pH adjustment to balance the corrosivity toward cop-
per and lead. Furthermore, evidence exists that chlorine can
accelerate the rate of copper corrosion. Fluoride dosage, when
added as hydrofluosilicic acid, also causes a pH decrease in
poorly buffered waters.
Section 5.2 discusses some of these conflicts in greater
detail.
5.1.2.4 Related Requirements
Different locations must comply with different regulatory
requirements. The considerations of the lead and copper rule
itself, as well as other water treatment objectives dictated by
other primary and secondary drinking water regulations, must
be balanced. Further, each primacy agency has the latitude
to impose other constraints that are thought to be effective in
the region or state. Certain treatment processes might be fa-
vored over others. Additional water quality objectives also
might exist.
Major industrial/commercial water users provide a signifi-
cant economic base to a community. These users can be seri-
ously affected by major changes in water treatment and water
quality. Therefore, a utility might be constrained by, or at least
must take into serious consideration, the compatibility of a
water treatment with current users. The utility might select a
method to which the users can adjust, given equivalent health-
based performance.
It is a regulatory requirement that optimal lead and copper
control, once in place, must be properly maintained, as demon-
strated by meeting specified treatment goals. There is more than
one way to achieve a water quality objective. Since the utility
has to meet goals agreed to with the primacy agency, it is in
the utility's best interest to choose the most mechanically reli-
able, safest, and most operationally consistent method.
63
-------
A utility is best served by choosing the least costly among
otherwise equivalent treatment approaches. It is sometimes dif-
ficult to obtain accurate cost projections for a fully imple-
mented treatment from bench and pilot-plant scale studies. A
significant difference is that, when full-scale treatment is im-
plemented, large quantities of bulk chemicals can be obtained
through a bidding process. This can give the utility the ability
to get a large quantity of a chemical much more cheaply than
most tabulated price estimates. For example, a single barrel of
silicate might appear very expensive, but when vendors begin
competing with other vendors for a long-term supply of bulk
chemical (e.g., railroad car scale), relative prices are often much
lower.
5.1.3 Treatment Chemicals
For each chemical control strategy, a variety of specific
chemicals is available. The chemicals can be obtained from
water treatment chemical specialists, often having proprietary
formulations for inhibitors. They also can be obtained from
industrial chemical manufacturers and their distributors. A use-
ful source for chemical suppliers and available products is
Standard 61 from the National Sanitation Foundation.1 It is a
tabulation of water treatment chemicals tested by a standardized
procedure for contamination by elements or compounds that
are regulated in drinking water for health concerns.
5.1.3.1 pH Adjustment
For pH adjustment, the most useful chemicals are lime
(CaO), slaked lime (Ca(OH)2), caustic (NaOH, KOH), and so-
dium silicate. Lime, slaked lime, and caustics have been dis-
cussed widely in the literature and historically have been the
major ways to adjust pH. Many utilities, however, particularly
smaller ones, continue to have consistent problems with pH
control using these chemicals. Historically, sodium silicate has
not been used for the specific purpose of pH adjustment How-
ever, its properties easily lend it to this application. Sodium
silicate might have several advantages over the other four
chemicals. It is easy to feed consistently, using relatively simple
pumps. It is at least as safe to use as any of the other chemicals
and possibly safer for the operators to handle. Its desirable
properties for source water iron and manganese control might
make it possible to accomplish more than one treatment objec-
tive simultaneously.
Type N® has been the most commonly used silicate, be-
cause it has one of the highest SiO2:Na2O ratios. For pH ad-
justment, however, a formulation having a lower SiO2:Na2O
ratio would be advantageous.
5.1.3.2 Alkalinity Adjustment
For alkalinity adjustment, appropriate chemicals are lime,
slaked lime (CaO, CaOH), caustic (NaOH, KOH), sodium sili-
cate, sodium bicarbonate, sodium carbonate (soda ash), and
sodium silicate.
These chemicals can work as pH adjusters (indirectly in-
'National Sanitation Foundation, 3475 Plymouth Road, Ann Arbor, MI 48113-
0140,313-769-8010.
creasing alkalinity), alkalinity adjusters (indirectly increasing
pH), or both.
5.1.3.3 Inorganic Carbon Adjustment
Only two chemicals are widely used for the supplementa-
tion of inorganic carbon (DIG): sodium bicarbonate (NaHCO3)
and soda ash (sodium carbonate, Na3CO3). Both will provide
some increase in pH, with soda ash having a greater effect than
sodium bicarbonate. The magnitude of the pH effect will de-
pend on the original water chemical characteristics.
5.1.3.4 Hardness Adjustment
Only two chemicals are ordinarily used to provide hardness
(calcium) addition: Lime (CaO) and slaked lime (Ca(OH)2).
Both also increase the pH. Frequently, the effects are confused.
These chemicals, at the proper dosages, can create condi-
tions that provide supersaturation of calcium carbonate in the
bulk water solution or at the pipe surface. By increasing pH,
the chemicals can increase buffering intensity in some pH re-
gions. Sometimes the buffering intensity increase can inhibit
hydroxide ion production by heterogeneous (calcite saturation)
buffering (2). Except under these conditions, little evidence
exists that calcium content has a direct role in reducing lead or
copper leaching.
5.1.3.5 Corrosion Inhibitors
Four classes of chemical inhibitor formulations are useful
for lead control:
• Sodium silicate (maximizes SiO^NajO ratio)
• Zinc orthophosphates
• Generic orthophosphates
• Blends of ortho- and polyphosphates
When selecting the sodium silicate formulation for use as an
inhibitor (instead of as a pH adjuster), the key basis for selec-
tion is to obtain the maximum SiC«2:Na2O ratio. For this pur-
pose, the silica concentration is the active agent. Dosages of
sodium silicate for lead and copper control can be perhaps 18
to 30 mg/L SiO2, which is much greater than the dosage usually
suggested in the literature. The extent to which the dosage can
be lowered to provide an adequate "maintenance" dosage has
not been studied extensively. Experiments by EPA suggest that
high sodium silicate dosage might be more useful than ortho-
phosphate dosage to reduce copper leaching from copper pipe,
but pH effects cannot be totally ruled out
The effectiveness of the use of orthophosphate and blended
phosphates for the control of copper leaching is less clear. Some
researchers report a decrease in the corrosion rate for copper
when orthophosphate is used; few studies, however, have been
conducted thus far that show orthophosphate addition at real-
istic concentrations (0 to 5 mg/L PO4) to reduce copper disso-
lution conclusively beyond that attributable to pH adjustment
64
-------
alone. Experiments by EPA have shown 3 mg/L PO4 orthophos- the- formulation, blended phosphates should be a viable treat
phate to be effective in reducing the leaching of lead from brass
at pH 7.5 in a moderately hard water with a DIG concentration
of approximately 10-12 mg/L C.
The family name "zinc orthophosphate" applies to a wide
range of commercial formulations. The chemicals are usually
acidic blends of zinc sulfate or zinc chloride, with phosphoric
acid or a dihydrogen salt of sodium or potassium (e.g.,
NaH2PO4). Sometimes, a deoxygenating or dechlorination
agent, such as sodium bisulfite, is added to decrease water
aggressivity. The formulations characteristically have different
ratios (as mg/L) of Zn:PO4, ranging from 1:10 to 1:1 hi the
most common commercial products. The ratio selection de-
pends on the necessity for zinc in the system (for example, to
protect asbestos-cement pipe), the orthophosphate level desired
for lead control, and the solubility of zinc in the background
water chemistry conditions (pH, orthophosphate concentration,
and DIG concentration).
The second family of orthophosphate chemicals are the
"generic orthophosphates," including industrial chemicals such
as:
H3P04
• NajHPO,, (or Kj)
• Na3PO4 (or K3)
From the standpoint of lead control, there should be essentially
no significance to whether the salt is based on sodium or po-
tassium. Mixtures of the chemicals are also possible. The use
of orthophosphoric acid (HsPCH) in conjunction with pH ad-
justment has been widely used in Britain (4,5,11,12).
The family of "blended" phosphates is highly diverse. In
general terms, the family includes:
• Orthophosphate salt plus Na- or K-pyrophosphate (P2O7 )
• Orthophosphate salt plus Na- or K-tripolyphosphate
• Orthophosphate salt(s), plus mixture of linear polyphos-
phates
Each possible polyphosphate chain has slightly differing
properties of affinity for different metals and resistance to
breaking down into a molecule of shorter length and an ortho-
phosphate group. Commercial products usually are formulated
to control a background water problem, such as iron or man-
ganese oxidation or calcium carbonate encrustation. The ortho-
phosphate component helps in film formation on metals that
form relatively insoluble surface films, such as lead and zinc.
Little objective data have been published on lead and cop-
per control using blended phosphates in different water chem-
istries. In principle, however, depending on the exact nature of
ment alternative, especially for lead.
When pH effects and reversion to orthophosphate are ac-
counted for, no evidence exists that straight polyphosphate ad-
dition is a desirable strategy for the control of lead and copper.
In fact, considerable data exist to show detrimental effects (7),
especially under the low pH conditions (pH 6-7) that are often
optimal for source water iron sequestration.
5.1,4 Summary
Many approaches are available for lead and copper corro-
sion control. The selection of the best choice might be limited
by several factors, particularly:
• Source water characteristics
• Secondary impacts
— On other drinking water parameters
— On wastewater treatment efficiency
— On discharged waters
• Need to control corrosion of other materials (e.g., asbestos-
cement, iron, cement mortar-lined pipe)
• Relative cost of equivalentiy performing treatments
A source of confusion in selecting a control strategy to
optimize corrosion control for a distribution system is that cop-
per and lead might not respond equivalentiy to each strategy.
Additionally, except for new construction areas with fresh cop-
per and brass plumbing pipes and materials, distribution sys-
tems have had scales and corrosion product buildups for many
years. Therefore, implementation of new corrosion control
strategies might cause short-term problems. This problem is
illustrated by the scenario given in Figure 5-11.
In Figure 5-11, point A represents the starting point on a
lead solubility diagram for a system currently applying pH
10,000
1,000
Figure 5-11. Path of lead response to treatment changes.
65
-------
adjustment to approximately 8.8 as its control strategy. At this corrosion in
fnr man
rty»
point, the lead leaching is being controlled by a mixed surface
coating including basic lead carbonate (hydrocerussite). Con-
sider the case in which the same utility decides to change its
corrosion control approach to the addition of 1 mg/L P04 or-
thophosphate at a pH of approximately 7.5, to improve disin-
fection effectiveness and to reduce the rate of trihalomethane
formation. Because the basic lead carbonate film is reversible,
and because the lead orthophosphate film formation rate on an
old pipe surface is possibly somewhat slower than the dissolu-
tion rate of the existing basic lead carbonate, the stability of the
system might follow the arrows along the solubility curve for
Pb3(CO3)2(OH)2 to the solubility for pH 7.5 (point B). The
solubility at this pH is much greater than at the original pH of
8.8, causing a transitional period in which the existing film is
destabilized. Until there is adequate contact time with the or-
thophosphate to reestablish a stable new lead orthophosphate
film (point C), lead levels might be higher and more erratic
than they were originally.
Very little published information exists on the stability of
lead corrosion films, as well as their formation and dissolution
rates under realistic distribution system conditions. Until these
gaps hi the research are filled, utilities and their advisors must
carefully consider unintended risks to public health, while
working to optimize corrosion control and simultaneously meet
other regulatory needs.
Although treatment strategies exist that make it possible
for utilities to comply with the action levels in the new lead
and copper rule, additional optimization might soon be neces-
sary. This problem might be caused by the increasingly strin-
gent wastewater effluent guidelines. Ambient, normal domestic
and commercial plumbing corrosion might ultimately contrib-
ute enough lead and copper to the wastewater to cause difficulty
in meeting those regulations. The problem will be even more
likely to occur as industrial discharge becomes a smaller frac-
tion of the contaminant load into the wastewater system. Utili-
ties will come under more pressure to minimize both lead and
copper levels beyond those required from the drinking water
regulatory standpoint.
At this time, inadequate systematic research exists to pro-
vide specific guidance for dosages and water chemistry adjust-
ments to guarantee the best selection of chemicals and water
chemistry conditions to ensure the minimization of lead and
copper levels in drinking water. The information in this chapter
and Chapter Four, however, should provide a starting point
from which to begin the evaluation process and choose among
the numerous alternatives available to best fit the overall needs
of a water utility.
5.2 Secondary Effects and Conflicts with Lead
Corrosion Control Strategies
The American Water Works Service Company (AWWSC)
is a large, private water company that has approximately 121
individual operating water systems merged into 21 companies
located throughout the country. The AWWSC has had the bene-
fit of dealing with a variety of waters and has been controlling
corrosion is not something new that arose because of the lead
rule. There are basically three ways to control lead: controlling
mineral stability, using an inhibitor, and not using any lead-
based materials in plumbing or distribution systems.
5.2.7 Carbonate Passivation
Most waters have some dissolved inorganic carbonate
(DIC), and by raising the pH, the amount of bicarbonate
(HCO3") and carbonate (CO3~2) can be increased. The CO3'2
reacts with lead to form stable insoluble carbonate films. When
the pH is raised, a decision must be made about where pH will
be adjusted within a treatment plant. Using lime presents quite
a problem if alum is used as a coagulant, because the solubility
of aluminum is very dependent on pH. If the pH is above 7, a
large fraction of aluminum will carry over through the treatment
plant to the clearwell. Another option is to add lime just prior
to filtration, after coagulation with alum, but the lime deposi-
tion that can occur in the filter is potentially a serious problem.
Injection of lime ahead of granular activated carbon should be
avoided. The safest way to avoid deposition in the clearwell is
to use caustic soda or caustic potash.
Careless raising of the pH can cause excess metal carbon-
ates to accumulate, particularly at the high-service pumps
where there is high velocity and high pressure and immediately
downstream of the high-service pumps.
Many industrial customers cannot tolerate elevated levels
of minerals, a high pH, or high concentrations of carbonates.
Proposed water quality modifications should be discussed with
large industrial customers.
The effect of pH on chlorine's ability to disinfect is very
important (Figure 5-12a), particularly now as many water sys-
tems are trying to meet the requirements of the Surface Water
Treatment Rule (SWTR). As the pH is raised, the stronger
oxidizer hypochlorous acid is converted to the weaker hypo-
chlorite ion. The higher the pH, the less effective the chlorine
10
11
Figure 5-12a. Distribution of HOCI and OCI" in water as a function of pH.
66
-------
will be. By raising the pH, the disinfection process might set
be able to meet the contact tunes (CTs) that are required by the
SWTR. The CT required by EPA increases by 50 percent when
the pH is increased from 7 to 8. To compensate for a pH
increase in the disinfection process, the free chlorine, CT, or
both will need to be increased.
Another potentially major problem of corrosion control is
the effect on disinfection by-products. The trihalomethane for-
mation potential (THMFP) increases as pH increases (Figure
5-12b). In one study, a 40 percent increase in trihalomethanes
was observed as a result of raising the pH from 7 to 8. That is
bad news regarding THMs; it is not the case, however, with
every disinfection by-product. All of them are pH dependent,
but some have the reverse trend and actually decrease with
increasing pH. EPA will be regulating many disinfection by-
products, and any changes in pH levels might affect future
600 f—
500 —
400 —
300 —
200 —
nant level (MCL) for sodium (Massachusetts, for example, has-
an MCL of 20 mg/L). Because of sensitivity in those states
about adding any sodium, problems can occur with the addition
of soda ash or caustic. It is preferable to raise pH with potas-
sium hydroxide rather than sodium hydroxide.
5.2.2 Corrosion Inhibitors
Zinc orthophosphate has performed best in AWWSC's dis-
tribution system. However, wastewater treatment plants might
object because there is a limit on zinc in land application of
sludge for composting. Another problem of this method in-
volves phosphates. Phosphates have been controlled for many
years in this country, and wastewater treatment plants have a
limit on phosphates. Therefore, these plants might have a prob-
lem meeting their own discharge limits with phosphate addi-
tion. Phosphates are nutrients essential for sustaining growth of
algae. If the system has any open reservoirs, particularly in a
warm climate, a summertime water problem of algal growths
will occur. Many corrosion inhibitors have a narrow pH range
in which they are effective. The pH range in which to use
phosphates to maximize their effectiveness is 7.0 to 8.0. Many
systems will need to either increase or decrease their pH to stay
within that range. For example, a lime-softening plant that op-
erates at a pH of 9.0 would need to lower its pH for phosphate
to be effective.
Compatibility with other chemicals is also important. In
some cases, a metal-phosphate precipitate forms. That can oc-
cur with aluminum when more than 0.1 mg/L of aluminum is
carried through the treatment plant and clearwell. The phos-
phate can combine with the aluminum, and aluminum phos-
phate compounds will precipitate in the clearwell and in the
distribution system. The amount of precipitate depends on how
much aluminum is present
Figure 5-13 shows the corrosion rate as a function of pH;
increasing the pH above 7.5 can hi some cases increase the
corrosion rate. Figure 5-14 shows the potential impact of cal-
cium hardness on phosphate addition. Above the curve, precipi-
Corroslon rate
mils/yr
60-]
100 —
u
3
100 —
200 I—
Figure 5-12b. Effects of pH and oxidant dosage on the formation of
TOX and THMs (CHCb) at 20°C in distilled water solu-
tions of 5 mg humto acid/L.
compliance.
There are other difficulties with trying to attain carbonate
passivation. All systems have trouble feeding lime, but options
exist for feeding lime that are virtually problem-free. Liquid
chemicals such as sodium hydroxide have very few feed prob-
lems. Soda ash works well in 15°C water, but trying to dissolve
soda ash in 5°C water will result in most of the carbonate
accumulating in the bottom of the feeder.
The addition of sodium itself also can present a problem.
Many years ago, EPA suggested a possible standard for sodium
and some of the states attempted to meet this standard. Some
of the New England states have a primary maximum contami-
50-
40-
30-
20-
94%
No Treatment
Percent Corrosion Inhibition
97% 98%
76%
5.0
8.0
Figure 5-13. Corrosion rate vs. pH, 114-hour laboratory test with aer-
ated tap water.
67
-------
8.2
8.0- -
PH
7.8-
7.6-
7.4-
7.2
100 200 300
Calcium Hardness
400
500
Figure 5-14. Tricalcium phosphate saturation.
tation of tricalcium phosphate will occur. For example, if the
calcium hardness is at 300 and the pH is at 7.4, tricalcium
phosphate will precipitate.
In addition, the sequestration properties associated with
phosphates should be considered. Polyphosphates under the
appropriate chemical conditions sequester iron and manganese.
One danger is that polyphosphate might sequester lead, which
will actually make the lead concentration increase. Iron and
manganese might consume the phosphate or polyphosphate that
is added, so the residual at the end of the system would be very
low.
5.2.3 Materials
A ban on lead solder now exists, and alternative materials
(antimony and silver solder and plastic pipes) must be used.
Some lead solder, however, is still in use. The substitutes—an-
timony and silver solder—also might have problems. On July
25, 1990, EPA proposed a maximum contaminant level goal
(MCLG) and an MCL for antimony. EPA proposed an MCLG
for antimony of 3 mg/L, with a possible MCL of 5 mg/L.
Studies need to be performed to determine the quantity of an-
timony that can be leached into water from antimony-based
materials.
Plastic pipe, plastic faucets, and other items made from
plastics, such as PVC and polyethylene, can be used in lieu of
lead, brass, or copper pipe. All of these also have some inherent
problems. Some plastic pipes are made with lead, and although
these are not approved in this country, imported pipe often is
made with a large amount of plasticizers, phelate being the most
frequently used. Plasticizers can release vapors that permeate
the pipe and enter the water column. Solvents can penetrate
through all plastic pipe and enter the water column while it is
under pressure.
5.2.4 Conclusions
All systems are going to be faced with looking at treatment
options and with trying to optimize lead control. While the new
regulation minimizes lead in water systems, every system needs
-to eonsio^-seeondaiy^ffeetsr-The-lead-control--strategyjthat-a
water system selects and presents to the state as the optimal
solution might be one that creates problems regarding secon-
dary effects. These possible secondary effects must be brought
to the state's attention.
5.3 Full-Scale Performance Testing of Sodium
Silicate to Control the Corrosion of Lead,
Copper, and Iron: York, Maine
5.3.7 Introduction
In Summer 1991, the York Water District (YWD) in Maine
placed a 4 million gallons per day (mgd) water treatment facil-
ity into service to provide coagulation, clarification, filtration,
and disinfection of its surface water supply. The plant was
designed to meet the requirements of the SWTR. In common
with other surface water treatment plants in New England, the
water produced by the plant is soft (Ca <1 mg/L), low in
alkalinity (<10 mg/L as CaCO3), and has a moderately high pH
(8.3 to 8.8). As this generally corrosive water passed through
me distribution system, it picked up significant quantities of
iron from unlined cast iron pipe. Consumers served from cast
iron water mains complained of a red water problem. Samples
were collected from these sites to verify the presence of iron,
and the iron concentration in these samples ranged from 0.4 to
1.9 mg/L.
Although the plant was designed with the ability to feed
polyphosphate to control the red water problems, the appropri-
ateness of this and other treatment chemicals was reviewed to
address the anticipated requirements of the lead and copper
rule. Zinc orthophosphate and silicate addition also were evalu-
ated as treatment strategies. Calcium carbonate saturation was
not considered a feasible or practical option, because it would
involve the construction of additional feed systems to introduce
both calcium and carbonate into the water.
Polyphosphates, although well-known for their ability to
control red water problems by sequestering iron, were deemed
inappropriate as a method to control lead- and copper-based
corrosion. To control iron, polyphosphates generally require a
pH in the 7.2 to 7.6 range, which is not optimal for control of
lead or copper. Furthermore, polyphosphates have the ability to
complex with lead and copper, potentially causing the concen-
tration of these metals to increase (7). Zinc orthophosphate was
considered for its ability to control lead by forming sparingly
soluble lead orthophosphate films (14), but it is unable to pro-
vide a mechanism for control of iron corrosion. Also, there was
concern that the zinc would be concentrated in the sludge gen-
erated by the community wastewater treatment facility. The use
of sodium silicate reportedly has been a common strategy for
low-hardness waters and has been favored for its potential to
form a surficial coating on piping systems (15). In addition,
silicate has a large capacity to disperse iron colloids, thus mask-
ing the red water problems (16). Several utilities in Maine with
low alkalinity (<15 mg/L as CaCO3) and low hardness (<5
mg/L as CaCO3) have reported that sodium silicate was ex-
tremely effective in eliminating red water complaints. An ad-
vantage of silicates over polyphosphates is the pH range in
68
-------
sj^ectiyj^
lems. Polyphosphates can sequester iron at a pH generally <7.5,
whereas silicates are effective in controlling red water problems
at a higher pH (>8). The higher pH that can be used with silicate
treatment is also more appropriate for controlling the dissolu-
tion of lead and copper. A well-known advantage associated
with sodium silicate is that it does not contain zinc. Based on
these considerations and system constraints, sodium silicate
was recommended for full-scale performance testing.
With assistance from an engineering firm, the YWD de-
signed a water quality monitoring program to track metal con-
centrations in response to the addition of sodium silicate over
an extended period of time (18 months). Twelve sampling sites
were identified throughout the distribution system to account
for spatial variations in water quality. All sampling sites were
cold water faucets located within buildings. First- and second-
draw samples were collected from all 12 sites on the same day
every 2 months. The first- and second-draw samples were ana-
lyzed for lead, copper, iron, calcium, and silica. A third sample
was collected immediately after the second and analyzed for
pH and alkalinity. The monitoring data collected over the
course of 1991 are discussed in the following sections.
5.3.2 Findings
• The finished water produced from the YWD filtration plant
without the application of sodium silicate has low alkalinity
(8 to 10 mg/L as CaCO3), moderately high pH (8.3 to 8.8),
low turbidity (<0.10 MTU), low color (<10 CU) and is very
soft (Ca <1 mg/L; Fe <0.05 mg/L). The water was corrosive
toward lead and iron, as it produced an average lead level
of 83 ± 145 |ig/L in first-draw samples and iron levels in
the range of 0.33 ± 0.55 mg/L from first- and second-draw
samples. The finished water was less corrosive toward cop-
per; the average copper level from first-draw samples was
0.15 ±0.13 mg/L.
• Periods of 2 to 3 years might be required before the impacts
of silicate addition can be determined, due to annual cycles
in temperature and flow rate.
• The low buffering capacity of the plant water and variations
in the coagulation process resulted in large pH fluctuations
in the water exiting the filters. Sodium silicate fed into the
filtered water served essentially two functions: to adjust the
pH and to add silica to the finished water. As a result, it was
extremely difficult for the operator to maintain a constant
finished water pH and silica dosage.
• The alkalinity and pH were significantly lower at dead ends
of the distribution system, especially when the dead-end
lines were unlined cast iron. These areas consistently had
lower silica concentrations and higher concentrations of cor-
rosion products.
• Lead levels averaged 83 ± 145 (ig/L during the initial sam-
pling event when sodium hydroxide was being applied to
finish the water during December and the first week of
January 1991. After feeding sodium silicate in lieu of sodium
hydroxide, the average lead levels in first-draw samples de-
__CTea&ed_and_slabilizedJ^2<5J-^
May to December 1991.
• Red water complaints received by the YWD when sodium
hydroxide was being fed were eliminated completely with
the application of sodium silicate. Iron concentrations in the
samples collected throughout the distribution system ranged
from 0.10 to 1.9 mg/L before silicate treatment, and from
0.10 to 1.37 mg/L after treatment. It is likely, therefore, that
silicate was sequestering iron.
• Iron concentrations showed only a slight reduction over time
in response to treatment with silicate.
• Copper levels in the first-draw samples before application
of silicate were relatively low, averaging 0.15 ± 0.13 mg/L
and ranging from 0.06 to 0.48 mg/L. Application of sodium
silicate reduced these levels slightly.
• Silica concentrations decreased as the water passed through
the distribution system, suggesting that silica was coating
the surface of pipes. Also, the average silica concentration
in the first-draw samples was lower during each sampling
event than the average silica concentration in the second-
draw samples, suggesting that forms of dissolved silica were
coating the internal surfaces of plumbing.
• With the average maintenance silica dosage of 11 mg/L used
in this evaluation (startup period excluded), the chemical
cost to the YWD is $8.12 per million liters.
5.3.3 Recommendations
• If silicates are used to control corrosion in soft, low-alkalin-
ity waters, careful consideration must be given to the design
of feed systems to ensure that a constant dosage of silica is
provided. Therefore, it might be necessary in certain situ-
ations to adjust pH separately by the addition of another
chemical, such as potassium or sodium hydroxide.
• In water with low alkalinity (<10 mg/L as CaCO3), the use
of silicates in conjunction with carbonate (alkalinity in-
crease) adjustment should be investigated. Alkalinity could
be supplied by silicates as long as the pH is raised into the
9.0 to 10.0 range. Increasing the alkalinity would minimize
the pH reductions that occurred at the ends of the system.
• Studies should be conducted under controlled conditions to
determine relationships among hardness, DIG, pH, existing
films, silica dosage, and effectiveness of treatment.
• Full-scale water quality monitoring programs aimed at de-
termining the effectiveness of silicate addition should be
performed over a period of several years.
• When silicates are used as a means of corrosion control, pH,
alkalinity, and silica levels should be monitored at the ex-
tremities of the distribution system.
69
-------
5.3.4—Methodology
5.3.4.1 Description of the Facilities
The source of water for the YWD is a shallow (<10 m)
pond. The facilities that process the water are an intake facility
at the shore of the pond and a filtration facility. Water flows by
gravity from the intake facility to the filtration facility. Al-
though the intake facility contains equipment to permit addition
of chlorine and potassium permanganate, these chemicals are
not routinely added.
Water entering the filtration facility is injected with alumi-
num sulfate and sodium hydroxide for coagulation. After being
coagulated, the water enters an upflow clarifier, consisting of
plastic media retained by a stainless steel screen. The media
retain a portion of the coagulated material, and the remaining
residual paniculate matter is retained on a mixed-media filter.
Water exiting the mixed-media filter is chlorinated for disinfec-
tion before it enters a 300,000-gallon contact basin/clearwell.
The pH of the disinfected water exiting the clearwell is raised
to between 8.3 and 8.8, prior to the addition of ammonia gas,
to maximize the formation potential of monochloramine. When
the trial application of sodium silicate was initiated, it was fed
through the sodium hydroxide feed system.
The distribution system consists of approximately 40 per-
cent unlined cast iron pipe and 60 percent cement-lined cast
and ductile iron pipe. The unlined cast iron pipe is approxi-
mately 50 to 100 years old. There are no known lead service
lines or asbestos-cement pipe in the system. York is a coastal
tourist community with the population served by the YWD
ranging from 5,000 in the winter to approximately 10,000 in
the summer. The large population fluctuation causes the aver-
age daily flow rate to range from approximately 1.3 mgd in the
whiter to 3 mgd in the summer.
5.3.4.2 Study Objective
The objective of the evaluation was to determine the effec-
tiveness of sodium silicate in controlling iron, lead, and copper
corrosion in the YWD's distribution system and within residen-
tial home plumbing systems. Effectiveness, in this case, means
noticeable reductions in the concentrations of the referenced
corrosion products over a period of 18 months. This report
covers data collected over the first 12 months of monitoring.
5.3.4.3 Treatment Scheme
The sodium silicate solution used in the evaluation was
Type N® (PQ Corporation, Philadelphia, PA), which has a silica
(SiO2) to sodium oxide (Na2O) ratio of 3.22:1. It was selected
because it was the least expensive available silicate solution in
the region and because it has a relatively high SiO2:Na2O ratio.
The silicate dosages used in this evaluation were based on
recommendations from the manufacturer and on information
available in the literature (15,17). The goal was to follow the
present practice of applying silica to control corrosion in water
distribution systems. Over the first 2 months of the monitoring
program, a silica dosage of 16 to 20 mg/L as SiO2 was used.
For the remainder of the monitoring program, the silica dosage
was lowered to 8 to 12 mg/L as SiO2.
-5,3 A4—Monitoring-Erogram-De^ign
The main objective of the monitoring program was to gen-
erate sufficient data to determine the effectiveness of sodium
silicate in reducing levels of principal corrosion products, in-
cluding lead, copper, and iron. Another goal was to gain an
understanding of the potential mechanism of silicate corrosion
inhibition (e.g., surficial coating) by monitoring silica concen-
trations throughout the distribution system. To meet these ob-
jectives effectively, a monitoring program was designed to track
pH, alkalinity, calcium, lead, copper, and iron levels at 12 points
throughout the distribution system over an 18-month period.
Sampling events consisted of collecting three samples from
each monitoring location on the same day.
Because water system personnel could gain regular en-
trance to only a limited number of buildings, a survey was
conducted to identify and select individual homeowners to par-
ticipate in the monitoring program. The selection of sites was
based on the ability of the participating residents to understand
and perform the prescribed sampling procedures effectively for
the period of the monitoring program. In addition, the locations
were apportioned throughout the distribution system, covering
both the center and the ends of the distribution system (Figure
5-15). An extensive materials survey to identify specific sam-
pling locations based on sources of lead and copper was not
performed prior to the monitoring program.
In York, annual cycles in water flow through the distribu-
tion system and in temperature represent important temporal
variations. It was necessary, therefore, to monitor water quality
changes over a period of 18 months. Sampling was conducted
every 2 months to account for changes in flow and temperature.
5.3.4.5 Sampling and Analytical Procedures
Sampling Procedures. First-draw and second-draw sam-
ples were collected from taps from 12 buildings throughout the
distribution system (Figure 5-15). First-draw samples were col-
lected after the water was allowed to stand motionless for 6 to
12 hours. Second-draw samples were collected after the tap had
been flushed for a period of 5 minutes. The first- and second-
draw samples were collected hi 250 mL bottles, and each was
analyzed for lead, copper, iron, calcium, and silica. A third 250-
mL sample was collected immediately after the second-draw
sample and was analyzed for pH and alkalinity. The three sam-
ples were collected on the same day from each of the 12 sites
to relate metal concentrations to the referenced water quality
parameters.
pH and Alkalinity. Samples for pH and alkalinity were
measured in the laboratory within 24 hours of the time of
collection. The pH was measured with an ORION SA250 pH
meter. The meter was calibrated with pH buffer standards at pH
4, 7, and 10. The meter was recalibrated at the end of a group
of analyses to check for instrumental drift. Alkalinity was de-
termined by EPA (1983) Method No. 310.1 using 0.02 N
H2S04.
Lead, Iron, Calcium, and Copper. Upon arrival at the labo-
ratory, samples for lead were acidified to pH <2 with concen-
trated nitric acid. Lead samples were analyzed on a Perkin
70
-------
Lined Pipe
Unllncd Cut Iron Pipe
Monitoring Location
Filtration Plant
Figure 5-15. Map of the York Water District distribution system.
Elmer 5100 PC Atomic Absorption Graphite Furnace according
to Standard Methods (1989) No. 3113 B. Samples for iron,
calcium, and copper were analyzed on a Perkin Elmer Model
No. 460 Flame Atomic Absorption Spectrophotometer, accord-
ing to Standard Methods No. 3500 B. Field spikes and blanks
were performed during each analysis to determine the accuracy
of the method.
Silica. Silica analyses were conducted using Inductively
Coupled Plasma (ICP) according to EPA (1983) Method No.
200.7.
Data Analysis. In the case of small sets of data, including
outliers can result in a bias in the calculated mean. Therefore,
sets of lead data from every sampling event were subjected to
the Dixon Test to eliminate outliers.
5.3.5 Results and Discussion
The data collected for the evaluation of silicates are pre-
sented in the following two sections. First, treatment plant op-
erating data over the 12-month period are discussed. Second,
the results of the distribution system monitoring program are
presented
5.3.5.1 Plant Operating Data
Finished Water Quality Data. Table 5-2 summarizes the
average annual finished water characteristics at the YWD fil-
tration facility during the monitoring period. In general, the
water is corrosive toward lead and iron due to its low alkalinity.
With the exception of temperature, the finished water quality
parameters do not vary significantly on a weekly or annual
basis.
Table 5-2. Average Finished Water Quality Summary
Parameter
pH
Alkalinity (mg/L as CaCOa)
Turbidity (NTU)
Temperature (°C)
Iron (mg/L)
Manganese (mg/L)
Aluminum (mg/L)
Mean
8.5
8.0
0.06
13.0
0.03
0.06
0.05
Standard Deviation
±0.29
±1.65
±0.01
±3.0
±0.01
±0.02
±0.04
Temperature. Temperature can have a pronounced effect on
the rate of corrosion, m general, as the temperature increases,
so does the corrosion rate of most materials. As illustrated in
Figure 5-16a, the temperature in the finished water increased
from 4°C during the winter to 24°C in the summer months.
Therefore, the rate of corrosion due to temperature effects
would be highest in the summer months.
Flow Rate. The average velocity of the water carried
through a distribution system should increase, in general, as
plant flow rate (output) increases. Velocity is an important
physical factor that affects the rate of corrosion. Slow velocities
within a distribution system cause water to be stagnant; often
a marked decrease or increase in pH is observed. Velocity, as
it relates to inhibitor-based corrosion control, is important in
sustaining a passivating film on a pipe surface. As velocity
increases, so does the rate at which a given mass of inhibitor
comes in contact with a given unit surface area of pipe.
71
-------
The uantit of water roduced
fignifi'''an*1y from
by
th^
nf finished water
Th?
winter to summer (Figure 5-16b), due to seasonal population
patterns. This variation had a tendency to cause stagnant areas
during the winter months, which resulted in lower pH values
at dead-end monitoring locations.
30,
25..
(a)
N«OH I Sodium Sllleata
V
\
\
-4 h
Dae Jan Fab Mar Apr May Jun Jul Aug Sap Oct Hm Dae
f 300 +
a
NaOH Sodium SllleaW
Dw Jan Fab Mar Apr May Jun
H r-
Aug Sap Oct Nm Dae
Figure 5-16. Temperature of the filtration plant finished water (a) and
monthly water production (b).
Silica Dosage. The monthly average silica dosage and raw
water silica concentrations over the course of a 12-month moni-
toring period are presented hi Figure 5-17. The average silica
dosages were determined by dividing the total volume of silica
dosages used in this evaluation (9 to 16 mg/L) were similar to
dosages (12 to 20 mg/L) at a nearby utility with similar water
quality conditions.
After reviewing the distribution system data in August, it
was noted that the pH at remote points in the distribution sys-
tem was low (<7.2). To raise the pH at these locations, the feed
rate of sodium silicate was increased in September and October.
As a result, the silica dosage increased (Figure 5-17) over the
same time period. The sodium silicate solution, therefore, was
performing two functions: to raise the pH of, and to add silica
to, the plant finished water. The operating data suggest that the
feasibility of feeding a more alkaline sodium silicate solution
(lower SiO2:Na2O ratio) or accomplishing pH adjustment sepa-
rately with another chemical, such as sodium or potassium
hydroxide, should be investigated.
5.3.5.2 Distribution System Monitoring Data
pH. During the period when the finished water was ad-
justed with sodium hydroxide, prior to application of sodium
silicate, the average pH from the monitoring points was 8.34
± 0.26. When the average startup dosage of approximately 16
to 20 mg/L as SiO2 was being administered, the pH from the
sites averaged 8.38 ± 0.14. After the initial startup dosage was
lowered to a maintenance dosage of 10 mg/L as SiO2 during
late March, the pH dropped to an average of 7.75 ± 0.10 for
the remainder of the monitoring program (Figure 5-18).
5. 10.
s-
(a)
Sodium Sllcat*
r-
Dae Jan Fab Mar Apr May Jun Jul Aug Sap Oct Nov Dae
20
IS. .
3
• •Doug*
O O Raw Water
.•—•-
H h
—\ 1 1 f 1 1 1 1
Dae Jan Fab Mar Apr May Jun Jul Aug Sap Oct Nov Dae
Figure 5-17. Average monthly silica dosages and raw water silica con-
centrations.
•a 15 • •
I
(b)
NaOH I Sodium Silicatt
H K-
-4 1-
DM Jan F»b Mar Apr May Jun Jul Aug Sap Oct Nov Dae
Figure 5-18. Average pH (a) and alkalinity (b) from the distribution sam-
pling events.
72
-------
At the dead ends of the system, the pH (7.52 ± 0.38; n =
3) was lower than the pH (8.17 ± 0.05; n = 8) at central points
within the distribution system. Lower pH values observed are
likely due to the release of metals such as iron, and subsequent
hydroxide-ion uptake, which frequently occur in stagnant areas.
The lower pH values are generally consistent with lower silica
concentrations found in the same regions (see the following
discussion on silica).
Alkalinity. The alkalinity typically ranged from approxi-
mately 5 mg/L as CaCO3 at dead-end locations to 10 mg/L at
most other points within the system. The average alkalinity
remained relatively constant throughout the monitoring period,
with the exception of a slight rise during February when the
startup dosage of silica was being administered (Figure 5-18b).
The increase in alkalinity was probably due to the presence of
the anionic silica species, H3SiO4.
Silica. From the distribution system monitoring data, it can
be seen that the silica concentrations in the center of the system
were higher (17.8 ± 0.53 mg/L as SiO2) than at the ends of the
system (16.0 ± 1.2 mg/L) (Figure 5-19a). These data suggest
that silica was being adsorbed onto pipe surfaces as the water
moved through the system. Silica has the ability to adsorb onto
metal-oxide surfaces (18,19). Potential evidence of this type of
adsorption was observed in this study as the average silica
concentration was lower (15.6 ± 1.5 mg/L; n = 3) at sampling
sites located on unlined cast iron mains than at sites located on
other types of pipe (17.5 ± 0.71; n = 9) (Figure 5-19a).
The calculated means of the first- and second-draw sam-
ples were compared; they displayed evidence of silica adsorp-
tion onto the surfaces of home plumbing systems (Figure
5-19b). Although these data suggest adsorption of silica was
occurring, it cannot be confirmed without X-ray diffraction
analyses.
Lead. Figure 5-20 shows the variation in lead concentration
of first-draw samples over the monitoring period. Prior to ap-
plication of sodium silicate, the lead levels ranged from 6 to
488 |0.g/L and averaged 84 ± 145 (ig/L. Over the period of May
through December, when the lead levels were relatively stable,
the lead concentrations ranged from 5 to 166 \ig/L and averaged
26 ± 22 ng/L (Figure 5-20a). These lead levels are relatively
high, considering that 11 of the 12 buildings were constructed
before 1981. The other building was constructed in 1990 and,
as a result, contained pipes with lead-free solder. Since the
first-draw sample volume was 250 mL, it is likely that the major
source of lead is from brass fittings.
The average lead concentrations were consistently lower
during the time when the sodium silicate was being fed. When
12- '
10'
Uiwd
Pip.
Unllned
Cut-Iron
150
100- •
Q
£
so- •
(a)
Sodium Silicate
H r-
-I r-
-r-
-I r-
r-
OK Jan fab Mar Apr May Jun Jul Aug Sep Oet Nov Dec
400
300 +
100'
Second Draw
flrtt Draw
(b)
I III
May
Jul
Sep
D«c
Figure 5-19. Silica concentrations from selected sites within the distri-
bution system (a) and in first- and second-draw samples
(b).
8
7
6
| 5 '
«l
s 4-
u
2'
l"
(b)
• >i/N/< n > 25 jig/L
K n>50jig/L
NaOH | Sodium Silicate
1
'
'
X
X
X
X
X
X
x
x
1
X
X
X
X
X
X
X
X
x
x
Y
1
X
X
X
X
X
X
I
X
X
X
x
X
X
I
X
X
X
X
X
X
X
x
X
JLJL
May
Sep
Figure 5-20. Average lead concentrations in the first-draw samples (a)
and the number of samples exceeding specified concen-
trations in first-draw samples (b).
73
-------
the number of samples exceeding :>50 |ig/L as lead and >25 been, observed to othef corrosion monitoring programs under
(ig/L as lead (Figure 5-20b) were compared before and after
treatment, however, only a slight improvement was observed
with the addition of sodium silicate. Second-draw samples,
collected after flushing for a minimum of 3 minutes, were
typically below the detection limit.
The highest lead concentrations were consistently found in
samples collected at monitoring points on dead-end unlined cast
iron mains, probably because of the lower pH values witnessed
at these locations. Typically, the pH at these locations ranged
from 6.6 to 7.2 compared to other sampling locations, where
the pH was 7.6 to 8.5.
In general, some sites showed a consistent reduction in lead
concentration; at other sites, the concentrations either remained
relatively constant or increased. This result is to be expected
since the source of lead (e.g., dezincification of brass, or dis-
solution of lead-tin solder) and types of films present will vary
significantly depending on the specific location of the site. In
particular, the dezincification of brass fittings, which was prob-
ably the major source of lead at most of the sites, can respond
erratically to silicate treatment (20).
Iron. As shown in Figure 5-21, the iron concentration over
time, after silicate addition, gradually decreased, and then in-
creased, probably in response to low flow rates during the
following fall and winter months. Each point on the figure
represents the average iron concentration of 12 first-draw and
12 second-draw samples.
Dec Jin Fib Mir Apr Miy Jun
Aug Sip Oct Nov Die
Figure 5-21. Average iron concentrations in the first- and second-draw
samples.
During the last 6 months of 1990, the York Water District
received approximately 15 red water complaints. Silicate treat-
ment eliminated these complaints over the 12-month trial ap-
plication. Iron concentrations ranged from <0.10 to 1.87 mg/L
before treatment, and <0.10 to 1.37 mg/L after treatment; there-
fore, it is likely that the paniculate iron was being sequestered
by dissolved silica. The ability of sodium silicate to sequester
oxidized forms of iron in soft, low-alkalinity water has been
well documented (16).
Copper. Average first-draw copper concentrations from the
six sampling events were especially low (Figure 5-22), as has
similar water quality conditions (21). A possible reason for the
low copper levels is that the first-draw sample volume was 250
mL; as a result, a large portion of the sample volume was
contained within brass fittings and was not in contact with
copper pipe.
The copper levels decreased during the initial sampling
events but later increased during the winter (Figure 5-22). The
increase was primarily due to a drop in pH at two monitoring
stations located on dead ends. At dead-end monitoring stations
located on unlined iron pipe, the copper concentration averaged
0.39 ± 0.04 mg/L, and at all other locations averaged 0.05 ±
0.02 mg/L. When the average copper concentrations are deter-
mined excluding dead-end monitoring points, there appears to
be a slight reduction hi copper levels from the application of
silicate over time (Figure 5-22).
• • All Sites (n»11)
O O Diad-End Sites Excluded (n > 9)
0.20' • NiOH I Sodium SHIcitt
0.00
Die Jin Fib Mir Apr May Jun Jul Aug Sip Oct Nov Dec
Figure 5-22. Average copper concentrations In the first-draw samples.
5.3.5.3 Treatment Costs
Given the average maintenance silica dosage of 11 mg/L
administered between April and December, the cost of sodium
silicate is $8.12 per million liters. This figure is based on bulk
deliveries £15,142 L) of Type N® liquid sodium silicate and
a bulk chemical cost of $21.30/100 kg ($73.70/100 kg as Si02).
5.4 Assessing Zinc Orthophosphate vs. pH
Adjustment: Champlain, Vermont
5.4,1 Introduction
Champlain Water District (CWD) is a regional water sup-
plier in northwestern Vermont chartered by legislative action hi
1971. As a municipal district, its primary purpose is the supply
of potable water. At the tune the CWD was chartered, commu-
nities in the greater Burlington area were using a variety of
water sources. These existing sources were deficient in quality
and/or quantity and demand was being increased by a fast-
growing economy and population. CWD presently is composed
of eight member communities: South Burlington, Shelburne,
Williston, Essex, Colchester, Winooski, Milton, and the Village
of Jericho. Because of political divisions within member com-
74
-------
rwn pn»cpntiy
tn th"
1~? - 16-month study, begun in April 1981, usod mild steel coupons
water systems: South Burlington, Shelbume, Williston, Essex
Junction, Essex Town, Colchester Town, Colchester Fire Dis-
trict #1, Colchester Fire District #3, Winooski, Milton, Jericho
Village, and the Mallets Bay Water Company. The total popu-
lation served is 50,000 to 55,000 and the average daily flow in
1990 was 8.3 mgd, with a peak day of approximately 12 million
gallons.
The CWD's treatment and supply system went on line in
March 1973 and consists of three major components: (1) raw
water intake and pump station, (2) water treatment facility and
plant storage, and (3) the finished water pumping and transmis-
sion network of CWD-owned lines and storage facilities. The
distribution network encompasses both a low-pressure and a
high-pressure component. The CWD's transmission and storage
network was interconnected with existing distribution systems
of its member towns. The raw water source for the CWD is
Lake Champlain. The intake is located in the northern channel
of Shelbume Bay as it passes into the broad lake, and is located
at a depth of 75 feet, 2,480 feet from the Red Rocks Park
shoreline. Lake Champlain is characteristic of many New Eng-
land surface waters. CWD finished water has a moderate alka-
linity (approximately 50 mg/L as CaCO3), moderate hardness
(approximately 75 mg/L as CaCO3), and a pH of approximately
7.2. These properties, combined with typically saturated O2
levels, are conducive to forming an aggressive water. Calcula-
tions indicate a Langelier Saturation Index (LSI) of -1.39 to
-0.96 (0 to 20°C), indicating a significant CaCO3 undersatura-
tion. The Aggressiveness Index (AI) also was used to evaluate
corrosion potential and resulted in a value of 10.3. This falls
into the moderate to high range (>12 is considered nonaggres-
sive, 10 to 11.9 moderately aggressive, and <10 highly aggres-
sive).
Further verification of corrosivity was evidenced by the
visual inspection of the diatomaceous earth (DE) filtration pip-
ing dismantled during a plant expansion in 1982. This construc-
tion replaced the high service pumping units and required
removal and replacement of suction and discharge piping in-
stalled in 1972. The older piping was examined, and tuburcu-
lation and pitting measuring 1/4 to 3/8 inches throughout the
interior diameter of the pipe wall were observed. The piping
material was bare, unprotected steel. Although this type of pipe
was used in the CWD plant facility, it is very uncommon in the
distribution systems of CWD and its member towns. Assess-
ment of these findings warranted further investigation, based
on economics, health, and expectation of stricter federal regu-
lation of corrosion by-products. A corrosive water would be
costly to the CWD because of its large investment in water
storage tanks and distribution and transmission piping. Con-
sumers also would be affected economically through deteriora-
tion of domestic plumbing and water-related appliances.
Additionally, consumer health could be at risk as a result of
corrosion by-products leaching into drinking water.
Initial data at the CWD indicated that corrosion, with its
potential ramifications, needed to be studied further. To this
end, a pilot study was designed to establish corrosion rates of
metal coupons using CWD finished water and to help predict
the effectiveness of different treatment techniques. This initial
to establish the corrosion rate of the CWD finished water vs.
finished water treated with zinc orthophosphate (ZOP). A cor-
rosion rate of 9.61 mils per year (mpy) was obtained and is
considered to be in the moderate to severe range. This was
based on 10 tests conducted for an average of 24.3 days each.
The result of adding ZOP at a dose of 1 mg/L as zinc (product
has a 1:1 ratio of zinc to orthophosphate) was an average 78.8
percent reduction (range, 67.9 to 86.6 percent) in the corrosion
rate to 2.04 mpy (range, 1.15 to 3.71 mpy). This initial research
was expanded to include lead coupons, coupons in the distri-
bution system, and a bench-scale comparison of elevated pH
treatment technique to the use of ZOP. Expansion of the bench-
scale research also permitted the assessment of combining ZOP
addition with pH elevation. Analysis of the resulting data indi-
cated that implementation of a corrosion control treatment pro-
gram would be beneficial to CWD and its consumers on both
an economic and a health basis. The CWD Board of Commis-
sioners approved the expenditure to design and implement the
use of ZOP as a corrosion inhibitor on May 27, 1986, and the
process was on line April 28, 1987.
5.4.2 Materials and Methods
5.4.2.1 Materials
Metal coupons (1/2 inch x 3 inches x 1/16 inch) based on
the ASTM Standard D 2688-70 and NACE standard TM-01-69
were used to study corrosion rates and potential reduction in
corrosion due to (1) ZOP addition (pH = 7.0-7.2), (2) pH
elevation (to approximately 8.0), and (3) a combination of these
two treatments. The methods established the corrosivity of
water by measuring the weight loss of various metal coupons.
The rate of corrosion of a metal immersed in water is a function
of the tendency for that metal to corrode and the tendency of
the water and the materials it contains to promote (or inhibit)
corrosion. The relative corrosivity of water can be determined
by comparing the corrosion rate of a material in water with a
corrosion rate of the same material in another water. Mild steel
(SAE Steel [1010]) coupons were used from April 1984 to the
present. Use of lead coupons was incorporated into the study
in December 1988.
Technical Products Corporation (TPC) (formerly Virginia
Chemical Inc.) supplied the metal coupons and the 2902 and
2900 Corrosion Test Units used hi the bench-scale studies. The
2902 unit consists of three connected plexiglass cylinders on a
base. A plexiglass rod extends down from the cover that allows
for coupon attachment using a nylon nut and bolt. Cylinders
are approximately 9 inches high and 2 3/4 inches in diameter.
Water enters at the base of the first cylinder housing the pre-
weighed control coupons, then flows over the coupons into the
center cylinder. Here a Diaz AccuPlus® peristaltic metering
pump adds the ZOP corrosion inhibitor (Virchem® 932) from
a 5-gallon polycarbonate bottle. The ZOP-treated water then
flows into the bottom of the last cylinder over a second set of
pre-weighed coupons and exits at the top of this cylinder to
waste. The water flow was regulated (approximately 0.5 gal-
lon/minute) as was the rate of Virchem® 932 feed solution
(approximately 26 to 28 mL/hr) to maintain the desired 1 mg/L
(ppm) zinc concentration. Adjustments to the flow rate and feed
75
-------
solution concentration were made as required tn maintain thfi scale study to establish corrosion rates of the finished high
desired 1 mg/L zinc level. The Virchem® 932 feed solution was
prepared by mixing 900 mL of concentrated Virchem® 932 with
5 gallons of effluent water from the direct filtration process.
The prepared, weighed (to 0.1 mg) metal coupons were placed
in contact with flowing water for a period of 24 days. Upon
removal, coupons were submerged in acetone for 1 to 2 min-
utes, removed, and allowed to air dry before mailing to TPC.
TPC reweighed the coupons after processing and computed
corrosion rates based on weight loss and exposure time. Cor-
rosion rates expressed as mils per year (mpy), equivalent to
0.001 inch, were determined. TPC supplied all coupons, prepa-
ration of coupons, Virchem® 932, and weight loss and corro-
sion rate analysis.
Virchem® 932 is a liquid synergistic corrosion inhibitor
developed for use in potable water and designed to control
corrosion of contacted metal surfaces. It also has demonstrated
corrosion protection of asbestos-cement pipe in studies con-
ducted by EPA. Dissociated zinc and phosphate ions (at a 1:1
ratio) are provided by Virchem® 932 with the zinc concentra-
tion being analyzed to control the desired amount of ZOP ad-
dition. TPC reports Virchem® 932 to have the following
characteristics: color—clear, odor—none, density—10.6 Ib/gal,
specific gravity (@70°F)—1.273, solution pH—0.8, and zinc
content—0.83 Ib/gal.
The 2900 Single Cell Corrosion Test Unit has the same
dimensions and shape as the cylinders of the 2902 unit but
consists of only one cylinder with an inlet at its base and an
outlet near the top. Two of these units were used in the bench-
Station #4
ZOP Addition
Station #1
Filtered
Cntrl
service water (HS) and nontreated effluent water from the direct
filtration filters (DFs). DF water has been prechlorinated (0.60
to 0.80 mg/L) and filtered after the addition of coagulants (alum
and a polymer).
Coupons were inserted into water main distribution lines
using Corrosion Coupon Probe Assembly 2901 supplied by
Technical Products. The coupons were left in place for a period
of 83 to 142 days. The assembly fits onto a standard 1-inch
corporation stop (outside diameter = 1.25 inch). The insertion
rod is adjustable so that the faces of the coupons are parallel
to the water flow and near the center of the distribution line.
The probe assembly consists of three main parts: insertion rod,
bonnet, and body. The insertion rod is constructed of stainless
steel with a molded nylon tip, nut, and bolt that holds the
coupons and a movable stainless steel collar held in place by a
set screw. The bonnet is bronze and contains a brass packing
gland with asbestos packing. The packing gland prevents leak-
age and holds the rod in place after insertion. The body is made
up of a short nipple and a 1.25-inch inside diameter. NPT
coupling that screws onto the corporation stop.
5.4.2.2 Methods
Coupon Studies. The original 2902 triple cell unit (station
#2) in the pilot bench study used HS water (finished water
being supplied to the distribution system), mild steel coupons,
and a 1 mg/L zinc concentration added via the Diaz pump
(Figure 5-23). After the plant began using Virchem® 932 and
ZOP in the distribution system, two 2900 single cell units with
Station #2
ZOP Addition
Cntrl
Filtered
ZOP
Station #3
CWD
Finished _
(0.3-0.4 mg/L ZOP)
ApH=8.0
ZOP@1 mg/L
; Steel Coupon
= Lead Coupon
Figure 5-23. Coupon studies on corrosion rates in four cell units.
76
-------
mild steel coupons were set up in the laboratory. Station #1 was acid, final chloriaatiuii, aud plant ZOP addition) as the base
supplied by DF water (no ZOP) and the other 2900 cell was
supplied with HS water (station #3), which now contained
Virchem® 932 at a 0.3 to 0.4 mg/L as zinc concentration. To
monitor and evaluate the effects of Virchem® 932 in the distri-
bution system, 2901 probes were installed with mild steel cou-
pons in three distribution mains. By comparing the corrosion
rate of these coupons to the corrosion rate of the DF station #1
coupons in the laboratory (pre-ZOP addition), a percent reduc-
tion in the coupon corrosion rate was calculated.
The construction of an interconnection between the CWD
and the neighboring city of Burlington (which had elevated its
water's pH for corrosion control) offered an opportunity for a
comparison study in the CWD's laboratory (Figure 5-24). The
CWD and Burlington Water Resources (BWR) both use Lake
Champlain as their source water; therefore, raw water charac-
teristics show only minimal differences. The test cells were set
up to allow a comparison of ZOP treatment against BWR's
technique of raising the pH (to 8.0) to precipitate a CaCO3 film.
The design allowed the use of CWD DF water (post pre-chlori-
nation, coagulation, and filtration, yet prior to hydrofluosilicic
control unit at station #1 (Figure 5-25). The corrosion rate in
this DF control unit allowed for calculated reductions hi corro-
sion rates using three techniques: (1) adding ZOP, (2) raising
the pH (to 7.9-8.2) to precipitate CaCO3 (^H), and (3) adding
ZOP after elevated pH (7.9-8.2) adjustment C^pH + ZOP). Once
these rates were established, a comparison analysis was possi-
ble among any combination of these three different corrosion
reduction techniques.
A single cell 2900 unit at station #1 was used for the base
control coupons using CWD DF water. At station #3, ZOP was
added to CWD DF water and at station #4, to BWR water, using
triple cell units. The inlet cylinder to station #4 contained the
elevated pH coupons, and the outlet cylinder contained the ApH
+ ZOP treated coupons. The inlet cylinder to station #2 used
DF water, which contained base control coupons (duplicating
the 2900 DF coupons), and the outlet cylinder (after ZOP ad-
dition) contained the ZOP-only treated coupons. Therefore,
much of the time, duplicate coupons were being exposed to DF
water. During this time the corrosion rate from station #1 was
used as the base rate in calculating the rate reduction for the
Raw Water
Burlington Water
Treatment Plant
1
Champlain Water
Treatment Plant,
Raw Water Tank
Prechlorination
Flocculatlon
Filtration
Station #4
2902 Triple Cell
Station #1
2900 Single Cell
Base Control
Coupons
Station *2
2902 Triple Cell
Station #3
2900 Single Cell
High Service
Coupons
Station #2A
2902 Triple Cell
Figure 5-24. Comparison of municipal and regional water treatment using the same source waters (Lake Champlain, Vermont).
77
-------
Prechlorlnation Coagulants
Lake
Chamolain
1 ,
Vf 0.5 MG "Wf
^ Tank J (_
Ads
Cli
Fluoride
Postchlortnation
Zinc Orthophosphate
Figure 5-25. Schematic of Champlain Water District water treatment process.
distribution coupons (2901 units). The corrosion rate from the
inlet cylinder of station #2, using identical DF water, was used
as the base rate for comparing corrosion rate reductions in both
triple cell units (stations #2 and #4). A change in procedure was
made on June 1, 1989, to better reflect differences in corrosion
rates between CWD and BWR waters due to pH. The DF water
going to station #2 (ZOP only treatment) was changed to fin-
ished CWD water (post-chlorine, fluoride, and plant ZOP ad-
dition). These additions drop the pH of CWD DF water from
approximately 7.5 to approximately 7.2. Also, station #3 (a
single cell) coupons now were represented by the inlet cell of
station #2 using CWD finished (HS) water. (See Figure 5-26
for the laboratory coupon procedure change.)
Consumer Tap Sampling. The second approach used to
assess corrosion was sampling at consumer taps for specific
corrosion by-products. Over the course of approximately 2
To better reflect the actual differences in corrosion
rates between BWR and CWD waters, the source water
going into the CWD triple cylinder has been changed.
Previously, DF water (postfiltration but pre-plant ZOP and
C12 addition) has been used. Finished distribution water
(postchlorination and plant ZOP addition) will now feed to
this triple cylinder, which will represent the reduced pH
caused by these additions (typically pH 7.71 vs. pH 7.28).
The inlet cell will be used as the HS coupon value
and the ZOP solution being added will be adjusted to
continue to yield a 1 ppm zinc concentration. The required
number of test coupons will be reduced as the single
cylinder used for the HS can now be eliminated. The
"basic rate" will be the single cylinder "DF Lab" unit.
Historical data of the single cylinder and the first cyl-
inder of the triple cell (both DF water) show no variation
in corrosion rates. This change will more accurately reflect
CWD's distribution water and the comparison of BWR's
pH adjustment technique to CWD's ZOP treatment for
corrosion control.
Figure 5-26. Champlain Water District laboratory coupon procedure
change (06/01/89).
years, 16 different locations were sampled, with a total of 154
samples collected as of January 1990. Initial sampling included
first-draw samples (initial water from a tap after an extended
period of non-use, collected typically in the morning); a 2-min-
ute flush sample; and a 6-minute flush sample. Originally, the
length of non-use or stand time was not recorded. Collection
of the 6-minute flush sample was discontinued after the third
sampling (July 1988), because samples showed no reduction in
metal concentrations compared to the 2-minute flush samples.
Higher lead level sites were sampled more frequently. The met-
als originally tested for were iron, zinc, copper, and lead. The
iron, zinc, and copper tests were performed in the CWD labo-
ratory using a Hach DR/3000 Spectrophotometer (Hach Co.,
Loveland, Colorado). Copper was analyzed using Procedure
Code C.I2, Bicinchoninate Method; total iron using Procedure
Code 1.4, Ferrozine Method; and zinc using Procedure Code
Z.I, Zincon Method. Iron testing was discontinued after the
August 1988 sampling because extremely low levels were
found in all samples. The Vermont Department of Health Labo-
ratory in Burlington, Vermont, conducted the lead analysis. The
CWD laboratory performed all pH measurements using an A1-
tex Model 71 pH meter and a Hach Model 44300 combination
electrode.
Most sample volumes were 1 liter, but samples collected
in May 1989 were 250 mL, and samples collected in October
1989 were 1-L samples except at three locations. The first-draw
sample was broken down into two fractions, a 125-mL portion
followed by a 875-mL portion. The reported 1-L first-draw lead
concentration was calculated from the first two samples. Only
the 875-mL sample was used in testing for other metals, be-
cause of the volumes required. This sampling protocol was
followed to determine the lead contribution by faucet fixtures,
because these three locations had shown elevated lead levels.
The 125-mL sample primarily represented the water contained
in the faucet fixture.
Samples were collected in December 1989 from these
same three locations. To further identify the source of lead, five
samples were collected without flushing between sample col-
lection. Again, a 125-mL sample was taken followed by a 875-
mL sample. Then, a series of three 1-L samples were collected
without flushing between samples. Each liter represents ap-
proximately 25 feet of 1/2-inch copper pipe.
78
-------
5.4.3 Results
equally effective aud avetaged a 44 percent reduction in corro-
sion rates (Table 5-4). Elevated pH alone reduced corrosion
rates in only two of the five runs and increased corrosion rates
in three runs. The lead base-control coupons in the DF filtered
water averaged a corrosion rate of 1.25 mpy with a range of
0.76 to 1.69 mpy.
Table 5-4. Corrosion Rate Reductions of Laboratory Lead Coupons
84 days exposure time
5.4.3.1 Deviations
Certain inherent and operational deviations occurred:
1. Row rates, and therefore ZOP concentrations, to the
triple cells at stations #3 and #4 fluctuated and were
adjusted periodically to maintain a 1 mg/L concentra-
tion as zinc. Row variations were always as reduced
flows (increased ZOP concentration).
2. Water temperature of the laboratory coupons was
higher than the water temperature of the distribution
coupons. This would be expected to yield a higher
corrosion rate in the laboratory coupons, resulting in a
positive error in the percent of corrosion rate reduction
for the field coupons.
3. The pH of the DF filtered water supplied to the base
control coupons at station #1 (single-cell unit) was 0.3
to 0.4 (7.5 to 7.6) units higher than the distribution
water. The lower pH in the distribution water is because
of the addition of fluoride (1 mg/L), postchlorination
(approximately 1.8 mg/L as free chlorine) and plant
ZOP addition (0.3 to 0.4 mg/L) to DF water. This
higher pH water also was fed to the 2902 triple cell
representing ZOP-only treatment until the procedure
change of June 1, 1989.
4. Coupons in the distribution system were exposed to
much higher water velocities than were the laboratory
coupons.
5. Lead coupons are extremely soft and were subject to
abrasion during insertion into distribution mains.
5.4.3.2 Laboratory Coupon Analysis
Steel coupons in the laboratory study treated only with
ZOP showed the most consistent and highest average percent
reduction in corrosion rate (Table 5-3). Elevated pH plus ZOP
addition also showed good reductions in corrosion rates, al-
though not as high as the ZOP-only treatment. Elevated-pH
treatment only increased the corrosion rate in four of the five
runs with an average corrosion rate increase of 8 percent.
Table 5-3. Corrosion Rate Reductions of Laboratory Steel Coupons
84 days exposure time
Treatment
ZOP only
ZOP + *pH
*pH only
% Reduc-
tion*
84%
76%
(8%)
Range
62% - 95%
20% - 92%
(16%) - 5%
Avg.
Mpy
1.25
1.94
7.62
Avg.
pH
7.1
8.0
8.0
Range
44 - 3.45
0.47 - 7.20
5.04 - 10.46
'Based on comparison to raw water (pH = 7.0 - 7.2)
The base control coupons hi the DF filtered water averaged
a corrosion rate of 7.00 mpy, with a range of 4.81 to 9.00 mpy,
over 84 days of exposure.
All treatment techniques for lead coupons in the laboratory
showed a lower percent reduction and mpy rate than the steel
coupons. ZOP-only and elevated-pH-plus-ZOP additions were
Treatment
ZOP only
ZOP + *pH
ApH only
% Re-
duction*
44%
44%
2%
Range
17% - 63%
3% -71%
(39%) - 25%
Mpy
0.70
0.67
1.20
Avg.
PH
7.1
8.0
8.0
Mpy Range
0.46 - 1.30
0.49 - 1.23
0.85 - 1.87
'Based on comparison to raw water (pH = 7.0 - 7.2)
5.4.3.3 Distribution Coupon Analysis
Coupons placed at the four distribution sites (station #3
single-cell laboratory location is included here, because the
water used was finished water as supplied to the distribution
system) yielded the results shown in Tables 5-5 and 5-6.
Table 5-5. Corrosion Rate Reductions for the Distribution System
Steel Coupons
Location
High Service
Essex West
Kellog Rd.
DE Header
Overall Avg.
% Re-
duction*
53%
64%
47%
78%
61%
Range
8% - 86%
(2%) - 90%
20% - 67%
64% - 91%
23% - 84%
'Based on comparison to raw water (pH =
Table 5-6. Corrosion Rate
Lead Coupons
Location
High Service
Essex West
Kellog Rd.
DE Header
Overall Avg.
% Re-
duction*
39%
31%
30%
43%
36%
Reductions
Range
10% - 67%
(8%) - 51%
(26%) - 59%
24% - 56%
0% - 58%
Mpy
3.09
2.75
4.26
1.62
2.93
= 7.0 - 7.2)
Mpy
Range
1.02-6.92
0.49 - 7.64
1.55-6.38
0.98 - 3.45
1,01 -3.45
for the Distribution System
Mpy
0.74
0.76
0.78
0.63
0.73
Mpy
Range
0.42-1.21
0.30-1.20
0.25-1.28
0.27 - 0.86
0.31 -1.14
•Based on comparison to raw water (pH = 7.0 - 7.2)
The average control coupon corrosion rate for the DF fil-
tered water was 7.08 mpy, with a range of 4.55 to 9.51 mpy,
over 84 days of exposure.
The average lead control coupon corrosion rate for the DF
filtered water was 1.11 with a range of 0.84 to 1.29 mpy.
5.4.3.4 Consumer Tap Analysis
A total of 154 samples over a 2-year period were collected
from 16 different locations. Sampling was conducted with the
following three frequencies per location: three locations were
79
-------
sampled once, one location was sampled three times, two loca-
tions were sampled four times, three locations were sampled
five times, three locations were sampled six times, and one
location was sampled seven times.
First-Draw and 2- and 6-Minute Sampling. Thirty-five
first-draw and 2-minute flush samples were analyzed for total
iron and averaged 0.037 mg/L and 0.047 mg/L, respectively.
The 24 6-minute flush samples averaged 0.025 mg/L iron.
Forty-four first-draw and 2-minute flush samples from ZOP-
treated water were analyzed for zinc and averaged 0.422 mg/L
and 0.317 mg/L, respectively. The 18 6-minute flush samples
averaged 0.280 mg/L zinc. Fifty-seven first-draw and 56 2-min-
ute flush samples were analyzed for copper and averaged 0.343
mg/L and 0.079 mg/L. The 25 6-minute flush samples averaged
0.055 mg/L.
Fifty-seven first-draw and 2-minute flush samples were
analyzed for lead and averaged 37 |ig/L and 2 (ig/L. The aver-
age for the 25 6-minute flush samples was 1 (ig/L.
Sequential Tap Sampling. To identify lead sources in loca-
tions showing the highest lead levels, samples were collected
without flushing between samples. To represent water standing
in faucet fixtures, 125-mL samples were collected, followed
immediately by successive samples, representing water stand-
ing in the plumbing. The highest lead levels were from the 125
mL samples (Table 5-7).
Table 5-7. Lead Concentrations in Sequential Samples, |ig/L
Location* 125 ml 875 ml 1L 1L
1L
#10
#12
#13
#16
49
190
211
55
73
21
08
92
55
35
33
14
—
52
40
—
32
—
12
15
—
16
—
10
11
—
5.4.4 Discussion
The pathways and causes of corrosion, and the influence
of various factors on the corrosion process, are enormously
complex. Taking information from controlled research condi-
tions to applied field applications, where several parameters
continuously react and can vary regionally along with treatment
processes, sources, and season, is extremely difficult. EPA's
extended efforts to promulgate the lead and copper rule are
evidence of the complexities of addressing corrosion and its
by-products. In addition, several different materials typically
make up a distribution system and household plumbing, each
with its own dissolution characteristics.
One portion of this study, using controlled laboratory con-
ditions, measured the weight loss of coupons to determine cor-
rosion rates. Weight loss of lead and steel coupons (reported as
a mpy corrosion rate) compared three treatment techniques
(ZOP, elevated pH, and ZOP plus elevated pH). Baseline rates
were determined by coupon weight loss in water exiting the
direct filtration process (prior to any corrosion treatment). The
laboratory bench-scale results also were compared to weight
loss of coupons placed in the distribution system. The distribu-
tion coupons assessed ZOP's effectiveness under actual field
conditions.
The second stage of this study involved measuring certain
corrosion by-product levels at consumer taps. Unfortunately,
these levels were not measured at the consumer taps prior to
the CWD implementing its corrosion control program (ZOP).
Thus, a comparison of these values is not available. The result-
ing information, however, has been valuable in identifying lead
sources and in educating consumers on how to minimize their
exposure to lead.
5.4.4.1 Steel Coupons, Laboratory
In a similar laboratory study by Mullen and Ritter (22)
using mild steel coupons, corrosion rates of three different treat-
ment techniques were analyzed. Raising pH with caustic soda
to reach the pH of saturation reduced corrosion by approxi-
mately 13 percent, addition of sodium zinc glass phosphate at
2.0 mg/L reduced corrosion by 13 percent, and addition of ZOP
at 2.5 mg/L (0.5 mg/L as zinc) reduced corrosion by 55 percent.
Below 16°C, pH elevation increased the corrosion rate by 22
percent and, above 16°C, pH adjustment with caustic soda re-
duced the rate by 5 to 32 percent. The combination of ZOP plus
pH adjustment reduced corrosion by 79 percent. Below 13°C,
when pH was increasing corrosion, ZOP without pH adjustment
was reported more effective than ZOP plus pH adjustment. This
might be because pH adjustment with ZOP brought the pH (7.8
- 8.0) outside the optimal range for ZOP. The filtered effluent
pH used for ZOP addition was 6.8, with plant effluent after pH
treatment being 7.8. A 63 percent reduction hi corrosion rates
was reported for distribution coupons for a comparable time
and temperature period.
It is possible that the additional reduction in corrosion rate
reported in Mullen and Ritter's study by raising pH plus ZOP
addition was because the relatively low pH (6.8) was below the
optimal range for ZOP. The CWD's study showed no additional
reduction by raising the pH from 7.5 to 8.0.
These two studies, and others, make it obvious that any
corrosion control treatment program that does not account for
other water quality characteristics might not result in successful
corrosion control. Pisigan, Jr. and Singley (23) noted that the
composite effects of pH and alkalinity combined into one pa-
rameter, buffer capacity, might be more useful in assessing the
corrosive behaviors of water. Significant differences in cor-
rosivity of two waters having similar qualities have been attrib-
uted by Loewenthal and Marais to higher buffer capacity of one
water compared to that of the other (24).
Conductivity also has been found to have a positive rela-
tionship to corrosion rates (23). As more ions (Na+ and HCO3")
are introduced into aqueous systems (to raise alkalinity and/or
pH) the ionic conductivity increases and enhances the corrosive
attack on metal. Other research also has shown that in some
circumstances the dominant effect of adding alkalinity might
be to increase corrosion by increasing conductivity (25). Dif-
ferent materials show different corrosion rate responses to a
80
-------
change in the water's chemistry. Stone et al. (25) found that an
increase in pH from 6.0 to 8.0 decreased new copper corrosion
by 50 percent and aged copper even more, while no change in
the corrosion rate of a zinc electrode occurred over a pH range
of 5.0 to 9.0.
While the use of mild steel (or other) coupons might be
helpful in the general assessment of a water's corrosivity, a
comprehensive analysis must consider the response of all dis-
tribution materials (none of which include mild steel) as well
as that of home plumbing materials. In the case of the latter,
corrosion by-products of components containing lead (brass
fixtures, leaded solder, and pipe) might be the dominant factor
that dictates a particular corrosion control treatment. It is criti-
cal, however, to consider additional factors when changes in a
water's chemistry are made.
5.4.4.2 Lead Coupons, Laboratory
The solubility of Pb"1"2 can be greatly reduced by increasing
pH into the range of 9.0 to 10.0 (26). Often, even hi a low
alkalinity, enough DIG is present for protective film formation.
The lack of significant reductions in corrosion rates for CWD
coupons exposed to the 8.0 pH adjusted water would indicate
that factors other than pH are the dominant rate-controlling
factors.
Substantial reductions in the theoretical solubility of Pb"1"2
were computed for a system containing several levels of ortho-
phosphate at a DIG concentration of 32 mg CaCOjfL (28). The
results indicated that a substantial reduction of lead solubility
could take place when the pH is increased from 7.0 to 9.0 with
an orthophosphate concentration of only 0.5 mg/L. (The theory
that zinc in a ZOP formulation combines with lead to form a
protective film has not been proved.) The passivating action of
orthophosphate depends, at least, on the pH, DIG concentration,
phosphate concentration, and temperature (10). The CWD re-
sults showed significant reductions in lead coupon corrosion
rates in the pH range of 7.2 to 7.5 with no further reduction
when the pH was raised to 8.0. This shows that the optimal
range for ZOP in this system was 7.2 to 7.5. A major advantage
to corrosion control methods that do not substantially raise the
pH is lower organic halogen formation rates (THMs). Also, the
increased disinfection efficiency of free chlorine at lower pH
values has been well documented (26).
5.4.4.3 Distribution Coupons
At least three important variables—ZOP concentration,
temperature, and flow rate—make it difficult to identify the
cause of lower corrosion rate reductions of both steel and lead
coupons in the distribution system. The laboratory coupons
were exposed to higher temperatures and ZOP concentrations
(1 mg/L as zinc) but to greatly reduced flow rates as compared
to distribution coupons. Increased corrosion rates of mild steel
and copper due to high flow rates have been noted in other
studies (23). Low flow rates, typical of home plumbing, were
found not to affect corrosion rates (25). The flow rate in the
laboratory cylinders was less than that found in home plumb-
ing. The expected net result, because of flow differences, would
be for the distribution coupon rate to be higher. Based on lead
solubility, the higher ZOP concentration dosed to the laboratory
coupons would be expected to reduce corrosion rates to a
greater degree. The higher laboratory water temperature should
increase corrosion rates compared to the colder distribution
water. The degree to which two of these factors, flow rate and
temperature, affected the corrosion rates is not possible to de-
termine. Some insight, however, might be gained as to the effect
of ZOP concentration by comparing the distribution location
designated as HS at station #3 and the laboratory location des-
ignated ZOP-only station #2. During the four test periods over
a 7-month span, the 1 mg/L dosed steel coupons averaged 1.46
mpy and lead coupons averaged 0.72 mpy. This compares with
3.35 mpy for steel coupons and 0.74 mpy for lead coupons at
the 0.3-0.4 mg/L concentration. (All time periods were the
same.) Comparison of the two laboratory test cylinders, both
using water exiting the direct filtration plant, showed equivalent
corrosion rates. Station #1 steel averaged 7.00 mpy (range 4.81
to 9.00) to station #2 at 7.08 mpy (range 4.55 to 9.51); station
#1 lead averaged 1.11 mpy (range 0.84 to 1.29) to station #2's
1.25 average (range 0.76 to 1.69). The higher ZOP concentra-
tion used in the laboratory was based on the recommendation
of the chemical supplier, whose previous experience indicated
that this adjustment yielded laboratory results that corre-
sponded to distribution environments.
Although these studies are helpful in designing and moni-
toring a corrosion control strategy, one should be aware that
because of lead's toxicity, corrosion control in systems incor-
porating lead-containing materials must target only the lead
levels in the water rather than a reduction in corrosion rates.
Therefore, lead control programs are somewhat different from
corrosion control programs normally designed for other metals
such as copper, iron, or galvanized steel, where there is more
concern about the lifetime of the plumbing materials.
5.4.4.4 Consumer Tap Analysis
Iron. Minimal amounts of exposed iron are in the CWD
distribution system and most CWD households. The CWD
rarely experiences any iron-related consumer complaints, as
evidenced by the extremely low iron concentrations reported.
The iron analysis was dropped after consistently low levels
were established. Iron, currently classified as a secondary con-
taminant, has a Secondary Maximum Contaminant Level
(SMCL) of 0.3 mg/L. This SMCL currently is being met, with
an average iron concentration of 0.04 mg/L for all first-draw
samples.
Zinc. The 44 first-draw samples from ZOP-treated water
had an average zinc concentration of 0.404 mg/L. This value
is higher than the amount attributable to ZOP addition. The
additional zinc probably was introduced from galvanized piping
and brass fixtures. The 0.266 mg/L average of the 18 6-minute
flush samples best represents the residual zinc concentration in
the distribution system. The corresponding 18 first-draw sam-
ples had an average zinc concentration of 0.429 mg/L.
Zinc is listed as a secondary contaminant with a SMCL of
5 mg/L. Even at substantially higher ZOP addition rates, the
total zinc concentration due to ZOP and corrosion by-products
would be well below the 5 mg/L SMCL.
81
-------
One concern about ZOP application is the possible zinc
buildup in municipal wastewater sludge. Sludges used in land
application or disposed of in landfills typically are regulated as
to maximum allowable metals concentration. The increase in
sludge zinc at one of the largest area facilities has not adversely
affected sludge disposal or usability. This plant has noted a
reduction in sludge copper levels. Other smaller wastewater
treatment plants also have not experienced any detrimental zinc
increases. One facility reported a reduction in sludge per acre
that could be applied due to maximum metal concentrations in
one sludge application. This case is the only one reported since
ZOP addition began in 1987. Given the requirements of the lead
and copper rule, the likelihood of the state approving a less
effective treatment technique because of a reduction of sludge
application seems remote. The additional phosphorus loading
at wastewater treatment plants due to ZOP application has not
resulted in any known adverse effects. Phosphorus is consid-
ered beneficial in sludge used in land application, but it might
pose a problem because of strict effluent phosphorus limita-
tions.
Copper. The average copper level in the 57 first-draw sam-
ples was 0.34 mg/L. Only two samples exceeded the new
MCLG for copper of 1.3 mg/L. The CWD foresees no problem
meeting the copper MCLG.
Lead. Site selection for the consumer tap analysis was
made before the lead and copper rule guidelines were estab-
lished. Sites selected are more representative of the CWD con-
sumer base and are not necessarily "high-risk" sites as specified
in the final rule. Sites consistently showing low lead levels were
sampled less often than sites having higher levels, which re-
sulted in a higher overall average for all sites.
Sample sites that showed low lead levels were relatively
consistent and concentrations did not vary greatly (Table 5-8).
Sites with levels >20 ng/L often showed a wide range of con-
centrations. The variability in stand times and one sampling
Table 5-8. Average Lead Concentrations at Consumer Taps
Cone. ng/L (number of samples)
using a 250-mL volume is not believed to be the cause of these
fluctuations. In other studies, particulates containing lead were
believed to be the source of similar variations. Large concen-
tration variations appear in many reported studies. The other
metal concentrations monitored in this study showed much
smaller fluctuations. Small changes in water chemistry that
significantly affect lead solubility compared to other metals, or
the mechanism by which lead corrosion by-products are intro-
duced, might be unique. Sites #12 and #16 had both high lead
levels and the highest copper levels. First-draw copper levels
at these sites were several times higher than all other sites
sampled. Site-specific factors are thought to be influencing cor-
rosion rates at these two sites. Electrical grounding to water
lines is known to affect corrosion rates, as is the joining of
dissimilar metals. Site #12 also showed the highest lead levels
in 125-mL samples (190 and 211 M-g/L). The faucet fixture has
a "goose neck" style of spigot. Sample sites #9 through #16 are
commercial locations and account for all of the highest lead
level sites. No identifiable cause for differences in lead levels
between residential and commercial locations is apparent.
Ten of 16 sites (63 percent) had average first-draw lead
concentrations of <15 |ig/L. An average value of 31 |ig/L was
obtained when each location's average was used to calculate
the overall average (based on 57 first-draw samples and using
2.5 ng/L for samples below the practical quantitation level
[PQL] of 5 Hg/L). Seventy-seven percent of all 2-minute flush
samples were below the PQL of 5 |ig/L. Only three samples (5
percent) were above 15 Hg/L, with respective values of 16, 18,
and 24 \igfL. The average lead concentration for all 2-minute
samples was 4 ^ig/L (using 2.5 \LgfL for results reported as <5
Hg/L).
In light of the variability commonly reported in first-draw
concentrations, a ±30 percent accuracy factor in analysis, and
the action level having to be met each monitoring period (as
compared to the THM regulation, which uses a rolling 12-
month average), a majority of PWSs are likely to exceed the
action level eventually.
Location #
1st Draw 2 Min. Rush 6 Min. Flush 5.4.4.5 Asbestos-Cement Pipe
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
<5(1)
17(1)
3(4)
<5(4)
4(5)
<5(3)
10(5)
7(2)
3(2)
36(6)
<5(1)
116(5)
76(6)
108 (2)
1(5)
102 (5)
<5(1)
24(1)
<5(4)
<5(4)
<5(5)
<5(3)
3(5)
<5(2)
<5(2)
<5(6)
<5(1)
5(5)
6(6)
9(2)
<5(5)
1(5)
—
—
2(3)
<5(3)
3(2)
<5(2)
3(2)
—
—
<5(3)
<5(1)
<5(2)
<5(3)
—
<5(2)
7(2)
A substantial portion of the CWD and/or town distribution
mains contain asbestos-cement (A-C) pipe. This study did not
address the response of A-C pipe to the different treatment
techniques. Other studies, however, have analyzed ZOP effects
on A-C pipe. Mah and Boatman (27) reported that a mixture of
lime and ZOP was the only inhibitor of six tested that was
believed to be beneficial in protecting A-C pipe. This mixture
consisted of lime and orthophosphate (5.0 mg/L) plus zinc (0.3
mg/L). In contrast to materials containing lead, zinc deposits
(not phosphorous) were found on the surface of the A-C pipe
after 218 days of exposure. This finding was contrary to the
belief that ZOP deposited a film containing zinc and phospho-
rus such as Zn3(PO4)2 or Z^lTO^MK^O. Subsequent com-
putations of chemical equilibria showed these results to be
reasonable and predictable. Schock and Buelow (28) reported
that orthophosphate salts of zinc provided substantial protection
to A-C pipe when added at proper concentrations and pH
ranges. Zinc was found to be the active agent in coating the
82
-------
pipe and protecting against asbestos fiber release and water
attack.
Tests at pH 8.2 by EPA's Drinking Water Research Division
(DWRD) using lead and A-C pipe in the same recirculation
system showed that ZOP provided corrosion inhibition for both
types of piping material due to the action of the orthophosphate
ion. The lead solubility in these systems was found to be gov-
erned by the formation of lead orthophosphate compounds
rather than ZOP. Protection of distribution materials also would
depend, at the least, on pH and dissolved carbonate concentra-
tions.
5.4.5 Conclusions
• A review of corrosion studies shows that a successful cor-
rosion control treatment for a particular water might be in-
effective in another water, or might even increase corrosion
of certain materials in contact with that water.
• Each distribution and residential plumbing material has its
own dissolution characteristics.
• Each contacted material might be affected differently by
water quality constituents (chemical and physical), external
factors (electrical grounding), and dissimilar materials con-
tacting each other.
• Past corrosion control studies and treatments by utilities
probably have been broader in scope than future studies
designed to comply with the lead and copper rule might be.
Future studies required by the lead and copper rule might
mistakenly be narrowly targeted at reducing lead corrosion
from sources identified in a specific water system.
• Identical materials of different ages can respond differentiy
to identical water chemistries.
• Lead pipe and/or lead coupons respond differently to a water
than does leaded solder in copper plumbing. If use of lead
coupons is anticipated, corrosion rates correlated with con-
sumer tap lead levels might be beneficial.
• The use of steel coupons in bench-scale or distribution lines
probably has little or no benefit in assessing lead corrosion
responses. Correlating corrosion rate reductions in steel cou-
pons to corrosion by-product levels at consumer taps cur-
rently is not possible. Steel coupons might be useful in
helping to assess the general corrosivity of a water.
• Some of the factors that should be considered in choosing
the best lead control treatment strategy include distribution
materials; type of storage facilities; commercial customer
uses and needs; potential impacts on wastewater treatment
plants; disinfection by-product levels; and EPA, state, and
local regulatory requirements.
• To date, no lead service lines or interior lead plumbing have
been identified in CWD households.
• The primary source of lead is leaded solder and faucet fix-
tures in consumer plumbing.
• Sequential sampling at locations with elevated lead levels
showed that faucet fixtures at these locations contributed
significantly to the high lead levels.
• Selected sample sites represented a broad cross section and
included residential and commercial structures. Included in
the 16 sites were eight commercial locations, one mobile
home, one residence with lead-free solder, one residence
built after 1982 with lead solder, and five residences built
prior to 1982 with lead solder. Only one of these sites met
the requirements of the lead and copper rule.
• The absence of lead service lines and interior plumbing in
CWD households limits all sample sites to homes that were
built, or that replaced interior plumbing, between 1983 and
1986.
• All the highest average lead levels were at commercial lo-
cations.
• Sites showed either consistently low (<10 Hg/L) lead levels
or significant variations.
5.4.6 Recommendations
• Initiate the materials survey and establish monitoring sites
as specified in the lead and copper rule as soon as possible.
• Set up an accelerated sampling program from the identified
sample base.
• Assess the probability of the action level being exceeded in
any one monitoring period.
• Review literature further as to zinc concentrations effective
in A-C pipe corrosion control.
• Establish asbestos fiber levels from A-C pipe.
• Based on literature review, asbestos fiber levels, and docu-
mentation of lead levels at lead and copper rule sample sites,
test on a full-scale basis a ZOP formulation that maintains
the needed zinc concentration for A-C pipe protection and
that provides a higher orthophosphate concentration (ap-
proximately 0.5 to 1.0 mg/L).
• Determine the effectiveness of the new ZOP formula and
propose a corrosion control treatment program to the state.
5.5 Reducing Corrosion Products in Municipal
Water Supplies: Chippewa Falls, Wisconsin
5.5.1 Background
The ground water from the Chippewa River Valley in west-
central Wisconsin is naturally soft (hardness = 80 mg CaCO^L)
and generally of good quality, both chemically and bacterially.
83
-------
The low pH (6.5) of the water makes it aggressive to metal
plumbing.
In July 1984, during routine testing of its water supply, the
City of Eau Claire found lead concentrations at South Junior
High averaging 285 |ig/L with some 100-mL samples contain-
ing lead concentrations as high as 1,000 |J.g/L. At that time, the
applicable federal drinking water standard for lead was 50
The discovery of elevated lead levels at the junior high
school began a chain of events that resulted in a detailed sam-
pling program throughout the area by the Wisconsin Depart-
ment of Natural Resources (WDNR) and the subsequent
request to the City of Chippewa Falls to centrally treat its water
source to reduce corrosion products. The situation quickly be-
came a heated local issue. Chippewa Falls was displeased more
than the other area communities, not only because additional
expense would be necessary to provide water treatment, but
also because the city's slogan is "Home of the World's Purest
Water." The local residents did not want corrosion control
chemicals added to their water supply under any circumstances.
5.5.2 Water System
Water is supplied to the City of Chippewa Falls from seven
wells. Five wells are located on the east side and two wells are
on the west side of the city (see Figure 5-27). Water is distrib-
uted to residential, commercial, and industrial customers
through a pipe network of approximately 73 miles. Water pres-
sure is provided by three elevated storage tanks with a total
capacity of 2.25 million gallons.
5.5.3 Regulations
As an operator of a municipal water supply system, the
city is regulated by the Wisconsin Administrative Code, Rules
of the Department of Natural Resources, Environmental Pro-
tection (cited as NR Code). Specific regulations (prior to the
promulgation of EPA's new lead and copper rule) included the
following:
• NR 109.11 establishes a maximum lead concentration of 50
• NR 102. 12 specifies that samples taken for compliance be
collected at the customers' tap.
• NR 109.14 allows the WDNR to require the water supplier
to implement corrosion control measures.
• NR 109.60 specifies a secondary standard (aesthetic limit)
of 1.0 mg/L for copper.
Sampling in Fall 1984 indicated that some buildings in the
service area were exceeding the 50 |J.g/L standard for lead con-
tent at the water tap when the first 250 mL was withdrawn in
the morning. Measured levels within the distribution system did
not reveal elevated lead concentrations prior to entering service
lines, indicating corrosion occurring within the service piping.
The Langelier Index is based on a chemical analysis of the
water supply and is an indication of the water's tendency to
precipitate or dissolve calcium carbonate (see Section 5.1.1). A
negative value indicates a tendency for the water to dissolve
calcium carbonate, whereas a positive value indicates a ten-
dency to precipitate calcium carbonate. Chippewa Falls has an
index value of -2.2.
Many water samples collected at the customer's taps (250
mL) also have exceeded both the state's 1 mg/L limit and the
new federal MCLG of 1.3 mg/L for copper.
The WDNR contacted city officials in December 1984 and
requested that corrosion control methods be implemented. In a
letter dated January 1985, WDNR clarified its earlier position
and required the City of Chippewa Falls to "centrally treat its
water source to reduce corrosion products."
Chippewa Falls hired a consultant to study the lead prob-
lem and evaluate alternative treatment methods. Work was be-
gun in late March 1985 to develop information and present-
technical solutions. Six areas that were studied and that are
described briefly in the following sections include:
• Hot water flushing of service lines
• Aging study on corrosion activity
• Centralized treatment
• Pilot test area
• Implementation
• Operating results
5.5.4 Hot Water Flushing
Research indicates that lead levels at the water tap tend to
decrease over a period of years. Two theories offer a possible
explanation for this reduction. One idea is that the tinning flux
used by plumbers during construction dissolves over a period
of time and slowly leaches into the water supply. A second idea
is that the piping system tends to become coated with metallic
oxidation products that prevent rapid dissolution of the lead
solder used in copper piping systems.
To test the first theory, three newly constructed homes were
chosen for testing. Early laboratory work revealed that flux
rapidly dissolves at a water temperature above 140°F. To see if
dissolving the flux would reduce lead levels, a series of four
hot water flushes was made on each of the test homes at ap-
proximately 3-week intervals. A portable hot water heater was
connected to the cold water system in each house. A hot alkaline
soap solution was circulated through the plumbing system for
several hours to dissolve residual flux. With the cooperation of
the homeowner, a first-draw tap water sample was collected on
the day preceding and the day following each flushing. A fol-
lowup sample also was collected 3 weeks after the fourth flush-
ing. The water samples were sent to a commercial testing lab
for lead analysis. The results are presented in Table 5-9.
84
-------
CHIPPEWA FALLS
POSTAL ZIP CODE 54729
Eddy Well Field
(5 Wells)
Figure 5-27. Well locations, Chippewa Falls, Wisconsin.
85
-------
Table 5-9. Hot Water (140° F) Rushing Results* (lead in
Cycle No. Home A Home B
Home C
Before one
After one
Before two
After two
Before three
After three
Before four
After four
Followup
660
330
610
140
22
12
120
52
60
400
200
230
96
180
540
150
230
110
730
820
760
177
850
590
510
230
610
*250-mL sample size.
Test results show a lower lead level in the final samples
than originally measured. Consistent results were not obtained,
however, and only general trends can be evaluated. It is evident
that levels below 50 ^.g/L were not obtained through the hot
water flushing program. Further work in this area was not done.
5.5.5 Aging of Service Pipe
To determine if lead and copper concentrations decrease
with time, city residential customer taps were sampled. Six
sampling groups were developed based on the age of the home.
The ages and sample size by group were:
• Less than 1 year—9 homes
• 1 to 2 years—10 homes
• 3 to 5 years—10 homes
• 5 to 10 years—10 homes
• 10 to 20 years—12 homes
• Greater than 20 years—12 homes
A minimum of nine homes in each age group was sampled for
first morning water drawn from the kitchen tap. Samples (250
mL) were sent to a commercial testing laboratory.
A total of 63 samples was collected. Ten of the samples
exceeded the state's 50 M-g/L standard for lead. All but one of
these homes with elevated lead levels were less than 2 years
old. The tenth sample was from a home more than 20 years of
age, which had been replumbed recently. Based on the sam-
pling, it appears that the elevated lead levels diminish over a
2-year period.
Copper levels also were measured in 63 samples. Only 11
of the 61 samples collected were below 1.3 mg/L. Six of these
low levels occurred in homes more than 20 years old. The other
5 samples testing at low levels were distributed among the
remaining sample groupings. Copper concentrations did not
appear to decrease with time.
5.5.
-------
was scheduled. At this election, the voters considered the fol-
lowing referendum question:
Shall the city of Chippewa Falls centrally treat its water with
chemicals to lower its corrosivity in order to meet state
drinking water health standards as to lead and taste and color
standards as to copper?
On October 8,1985, the citizens of Chippewa Falls showed
that they were in agreement with the Mayor and Council, with
343 voting "Yes" and 1,508 voting "No" to chemical addition
to control corrosion in the water system.
5.5.9 LegalAction
The legal channels were explored because the Mayor and
Council were not convinced that systemwide treatment was the
proper course of action. The Wisconsin Department of Industry,
Labor, and Human Relations (DILHR) had issued an emer-
gency order (9/25/84 to 2/22/85) that banned the use of lead
solder. DILHR reasoned that if the sources of the elevated lead
were in fact the 50/50 lead solder and the flux used in the
soldering process and this source was removed, the lead levels
should drop within a few years. Only a small percentage of
homes experienced elevated lead levels, and the majority of the
citizens were not in favor of chemical treatment. Furthermore,
the city was willing to test anyone's water, and if elevated lead
levels were detected, to furnish bottled water for drinking.
In addition, raising the pH by chemical addition to the
water would result hi economic burdens. First, there would be
a capital investment for the chemical treatment facilities and
annual operation and maintenance costs. Second, two major
industries, Leinenkugel Brewery and Cray Research, were con-
cerned when the discussion of treatment of the city water sup-
ply began. If the pH had been increased to 8.5 to 9.0 as
originally thought, both industries would need to lower the pH
for some of their applications.
The city requested a contested case hearing and on October
1, 1985, a "Notice of Prehearing Conference" was issued by
the Division of Hearings and Appeals. On October 22, 1985, a
prehearing conference was to be held for the purpose of iden-
tifying all parties to the proceeding, to simplify the issues that
would ultimately be contested at the hearing, and to establish
appropriate schedules for the presubmission of documentary
evidence and for prehearing discovery. No testimony would be
heard at the prehearing conference; however, a date would be
set for the hearing on the merits at the conference.
On October 14, 1985, the city met with its special legal
counsel and its expert on corrosion, Vernon L. Snoeyink, PhD.,
of the Department of Civil Engineering, University of Illinois
at Urbana-Champaign. During this meeting with Dr. Snoeyink,
the problem and potential solutions were discussed. This dis-
cussion led to the idea of proposing a "pilot study" to the
WDNR.
On October 21, 1985, a meeting between WDNR and the
city was held at the Governor's request in his office in the state
capitol. During this meeting, it was agreed to implement a pilot
study. It also was agreed that, during the pilot study, the con-
tested case hearing would be held in abeyance. Also, during the
pilot study, the city would continue to supply bottled water on
request to those homes where tests indicated lead levels in
excess of the health standard
5.5.10 Pilot Study
With water supplied from seven wells through four pump
houses on two sides of the city, central treatment, in fact, would
require treatment at multiple locations.
Since any treatment method would involve construction to
house the needed equipment, it was necessary to know what
the space requirements would be. The total treatment required
had to be determined before construction and equipment pur-
chases began. A pilot study was desirable to verify whether the
addition of caustic soda would sufficiently reduce the lead and
copper to comply with the current standard and the proposed
standards. The pilot system would determine the levels of lead
and copper that could be reached with caustic soda alone. If
additional treatment was needed, an orthophosphate could be
added and its effects determined.
Dr. Snoeyink believed that the pilot system also was de-
sirable to see what the effects "would be on homes with galva-
nized (pipe) services. With the change in water quality, the
corrosion of galvanized services should be less than without
any treatment, but that had to be verified. Also, a possibility
existed that the treated water might release the scale built up
in galvanized services and actually cause a poorer water quality
as the pipes were cleaned.
The size of the pilot project was reviewed and the first plan
considered was to select and treat buildings known to show
high lead levels. This plan was not believed to be a feasible
alternative, because it would not accurately simulate what was
being done with the whole system. If individual homes were
treated, it would not be very easy to control the feed rate.
A map of the city was studied and an area on the south end
of the city was selected (see Figure 5-28). By closing one valve
on a 20-inch water main, the total flow from the well would be
directed to the test area. This area is controlled with only three
small water main outlets, all near the northeast comer of the
area. The plan was to feed caustic soda with equipment that
ultimately would be used at this well when treatment was in-
stalled.
5.5.11 Goals of the Pilot Study
If the addition of caustic soda resulted in meeting the drink-
ing water standards for lead and copper, the city agreed to
implement systemwide treatment. If caustic soda was not ef-
fective in meeting standards for lead and copper, the city would
add an orthophosphate with the caustic soda (probably at re-
duced concentrations) for up to 3 additional months with moni-
toring. If the WDNR or the use of orthophosphates required the
addition of other chemicals (such as chlorine) or if the ortho-
phosphate caused adverse operational effects, then the city was
not obligated to perform systemwide treatment. If either of the
87
-------
CH1PPEWA FALLS
POSTAL ZIP CODE 54729
Figure 5-28. Pilot test area, Chippewa Falls, Wisconsin.
88
-------
above situations occurred or if the health limits were not at-
tained, both parties agreed to resume negotiations.
5.5.12 Implementation of the Pilot Study
Prior to starting the chemical addition, a test program
would be started to get baseline results to help determine the
effects of the pH adjustment Samples would be collected on a
weekly basis for 1 month prior to starting chemical addition.
WDNR and the city jointly selected 10 sites with copper plumb-
ing and 4 sites with galvanized plumbing. These homes had all
shown elevated levels of lead and/or copper.
During the month that baseline data were being gathered,
the chemical feed equipment and a day tank were purchased
and installed After starting chemical feed and closing the valve
to restrict the flow from the test well into the test area, problems
developed with trying to control the pumps to maintain equal
water levels in all elevated tanks. By partially closing the dis-
charge valve of the test well to reduce the volume of water
delivered to the test area, the water levels in the tanks could be
controlled. This action increased the operating pressure, and the
feed pump selected did not feed accurately at the increased
pressure. A different feed pump had to be installed. It took about
2 weeks to get an even pH and work out mechanical problems.
Because of these difficulties, the pH was not raised to the
desired 8.5 level. By mid-December, however, weekly test re-
sults were showing that lead levels and copper levels were
meeting drinking water standards without reaching a pH of 8.5.
On December 19,1985, the city and WDNR personnel met
to review test results. Test results showed that in the test homes
where the pH had been raised from 6.5 to 7.8, lead and copper
levels were below health standards.
Figure 5-29 shows the results from the tests at 461 A Street.
This site had a copper water service and copper plumbing and
had shown elevated lead and copper levels before the pilot
study. In May and September 1985, the lead had been at 310
and 490 Jig/L, respectively. This figure shows that with the pH
below 7.0, the health standards were being met. The other test
sites showed similar results. In fact, after November only three
sites had any samples that exceeded 15 |J.g/L.
pgft. 100
11/13 11/27 12/10 12/23 1/5 1/20 2/4 3/4
4 pH
•pH
Copper x 100
Figure 5-30 shows the results from the tests at 467 Chip-
pewa Street. This site is one of four with a galvanized water
service and galvanized plumbing. It appears that the addition
of caustic soda had no noticeable effect on the levels of iron or
zinc in the water.
4 PH
1/20
pH -*- Iron
-x-ZInc
Figure 5-29. pH, copper, and lead at the 461 A Street copper services
during pilot study.
Figure 5-30. pH, iron, and zinc at the 467 Chippewa Street galvanized
service.
Based on the lower lead and copper results at a lower pH,
the WDNR agreed to allow chemical feed to continue at re-
duced feed rates and to study the effects. If needed, the feed
rates could be raised until allowable lead and copper levels were
attained.
5.5.13 Decision to Treat
The following factors prompted the city to proceed with
chemical addition on a systemwide basis:
1. The use of 95/5 solder in new homes was not successful
in attaining lead levels meeting the primary (health)
drinking water standards.
2. Indications were that the state's secondary (aesthetic)
drinking water standard of 1.0 mg/L for copper would
soon be changed to a primary (health) standard of 1.3
mg/L.
3. It was proposed to reduce the state's primary standard
for lead from 50 ng/L to 20 ^g/L.
4. Meeting the standards for lead and copper were attain-
able with the pH at 7.0 and no chlorination, producing
no noticeable change to the water.
On March 4, 1986, the City Council adopted a resolution
stating that because the pilot project for addition of caustic soda
had been successful, systemwide treatment (50% NaOH)
should be implemented.
5.5.14 Implementation of Central Treatment
On April 19,1986, a letter from the city to WDNR advised
the department that the schedule below had been approved by
the Chippewa Falls Common Council on March 18, 1986, and
that the city should proceed on this schedule:
89
-------
• Open bids on Apiil 17, 1986.
• Board of Public Works reviews and makes recommendation
on April 21, 1986.
• City Council awards contract on May 6, 1986.
• Construction begins on May 15, 1986.
• Contractor completes work; city begins equipment installa-
tion on July 15, 1986.
• City completes equipment installation on August 15, 1986.
The Council authorized the preparation of plans and speci-
fications as well as advertising for bids for the building addi-
tions to the pumphouses to accommodate water treatment
equipment. The initial schematic plans submitted to WDNR
were approved Public Utilities staff prepared bid plans and
specifications for the building additions. Bids were opened on
April 17, 1985, and an award was made on May 6, 1986.
Construction began in late May and was completed in early
August 1986. Public Utilities staff sought quotations on the
needed equipment and tanks. All equipment was purchased and
installation, with the exception of electrical work, was com-
pleted by Public Utilities staff.
5.5.15 Facilities Constructed
At the East Well Field, where five pumps are located, a 20
ft x 22 ft addition was built onto the existing pumphouse. Inside
are housed two 1,600-gallon storage containers for bulk caustic
soda. A separate chemical feed and day tank are provided for
each well pump. The chemical feed pumps are electrically in-
terlocked to the matching well pump.
At each West Well Field pumphouse, a 10 ft x 22 ft addition
was constructed. Each of these buildings contains a 1,000-gal-
lon storage tank along with a chemical feed pump and day tank.
At each of the installations, the main storage tanks are within
a containment area of sufficient size to hold the contents of the
tanks.
All installations also have:
• A transfer pump to move the chemical from the storage tanks
to the day tanks.
• Connections for transfer of caustic from transport to the
storage tank.
• Water supply for flushing and safety eye wash stations.
• A stand-by chlorine feed system including a day tank and
pump interlocked to each well.
As part of the central control system, a temperature alarm
was added (because of the high freezing temperature of caustic
soda) along with a flooding alarm. If the liquid level on the
floor rises 1/8 inch above the floor, an alarm will be sounded.
All alarms are transmitted back to the wastewater treatment
plant, where there is 24-hour coverage. The estimated costs
compared to actual costs are listed in Table 5-10.
Table 5-10. Construction Costs
Estimated
Actual
Building Additions
Equipment, Tanks, Piping, Misc.
Installation Costs
TOTAL
$38,300
31,100
8,800
$78,200
$44,127
33,172
2,821
$80,220
The engineering report also estimated $11,900 for design
costs. With the utility staff doing this work, these costs were
included in the normal operating budget and not included
above. Installation costs also are distorted because staff labor
costs are not included above.
5.5.16 Monitoring
The WDNR required monitoring of the treatment, and in
September 1986, the city proposed a monitoring program to
WDNR. The proposal was based on input from the WDNR
district engineer on the frequency of sampling and on the analy-
sis of several parameters. The city's proposal was as follows:
• In conjunction with sampling for bacteria, collect pH sam-
ples at the 15 sites sampled each month.
• On a daily basis, monitor the pH at the wastewater treatment
plant laboratory.
• Three times per week, monitor the pH at the individual
wells.
• Select 10 sites for the monitoring of copper, lead, and pH
of first-draw water to evaluate the effects of treatment in
reducing corrosion products.
• Implement a sampling schedule as follows:
— For the first 3 months, sample and analyze on a monthly
basis.
— For the next 9 months, sample and analyze on a quar-
terly basis.
— Thereafter, sample and analyze on an annual basis.
• Continue to use the same laboratory for the copper and lead
analysis, thus avoiding the need to split samples with the
state laboratory to verify accuracy.
In October 1986, the WDNR Area, District, and Central
offices reviewed and approved this monitoring program without
change. They indicated that modifications to the monitoring
program might be necessary based on monitoring results and
the evaluation of treatment effectiveness.
90
-------
5.3,1 / sampling protocol
The sampling protocol up to this stage was to collect three
250-mL samples in the morning after the water had been in the
pipes overnight. The first sample was taken in the morning
before any water was used. The second sample was taken after
the water was run for approximately 2 minutes or until the
water felt cool. The third sample was taken after the water was
run for approximately 5 minutes after the first sample. The first
sample reflected the water in the faucet assembly, the second
sample reflected the water in the house plumbing, and the third
sample reflected the water in the distribution system.
The sample from 1301 Waldheim Road taken on October
25, 1984, shown in Table 5-11, is typical of most results. It
shows that as the water was run, the lead levels dropped. This
finding indicated that the elevated lead seemed to come from
the faucet assembly, and a lesser amount of lead from the house
plumbing.
Table 5-11. Lead Levels in the Samples Collected at 1301 Waldheim
Road (ug/L)
Lead
First Draw
Second Draw
Third Draw
400
8
4
During early sampling, there was some discussion about
whether the first-draw sample should be used for determining
compliance with the health standard, or if an average of the
three samples should be used. With an average, a first-draw
sample could be well in excess of the 50 ng/L limit and the
average would still be less than the limit.
In July 1987, WDNR tried to compare the three 250-mL
sampling procedure with a two 1,000-mL sample routine. The
250-mL and 1,000-mL samples were taken at a home on suc-
cessive days in the morning before any water had been used.
Table 5-12 shows a comparison at the same home as above.
These results also are compared in Figure 5-31 for the lead
results and Figure 5-32 for the copper results. The 1,000-mL
sample appears to be about an average of the first-and second-
draw of the 250-mL.
Table 5-12. Lead and Copper Levels in the Samples Collected at 1301
Waldheim Road
250 mL
1,OOOmL
Rrst Draw
Second Draw
Third Draw
Lead
18
<3
<3
Copper
(mg/L)
3.3
3.8
0.32
Lead
6
<3
Copper
(mg/L)
5.2
4.0
On the same days, the same sampling procedures were
used hi another home (at 47 Stump Lake Road), shown in Table
5-13. In that case, the first-draw had a lower lead level than the
second-draw using the 250-mL sample. However, the first-draw
1;
10
250 mL & 1,000 mL Sample
1«t Draw 2nd Draw 3rd Draw
H|7/8/S7-250mL EM 7/9/87 -1.000 mL
Figure 5-31. Lead levels in samples collected at 1301 Waldheim Road.
250 mL & 1,000 mL Sample
lit Draw 2nd Draw 3rd Draw
•I 7/8/87 • 250 mL EM 7/9/87 • 1,000 mL
Figure 5-32. Copper levels in samples collected at 1301 Waldheim
Road.
Table 5-13. Lead and Copper Levels in Samples Collected at 47 Stump
Lake Road
250 mL
1,000 mL
First Draw
Second Draw
Third Draw
Lead
(WJ/L)
17
31
4
Copper
(mg/L)
0.67
1.50
0.93
Lead Copper
(ng/L) (mg/L)
20 0.82
8 0.56
with a 1,000-mL sample still appeared to be an average of the
first-and second-draw lead levels with the 250-mL samples.
At a third location where this procedure was used, 1100
West River Street (Table 5-14), the lead results were different.
The results indicated elevated lead levels in all samples, with
Table 5-14. Lead and Copper Levels in Samples Collected at 1100
West River Street
250 ml
I.OOOmL
First Draw
Second Draw
Third Draw
Lead
(W/L)
81
28
16
Copper
(mg/L)
.86
1.9
.29
Lead
(W/L)
62
63
Copper
(mg/L)
.82
.51
91
-------
some exceeding the health standard. The high lead levels at this
location should be considered differently, because this location
is not a residence. It is the Water Department Maintenance
Building, which has a large plumbing system with little water
usage except when water meters are being tested.
In all three of the above cases, the data indicated that the
copper levels were elevated for all water sitting in the house
plumbing. These levels varied depending on time as well as on
water chemistry.
Based on these data, although the existing standards were
not violated at every location, corrosion products were clearly
present at elevated levels. As a result, WDNR asked that the
city:
• Raise the pH in the system to 8.0 or above. This was still
not at the saturation point but would be closer. WDNR was
willing to allow the city to operate at a lower pH provided
that the treatment was effective.
• If it should wish to attempt some other method to reduce the
level of corrosion products, submit that proposal to WDNR
by October 1, 1987.
• Begin monthly sampling for lead, copper, and pH at the 10
selected locations once the treatment scheme was imple-
mented.
Figures 5-33 and 5-34 show the past 2 1/2 years' results
from 1301 Waldheim Road and 43 to 45 Stump Lake Road. At
1301 Waldheim Road, the last exceedance of the 50 |ig/L limit
was in October 1988. The last time this site exceeded the new
15 |ig/L limit was May 1990. At Stump Lake Road, which is
an eight-unit condominium, the last exceedance of the 50 (ig/L
limit was in April 1989. The last time it exceeded the 15 |0.g/L
limit was in May 1990. The third site, 1100 West River Street,
has not exceeded the 50 |j.g/L limit since February 1989 and
has not exceeded the 15 |J.g/L limit since August 1990.
Oet Jin Apr Jul Oct J«n May Aug Oct Ftta Miy
Oct Jin Apr JJ Oct Jin May Aug Oct Fit) M«y
• PH
•Lnd
• CopptrxlO
Figure 5-33. pH, lead, and copper at 1301 Waldheim Road.
On December 15, 1987, the city proposed that data from
the 10 selected locations were not providing significant infor-
mation to warrant continuing. A proposal was made to monitor
•pH
' CopperxlO
Figure 5-34. pH, lead, and copper at 43-45 Stump Lake Road.
at only the two locations that showed some significant levels
of lead and copper: 1301 Waldheim Road and 43 Stump Lake
Road. If these samples indicated that the treatment could attain
the desired results, then the city would require pH monitoring
only.
On January 20, 1988, the city and WDNR met to discuss
the corrosion product monitoring. The WDNR District Engi-
neer summarized the meeting as follows:
• Monitoring Frequency and Location. Most of the sites that
are being monitored show lead levels well below the stand-
ard. The city believed that these additional data serve no
purpose because little has changed over the past year. Cop-
per levels also were down significantly from where they
were prior to treatment. The city believed that since the
standard for copper at that time was not a health-based stan-
dard, any reduction was an indication that the treatment was
working. The recommendation was that the city continue to
monitor at Stump Lake Road, 1301 Waldheim Road, and the
Water Department shop (1100 West River Street) on a quar-
terly basis for lead and copper.
• Optimum pH. The industries in the city were reporting
problems from the higher pH and customer complaints had
increased. These issues, coupled with the cost of treatment,
prompted the city to look for an optimum pH level to main-
tain. Using copper levels as an indicator and choosing an
arbitrary level of 1.0 to 1.3 mg/L as appropriate, it appeared
that a maximum pH of about 7.8 would be effective. Real-
izing that the pH tends to vary within the system, the city
proposed to aim for apH of 7.5 to 7.8 throughout the system.
The recommendation was an operating pH of 7.5 to 7.8.
• pH Variations. Sampling data often showed a wide vari-
ation of pH within the system on any given day. The city
would initiate some bench testing to determine a cause for
this variation.
• Summary. It appears that the caustic addition was having a
measurable effect on the level of corrosion products. It was
suggested that the pH be maintained between 7.5 to 7.8 and
monitored at three locations quarterly. If significant levels
92
-------
'1C StU.1.1 IILJL
LllUUICl CYCULUcluOLl IrOUIUUC
performed.
On February 8,1988, the city was notified that monitoring
at the three locations for lead, copper, and pH should continue
and that the pH should be maintained at a minimum of 7.7 in
the system.
During 1987, only one site had a lead level that exceeded
50 ng/L. This was 1100 West River Street, the Water Depart-
ment Maintenance Building. Two of the 10 sites had exceeded
the new 15 \ig/L limit but were under the 50 \ig/L limit at that
time.
5.5.18 Feed Rates
Based on the results observed during the pilot study, the
feed rates were set to attain a pH of 7.0 rather than the design
pH of 7.9 or the calculated pH of 8.5 to reach the Langelier
Saturation Index.
The feed rate at startup, based on the pilot study, was about
20 to 21 gallons of caustic soda per million gallons of water.
When the systemwide treatment was started, the feed rates were
up to 24 to 26 gallons per million gallons. When WDNR re-
quired that the pH be raised to 7.7, the feed rates had to be
raised to average about 28 to 30 gallons per million gallons.
The average feed rates and average daily pumpage are listed in
Table 5-15 by year.
Table 5-15. Chemical Feed Rates of Caustic Soda
Year Avg. MGD
Caustic Soda
Gal/MG
1986
1987
1988
1989
1990
3,504
4,058
4,156
3,951
3,854
24.5
24.0
26.5
30.8
28.0
5.5.19 Operation and Maintenance Costs
The annual operation, maintenance, and chemical costs for
the past 5 years are listed in Table 5-16 and displayed in Figure
5-35.
The operation costs have decreased as the system problems
are worked out and as less monitoring and testing is conducted.
The caustic soda costs have increased drastically. This increase
was due to the higher feed rates and was also due to rising costs
for caustic soda.
Table 5-16.
Year
1986
1987
1988
1989
1990
Annual Operation and
Operation
$ 8,133.24
21,422.80
14,579.74
14,597.68
7,814.17
Maintenance Costs
Maintenance
$1,678.81
1,678.81
236.33
687.77
2,089.53
Caustic
Soda
$10,910.00
20,990.00
38,325.00
57,345.00
51,080.00
1986 1987
• OPERATIONS JS
1988
I MAINTENANCE
1989 1990
HH CAUSTIC SODA
Figure 5-35. Annual operation and maintenance costs for the chemical
feed system.
Originally, in late 1986, the price for caustic soda (sodium
hydroxide) was $155 per anhydrous ton. Since then, the cost
has increased as shown in Table 5-17. This increase is shown
graphically in Figure 5-36. In 1989, there was a shortage of
caustic soda and suppliers established quotas for existing cus-
tomers and would not take any new customers. As a result of
the monthly quotas, shortages existed in some of the wells,
which caused pH adjustment to cease.
Table 5-17. Caustic Soda Costs
Month/Year of Increase
Cost Per Anhydrous Ton
1986
September 1987
December 1987
May 1988
August 1988
October 1988
January 1989
April 1989
July 1989
September 1989
May 1990
October 1990
g aoo
$155.00
195.00
215.00
245.00
300.00
330.00
350.00
420.00
410.00
375.00
400.00
420.00
^^/ ^^
^^
^
Aug Sap Due May Aug Oct Jan Apr
1936 1987 1988 1989
MONTH OF PRICE CHANGE
Aug Sep May Oct
1990
Figure 5-36. Cost of caustic soda per anhydrous ton.
93
-------
5.6 Evaluating a Chemical Treatment Program to
Reduce Lead in a Building: A Case Study
For many people, a significant fraction of total daily water
intake comes from the workplace, such as an office building, a
factory, or a school. Most people spend about one-third of their
work days in the building environment, which is being scruti-
nized for health and safety factors by employees and employers.
Buildings have the same types of plumbing materials as resi-
dences and are subject to the same types of potential problems.
Corrosion of the distribution system components results in the
leaching of lead and copper into the drinking water. Buildings
also have unique situations and problems. A large building
might have hundreds or even thousands of water taps and might
serve a population larger than a small community. Because of
concerns about the quality of drinking water hi buildings, more
and more tenants are getting their drinking water from water
coolers, drinking water fountains, and water taps that have been
analyzed for levels of lead and copper. If these levels are high,
solutions are being investigated and implemented at the site.
A building that EPA has studied is a research facility con-
structed 5 or 6 years ago in the Washington, DC, area. Because
of a variety of construction problems, several years and $10 to
$15 million were spent correcting structural and other defects
in the building. When the drinking water was sampled, how-
ever, elevated lead levels were found throughout the building.
The water had the characteristics shown in Table 5-18. The pH
was in the mid-7s, the alkalinity was 37, and hardness was at
about 50. As Table 5-19 shows, the flush samples ranged from
less than 5 to about 81. Sixty to 70 locations were sampled. A
4-day static test was made and the lead levels ranged from 63
to more than 100
Table 5-19. Lead Levels in Samples of Flushed and Static Water from
Various Locations
Table 5-18. Water Quality Characteristics
PH
Alkalinity
Total Hardness
Calcium
Magnesium
Iron
Manganese
Chloride
Sulfate
Fluoride
Silica
7.53
37 mg/L (as CaCOs)
46.9 mg/L
14 mg/L
2.9 mg/L
0.11 mg/L
<0.05 mg/L
14 mg/L
13 mg/L
0.62 mg/L
1.3 mg/L
It was decided that a possible solution might be to flush
the water system, putting 7 million gallons through the building
in 4 to 5 days. Presumably, the flushing would remove some
of the lead and perhaps age the system. But flushing 7 million
gallons of water through a system (requiring 6 months under
normal circumstances) does not necessarily age the system.
After flushing, two sequential samples showed lead levels rang-
ing from 11 to more than 200 p.g/L.
Pb(ug/L)
Sample Locations
Flushed (1 5 min.)
(12-14-90)
4-Day Static
(12-26-90)
Flushed (2-8-91)
1
2
Static (2-12-91)
1
2
Fountain
81
63
72
36
40
13
Lab Sink
>5
183
27
11
96
242
Utility
Closet
>5
63
47
21
121
69
Locker
Room
Sink
>5
101
246
21
189
1,480
Several weeks later samples were collected again, and the
lead levels were as high as 1,000 |J.g/L. According to the speci-
fications, the building did not contain lead solder. As shown in
Table 5-20, the percentage of lead from about 20 different
locations varied from very low to approximately 50 percent,
indicating 50/50 lead solder. It was apparent that the contractor
had not followed specifications and had used lead solder at least
in part of the building.
Table 5-20. Percentage of Lead in Solder Samples
Sample
Percentage of Lead
1
2
3
4
5
6
7
8
9
10
51.2
0.17
58.0
0.13
34.1
0.15
39.3
49.1
46.4
42.0
EPA's Drinking Water Research Division in Cincinnati,
Ohio, was contacted at this point, and the division explained
some potential problems: lead solder joints, brass valves, and
brass fittings. The building had more than 600 outlets and thou-
sands of lead-soldered joints. Usually, three or four soldered
joints are associated with each fixture. The main lines might
have seven or eight additional soldered joints. Several large
brass valves located on the incoming water lines could cause
potential problems at the water coolers, which now carry warn-
ings.
The primary question was: "What are the main sources of
lead?" EPA believed that the main sources would be the brass
fixtures and the lead solder. The second question was: "Can
you determine the amount of lead from each source?" The
94
-------
intent was to go back to the contractor and recover some funds
to help pay for a solution or to be compensated for the lead
problems. Facility personnel had not been fully aware of the
lead levels that could result from brass fixtures. EPA sampled
about 40 taps and found the same elevated lead levels. EPA
collected 250-mL samples of 24-hour standing water and the
lead levels varied extensively. The values EPA obtained were
different from the initial samples, even though the same sites
were sampled. A total of 12 sequential samples were collected
at about six locations that were likely to produce high values.
The first 2 samples were 60-mL samples and the remaining 3
to 12 samples were 125-mL samples. The samples were ana-
2SOO
5678
SAMPLE SEQUENCE f
I 300-
5678
SAMPLE SEQUENCE t
10 11
Figure 5-37. Lead (a) and zinc (b) concentrations in samples collected
sequentially (Room 3329).
lyzed for lead, zinc, and copper. Figures 5-37 and 5-38 show
the results of the data for zinc (part of the brass in the brass
fixture) and lead. High levels of zinc were noticed in the first
two samples at 60 mL each and represent the water contact with
the fixture. If the only source of lead was the brass fixture (brass
generally contains about 7 or 8 percent lead) then the concen-
tration pattern of lead would be similar to that of zinc. As shown
in Figure 5-37, higher lead levels were found at the 6th and 7th
sample and at around the 9th through 12th samples, indicating
that lead was coming from the solder. On the other hand, if no
lead was found in the brass and the only source of lead was the
solder, the first few samples would have had low lead levels
and the following samples would have had high lead levels.
With these data, it was apparent that the fixtures, solder, and
some brass valves were contributing to the high lead levels.
5 6 7 B
SAMPLE SEQUENCE*
11 12
Figure 5-38. Lead (a) and zinc (b) concentrations in samples collected
sequentially (Room 1618).
Potential solutions were presented to the building manage-
ment. The people who worked in the building were well aware
of the problem and were urging management to replace all of
the plumbing. Management estimated that this would cost $2
million to $4 million and probably would take 6 months to 1
year. Although new plumbing was a potential solution, it was
ruled out because of the time it would take to implement this
solution. Apoint-of-use (POU) manufacturer was contacted and
there was some discussion about installing POU systems at
each tap. It was impractical, however, to place POU devices at
each of more than 600 taps. POU devices could be placed
within the lines, taking out the lead from the solder, but the
problem with the brass valve would still exist. As a result, POU
devices were not considered a practical solution.
The third potential solution, chemical treatment (such as
using corrosion inhibitors) was selected. A research plan was
developed to determine which chemical treatment scenario
would be most effective for the system. The research plan had
two phases. First, the effect of water usage on lead leaching
was evaluated. This new building had not been used and EPA
was convinced that, with water usage, the lead levels would
decrease. Second, the effect of adding a corrosion inhibitor was
evaluated. Three inhibitors were selected for evaluation: zinc
orthophosphate, "calcium" orthophosphate (manufacturer des-
ignation), and sodium silicate. In discussing these, some of the
management people, particularly the technical people, objected
to all three. They continued to favor having all the plumbing
replaced. Some were concerned about the potential for high
95
-------
sodium and others were concerned about zinc. Eventually, the
building personnel were convinced that action should be taken
and that it was likely that the water utility would add a corro-
sion inhibitor in the future.
Half of the building is devoted to laboratories, and as a
result, a total of eight sections of the building could be isolated.
A program was set up in which water would be run 5 days a
week for 30 minutes at a time, four times a day, with an hour
and a half between each time. Standing samples (16 hours) were
collected twice a week. All of these samples were 250-mL
samples. A meter was placed on the line leading into each
isolated section. Tap water from nine laboratories was sampled
in each, wing.
Baseline data, produced by collecting flushed water sam-
ples, are shown in Table 5-21. On the ground floor, the lead in
I
£
100
20 30 41 61 62 72 83 107
Figure 5-39. Water usage study—lead concentrations over time in
Room G402.
Table 5-21. Lead Baseline Data Collected at the Ground Roor and at
the Third Floor
Results - Initial (Pb (ug/L) Tests
Ground Floor
Third Roor
Tap
Flush
Standing
Rush
Standing
1
2
3
4
5
6
7
8
9
Avg.
11
8
3
5
4
6
2
6
2
5
131
112
50
291
117
330
99
125
109
152
11
5
5
14
6
3
102
7
4
17 (6.9)*
46
583
79
101
167
118
135
309
75
179
*6.9 is the average value excluding tap 7.
the samples averaged about 5 ng/L; on the third floor, the
average was about 7 \igfL. Standing samples were collected on
the ground floor and the third floor. The lead levels had a wide
range, but the average produced at the taps on the ground floor
was 152 jag/L, and on the third floor, 179 ng/L. During the
operation of the system, the lead levels had varied (Figures 5-39
and 5-40). The chemists who performed the analysis at the
laboratory in Cincinnati indicated that most of the samples with
high lead levels had some particulate material in the sample,
indicating that the system was still being flushed. There also
might have been some particulate in the line from solder being
broken off.
After about 80 days, there appeared to be a consistent
reduction in lead levels. Figure 5-41 shows average values for
the ground floor (nine taps). Until the 62nd day, the lead levels
varied greatly. Then the lead levels declined, probably due to
aging or film developing on the insides of the pipes. Figure
5-42 shows the average values for the third floor. The third
floor is the top floor; those rooms have not been used. Some
administrative and maintenance people had been using several
eoo
500
•->
a400
£ 300
200
100
0
Figure 5-40. Water usage study—lead concentrations over time in
Room 3325.
I
£
0 9 20 30 41 51 62 72 83 107
Figure 5-41. Water usage study—average lead concentrations from the
ground floor.
laboratories on the ground floor. It was assumed that these
individuals had used some of the taps in these laboratories,
which might explain why the ground floor lead levels seem to
be slightly lower than the third floor lead levels.
96
-------
200
180
=1- 100
so
ground floor. The goal is to have all taps eventually below 15
0 9 20 30 41 51 82 72 83 107
DAYS
Figure 5-42. Water usage study—average lead concentrations from the
third floor.
The number of samples collected at the taps that are below
50 \ig/L is increasing (Figures 5-43 and 5-44). Building per-
sonnel must make a decision about when they believe the water
is safe or potable. The third floor response to treatment is
slightly slower than that of the ground floor. It is suspected that
the slowdown results from the regular water usage on the
5.7 Iowa's Lead in Schools' Drinking Water
Program: More Than Just a Monitoring
Program
5.7.7 Introduction
The Lead Contamination Control Act of 1988 (LCCA) was
enacted on October 31, 1988. The passage of this act was
prompted by concerns that children were being exposed to
excessive levels of lead in drinking water in schools, pre-
schools, and daycare centers. The water at these locations was
of particular concern for three reasons. First, several models of
water coolers found in schools at this time were known to have
lead-lined storage tanks that contributed high levels of lead to
the water. Second, the pattern of water usage in these buildings
meant that water could sit in contact with any lead in the
plumbing for an extended period of time, leading to high lead
levels in the drinking water. Finally, children are more likely
than adults to suffer adverse health effects from exposure to
lead.
00 13 23 34
SS tS 78 M 97
DAYS
5.7.2 Requirements of the LCCA
The LCCA placed requirements on EPA, the Consumer
Product Safety Commission (CPSC), the states, the Centers for
Disease Control (CDC), schools, preschools, and daycare cen-
ters. The law also prohibited the sale of water coolers that are
not lead-free.
EPA was directed to distribute a list of water coolers that
were not lead-free and a guidance document and water testing
protocol to the states by February 1989. EPA also was directed
to make grants to states for the purpose of helping schools,
preschools, and daycare centers to test their water for lead and
to solve problems. These grants were never funded by Con-
gress.
Rgurc 5-43. Water usage study—number of samples with less than 50
u.g/1 and 15 ug/L lead from the ground floor.
13 23 34 44 55 SS 76
DAYS
97
Rgure 5-44. Water usage study—number of samples with less than 50
ug/l and 15 ng/L lead from the third floor.
The CPSC was directed to initiate a recall or other correc-
tive action for water coolers with lead-lined tanks by October
31, 1989.
The states were directed to distribute the EPA guidance
information, the list of certified laboratories, and the list of
water coolers that were not lead-free. The states also were
directed to establish programs by July 31, 1989, to assist
schools, preschools, and daycare centers in testing their water.
The state programs were directed to ensure that schools, pre-
schools, and daycare centers would take steps to eliminate lead
contamination from coolers that were not lead-free by January
31, 1990.
The law is somewhat unclear as to whether schools, pre-
schools, and daycare centers are required to test their water for
lead. The requirement actually is for the states to ensure that
the testing is done rather than for the institutions to perform
testing. If institutions do test their water, however, they are
97
-------
required to notify the public that the test results are available
for their inspection.
The CDC was directed to provide grants for prevention of
childhood lead poisoning. These grants were funded by Con-
gress.
5.7.3 More Than a Monitoring Program
Two years after implementation, Iowa's program was per-
ceived to be simply a "monitoring" program, that is, a program
to monitor the levels of lead that the schools found in water.
The LCCA directed, however, that lead levels above the action
level of 20 |ig/L (250-mL sample size) be reduced to safe
levels; if the monitoring showed problems, solutions were
needed. The distinguishing feature of Iowa's program has been
that it helps schools, preschools, and daycare centers with wide-
spread contamination problems to find solutions.
5.7.4 Implementation in Iowa: Monitoring Results
In Iowa, this program was assigned by the Governor to the
Iowa Department of Public Health (IDPH). IDPH assigned the
program internally to the Health Engineering Section. This sec-
tion provides technical assistance to the public and to local
health officials in the areas of plumbing, health effects of drink-
ing water contamination, well construction, and other related
subjects. No money was allocated specifically for this program.
An interagency effort and existing technical assistance pro-
grams were used, however, to overcome this lack of specific
funding.
The program was initially implemented through three mail-
ings made to all schools, preschools, and daycare centers. An
initial informational mailing was sent in May 1989. A followup
survey was sent in October 1989, and a second informational
mailing was sent in July 1990. The Iowa Departments of Edu-
cation and Human Services provided mailing lists for schools,
preschools, and daycare centers and shared the printing and
mailing costs for all mailings.
The first informational mailing (May 1989) contained a
memo offering assistance from IDPH, a list of Iowa laboratories
certified to test for lead in water, a list of coolers that were not
lead-free, and the EPA booklet, Lead in Schools' Drinking
Water, which contained the water testing protocol. The fol-
lowup survey asked schools, preschools, and daycare centers
to let IDPH know whether they were finding any coolers that
were not lead-free, whether they were testing their water for
lead, and if they were testing, what lead levels were found. The
second informational mailing (July 1990) contained an updated
list of coolers that were not lead-free, a question-and-answer
sheet to help alleviate some of the confusion revealed by the
responses to the followup survey, and a second followup survey
to be filled out and returned to IDPH.
The responses to the initial followup survey (October
1989) revealed that schools, preschools, and daycare centers
were confused about the requirements of the law and about the
level of lead that was to trigger action on their part to lower
the lead levels. Some thought that the law mandated water
testing, while others thought that it was a voluntary program.
Some schools thought that it was sufficient to test only a few
outlets rather than all drinking water outlets. In some cases,
schools took no action to reduce the high levels of lead that
were discovered. In addition, there was confusion about the 50
|ig/L MCL as opposed to the 20 \iglL action level that the
schools were directed to use. An effort was made to clear up
this confusion with the question-and-answer sheet that was in-
cluded in the second informational mailing. In addition, the
University Hygienic Laboratory agreed to send out a special
notice with test results to schools, preschools, and daycare cen-
ters to inform them of the 20 \lg/L action level and to direct
them to call IDPH with any questions.
In addition to indicating confusion, the responses to the
followup surveys indicated that schools, preschools, and day-
care centers were finding problems and needed help in solving
them. Because the program was assigned to a section of the
health department that already provided extensive technical as-
sistance to the public, it was natural that the program would
progress beyond a simple monitoring program to providing the
needed technical assistance.
5.7.5 Implementation in Iowa: Technical Assistance
Program
The technical assistance program consists of telephone
consultations to answer questions and limited assistance and
onsite investigations as needed to solve widespread contamina-
tion problems. The onsite investigation component of Iowa's
program is unique among the states. Local health officials and
water utilities are involved whenever possible. The investiga-
tions consist of a visit to the building to look at the plumbing,
take metal samples from solder and fixtures to screen for lead,
and determine where to take additional water samples to pin-
point the source of the problem and to provide a solution.
Extensive water testing often is required to find the source(s)
of lead and the solution. This water testing is provided free of
charge to the school through the state laboratory as part of the
investigation. To date, six investigations have been completed,
six more are under way or pending, and many more are needed.
5.7.6 Test Results from Iowa's Program
Statistics have been compiled from the returned followup
surveys to show the extent of lead contamination being found
by schools, preschools, and daycare centers. Two items were
considered when interpreting these results. First, these are re-
sults from 250-mL samples taken hi the morning before any
water is used. Second, not all institutions sampled all types of
sources, such as coolers, bubblers, and faucets. Some sampled
only coolers or only coolers and bubblers. The results of this
sampling showed that an unexpected number of schools, pre-
schools, and daycare centers found lead levels higher than the
20 |J.g/L action level.
There are 800 public school districts and private schools
in Iowa and 1,300 licensed preschools and daycare centers. A
followup survey was returned by 48 percent of the schools and
44 percent of the preschools, although many of these surveys
98
-------
were not complete. Some 34 percent of the schools and 25
percent of the preschools and daycare centers sampled drinking
water outlets for lead. A slightly smaller number (32 percent of
the schools and 23 percent of the preschools) actually reported
test results to IDPH. Of those reporting test results, 27 percent
of the schools and 8 percent of the preschools had at least one
source testing above the action level of 20 |ig/L. Because it was
anticipated that the standard for lead in water might be lowered,
the number of institutions reporting levels between 10 (ig/L and
20 (Ag/L was also recorded. Forty-one percent of the schools
and 13 percent of the preschools/daycare centers had at least
one source that tested in this range. A summary of the lead
results is presented in Table 5-22.
Table 5-22. Summary of Lead Levels Found by Institutions
Level Found (ug/L) Facilities Reporting
20
30
40
50
60
70
80
90
-30
-40
-50
-60
-70
- 80
-90
-100
>100
47
32
24
12
12
7
7
6
17
The points of water sampling where lead levels greater than
20 u,g/L were reported and the highest lead level recorded for
each type of sampling point are presented in Table 5-23. The
highest level reported on a followup survey was 3,700 |ig/L,
and the highest level found by the University Hygienic Labo-
ratory hi a sample sent in by a school was 10,000 |ig/L.
Table 5-23. Number of Facilities Reporting Lead Levels above 20
and the Highest Lead Levels Recorded from Those Facilities
Point of Sampling*
Cooler
Non-cooled bubbler
Faucet
Steam kettle
Facilities
Reporting >20
H9/L
45
28
44
4
Highest Levels
100ug/L
3,700 jig/L
1,100 ug/L
300 ng/L
*250-mL sample, overnight standing.
5.7.7 Example of a Solution: Finding a Solution for
New Hampton High School
In Summer 1989, New Hampton High School began sam-
pling for lead in water according to the EPA protocol. Officials
flushed outlets the day before testing to simulate normal use
during the school year. Three of 12 coolers tested higher than
20 |ig/L. Eight of the 9 remaining coolers tested higher than 10
|ig/L. All of the 8 faucets tested had levels higher than 20 \ig/L.
Additional tests were taken hi September 1989, shortly after
school started. Nine out of 9 faucets tested hghter than 20 Hg/L.
Eight of these were repeat samples. All of the lead levels were
lower than those found in the summer samples, even though
they were higher than 20
According to the EPA protocol, the next step was to take
flushed samples. For coolers, this test involved a 15-minute
flush. Two out of the three coolers tested had levels almost as
high after flushing as they did on first-draw samples. This result
indicated that the problem was likely to be in the upstream
plumbing. Flushed samples also were taken from the faucets.
Six out of 9 faucets tested lower than 10 jig/L after a 30-second
flush. The remaining three faucets tested between 10 ng/L and
20 M-g/L. This finding again indicates that the upstream plumb-
ing is contributing to the high levels.
The service connection and water main samples were all
low, indicating that the problem was within the building.
The school continued to follow the EPA protocol and at-
tempted to test the upstream plumbing. They took these samples
at shut-off valves upstream. The valve stem packings/seals were
partially dismantled to collect the samples. These samples had
lead levels 5 to 10 times higher than any first-draw or flushed
samples taken at the sources. The EPA protocol gave little
guidance for what should be done when the flushed samples
came back much higher than the first-draw samples. At this
point, the school contacted IDPH for assistance.
The IDPH onsite investigation revealed that the distribu-
tion system within the building was made primarily of galva-
nized pipe. The system was oversized in that it had 60 ft of
4-inch pipe for a school with approximately 400 students.
Analysis of solder and brass with a lead-in-solder test kit indi-
cated the likely presence of lead in some solder and brass at
the school. It appeared from the test results to date that the high
lead levels could be isolated largely to one part of the building.
(The EPA protocol recommends trying to isolate the levels.)
This line of reasoning was followed initially in the IDPH in-
vestigation, but it turned out that there was actually a different
reason for high- and low-testing areas.
Additional samples were taken at Iowa's expense, but the
results were confusing because there were large variations in
lead levels for water samples taken from the same source. The
results indicated that water corrosivity as measured by the Lan-
gelier Index should not be a problem. After all of the original
and additional samples were listed together by source and ana-
lyzed, a pattern of high and low readings according to the time
that the sample was taken appeared. The lead levels were high
when taken during the summer or just after school started when
water usage was low and stagnant water had not been com-
pletely flushed from the pipes. The lead levels were mostly
lower than 20 jig/L during periods of high water usage during
the school year. This result was confirmed by selected retesting.
The high lead levels found in the upstream samples are believed
to be due to contamination introduced from the valve bodies in
the abrasive action of dismantling them prior to the water sam-
pling. These valves are not believed to contribute to the high
99
-------
lead levels at tne points 01 initial sampling (.coolers, bubblers,
or faucets).
The major source of contamination is believed to be sedi-
ment in the galvanized pipe, compounded by the presence of
oversized pipe (because a large amount of water must be used
to completely empty the distribution system of stagnant water).
Smaller amounts of lead are contributed by lead solder and
brass faucets at the sources.
The-solution developed for the school is to flush all water
out of the building after more than a 2-day (weekend) vacation.
This solves the problem in all but four sources (three sinks and
one cooler.) Daily flushing, replacing brass faucets and/or lead
solder in the immediate vicinity of the faucets, or disabling the
outlets was recommended for these four sources.
5.7.8 General Observations from Investigations
The following observations have been summarized from
the investigations:
• Galvanized pipe should be suspected as the lead source
when contamination is widespread throughout a building
and when lead levels hi flushed samples are higher than
those in first-draw samples.
• Brass in faucets should be suspected as a source when most
of the faucets have high lead levels, most of the coolers have
low lead levels, and the lead levels in the 30-second flushed
samples from faucets are low.
• Lead solder should be suspected as a major source of lead
contamination when it can be seen from the outside that a
sloppy job of soldering was done and when the lead levels
in the flushed samples are high, but lower than those in the
first-draw samples.
• Schools can have high lead levels even if the water utility
monitoring in homes shows no problems, if the water is not
corrosive, or if the water utility uses corrosion inhibitors.
This situation is due to different sources of lead in homes
and school buildings, different water usage patterns in
homes and schools buildings, and the different sampling
protocols used by the schools and the water utility.
The future of Iowa's program is uncertain at this point. The
state will continue to provide technical assistance as time per-
mits. Additional followup with most schools, preschools, and
daycare centers is needed, however, to remind them to test their
drinking water and to ensure that any earlier testing was done
properly. This requirement undoubtedly will increase the need
for technical assistance to provide solutions for the lead con-
tamination problems that will be found. It is unlikely that ad-
ditional funding will come from the state of Iowa. One possible
source of federal funding exists, however, which Iowa currently
is pursuing.
5.8 References
1. APHA-AWWA-WPCF (1989). Standard Methods for Ex-
amination of Water and Wastewater. Washington, DC.
2. American Water Works Association (1990). Water Quality
and Treatment, Fourth Edition. McGraw-Hill, Inc., New
York, NY.
3. Amlie, R.F. andT.A. Berger (1972). Polarographic Analy-
sis of Lead (TV) Species in Solutions Containing Sulfuric
and Phosphoric Acids. Journal of Electroanalytical Chem-
istry. 36:427.
4. Breach, R., S. Crymble, and M.J. Porter (1991). Systematic
Approach to Minimizing Lead Levels at Consumers Taps.
Proc. AWWA Annual Conf. Philadelphia, PA.
5. Colling, J.H. et al. (1992). Plumbosolvency Effects and
Control hi Hard Waters. Journal of the Institute of Water
and Environmental Management. 6(6):259.
6. Gregory, R. (1990). Galvanic Corrosion of Lead Solder hi
Copper Pipework. Journal of the Institute of Water and
Environmental Management. 4:112.
7. Holm, T.R. and Schock, M.R. (1991). Potential Effects of
Polyphosphate Products on Lead Solubility in Plumbing
Systems. Journal of the American Water Works Associa-
tion. 83(7):74.
8. Kuch, A. and I. Wagner (1983). A Mass Transfer Model to
Describe Lead Concentrations in Drinking Water. Water
Research. 17(10):1301.
9. Millette, L. and D.S. Mavinic (1988). The Effect of pH
Adjustment on the Internal Corrosion Rate of Residential
Cast-Iron and Copper Water Distribution Pipes. Canadian
Journal of Civil Engineering. 15:79.
10. Schock, M.R. (1989). Understanding Corrosion Control
Strategies for Lead. Journal of the American Water Works
Association. 81(7):88.
11. Schock, M.R. and I. Wagner (1985). The Corrosion and
Solubility of Lead in Drinking Water, Chapter 4 in Internal
Corrosion of Water Distribution Systems. AWWA Research
Foundation. Denver, CO.
12. Sheiham, I. and P.J. Jackson (1981). Scientific Basis for
Control of Lead hi Drinking Water by Water Treatment.
Journal of the Institute of Water Engineers and Scientists.
35(6):491.
13. Snoeyink, V.L. and D. Jenkins (1980). Water Chemistry.
John Wiley and Sons, New York, NY.
14. AWWA Research Foundation (1990). Lead Control Strate-
gies. AWWA Research Foundation and American Water
Works Association, Denver, CO.
100
-------
15. Lane. R.W. et al. (1977). The Effect of pH on Silicate 23. Pisigan, R.A., Jr. and J.E. Singley (1987). Influence of
Treatment of Water in Galvanized Piping. Journal of the
American Water Works Association. 69(8):457.
16. Robinson, R.B. et al. (1990). Sequestering Methods of Iron
and Manganese Treatment. AWWA Research Foundation.
17. Kastanis, E.P. et al. (1986). Soluble Silicate Corrosion In-
hibitors in Water Systems. Materials Performance 25:19-
25.
18. Oliphant, R. (1978). Dezincification by Potable Water of
Domestic Plumbing Fitting: Measurement and Control.
Water Research Centre Technical Report TR-88.
19. Kingston, F.J. et al. (1987). Specific Adsorption of Anions.
Nature. 215:1459.
20. Huang, C.P. et al. (1975). The Removal of Aqueous Silica
from Dilute Aqueous Solution. Earth and Plant Letters.
27:265-275.
21. Karalekas, P. et al. (1983). Control of Lead, Copper, and
Iron Pipe Corrosion Pipe in Boston. Journal of the Ameri-
can Water Works Association. 75(2):92-95.
22. Mullen, E.D. and J.A. Ritter (1980). Monitoring and Con-
trolling Corrosion by Potable Water. Journal Health. May.
Buffer Capacity, Chlorine Residual, and Flow Rate on Cor-
rosion of Mild Steel and Copper. Journal of the American
Water Works Association. 79:(2):62-70.
24. Loewenthal, R.E. and G.V.R. Marais (1976). Carbonate
Chemistry of Aquatic Systems: Theory and Practice. Ann
Arbor Science Publications, Ann Arbor, MI.
25. Stone, A., C. Spyriclakis, M. Benjamin, J. Ferguson, S.
Reiber, and S. Osterhus (1987). The Effects of Short-Term
Changes in Water Quality on Copper and Zinc Corrosion
Rates. Journal of the American Water Works Association.
79(2):75-82.
26. AWWA Research Foundation and DVGW For-
schungsstelle (1985). Internal Corrosion of Water Distri-
bution Systems. Denver, CO.
27. Mah, M. and E.S. Boatman (1978). Scanning and Trans-
mission Electron Microscopy of New and Used Asbestos-
Cement Pipe Utilized in the Distribution of Water Supplies.
Scanning Electron Microscopy. (l):85-92.
28. Schock, MR. and R.W. Buelow (1981). The Behavior of
Asbestos-Cement Pipe Under Various Water Quality Con-
siderations: Part 2, Theoretical Considerations. Journal of
the American Water Works Association. 73(12):609.
101
-------
Appendix A
EPA/AWWA National Workshop on Control of Lead and
Copper in Drinking Water
Workshop Agenda
Monday, September 23,1991
8:45 a.m. WELCOME
Dale S. Bryson, Water Division, U.S. EPA Region 5
9:00 a.m. INTRODUCTION AND OBJECTIVE
Jon DeBoer, AWWA
9:30 a.m. LEAD AND COPPER REGULATION
Jeff Cohen, Office of Drinking Water, U.S. EPA
Harry Pawlowski, Office of Drinking Water, U.S. EPA
10:00 a.m. LEAD AND COPPER MONITORING PROGRAM USING EMPLOYEES AND CUSTOMERS
Jack DeMarco, Cincinnati Water Works
10:30 a.m. AT THE TAP MONITORING
Doug Neden, Greater Vancouver Regional District
11:00 a.m. BREAK
11:30 a.m. SELECTION OF AN ANALYTICAL LABORATORY
Jack C. Dice, Denver, Colorado Water Department
Noon LUNCH
1:30 p.m. CHARACTERIZING THE SYSTEM—BASELINE MONITORING
William G. Richards, Roy F. Weston, Inc.
2:15 p.m. IOWA'S LEAD IN SCHOOL'S DRINKING WATER PROGRAM: MORE THAN JUST A
MONITORING PROGRAM
Rita M. Gergely, Iowa Department of Public Health
2:45 p.m. BREAK
3:15 p.m. INTEGRATING WATER TESTING AND OCCUPANCY CERTIFICATION
Tom Bailey, Durham, North Carolina Department of Water Resources
3:45 p.m. EVALUATING CHEMICAL TREATMENT TO REDUCE LEAD IN A BUILDING:
A CASE STUDY
Thomas J. Sorg, Risk Reduction Engineering Laboratory, U.S. EPA
4:15 p.m. CORROSION CHARACTERISTICS OF MATERIALS
Vernon L. Snoeyink, University of Illinois
4:45 p.m. ADJOURN
103
-------
THP.sdny} September 1d> 1991
8:00 a.m. OVERVIEW OF CONTROL STRATEGIES
Mike Schock, Risk Reduction Engineering Laboratory, U.S. EPA
8:45 a.m. SECONDARY EFFECTS—CONFLICTS WITH CONTROL STRATEGIES
Richard H. Moser, American Water Works Service Co., Inc.
9:30 a.m. BREAK
10:00 a.m. OVERVIEW OF DIAGNOSTIC TOOLS FOR CORROSION STUDIES
Michelle M. Frey, Black & Veatch
10:30 a.m. IMPLEMENTING THE LEAD AND COPPER RULE AT THE STATE LEVEL
Lou Allyn Byus, State of Illinois EPA
11:00 a.m. STATE PERSPECTIVE
William F. Parrish, Jr., Maryland Department of Environment
11:30 a.m. LUNCH
1:00 p.m. SODIUM SILICATE FOR THE SIMULTANEOUS CONTROL OF LEAD,
COPPER, AND IRON-BASED CORROSION
Jonathan A..Clement, Wright-Pierce Engineers & Surveyors
1:30 p.m. ZINC ORTHOPHOSPHATE VS. pH ADJUSTMENT, AN OVERVIEW OF TEST
Al Ilges, Champlain, VT Water District
2:00 p.m. BENCH-SCALE CORROSION STUDIES AT BOULDER, COLORADO —
RESULTS AND EXPERIMENTAL SETUP
Brad D. Segal, City of Boulder Water Department
2:30 p.m. BREAK
3:00 p.m. REDUCING CORROSION PRODUCTS IN MUNICIPAL WATER SUPPLIES
John W. Allen, Chippewa Falls, Wisconsin
William F. Barry, Ayres Associates
3:30 p.m. SMALL WATER SYSTEM SOLUTION TO LEAD AND COPPER REGULATION
Victor Ertman, Cass Rural Water Users, Inc.
4:00 p.m. ARLINGTON'S EXPERIENCE WITH pH ADJUSTMENT AS ITS CORROSION
CONTROL STRATEGY
Travis Andrews, Pierce-Birch Treatment Plant, Arlington, Texas
4:30 p.m. METHODOLOGIES FOR ELECTROCHEMICAL CORROSION MEASUREMENT
Steve H. Reiber, University of North Carolina at Charlotte
5:15 p.m. OPEN DISCUSSION
Wednesday, September 25,1991
8:30 a.m. TWO CONCURRENT WORKSHOPS
DESIGN CONSIDERATIONS FOR PIPE LOOP TESTING
Anne Sandvig, Economic and Engineering Services, Inc.
S. Boris Prokop, Economic and Engineering Services, Inc.
Mike Schock, Risk Reduction Engineering Laboratory, U.S. EPA
DESIGN CONSIDERATIONS AND PROCEDURES FOR COUPON TESTS
Chester H. Neff, Illinois State Water Survey
John Ferguson, University of Washington
10:15a.m. WORKSHOPS REPEATED
Noon ADJOURN
104
-------
Appendix B
Units and Conversions
Metric to inch-pound units
LENGTH
1 millimeter (mm)=0.001 m=0.03937 in.
1 centimeter (cm)=0.01 m=0.3937 in.=0.0328 ft
1 meter (m)=39.37 in.=3.28 ft=1.09 yd
1 kilometer (km)=l,000 m=0.62 mi
AREA
1 cm2=0.155 in.2
1 m2=10.758 ft2=1.196 yd2
1 km2=247 acres=0.386 mi2
VOLUME
1 cm3=0.061 in.3
1 m3=l,000 L=264 U.S. gal=35.314 ft3
1 liter (L)=l,000 cm3=0.264 U.S. gal
MASS
1 microgram (p.g)=0.000001 g
1 milligram (mg)=0.001 g
1 gram (g)=0.03527 oz=0.002205 Ib
1 kilogram (kg)=l,000 g=2.205 Ib
Inch-pound to metric units
LENGTH
1 inch (in.)=25.4 mm=2.54 cm=0.0254 m
1 foot (ft)=12 in.=30.48 cm=0.3048 m
1 yard (yd)=3ft=0.9144 m=0.0009144 km
1 mile (mi)=5,280 ft=l,609 m=1.609 km
AREA
1 in.2=6.4516 cm2
1 ft2=929 cm2=0.0929 m2
1 mi2=2.59 km2
VOLUME
1 in.3=0.00058 ft3=16.39 cm3
1 ft3=1728 in.3=0.02832
1 gallon (gal)=231 in.3=0.13368 ft3=0.00379 m3
MASS
1 ounce (oz)=0.0625 lb=28.35 g
1 pound (lb)=16 oz=0.4536 kg
ft U.S. GOVERNMENT PRINTING OFFICE:1993-750-002/60952
105
-------
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268
Official Business
Penalty for Private Use
$300
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand comer.
If you do not wish to receive these reports CHECK HERE D;
detach, or copy this cover, and return to the address in the
upper left-hand corner.
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
EPA/625/R-93/001
------- |