United States Office of
Environmental Protection Drinking Water
Agency (WH-550)
Washington, DC 20460
«&EPA Corrosion in
Potable Water Systems
Final Report
-------
SumX No: C79-014
FINAL REPORT
CORROSION IN POTABLE WATER SYSTEMS
Contract No. 68-01-5834
Prepared by:
David W. DeBerry
James R. Kidwell
David A. Malish
SumX Corporation
P.O. Box 14864/1300 E. Braker Ln.
Austin, Texas 78761
February 1982
PROJECT OFFICER:
Mr. Peter Lassovszky
Science and Technology Branch
Criteria & Standards Division, ODW (WH-550)
U.S. Environmental Protection Agency
401 M Street, SW
Washington, D.C. 20460
-------
DISCLAIMER
This report has been reviewed by the U.S. Environmental
Protection Agency and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commerical
products constitute endorsement or recommendations for use
-------
ABSTRACT
The purpose of this investigation was to collect, review, evaluate, and
present existing information to determine whether a sufficient data base is
available to develop corrosion control regulations for the water works indus-
try as required by the Safe Drinking Water Act. To accomplish this objective,
an exhaustive literature search was completed which included a review of the
various materials used in the water works industry and their corrosion char-
acteristics. Results of laboratory and field research on each material as
related to corrosion in the water works industry are extensively reviewed and
data is presented as appropriate. Major emphasis is placed on assessing the
conditions of service and water quality characteristics in potable water
systems on the corrosion or deterioration of each material. A review of
corrosion monitoring and detection techniques is given which addresses the
various methodologies used to identify and evaluate corrosive waters. Avail-
able corrosion prevention and control techniques are also evaluated and pre-
sented. Additionally, case histories of corrosion control programs are
presented for examples. Finally, the information and data presented in these
reviews are compiled and presented in tabular form. These tables provide an
overall view of the nature of the corrosion problems in the water works indus-
try and can be used as a guide for the initial consideration of corrosion
control regulations.
Corrosion in potable water systems may be caused by either inherent fac-
tors or design, construction, or operational deficiencies. Most materials
used in potable water systems are susceptible to corrosion. The properties
of the material and the composition of the water are interactive in the
occurrence or inhibition of inherent corrosion. Corrosion susceptibility is
also influenced by many other factors including temperature, local flow rate,
pH, alkalinity, carbon dioxide, dissolved solids, and minor chemical con-
stituents of either the water or corroding material. Corrosion problems may
also be caused by poor choice of materials, coupling of dissimilar metals,
improper installation practices, incorrect design allowing unnecessary stag-
nant areas or crevices, release of unstable waters, or addition of additives.
Potable water distribution and plumbing systems may contain a variety of
materials. The environmental conditions noted above can have effects on
individual materials that differ both in type and degree and alleviation of
corrosion problems for one material or type of installation could create
problems with another or result in deterioration of overall water quality.
Results of this study show that the nature of corrosion and possible corro-
sion control alternatives are extremely complex. Limiting or predicting the
corrosiveness of water by the use of a universal corrosion index or parameter
is not feasible at this time. Instead it appears that corrosion control can
only be accomplished through a comprehensively applied program on a community
water system case by case basis.
m
-------
This report is submitted in fulfillment of Contract No. 68-01-5834 by
SumX Corporation under the sponsorship of the U.S. Environmental Protection
Agency.
IV
-------
TABLE OF CONTENTS
Section Page
1 INTRODUCTION 1-1
BACKGROUND 1-1
OBJECTIVES 1-4
2 CHEMISTRY AND WATER CHEMISTRY BACKGROUND 2-1
GENERAL ASPECTS OF CORROSION AND LEACHING IN
POTABLE WATER 2-1
TYPES OF CORROSION 2-2
CORROSION INDICES 2-3
GENERAL CORROSION BIBLIOGRAPHY 2-9
CORROSION INDICES BIBLIOGRAPHY 2-9
3 MATERIALS USED IN THE WATER WORKS INDUSTRY 3-1
PIPES AND PIPING 3-1
STORAGE TANKS 3-6
REFERENCES 3-9
4 CORROSION CHARACTERISTICS OF MATERIALS USED IN THE
WATER WORKS INDUSTRY 4-1
IRON BASED MATERIALS 4-1
Corrosion of Iron 4-1
Effect of Di ssol ved Oxygen 4-3
Effect of pH 4-5
Effect of Dissolved Salts 4-9
Effect of Dissolved Carbon Dioxide 4-11
Effect of Calcium 4-13
Effect of Flow Rate and Temperature 4-16
Effect of Other Species in Solution 4-17
Comparison of Cast Iron and Mild Steel 4-18
Corrosion of Galvanized Iron 4-19
Effect of Water Quality Parameters 4-19
Stagnant Conditions 4-22
Hot Water Corrosion 4-24
Stainless Steels 4-26
Passivity 4-26
Type of Corrosion and Effect of Alloy
Composition 4-27
-------
TABLE OF CONTENTS
(continued)
Section Page
Environmental Effects on Corrosion of
Stainl ess Steel s 4-27
Results in Potable Water 4-28
CORROSION OF COPPER IN POTABLE WATER SYSTEMS 4-28
General Considerations 4-30
Uniform Corrosion of Copper 4-31
Effect of 02 4-31
Effect of pH 4-32
Effect of Free C02 4-35
Effect of Temperature 4-36
Effect of Mi seel 1 aneous Parameters 4-36
Localized Corrosion of Copper 4-38
Causes of Pitting 4-38
Impingement Attack and Flow Rate Effects 4-40
Copper Al 11 oys 4-40
Corrosion of Brasses 4-40
Corrosion of Bronzes 4-42
Other Copper Al 1 oys 4-44
CORROSION OF LEAD IN THE WATER WORKS INDUSTRY 4-44
Effect of Flow Rate and Volume of Water Flushed 4-47
Effects of Di ssol ved Oxygen 4-49
Effect of Hardness 4-50
Effects of pH 4-51
Effects of pH and Hardness 4-53
Effects of Alkalinity 4-56
Effects of Temperature 4-60
Effects of Chiorination 4-60
Effects of Carbon Dioxide 4-61
Lead Release from Solder Joints 4-62
CORROSION OF ALUMINUM IN THE WATER WORKS INDUSTRY 4-63
Effects of Velocity 4-65
Effects of Temperature 4-66
Water Quality Effects 4-66
ASBESTOS-CEMENT PIPE PERFORMANCE IN THE WATER
WORKS INDUSTRY 4-76
Causes of Asbestos Fiber Release 4-79
Organic Rlease from Asbestos - Cement Pipe 4-88
CONCRETE PIPE 4-89
PLASTIC PIPE 4-91
VI
-------
TABLE OF CONTENTS
(continued)
Section Page
Polyvinyl Chloride (PVC) 4-92
Polyethylene 4-92
Polybutylene 4-94
Acrylonitrile-Butadiene-Styrene (ABS) 4-94
Polypropylene 4-94
Deterioration and Release from Plastic Piping 4-94
REFERENCES , 4-100
5 CORROSION MONITORING AND DETECTION 5-1
SPECIMEN EXPOSURE TESTING 5-2
ELECTRICAL TEST METHODS 5-6
CHEMICAL ANALYSIS FOR CORROSION PRODUCTS 5-10
REFERENCES 5-13
6 CORROSION PREVENTION AND CONTROL 6-1
MECHANICALLY APPLIED PIPE LININGS AND COATINGS 6-2
Hot Applied Coal Tar Enamel 6-2
Epoxy 6-3
Cement Mortar 6-4
TANK LININGS AND COATINGS 6-5
Coal Tar Based Coatings 6-5
Vinyl 6-6
Epoxy 6-6
Other Mechanically Applied Tank Lining 6-6
CORROSION INHIBITORS 6-8
CaC03 Precipitation 6-10
Sodium Silicate 6-13
Inorganic Phosphates 6-16
Miscellaneous Methods 6-19
ECONOMICS 6-20
Benefit/Cost Analysis.... 6-20
Trends and Costs of Mechanically Applied Linings
and Coatings 6-23
Costs of Corrosion Control by Chemical Applications... 6-25
CASE HISTORIES 6-33
Seattle 6-33
Carroll County, Maryland 6-36
Orange County, California 6-37
Additional Corrosion Control Practices 6-39
REFERENCES 6-40
vn
-------
TABLE OF CONTENTS
(continued)
Section Page
7 CONSIDERATIONS FOR CORROSION CONTROL REGULATIONS 7-1
REFERENCES 7-12
8 RECOMMENDATIONS 8-1
VI 1 1
-------
TABLES
Table Page
1 Correction Parameters Used With The Refined SI 2-5
2 Corrosion Indices 2-8
3 Materials Used for Transmission and Distribution Lines. 3-3
4 Materials Used for Service Lines 3-5
5 Pipe Used in U.S. Water Supply Distribution Systems by
Utilities Serving Over 2500 Persons 3-7
6 Chemical Composition and Corrosion Characteristics of
Waters Investi gated 4-14
7 Results of Correlation Analysis 4-16
8 Water Composition 4-23
9 Zinc, Cadmium, Lead Contents in Stagnation Water 4-23
10 Cadmium and Lead Content of Stagnant Water 4-24
11 Typical Average Analysis of New York City Reservoir
Water 4-29
12 Stainless Steels—Pitting in Reservoir Water 4-29
13 Conditions of Copper Dissolution Experiments of
Tronstad and Veimo 4-33
14 Copper in Initial Morning Sample of Water from
Copper Pipes 4-37
15 Ranges of Composition of Cu-Zn Alloys 4-41
16 Ranges of Various Compositions of Cu-Sn Alloys 4-43
17 Results of Worcester Lead Sampling Analysis Program 4-47
18 Lead in Water Taken from an Occasionally Used Tap
24 Hours after Last Flushing 4-48
19 Lead in Water Taken from a Tap Not Used for About
Six Months 4-48
20 Results of Investigation of Water Quality on Lead
Corrosion 4-54
-------
Table Page
21 Characteristics of Finished Water Supplied to Cambridge,
Roston, and Somervilie 4-53
22 Lead Concentrations in Water Stagnant for One Hour
in a New Simulated Household Copper Plumbing System 4-62
23 Composition of Typical Aluminum Alloys Used in
Fresh Waters 4-64
24 Water Velocity Effects on Pitting of Aluminum 4-65
25 Partial Compositions and Pitting Data for Seventeen
Fresh Waters 4-69
26 Effect of CaC03 on Maximum Pit Depths 4-74
27 Field Data Study - Data Collected and Results 4-81
28 Effects of Pipe Aging 4-82
29 Summary of Field Data Collected by Buelow Et Al 4-85
30 Water Quality Conditions and General Observations for
Small Scale Experiments 4-87
31 Cement-to-Calcium Oxide Ratio 4-90
32 Typical Physical Properties of Major Thermoplastic
Piping Materials 4-93
33 Typical Extraction Test Results 4-97
34 Concentration of MEK and THF in Water Samples at
Various Residence Times in PVC Pipe 4-98
35 Planned Interval Test 5-5
36a Estimated Concentrations of Compounds Detected in
The Water in The Bayou Cassette Ground Storage
Water Tank 6-7
36b Dosages and Preferred Water Parameters for Corrosion
Inhibitors 6-9
37 Analysis of A New Jersey Water 6-18
38 63-Day Corrosion Test on Cast Iron Using Treated New
Jersey Water 6-18
39 Copper Concentrations in Domestic Waters 6-21
-------
Table page
40 Comparison of Estimated Costs and Savings of Corrosion
Related Deterioration With and Without Corrosion Control
Treatment 6-22
41 Extremes of Various Estimates of Use of Tank Linings
and Coati ngs 6-24
42 Regional Preference for Water Tank Linings 6-24
43 Cost Estimates 6-26
44 Corrosion Control by Calcium Carbonate Stabilization... 6-27
45 Materials and Their Application in The Water Works
Industry 7-5
46 Significance of Corrosion on Deterioration of Various
Materials Used in The Water Works Industry 7-6
47 Preferred Water Quality and Conditions of Service to
Minimize Corrosion of Materials Used in The Water Works
Industry 7-7
48 Applications of Corrosion Control Mechanisms 7-11
49 Materials and Their Associated Corrosion Products 8-5
-------
FIGURES
Figure
1 Effect of 02 Concentration on Corrosion of Mild Steel
in Slowly Moving Water Containing 165 ppm CaCl2 4-4
2 Effect of 02 Concentration on Corrosion of Mild Steel
in Slowly Moving Distilled Water 4-4
3 Effect of Velocity on Corrosion of Mild Steel Tubes
Containing Cambridge Water 4-5
4 Effect of pH on Corrosion of Mild Steel 4-7
5 Effects of Test Duration on pH-Corrosion Relationships..4-7
6 Effect of pH on Corrosion Rate at Chloride-Alkalinity
Ratio of 0.4 4-8
7 Effect of NaCl Concentration on Corrosion of Iron in
Aerated Solutions 4-10
8 Effect of Chloride-Bicarbonate Salts Ratio on Corrosion
of Mi 1 d Steel 4-12
9 Relative Corrosion Rates of Mild Steel at Particular
Chloride-Bicarbonate Ratios With and Without Chlorine...4-17
10 Corrosion of Cadmium vs. pH 4-20
11 Effect of pH on Corrosion on Galvanized Steel Tubes 4-21
12 Corrosion of Galvanized Pipe in Potable Water Systems...4-22
13 Effect of pH on Corrosion of Copper 4-34
14 Effect of Time and pH on Corrosion of Copper 4-35
15 Effect of Oxygen on Corrosion of Lead Submerged in
Distilled Water 4-49
16 Effects of Synthetically Hard Water on Lead Corrosion...4-50
17 pH Effects on Lead Corrosion 4-51
18 Effect of pH on Lead Solubility 4-52
19 Percentage of Samples Exceeding Lead Standard 4-55
20 Percentage of Housholds Exceeding Lead Standard in One
or More Samples 4-55
-------
Figure Page
21 Lead (II) Soluble Species in Relation to pH 4-58
22 Lead (II) Soluble Species in Relation to Total
Carbonate 4-59
23 Temperature Effects on Lead Corrosion 4-60
24 Effects of Carbon Dioxide Content on Corrosion of Lead..4-61
25 Temperature Effects on Pitting of Aluminum 4-67
26 Temperature Effects on Pit Current 4-68
27 Weight Loss of Aluminum in Various Water Qualities 4-71
28 Effects of Copper on Weight Loss of Aluminum 4-71
29 Effect of Low Calcium Content on Weight Loss of
Al umi num 4-72
30 Weight Loss of Aluminum in Tap Water 4-72
31 Effect of Ca++/Cl~Ratio on Weight Loss 4-73
32 Effect of Calcium Carbonate on Weight Loss of
Aluminum Specimens 4-74
33 Corrosion of Aluminum after 22 Hours Exposure in
Solution Containing Chloride, Bicarbonate, and Copper...4-77
34 Schematic of 3 and 2 Electrode Circuitry 5-8
35 Empirical Relationship Between Initial Slope of
Polarization Curve and Corrosion Rate 5-8
36 Construction Cost for Lime Feed Systems 6-29
37 Operation and Maintenance Requirements for Lime Feed
Systems 6-30
38 Construction Cost for Recarbonation-Liquid C02 Source...6-31
39 Operation and Maintenance Requirements for Recarbonation -
Liquid C02 Source 6-32
-------
SECTION 1
INTRODUCTION
BACKGROUND
The Safe Drinking Water Act of 1974 (PL 93-523) was passed by Congress
to safeguard public drinking water supplies and to protect public health.
This act essentially states that National Drinking Water Regulations will be
established and enforced for all public drinking water supplies. As a
result, the Environmental Protection Agency (EPA) proposed National Interim
Primary Drinking Water Regulations (NIPDWR) which became effective June 24,
1977. These interim regulations have established maximum allowable concen-.
trations of various contaminants in drinking water supplies and require that
water suppliers sample and analyze the water on a regular basis. Contami-
nants identified by the NIPDWR for limitations in drinking water supplies
include bacteria, turbidity, radioactivity, trihalomethanes, ten inorganic
chemicals, and six organic pesticides. These regulations will be reviewed at
least once every three years and can be amended at any time.
Corrosion control regulations are addressed directly in the National
Secondary Drinking Water Regulations which state that potable waters should
be non-corrosive for protection of the public welfare. These Secondary
Regulations are fedora 1 guidelines only, but may be adopted and enforced by
individual states.
The EPA has determined that corrosive materials in the water works indus-
try can pose a serious threat to public health and, on August 27, 1980, issued
amendments to the NIPDWR which specifically outline monitoring for corrosion
related parameters. These regulations require public water systems to iden-
tify the presence of specific materials of construction within the distribu-
tion systems and to monitor and report corrosivity characteristics including
pH, alkalinity, hardness, total dissolved solids, and the Langelier Index.
Additional regulations will be developed and enforced by EPA that set forth
requirements for systems distributing corrosive waters to increase monitoring
for corrosion byproducts such as lead and cadmium.
Attempts to develop regulations controlling corrosion in the water works
industry is controversial owing to the complexity of the corrosion problem.
Major problems include the lack of legal definition for corrosivity, the lack
of a generally acceptable method for measuring corrosivity, and the lack of
corrosion control methods which are effective throughout the entire distribu-
tion system as well as compatible with other potable water supply objectives.
The problem is further complicated by the requirement to regulate corrosion
at the "free flowing outlet of the ultimate consumer of a public water
1-1
-------
system." However, in defining maximum contaminant levels, the N1PDWR ex-
cludes contaminants added to the water under circumstances controlled by the
user, except those resulting from corrosion of piping and plumbing caused by
the water quality. This regulation implies that the water supplier must be
cognizant of the pipe materials in the service lines and within the household
as well as the distribution system.
Corrosion in potable water systems may be caused by either inherent fac-
tors or design, construction, or operational deficiencies. Most materials
used in potable water systems are susceptible to corrosion in waters contain-
ing oxygen. This inherent susceptibility is often decreased by formation of
protective coatings on the bare material surface. The coatings are formed by
precipitation of substances such as calcium carbonate from solution or growth
of insulating corrosion reaction product films on the material surface.
Thus, the properties of the material and the composition of the water are
interactive in the occurrence or inhibition of inherent corrosion. Corrosion
susceptibility is also influenced by many other factors including tempera-
ture, local flow rate, pH, alkalinity, carbon dioxide, dissolved solids, and
minor chemical constituents of either the water or corroding material.
Corrosion problems may also be caused by poor choice of materials,
coupling of dissimilar metals, improper installation practices, incorrect
design allowing unnecessary stagnant areas or crevices, release of unstable
waters, or addition of additives. While in principle subject to correction,
it is necessary to assume that many of these defects will persist for some
time in present systems.
Potable water distribution and plumbing systems may contain a variety
of materials. The environmental conditions noted above can have effects on
individual materials that differ both in type and degree. It is possible that
alleviation of corrosion problems for one material or type of installation
could create problems with another or result in deterioration of overall water
quality.
Before corrosion control measures can be implemented, sufficient evi-
dence must be available which detects the presence of corrosion. Addition-
ally, the location and cause of the corrosion occurrence must be identified.
This information can be obtained through the initiation of a comprehensive
monitoring program of potential corrosion byproducts. However, under exist-
ing regulations, water utilities are required to monitor only once per year
for facilities using surface water sources and only once every three years
for facilities using groundwater sources. The only major corrosion products
required to be monitored are lead and cadmium. Thus, not only is the current
monitoring schedule insufficient to reliably detect corrosion, much less to
isolate the location and possible causes, but some corrosion recognized by the
consumer, in terms of economics and aesthetics, may not be recognized by the
water supplier.
The secondary drinking water regulations state that potable waters
should be non-corrosive. No attempt has been made to assign a numerical
value for a maximum contaminant level (MCL) for corrosion. A major problem
1-2
-------
in attempting to produce a "non-corrosive" water is the lack of a generally
applicable measure of water corrosivity. Several corrosive or "aggressive"
indices have been developed to rate and evaluate the corrosivity of potable
waters. These indices include the Langelier Index, the aggressive index,
the Ryznar Stability Index, the Larson Index, the Casil Index, the Riddick
Index, and the driving force index. Predictions using these indices are often
found to be in disagreement with field experience.
These discrepancies are often a result of the differences between theo-
retical deviations and actual conditions. For instance, indices predic-
ting CaC03 deposition do not incorporate factors for metastable conditions
nor do they reflect that the protective capacity of the deposit is a func-
tion of the physical and chemical conditions existing during its formation.
Development of a general index is also difficult because of the multiple
roles of chemical species in potable water. The most common and abundant
aggressive ions in potable waters are usually considered to be chloride and
sulfate, although they may aid formation of more protective calcium carbonate
scale. Also, although oxygen provides the main driving force for many types
of corrosion in potable water, there is evidence that a certain minimum
solution oxygen level is necessary for the initial formation of protective
calcium carbonate films on steel. Flow velocities are also important for the
formation of protective layers and there is some evidence that low levels of
natural organic or silicate inhibitors may be influential.
Direct tests have been performed using pipe sections, metal coupons, and
water quality analyses to determine corrosivity. These tests must be per-
formed over a fairly long period of time and may be prohibitively expensive
for small communities. The results of these tests are also often inconsis-
tent with observations of field equipment and often are of limited use in
isolating a specific problem.
The inconsistency between practical and theoretical measurements and
field observations can be attributed to the large variety of mechanisms by
which corrosion can occur. All factors which can influence or cause corro-
sion are not necessarily identified using indices or direct field tests.
For example, pitting corrosion can be initiated from oxygen being trapped and
not evenly distributed in areas within the distribution system. Corrosion
from erosion or impingement caused by excessive velocities at certain areas
or occasional sand discharges can remove protective films and accelerate
corrosion. Increased velocities often result from improper design of home or
commercial plumbing facilities.
Another problem exists with the wide range of materials and installation
practices used in the water works industry. A specific water quality may be
non-corrosive to some materials but corrosive to others. Most older distri-
bution systems were constructed of cast iron pipe. Materials used more
recently for distribution systems include welded steel, ductile iron, pre-
cast concrete, and asbestos cement. Asbestos cement is used most extensively
today because of its excellent physical properties. However, asbestos cement
1-3
-------
pipe is limited to 30 inches in diameter and cannot be used if large hydro-
static pressures are required. Steel pipes are often used for larger dia-
meter transmission mains and may be protected with a cement mortar coating.
In alkaline conditions, steel or cast iron pipes are sometimes lined with
coal-tar enamel or an asphaltic layer over the cement lining for protection.
Materials used for the construction of facilities which cannot be pro-
tected with coatings such as valves, small pipes, and pumps include lead,
copper, zinc, aluminum, and alloys such as brass, bronze, and stainless
steels. Many municipal plumbing codes are not restrictive and a wide variety
and mix of material types and installation practices are used for home plumb-
ing. This practice can be conducive to galvanic corrosion. Domestic corro-
sion can be also be aggravated by the use of home water softening units, infer-
ior grade hot water heaters, and incorrectly sized pipes.
OBJECTIVES
The purpose of this investigation is to collect, review, evaluate, and
present existing information to determine whether a sufficient data base is
available to develop corrosion control regulations for the water works indus-
try as required by the Safe Drinking Water Act. To accomplish this objective,
an exhaustive literature search was completed which included a review of the
various materials used in the water works industry and their corrosion char-
acteristics, corrosion monitoring and detection techniques, and corrosion
prevention and control strategies. Materials which are addressed are iron-
based materials, copper-based materials, lead-based materials, aluminum,
asbestos cement, concrete, and plastics. Results of laboratory and field
research on each material as related to corrosion in the water works industry
are extensively reviewed and data is presented as appropriate. Major empha-
sis is placed on assessing the conditions of service and water quality char-
acteristics in potable water systems on the corrosion or deterioration of
each material.
The review of corrosion monitoring and detection techniques addresses
the various methodologies used to identify and evaluate corrosive waters.
This discussion includes a review of the search for corrosion indices using
water quality monitoring data as well as direct monitoring using coupon
exposure tests. The limitations of each of the techniques are presented.
Available corrosion prevention and control techniques are also evaluated
and presented. These techniques include adjustments in water quality charac-
teristics, additions of corrosion inhibitors, and the application of various
pipe coatings. Additionally, case histories of corrosion control programs
are presented for examples.
Finally, the information and data presented in these reviews are com-
piled and presented in table or matricies form as a summary. These tables
provide an overall view of the nature of the corrosion problems in the water
works industry and can be used as a data base for the initial consideration
of corrosion control regulations.
1-4
-------
SECTION 2
CORROSION AND WATER CHEMISTRY BACKGROUND
GENERAL ASPECTS OF CORROSION AND LEACHING IN POTABLE WATER
In general, corrosion refers to the degradation of a metal by electro-
chemical or chemical reaction with its environment or by physical wearinp
away. Leaching implies the removal of a soluble constituent, not necessarily
metallic, by the action of a percolating liquid. While more often used in
relation to ground water systems, leaching has also been used to describe
some corrosive dissolutions. This report is concerned with both the physical
and chemical degradation of any material used in potable water systems that
impart additional species to the water. The nature of corrosion and degrada-
tion of materials varies to a large extent with the specific material and
environment. Detailed discussions of these interactions with domestic water
environments are given in Section 4. These discussions are primarily based
on the corrosion and water works literature. Although this literature is
voluminous, it is also rather chaotic in the sense that reported results are
often not accompanied by sufficient background data or are anecdotal or subject
to multiple conclusions. This situation is not surprising in view of the
highly complex physical-chemical nature of corrosion processes in general and
especially in large-scale field installations. Much of the pertinent literature
is oriented to engineering implications of corrosion; there is relatively
little information on leaching of constituents of materials with regard to
health implications. A number of modes of corrosion or degradation may exist.
This report is primarily concerned with those processes which occur on sur-
faces contacting the potable water and which impart substances to the water.
This excludes external corrosion and stress corrosion cracking; the latter is
rare in domestic water environments in any case.
The fundamentals of corrosion are given in several textbooks. (See
bibliography for this section.) The book by Fontana and Greene is recent and
lucid. The text by Butler and Ison has a stronger orientation to corrosion in
natural waters and the chemistry of natural waters. The 1948 edition of
The Corrosion Handbook by Uhlig compiles a very large amount of data obtained
prior to its publication; the more recent work by Uhlig is a useful introductory
text. The other references listed are useful for a somewhat different outlook
or additional detail on the broad subject of corrosion.
2-1
-------
TYPES OF CORROSION
Corrosion types may be classified as either uniform or localized. Uniform
corrosion is the loss of a more or less equal amount of material over the
surface of a pipe or other structure. It may proceed directly tn metal ions
which go into solution or by way of a solid reaction product such as metal
oxide, hydroxide, carbonate, or other compound. In this case the amount of
material imparted to solution may become limited by the solubility of the re-
action product compound or by its dissolution kinetics. There is also the
possibility of periodic sloughing off of particles or chunks of the reaction
product either due to erosion or stresses built up during growth of the layer
on the metal. Uniform corrosion is not often a major concern in domestic
waters from an engineering or maintenance standpoint. It is a more likely
concern from a water quality view since a fairly low uniform corrosion rate,
spread over considerable area, can impart more impurity to a given amount of
water than a few deep pits.
Localized, or non-uniform, corrosion results in relatively rapid attack
and penetration on small areas of metal surface while the remainder of the
surface is not affected. Such attack can affect the structural or hydraulic
integrity of equipment. Pitting of iron can develop the tuberculation
which decreases the flow capacity of pip^s. For these reasons, localized
corrosion is often an engineering or maintenance consideration. Since attack
can be rapid and may result in selective leaching of a metal from an alloy,
localized corrosion may also have environmental repercussions. It is note-
worthy that either changing to slightly more resistant materials or modifying
the parameters in a corrosive medium to reduce the uniform corrosion rate
may produce conditions in which localized corrosion is the predominant mode.
Localized corrosion types of interest in potable water systems include
pitting, galvanic, concentration-cell, and selective-removal corrosions.
Pitting is a general term that refers to the formation of a pit where local
anodic conditions exist relative to a nearby cathodic area. Tuberculation
occurs when oxides of corrosion products are deposited over or adjacent to
the pit. Galvanic corrosion results when two metals of different solution
potentials contact each other. The anodic metal will corrode, affording
"protection" to the cathodic metal. Concentration-cell corrosion occurs
when localized differences in the potential of a single metal exist. Con-
ditions creating this environment could include differences in acidity,
cation or anion concentrations, dissolved oxygen, or even temperature
fluctuations. Crevice corrosion is a form of concentration-cell corrosion
where the rate of oxygen reaching the metal surface is controlled by diffusion
in a confined area. Selective-removal corrosion would include dezincification,
the removal of zinc from brass, or graphitization, the removal of the iron
silicon metal alloy from cast iron, leaving graphite.
2-2
-------
CORROSION INDICES
The following presents a brief general discussion of historical attempts
to develop a corrosion index. While these indices developed have a potential
use, none can be considered as a single tool for formulating a responsible
corrosion control program.
In 1912, J. Tillmans proposed the carbonate saturation theory of pipe
protection. Since then, several theoretical and empirical approaches have
been made to determine a single parameter that would indicate the ability
of a given water to protect or corrode the carrying pipes.
In 1936, W. F. Langelier developed the Langelier Saturation Index (9).
The SI is based on the theoretical tendency of a water to deposit or dissolve
calcium carbonate. The index is derived from the solubility product of
calcium carbonate, the dissociation constant of water, the second dissociation
constant of carbonic acid (H+ + CO;
HC03~), and a stoichiometric equilibrium
the carbonate, bicarbonate, and hydroxyl
'3 •*•
between alkalinity and protons versus
species. By restricting the applicable pH to 6.5-9.5, a saturation pH,
is determined as:
<') + pCa + pAlk
PHS =
where pX = log (-i) and
Ca = calcium ion concentration in moles/liter
Alk = total alkalinity as equivalents/liter
Kg = second dissociation constant of H2C03, corrected
for ionic strength and temperature
K' = solubility product of CaC03, corrected for
ionic strength and temperature
The saturation index is: SI = pH - pH
It is a logarithm of the ratio of the hydrogen ion concentration that the
water must have if saturated with calcium carbonate to the actual hydrogen
ion concentration. A negative value indicates an undersaturat.ion of CaC03
and, hence, a tendency to dissolve any existing CaC03 coating. A positive
value indicates oversaturation, a tendency to precipitate CaC03 and to form
a protective layer.
2-3
-------
Correlation attempts between the calculated SI and observed corrosion
effects have often, but not always, been in agreement. In general, a cor-
rosion free system exists when the SI is greater than -0.5 for cold water
and 0.0 for hot water, provided the water is of moderate hardness and alka-
linity (7), where the Langelier presumption of calcite being the CaC03
phase is less apt to be discrepant from impure calcite or precipitation in-
hibition. The SI is similarly not suitable for use in soft, saline waters
where a low buffer capacity and ionic species such as chlorides may disrupt
the CaC03 equilibrium conditions. Parameters that must be evaluated to cal-
culate the SI are methyl orange alkalinity, pH, temperature, total dissolved
solids, and the calcium ion concentration.
When using the SI, it is imperative to remember that this parameter states
a difference between actual pH and a pH at which CaC03 equilibrium is theoret-
ically achieved for that water. It says nothing about the driving force or
tendency for the CaC03 to dissolve or crystallize in the given water. Dif-
ferences may result from temperature effects, specific uncharged or ionic
constituents in the water, or crystal growth inhibition. Increased temperatures
will decrease the solubility of calcium carbonate as well as increase reaction
velocities, and consequently establish local equilibrium situations much faster
than at normal temperatures. There may also exist a difference in pH between
the CaC03 saturation point and the point at which crystal growth actually
begins. This difference, called the metastable region, may increase in the
presence of other dissolved ions, especially those that form slightly soluble
salts with calcium or carbonates such as magnesium or sulfate, causing a poor
correlation between the calculated SI and actual conditions of solubility.
A commonly used refinement of the SI that includes temperature and ionic
strength corrections is
pHs = A + B - Log (Ca++) - log (alkalinity)
where A and B are constants derived from the following tables:
2-4
-------
Water Temperature mg/1 Total
C A Filterable Residue
0
4
8
12
16
20
25
30
40
50
60
70
80
2.60
2.50
2.40
2.30
2.20
2.10
2.00
1.90
1.70
1.55
1.40
1.25
1.15
0
100
200
400
800
1,000
9.70
9.77
9.83
9.86
9.89
9.90
TABLE 1. Correction °arameters Used With The Refined SI
Other corrosion indicators include the Ryznar, Larson, Driving Force,
Casil, Aggressive, and Riddick indices. The Ryznar Stability Index is defined
as
RI = 2pHs - pH
with pH and pH defined as before. A value of seven or greater indicates an
aggressfve water while 6 or less indicates a tendency to form scale (U).
This index may be used with moderate to hard waters, but is not applicable
to soft or saline waters for the reasons previously cited.
The Larson index attempts to measure the aggressive nature of specific
ions and is defined as:
C1 +
, T _
L1
Alk
where Cl and SOi, are the chloride or halogen concentrations and sulfate con-
centrations, repectively, and Alk is total alkalinity. All three are expressed
in mg/1 of equivalent CaC03. When this ratio of reactive anions to alkalinity
is greater than 0.5, the possibility of corrosive action exists. Unlike the
SI, this index does not refer to the solubility of CaC03 but rather to the
faster rates of corrosion of metals because of conductivity effects. It is
not applicable to water that is soft or has a low dissolved solids concentration.
2-5
-------
The driving force index is defined as:
DPI = (Ca++) X (C03=)/K: X 1010
o
DPI = 10SI
where the calcium and carbonate concentrations are expressed in mg/1 as CaC03,
and K' is the CaC03 solubility product, corrected for ionic strength and temp-
erature. This index is a ratio of the actual ion product to that which would
exist during equilibrium conditions. Values greater than 1.0 indicate a
tendency for deposition of CaC03 while values lower than 1.0 indicate a CaC03
dissolution condition.
The Casil Index is a modification to the calcium carbonate solubility
indices that accounts for the effect of other parameters for soft waters.
It is defined as
CI = Ca + Mg + HSiOs - An12°ns
where each concentration is expressed in milliequivalents per liter. Negative
values are considered indicators of very corrosive water, values between 0 and
0.1 indicate slightly corrosive waters, and values above 0.1 indicate non-
corrosive conditions.
The aggressive index was formulated to determine the quality of water
that can be transported through asbestos-cement pipe without adverse struc-
tural effects. It, however, does not incorporate temperature or TDS effects
nor does it indicate the tendency of the pipe to release fibers or allow
Ca(OH)2 leaching.
Aggressive Index = pH + log (AH) where
A = total alkalinity as mg/1 CaC03
H = calcium hardness as mg/1 CaC03
According to the AI, values greater than 12 define a nonaggressive water;
values less than 10 define a highly aggressive water; and values between 10
and 12 define a moderately aggressive water.
The Riddick Corrosion Index is an empirically based formula that weighted
several corrosion-influencing factors including dissolved oxygen, chloride ion
concentration, noncarbonate hardness, and silica. The Riddick Index is
RCI = ^ [C02 +1 (Hardness-Alk) + CT + 2N] (5]^) (sa?;D*Q2 )(12)
2-6
-------
where C02 is expressed as mg/1 CaC03
Hardness is expressed as mg/1 CaC03
C1~ is the chloride ion concentration as mg/1
N is the nitrate ion concentration as mg/1
D.O. is the dissolved oxygen as mg/1
Sat.D.O. is the saturated oxygen value as mg/1.
The results are interpreted as
0-5 extremely noncorrosive
6-25 noncorrosive
26 - 50 moderately corrosive
51 - 75 corrosive
76 -100 very corrosive
> 100 extremely corrosive.
The values obtained correlated well with the soft waters in the eastern part
of the U.S., but not to the harder waters found in the middle states.
A more recent attempt to index the corrosivity of waters resulted from
a combination of the ratio:
(COT)
which represents the CaC03 precipitation equilibrium in the reaction:
Ca++ + 2HC03 ^CaC03(s) + C02(g) + H20
and the Larson Index. When corrected for low hardness cases, the result is:
Y = AH. + B[C1~] + [SO^] exp(- ^) + C
where A = 3.5 x lO"4
B = 0.34
C = 19.0
H =
^
HC03")2
++n
(C02)
[Cl~], [SOtt"], [CaTT], [C02] are expressed in ppm
[HC03~] is expressed in ppm as CaC03
2-7
-------
A correlation between this index and the scale formed by waters of that con-
stituency is shown in Table 7. The scale was quantified by impedance mea-
surements and only three samples were analyzed. However if substantiated,
this index would indicate that the chloride and sulfate concentrations, while
conventionally regarded as corrosive factors, may actually assist in the
crystal growth of calcium carbonate and resultant pipe protection.
A summation of the indices is presented below.
TABLE 2. CORROSION INDICES
(numbers in parenthesis refer to corrosion indices bibliography, symbols are
explained in the test)
Langelier Saturation Index (9, 10)
S.I. = pH - {(pK^ - PK^) + pCa + pAlk }
Ryznar Stability Index (17)
R.I. = 2pHs - pH
Larson Index (11, 12)
. Cl
Alk
Driving Force Index (14)
DPI = (Ca++) X (C03=)/IC X 1010
s
Casil Index (13)
C.I. = Ca + Mg + HSi 03 -
Aggressive I Ddex (15)
A.I. = pH + Log [AH]
Riddick Corrosion Index (16)
R.C.I. =[C02+. (Hardness -Alk)+Cl
Feigenbaum, Gal-or, Yahalom combination (8)
Y = AH + B([CT] + [S04=]) exp(- -) + C
2-8
-------
GENERAL CORROSION BIBLIOGRAPHY
1. Butler, G. and H. C. Ison, Corrosion and Its Prevention in Waters,
Reinhold, New York (1966).
2- Fontana, M. G., and N. D. Greene, Corrosion Engineering, McGraw-Hill,
New York (1978).
3. Larson, T. E., Corrosion by Domestic Waters, Illinois State Water Survey,
Urbana, Bulletin 59 (1975).
4. Speller, F. N., Corrosion Causes and Prevention, McGraw-Hill, New York,
(1951).
5. Uhlig, H. H. (ed.), The Corrosion Handbook, John Wiley & Sons, New Yo1
(1948).
6. Uhlig, H. H., Corrosion and Corrosion Control, John Wiley, New YOK
(1963).
CORROSION INDICES BIBLIOGRAPHY
7. DeMartini, F. E., "Corrosion and the Langelier Calcium Carbonate Satura-
tion Index," JAWWA. Vol. 30, No. 1, pp 85-111.
8. Feigenbaum, C., L. Gal-or, and J. Yahalom, "Microstructure and Chemical
Composition of Natural Scale Layers," Corrosion, Vol. 34, No. 2, pp 65-70,
1978.
9. Langelier, W. F., "The Analytical Control of Anti-Corrosion Water Treat-
ment," JAWWA, Vo. 28, No. 10, pp. 1500-1521, 1936.
10. Langelier, W. F., "Chemical Equilibria in Water Treatment," JAWWA, Vol.
38, No. 2, pp 169-179, 1946.
11. Larson, T. E., and F. W. Sollo, "Loss in Water Main Carrying Capacity,"
JAWWA. Vol. 59, p 1564, 1967.
12. Larson, T. E., "Corrosion by Domestic Waters," Bulletin 59, Illinois
State Water Survey, Urbana, 1975.
13. Loschiavo, G. P., "Experiences in Conditioning Corrosive Army Water
Supplies in New England," Corrosion. Vol. 4, pp 1-14, 1948.
2-9
-------
14. McCauley, R. F., "Controlled Deposition of Protective Calcite Coatings
in Water Mains," JAWWA, Vol. 52, 1960.
15. Millette, J. R., A. F. Hammonds, M. F. Pansing, E. C. Hanson, and
P. J. Clark, "Aggressive Water: Assessing the Extent of the Problem,"
JAUWA. Vol. 72, No. 5, 1980.
16. Riddick, T. M., "The Mechanism of Corrosion of Water Pipes," Water Works
and Sewerage, p 133, 1944.
17. Ryzner, J. W., "A New Index for Determining Amount of Calcium Carbonate
Scale Formed by Water," JAUWA, Vol. 36, 1944.
2-10
-------
SECTION 3
MATERIALS USED IN THE WATER WORKS INDUSTRY
A variety of materials are used by the water works industry for the
construction of facilities for treatment, storage, and distribution of pota-
ble water supplies. The majority of materials are used for pipes and piping
and for water storage or pressure tanks. For many small installations, no
treatment facilities exist and the water utility facilities consist of only
pumps, pipelines, and storage/pressure tanks.
The various materials used by the water works industry are identified
and briefly described in this section. Emphasis is placed on those materials
used for pipes and piping and for water storage as any corrosion control regu-
lations or utility programs will be primarily governed by the performance
of these facilities and materials. Materials used for the construction of
water treatment facilities are essentially the same as those used for pipe-
lines and storage tanks and, therefore, are not neglected from this presentation,
PIPES AMD PIPING
Pipes used in the water works industry are categorized under four clas-
sifications, excluding household plumbing. These four classifications are
transmission lines, distribution mains, service lines, and in-plant systems.
Transmission lines are those pipes used to transport water from the
water resource to the treatment facilities or finished water from the treat-
ment facilities to a community distribution system. These pipes can be sig-
nificantly large and occasionally a tunnel is required if the maximum size
pipe available is insufficient for design. Transmission lines are usually
designed for gravity flow to avoid pumping costs and to reduce line pressures.
Design flow velocities should not exceed 5 fps, but sometimes range from 12
to 15 fps. Factors which should be considered when selecting a particular
material for a transmission line are corrosion resistance, structural quali-
ties, hydraulic characteristics, installation and field conditions, and
economics.
Distribution systems are those facilities used to carry water from the
transmission lines and distribute it throughout a community. The distribu-
tion system includes a network of pipelines or mains, distribution reser-
voirs, elevated storage tanks, booster stations, and valves. Components of
the distribution system include arterial mains, distribution mains, and a
valve system. Arterial mains, sometimes called trunk mains or feeders, are
used to connect transmission lines to the distribution lines. Arterials are
3-1
-------
normally placed in a loop arrangement to avoid dead ends. Distribution lines
are connected to the arterial loop forming a grid system. These lines are
used to serve communities or commercial areas and hook up to individual serv-
ice 1ines.
Materials commonly used for transmission lines and distribution mains
are asbestos cement, cast iron and ductile iron, concrete, plastic, steel,
and wrought iron (2). The advantages and disadvantages of the use of these
materials are presented in Table 3. Aluminum is also used for pipelines, but
to a lesser extent (1). Plastic and wrought iron are more commonly used in
service lines and in-house plumbing systems (4).
Service lines are small diameter pipes that connect the consumer to the
distribution main. The selection of a particular material for a service line
is influenced by required size, durability, water characteristics, corrosion
resistance, material availability, ease of installation, and economics.
These criteria as well as corrosive tendencies of the waters in the specific
area are usually reflected in the local plumbing codes. Because of its
excellent physical characteristics, lead was the earliest material used for
service lines. However, the use of lead is now being questioned because of
its cost and its tendency to dissolve in soft waters of low pH. Copper is
now more frequently selected for service lines and approximately 50 percent
of the water utilities in the U.S. use copper exclusively. However, plastic
pipe is becoming more popular (4).
Minimum size service lines range from 3/4 to 1.0 inches in diameter.
For larger residences with numerous baths, minimum size service lines will
range from 1-1/4 to 1-1/2 inches in diameter. Approximately one-half of all
service lines in the U.S. are owned by utilities and the other one-half are
owned by the customers. However, approximately two-thirds of all service
lines are installed by utilities (4).
The most commonly used piping materials for service lines are asbestos
cement, brass, cast iron, copper, galvanized iron, lead, plastics, steel,
and wrought iron. The hydraulic flow characteristic of all these piping
materials is good when initially installed. These flow characterictics
generally remain good for asbestos cement, copper, lead, and plastic. These
materials are listed and briefly characterized in Table 4.
Flexible materials used for service lines are usually connected directly
to the corporation cock on the main and to the stop valve within the house-
hold. Nonflexible materials require the use of a "gooseneck" connection to
the corporation cock and possibly some type of flexible connection to the
household plumbing system. Goosenecks are available in lead, copper (if
permitted by local plumbing codes), and flexible plastic (4).
Every type of piping material previously discussed is used for in-plant
piping systems. Other materials used include glass and rubber. Glass and
rubber are not usually used for conveying potable water within the plant, but
rather for other in-plant operational functions. Pipe materials used for
3-2
-------
TABLE 3. SEVERAL MATERIALS USED FOR TRANSMISSION AND DISTRIBUTION LINES (4)
Materials
Available
Size
Diam. (in.)
Advantages
Disadvantages
OJ
I
OJ
Asbestos Cement
Cast Iron
(cement-lined)
Ductile Iron
(cement-1ined)
Concrete
(reinforced)
Concrete
(prestressed)
4-36 Corrosion resistant; good flow
characteristics; light weight;
easy handling; low maintenance.
2-48 Durable and strong; good corrosion
resistance; easily tapped; flow
characteristics good.
4-54 Durable, strong, high flexural
strength; lighter weight than
cast iron; greater carrying
capacity for same external diam-
eter; fracture resistant; easily
tapped.
12-168 Durable with low maintenance;
good corrosion resistance; flow
characteristics good; resists
backfill and external loads.
16-120 Durable, low maintenance; good
corrosion resistance; good flow
characteristics; resists backfill
and external loads.
Low flexural strength in small
sizes; more subject to impact damage;
difficult to locate underground.
Subject to electrolysis and attack
from acid and alkaline soils; heavy
to handle.
Similar to cast iron.
May deteriorate in alkaline soil,
if cement type is improper, or in
acid soil if not protected.
Same as above.
(Continued)
-------
TABLE 3 (Continued)
Materials
Available
Size
Diam. (in.)
Advantages
Disadvantages
Steel
4-120
Plastics
ABS
PE
PVC
Wrought Iron
1/2-12
1/2-6
1/2-16
1/4-30
Light weight and easily installed;
high tensile strength; low cost;
good hydraulically when lined;
adapted to locations where some
movement may occur.
Smooth interior surface minimizes
pumping losses; chemically inert;
corrosion resistant, non-reactive
with water; light weight.
Tough; ductile; malleable; weld-
able and corrosion resistant.
Subject to electrolysis; external
corrosion in acid or alkaline soil;
poor corrision resistance unless
properly lined, coated, and wrapped;
low resistance to external pressure
in larger sizes; air-vacuum valves
imperative for large sizes; subject
to tuberculation when unlined.
Jointing sometimes difficult; tenden-
cy to creep; brittle at low tempera-
tures.
Generally rough surface; hot dip gal-
vanizing or non-metallic coating gen-
erally required.
-------
TABLE 4. MATERIALS USED FOR SERVICE LINES (4)
Material
Size
range
(in.)
Comments
Asbestos-Cement
Brass
Cast iron
Copper
Galvanized iron
Lead
Plastics
ABS
PE
PVC
Steel
Wrought iron
3,4,6
1/2 to 6
2,3,4,6
1/2 to 6
1/2 to 3
3/8 to 2
1/2 to 6
1/2 to 2
1/2 to 6
1/2 to 6
1/2 to 6
Corrosion resistant; not available below
3-in.; slip coupling joints.
Long life under normal conditions; cor-
rodes in acid soils; uses threaded
coupling joints; requires gooseneck
connection.
Corrosion resistant when lined and
coated; not available in small diameters;
rigidity and short length require joints
and gooseneck connection.
Direct connection to mains; corrosion
resistant; dissolves in soft v/ater with
high C02 content.
Not highly resistant to corrosion;
requires threaded joints and gooseneck
connection.
Direct connection mains; corrosion
resistant except in soft waters with
high COz; some tendency to creep or
crack unless properly formulated.
Available in three grades; strong,
extra-strong, double extra-strong;
not resistant to corrosion unless
cement-lined.
Same comments apply as for steel.
3-5
-------
in-plant systems for transporting potable waters have the same design and
are manufactured by the same process as previously discussed.
The extent of use of various materials for piping by water utilities
serving more than 2500 persons in the U.S. as of 1975 and for 1975 was sur-
veyed and compiled by Scott and Caesar and is summarized in Table 5 (3).
Approximately 75 percent of all water main piping in the U.S. as of 1975 was
cast iron and it accounted for approximately 46 percent of total pipe installed
in 1975. Asbestos cement pipe and steel pipe accounted for approximately
13 and 6 percent, respectively, of all water mains in place by 1975. The use
of steel pipe appears to be declining as it accounted for only approximately
3.4 percent of all water pipes installed in 1975.
STORAGE TANKS
Materials used for the construction of water storage tanks are wood,
fiberglass, concrete, and steel. Aluminum is sometimes used for construction
of storage tank roofs. The type of material selected is generally determined
by the required capacity, the specific use, and economics. Concrete is con-
sidered economical for large storage tanks with capacities ranging between
1.25 and 5 million gallons. Concrete tanks are estimated to constitute 15
percent of all new water tanks (2).
Steel tanks are considered more economical for tank capacities smaller
than 1.25 million gallons, and they are more adaptable for elevated use when
natural relief or topography is not available (2). However, steel tanks are
often lined with a protective coating, such as a coal tar, to minimize corro-
sion. Consequently, the maintenance costs of steel tanks is generally higher
than that for concrete tanks (2).
Currently there are approximately one-half million steel water storage
tanks in the U.S. ranging in capacity from 50,000 gallons to 10 million gal-
lons. It is estimated that over 1000 new steel water tanks are constructed
each year (2).
Wooden tanks generally have capacities ranging in size from 25,000 to
50,000 gallons, but can be as large as 250,000 gallons. Because these tanks
have limited capacities and are prone to leak, their use is limited (2).
Fiberglass-reinforced plastic tanks are also used for storage of potable
water, but to a limited extent owing to problems encountered in field con-
struction. Because of these construction problems, most of these tanks are
limited in capacity to that which can be shop fabricated and transported.
Fiberglass tanks have been shop fabricated up to capacities of nearly 50,000
gallons (2).
Steel storage tanks are usually lined with a water-impervious coating.
Traditionally, these coatings have been coal tar based enamels, but recent
difficulties in controlling fumes during application and reports that toxic
3-6
-------
TABLE 5. PIPE USED IN U.S. WATER SUPPLY DISTRIBUTION SYSTEMS
BY UTILITIES SERVING OVER 2,500 PERSONS (3)
TYPE OF PIPE
Cast Iron
Asbestos Cement
Steel
Reinforced Concrete
Plastic
Ductile
Galvanized Wrought Iron
Wood
co RCP Steel Cylinder
Black Galvanized Iron
Copper
All others and unidentified
MILEAGE
IN PLACE
(beginning
of 1975)
481,816
83,871
37,852
10,083
6,981
7,498
2,364
1,246
652
431
312
6,879
* OF
TOTAL
75.29
13.11
5.91
1.58
1.09
1.17
0.37
0.19
0.10
0.07
0.05
1.07
MILEAGE PERCENT OF PIPE MILEAGE IN PLACE AT BEGINNING OF
INSTALLED 1975 BY DIAMETER
(1975)
Under 6" 6"-12" 13"-24" Over 24"
6,847 15.7 76.7 6.6 1.0
3,743 9.7 86.3 4.0 0.1
505 53.4 29.5 10.7 6.4
517 0.2 4.1 43.4 52.3
1,826 62.0 37.3 0.6 0.1
1,388 * * * *
3 * * * *
* * * * *
8 * * * *
* * * * *
* * * * *
96
Source: Scott and Caesar, 1975 (3)
*Not Specified
-------
substances may be introduced to the waters by their use have resulted in the
use of epoxy and vinyl paints (2). Various estimates report that 50-90% of
all new water tanks are lined with vinyl and 10-50% are lined with epoxy (2)
3-8
-------
REFERENCES
1.
2.
P. Godard, "Corrosion Behavior of
Reference to Pipeline," Br. Corros,
3.
4.
Booth, F. F., Murray, G. A. W. and H.
Aluminum in Fresh Waters with Special
J_., Vol. 1, No. 2, 1965, pp. 80-86.
Goldfarb, A. S., Konz, J., and Pamela Walker, "Coal Tar Based Materials
and Their Alternatives," Interior Coatings in Potable Water Tanks and
Pipelines. The Mitre Corp., Mitre Technical Report MTR-7803, U.S. EPA
Contract No. 86-01-4635, January 1979.
Scott, J. B. and Adelaide E. Caesar, Survey of Water Main Pipe in U.S.
Utilities Over 2500 Population, Morgan Grampian Publishing Co.,
Pittsfield, Massachusetts, 1975.
Symons, G. E., Ph.D., "Water Systems, Pipes and Piping, Part I/Piping
Systems Design," Water and Wastes Engineering. Manual of Practice
Number Two, Vol. 4, No. 5, May 1976, pp. M3-M50.
3-9
-------
SECTION 4
CORROSION CHARACTERISTICS OF MATERIALS
USED IN THE WATER WORKS INDUSTRY
The corrosive behavior of specific materials when subject to the envir-
onmental conditions of potable water systems is presented in this section.
This information is compiled primarily from published results of laboratory
and field research. In general, most studies reviewed are consistent and in
agreement in identifying the conditions of service and water quality charac-
teristics which initiate and maintain the corrosion or deterioration of a
specific material. However, specific data presented by various investigators
is sometimes inconsistent or in disagreement. This inconsistency usually
results f m variations in the conditions of testing and/or reporting. It is
also noted that the literature often fails to fully describe or present the
details of the testing procedures which are often critical for assessing test
results.
The corrosion behavior of each material is discussed independently and
the presentation format for each material is dictated by the information
available in the literature. Emphasis is placed on presenting numerical
results of various corrosion testing and monitoring as this data serves as
the basis for considering specific corrosion control alternatives.
IRON-BASED MATERIALS
Iron-based materials are among the most common piping materials. They
are also subject to a variety of corrosion mechanisms that may occur in
potable water systems. This subsection discusses the various iron-bearing
metals that may be encountered.
Corrosion of Iron
The corrosion behavior of steel and cast iron materials in potable water
environments is highly complex. Many factors can be involved and are often
interrelated. The effects of several factors can vary from beneficial to
conducive to greater corrosion, depending on the specific situation. It is
often difficult to say what the main factor controlling the corrosion of
steel is, due to these subtle relationships. The following discussion out-
lines the basic corrosion mechanism of iron and then discusses these contrib-
uting factors.
The corrosion of iron and steel in waters is basically electrochemical
in nature. The actual metal loss is due to an oxidation of iron atoms on the
4-1
-------
metal surface to give ferrous ions which can go into solution and electrons
which stay with the metal:
Fe + Fe2+ + 2e~ (1)
In order for this process to proceed, the electrons must be taken up by a
reduction process which can take place on another part of the surface.
Usually this complementary process is the reduction of dissolved oxygen
(Eq. 2) or the reduction of hydrogen ions or related species (Eq. 3).
02 + 2H20 + 4e~ -»• 40H" ( 2)
2H+ + 2e~ •* H2 (3)
The oxidation and reduction reactions are parallel-coupled events which must
proceed at identical rates. The overall corrosion rate is limited by the
slower of the two coupled reactions. The oxidation reaction (Eq. 1) is
rapid in most media. The rates of the reduction reactions are limited in
natural waters by reactant concentrations, sluggish electrochemical kinetics,
or a combination of factors. Thus the overall uniform corrosion rate is
normally controlled by the rate of the reduction reaction, as amplified
below.
For corrosion to occur, the difference between the electrochemical
potentials for the oxidation reaction (occurring at anodic sites) and reduc-
tion reactions (occurring at cathodic sites) must be such that the overall
free energy change drives the reactions as written. This potential differ-
ence ultimately appears as a driving force which can be viewed as being
divided between the two reactions in such a way that the inherently slower
reaction receives the larger share of driving force. For iron corrosion with
hydrogen ion or water reduction, this overall driving force is relatively
small and decreases with increasing pH. On the other hand, the driving force
for iron corrosion with oxygen reduction is very large. The reduction of
oxygen is a complex electrode process which is inherently quite slow. The
detailed mechanism is not well known. In spite of the low inherent rate, the
driving force is so large that the reduction can be fast, and transport of
02 to the iron surface often becomes the rate limiting process.
The metal loss will be uniform or general over the surface as long as
the oxidation and reduction sites constantly shift in location and the frac-
tional coverage of sites is roughly the same. If an oxidation site becomes
small, fixed, and surrounded by a much larger reduction area, then localized
corrosion such as crevice corrosion or pitting can result. This localization
of an oxidation site can be caused by a variety of factors such as local
breakdown of a protective oxide film, presence of a crevice, a break in a
deposit on the metal, and so on. Uniform corrosion is favored by a clean
metal surface and ample supply of cathodic reactant. Thus in acidic solutions
where oxide films are not stable and the'concentration of hydrogen ions is
high, steel generally corrodes uniformly. Localized corrosion is often
favored by conditions which reduce the rate of uniform corrosion. Factors
influencing the type of corrosion are also related to the effects of oxygen
concentration, pH, flow rate, temperature and electrolyte concentration,
4-2
-------
which in turn do not operate independently. These factors are taken up
separately below, but their interdependence should be kept in mind.
It is tempting to attribute a majority of the factors involved in corro-
sion of iron in aerated near neutral pH waters to control by oxygen transport
(cathodic control). Stumm strongly suggests, however, that at long exposure
duration, formation of protective films on iron and other events may make the
rate of the anodic iron dissolution process an important factor (97).
Although cathodic control appears to explain a large number of environmental
effects, the latter possibility should also be kept under consideration.
Effect of Dissolved Oxygen--
Dissolved oxygen plays a key role in corrosion or iron in natural
waters, but its effects can be conflicting and partially dependent on other
environmental factors. In near neutral pH waters at ambient temperature,
dissolved oxygen provides the reduction reaction (Eq. 2) which sustains the
corrosion of iron. However, oxygen also plays a role in formation of semi-
protective iron oxide films on the metal, and the more protective films are
formed at higher oxygen concentrations. The presence of oxygen also appears
to be necessary for formation of protective layers on steel by calcium carbo-
nate deposition (7). Once these films are formed, however, oxygen provides
the main driving force for initiation of pitting (leading to tuberculation)
or other forms of localized corrosion.
The first part of this discussion is for conditions in the absence of
calcium carbonate or other external inhibiting species. At sufficiently high
pH values, iron oxide or hydroxide layers can be formed. The first of these
is probably ferrous hydroxide which can be formed by an overall reaction such
as Eq. 4.
Fe + H20 + 1/2 02 + Fe(OH)2 ( 4)
This solid is often found next to the metal surface and can act as a diffu-
sion barrier to oxygen. Further oxidation of this product yields hydrous
ferric oxide which comprises most of ordinary rust. An intermediate oxida-
tion stage, Fe3Oit-nH20 often forms as a layer between the ferric and ferrous
compounds.
In the absence of dissolved oxygen, the corrosion rates for both pure
iron and steel becomes negligible in near neutral pH water at room tempera-
ture. Corrosion rates may be high when the metal is first exposed to air-
saturated water, but the iron oxide films formed over a period of a few days
act as a barrier to diffusion of oxygen to the surface and a steady state
corrosion rate is obtained. This steady state rate is proportional to oxygen
concentration, as shown in Figure 1,since the oxygen diffusion rate is pro-
portional to its concentration. An oxygen concentration of about 6 mg/1
corresponds to air-saturated water. At still higher oxygen concentrations,
the uniform corrosion rate of mild steel may decrease abruptly, as shown in
Figure 2. This effect is apparently due to passivation of the iron which
involves either the oxidation of the normal ferrous hydroxide layer to one
which has better protective properties or the formation of a thin chemisorbed
oxygen layer on the metal surface. More oxygen is required in waters
4-3
-------
100
'01234 56
Cone, of dissolved 02, ml/liter
Figure 1. Effect of 02 concentration on corrosion of mild
steel in slowly moving water containing 165 ppm CaCL2, 48-
hour test, 25°C (107).
T3
E
«t
0)
4-1
OJ
c
o
I/)
o
o
o
< ys~ Corresponds to air saturation j
0246 10 15 20 25
Concentration of dissolved oxygen, ml/liter
Figure 2. Effect of 02 concentration on corrosion of mild
steel in slowly moving distilled water, 48-hour test, 25°C
(107).
4-4
-------
containing chloride ions and passivation of mild steel connot occur if the
chloride concentration is high enough. The breakdown of passivity in the
presence of moderate amounts of ch^rides (more than about 20 ppm, see below)
is often accompanied by severe pitting or crevice corrosion.
The effect of solution flow rate on mass transport of oxygen to the cor-
roding surface can also give rise to diverse effects. At moderate oxygen
concentrations and flow rates, increasing the flow rate increases the corro-
sion rate due to the increase in amount of oxygen transported' to the surface.
At higher flow rates, the surface oxygen concentration can become high enough
to cause passivation, provided the chloride content is not too high. These
effects are shown in Figure 3. Still higher flow rates, over 15 ft/sec, can
greatly accelerate corrosion by erosion of the protective films, combined
with fast transport of oxygen to the surface by turbulent flow. At the other
extreme, stagnant conditions are usually most conducive to pitting and other
forms of localized corrosion. The threshhold for chloride effects on passivity
is ill defined. As little as 20 ppm chloride may cause breakdown and pitting
but the threshhold may be higher depending on solution composition (vida infra).
0.05
246
Velocity, ft/sec
Figure 3. Effect of velocity on corrosion of mild steel tubes
containing Cambridge water, 21°C, 48-hour tests (37).
Effect of pH—
Conflicting results and conclusions have been obtained for the effect of
pH on corrosion of steel and cast iron in potable water environments. This
is probably because a number of effects are possible in the range of inter-
est, from about pH 6 to pH 10, and both the mode of corrosion as well as
degree of corrosion and structure of surface films may be affected. In
addition, all attempted correlations are done with respect to the bulk pH of
4-5
-------
the liquid phase, while the actual pH at the material surface can be affec-
ted by the corrosion reactions, surface iron hydroxide layers, other preci-
pitated layers, other chemical species in solution, and solution flow rate.
In the moderately acidic region, corrosion is generally uniform and the
rate increases rapidly with decreasing pH. The onset of this region is about
pH 4 to 5 if no other acidic species other than hydrogen ions are present,
and from about pH 5 to 6 if acidic species such as dissolved carbon dioxide
are present in appreciable quantities. Corrosion is accelerated in these
regions because hydrogen evolution can occur utilizing fairly concentrated
reagents (hydrogen ions or protonated anions). In addition, the oxide films
which act as diffusion barriers are dissolved. The latter effect also offers
greater access of oxygen to the metal surface which can be more important
than hydrogen evolution at the higher pH end of this range. Both hydrogen
evolution and oxygen reduction tend to raise the pH at the metal surface and
stabilize oxide films. The presence of acidic species such as dissolved C02
tends to mitigate this effect since the total acidity of such a system is
greater than that of a completely dissociated acid at the same pH.
In the near neutral pH range of most natural waters, oxygen transport
can in some cases dominate changes in pH, although in this region many com-
plex pH effects have been noted. The simplest type of behavior for corrosion
of steel by aerated water at 22°C is shown in Figure 4 (112). For this case,
in which either NaOH or HC1 were added to aerated water to adjust the pH, the
corrosion rate is constant from about pH 4 to 10. This is reasonable behav-
ior if the rate depends only on diffusion of oxygen to the metal surface
through a diffusion barrier of ferrous hydroxide or hydrated ferrous oxide
which is continuously renewed by the corrosion process. Also, in this bulk
solution pH region, the iron surface is always in contact with the saturated
ferrous hydroxide solution which should maintain a pH of about 9.5. Similar
results were obtained at 40°C except the constant corrosion rate plateau
(18 mpy) extends from pH 4.5 to 8.5. When the pH is adjusted with C02
instead of HC1, the rapid increase in corrosion rate occurs at pH 5.4 instead
of 4.1 (112).
Later studies have occasionally shown a maximum in corrosion rate occur-
ring in the pH range from about 6 to 9. Examples are given by Eliassen, et
al, along with controlled corrosion tests using synthetic "average" water
(84 ppm hardness, 57 ppm methyl orange alkalinity, and a pH after aeration of
8.05, pHs = 8.28) (25). These tests were done for steel pipe as a function
of flow rate and duration of exposure. A maximum in the corrosion rates was
observed in the pH range 6.5 to 7.5 for short-term tests (7-14 days exposure)
at flow velocities of 0.33 and 1.0 fps. At zero velocity a much lower corro-
sion rate was observed which was constant from pH 4 to 11. The authors note
that their short-term results are quite similar to others in the literature
which show corrosion maxima with pH (25). However, their "long-term" tests
(30-40 days) do not show maxima and tend to agree with the early results of
Whitman (112). These results are illustrated in Figure 5. These authors
also observed very severe pitting at a pH of 6.5, slight pitting at pH 8.0 to
10.0 and no pitting at pH 4.0 to 5.5. They generally agree with the model of
pH effects proposed by Whitman, but point out that variations may be caused
by changes in mode of corrosion and nature of corrosion products formed as a
4-6
-------
4 13 12 II |O 9 8 7 6 5 4 3 2
u
Figure 4. Effect of pH on corrosion of mild steel (112),
IN
O
a
a
'v.
TJ
-a
i
-------
function of pH and also time. The corrosion products they observed were
gelatinous and loosely attached to the metal in acid waters, but hard,
raised, and well attached at pH values greater than 8.0.
Some additional examples of more complex pH behavior have been given
by Larson (57, 59, and 60) and by Fontana (32). Larson's results were
obtained with waters containing bicarbonate and chloride. In one case, at a
CT/alk mole ratio of 0.4, a sharp maxima at a pH of 8.0 is seen which
increases with duration of exposure, in contrast to Eliassen's results (60).
These results are shown in Figure 6. In another case with a higher chloride
content, the corrosion rates increase by a factor of about 3 in going from
pH 7 to pH 8.5 (at a Cl"/alk ratio of 1.0). Fontana obtained results for a
sample maintained in a high flow rate location (39 ft/sec) in distilled water
at 50°C. A sharp peak in corrosion rate is observed at pH 8 which is about
10 times that observed at pH 6 or pH 10. The effect was attributed to an
atypical reaction product scale of granular Fe30it which formed at pH 8. In
regions of low attack the products were Fe(OH)2 and Fe(OH)3 which provided
better barriers to diffusion of oxygen and ions. The results are said to be
supported by power plant experience showing greater attack at pH 8 as com-
pared to slightly lower values (32).
-a
T3
(TJ
a:
c
o
1/1
o
o
cj
6.0
Figure 6. Effect of pH on corrosion rate at chloride-alkalinity
ratio of 0.4; duration: A, 16 days; B, 12 days; C, 8 days; D, 4
days; E. 2 days (60).
4-8
-------
An increase in corrosion rate with increasing pH (from pH 6.9 to 8.6)
has also been reported for ground cast iron samples in a variety of natural
and synthetic waters after 50 days exposure (97). The author attributes this
to a decrease in buffer capacity of the electrolyte (for a given alkalinity)
with increasing pH and to the pH dependence of relative local anode/cathode
areas. These views are discussed in the calcium carbonate protection section
below.
Some relatively recent electrochemical results may give additional in-
sight on the pH behavior. The fundamental corrosion reactions on iron have
been mainly studied in deaerated solutions at a pH less than 5 under condi-
tions where no solid reaction products are formed. The oxidation reaction
(Eq. 1) appears to proceed through intermediates involving hydroxyl ions in
such a way that its individual rate increases with increasing pH from about
pH 1 to pH 5 (46). Hydroxyl ion catalysis at near neutral pH might help
explain the disparate experimental results in natural waters.
Effect of Dissolved Salts— ++ ++
Dissolved salts (for example, ions such as Na , Cl", SCK , Ca , Mg ,
HC03~, C03=, etc.) can have a number of effects, some of which depend only
on the general effect of ions, while others depend on the individual chemical
species. Examples of the first class of effects are given below.
Increasing salt concentration increases solution
conductivity which can have several effects.
Increasing salt concentration generally decreases
the equilibrium concentration of dissolved oxygen
and C02.
Contrary to early predilections, solution conductivity itself has little
effect on most modes of uniform or localized corrosion. This is because the
local anodes and cathodes are so close together that the resistance offered
by the solution ds much less than equivalent electrochemical reaction rate
resistances. Solution conductivity can, however, affect the range over which
the effects of attack due to galvanic coupling of dissimilar metals is exten-
ded. Attack can extend for example, from about 1 cm from the joint in soft
waters to more than 10 cm in water with significant dissolved salt content.
Galvanic corrosion is discussed in a later section.
Changes in solution conductivity can have more subtle and significant
effects. Uhlig proposes that moderate increases in conductivity (by dis-
solved NaCl) from that of very soft or distilled water can lead to increased
corrosion rates due to formation of a less protective Fe(OH)2 film (107).
The corrosion of iron in aerated solution as a function of NaCl concentration
is shown in Figure 7. The film is less protective since it is formed further
away from the surface than in less conductive waters. This in turn is due to
coupling of local anodic (source of Fe2+) and cathodic (source of OH") areas
at greater distances in the more conductive solution. The initial increase
in corrosion rate might also have to do with some specific effects of
chloride and a number of postulates on these effects have been offered (31).
4-9
-------
Specific effects of chloride are discussed below. There is little direct
evidence for either the Fe(OH)2 film or chloride effect arguments.
1/1
O
O
u
D
CC
/r
!
!
3 3 J
""*\
> 1
"\
0 1
^
5 2
^
0 2
•-
5 3(
Cone NaCl, wt. %
Figure 7. Effect of NaCl concentration on corrosion
of iron in aerated solutions, room temperature (107).
It has recently been proposed that an increase in solution conductivity
may have the effect of actually promoting a more protective coating on iron
in the presence of calcium carbonate film forming precursors (28). Correla-
tion between the electrical impedances and protect!veness of scales and a
quantity related to the conductivity of the solution was found. Conductive
salts facilitate protective film formation, provided that sufficient tempora-
ry hardness exists. This effect is dicussed in more detail under the calcium
carbonate film section below.
The other overall effect of dissolved salts, that of decreasing the
solubility of dissolved gases, is also illustrated in Figure 7. Since the
dissolved oxygen content is the main controlling factor under the conditions,
"salting out" of oxygen causes a decrease in corrosion rate. This effect
only becomes pronounced at higher salt concentrations and is most likely
unimportant in natural fresh waters.
Chloride is the most deleterious individual ionic species normally
occurring in natural waters. Much of this appears to be due to the ability
of chloride to promote pitting by penetration or local destruction of other-
wise protective iron oxide films. It is quite difficult for a uniform, true
passive iron oxide film to be formed in the presence of chloride ions. Foley
lists five possible specific roles for chloride in iron corrosion (31). It
is difficult, however, to quantify the relation of corrosivity to chloride
concentration in natural waters. The treshold concentration of chloride,
above which pitting of iron is possible, is said to be about 10 ppm (31).
The role of sulfate is more nebulous, especially since sulfate does not
appear to have the film piercing properties that chloride displays. Larson
has done extensive studies of the effect of these salts on corrosion of steel
4-10
-------
under conditions simulating potable water environments (57, 58, 59, 60). He
concludes that the corrosivity of air-saturated domestic waters depends on
the following factors (60).
The proportion of corrosive agents (chloride and sulfate
ions) to inhibitive agents (bicarbonate, carbonate,
hydroxide, and calcium ions).
The concentration and degree of effectiveness of the
corrosive and inhibitive species.
The velocity of flow which affects the rates of diffusion
of both types of species to the surface.
He also notes that the relative effectiveness of the species involved is not
definitive, but may be influenced by one or more of the others. The inter-
dependence of the relevant factors is also emphasized by the observation that
"intermediate proportions" of corrosive to inhibitive species, which result
in incomplete protection (to uniform corrosion) at a particular solution flow
rate, are conducive to pitting and/or tuberculation. Representative results
obtained by Larson are shown in Figure 8 (60).
Effect of Dissolved Carbon Dioxide--
The effect of dissolved carbon dioxide depends in large part on solution
pH since this determines the relative amounts of "carbonic acid" (hydrated
C02), bicarbonate, and carbonate present. The carbonic acid form is aggres-
sive towards iron since it can serve as a relatively concentrated reactant
for hydrogen evolution at a relatively high pH (23). This effect is probably
not very important in aerated natural waters above a pH of about 6. Carbonic
acid can also act to dissolve calcium carbonate/hydrated iron oxide films and
thus remove protective diffusion barriers.
Bicarbonate is the predominant C02 species from about pH 7 to 10.
Larson has classified bicarbonate as a mild, but effective inhibitor of steel
in aerated natural waters, in the absence of calcium (60). This was a gen-
eral experimental result but was not explained physically. Recent work by
Davies and Burstein in concentrated bicarbonate solutions, borate buffered
at pH 8.8, indicates that the anodic dissolution of iron is accelerated by
bicarbonate due to formation of the complex Fe(C03)|~ (22)- Formation of
solid FeC03 along with Fe(OH)2 is also indicated under some conditions.
Pitting is attributed to the heterogeneous nature of the surface reaction
products and formation of the complex. This bulk solution is quite different
from the bulk bicarbonate content of natural waters, but the results may be
relevant to localized corrosion. The environment under a calcium carbonate
scale on iron or in a developing pit or tubercule could in some instances
have a high effective bicarbonate concentration and aid in initiation or
growth of localized corrosion. The effect on uniform corrosion on relatively
bare surfaces would still be expected to be minimal. Except as implied by
the foregoing, carbonate ions in normal domestic water are expected to be
essentially neutral except as they can act beneficially to form CaC03 films
or decrease acidity.
4-11
-------
-pi
100
80
0)
4->
CO
eo
C
O
in
O 40
L-
O
O
20
Alkalinity (as CaC03)
• 75-100 mg/1 o 150-180 mg/1
X 120-135 mg/1 o 250-260 mg/1
0.1
0.2
0.3 0.4 0.5
Equivalent Ration C1~/HC03
0.6
0.8
Figure 8. Effect of chloride-bicarbonate salts ratio on corrosion of mild steel (60),
-------
Effect of Calcium—
Much of the protection, both natural and man-made, for iron-based
materials in potable water environments is attributed to the formation of a
calcium carbonate film or scale on the surface of the material. Presumably,
this film provides a diffusion barrier to oxygen, thus further limiting the
oxygen reduction rate which is usually rate controlling in natural aerated
waters. Various indices have been proposed to approximate the tendency of
calcium carbonate to deposit or to dissolve in natural waters, as discussed
in Section 2. The actual mechanism of protection, however, is much more com-
plicated than simple deposition of a layer of CaC03. The saturation indices
are often useful as guides, but are too indirect to be applied indiscrimin-
ately. Langelier was one of the early proponents of applying a CaC03 satura-
tion index to corrosion control and described both an index and its correla-
tion with results obtained in New York City pipe corrosion tests (33). He
also presented refinements and reviewed early applications (54). The methods
do not provide a quantitative measure of the amount or rate of corrosion or
CaC03 deposition. By 1954, Larson noted that water works practice indicated
that the saturation index does not necessarily show corrosivity.
Inhibition by calcium carbonate appears to be intimately connected with
the corrosion reactions on iron or steel. In a review of treatment methods
for desalination product water, Bopp and Reed emphasize that sufficient
dissolved oxygen (they quote a minimum of 5 ppm) is needed for the proper forma-
tion of protective CaC03 (+ iron oxide) films (7). Untreated product water
will rapidly attack iron and other normal materials of construction of muni-
cipal systems. McCauley studied the properties of CaC03 coatings formed on
cast iron under different conditions (67). In general, development of an
adherent layer required the initial deposit of dense material well bonded to
the metal. Even if this initial layer was very thin, a tenacious coating
could be developed. The films formed in static tests produced poorly bonded
mixture of calcium, ferrous carbonate and iron oxide in a porous layer of
rust. Adherent, durable layers were usually formed under high flow rate con-
ditions on corroded samples. The presence of colloidal CaC03 was beneficial.
These adherent layers developed were largely hydrous ferric oxide (in the
form of limonite) with 5 to 40 percent calcite (CaC03). Siderite (FeC03) was
usually observed close to the metal surface in ridges (67).
Larson reports that calcium, independent of saturation index, is a mild
but effective corrosion inhibitor of machined cast iron at least in the
presence of sufficient alkalinity (60). The corrosion rate depended on cal-
cium concentration, practically independent of flow velocity from 0.08 to
0.85 fps, pH, minor variations in chloride to alkalinity ratios, and pres-
ence or absence of chlorine, chloramine, and silica. It was found that a
certain length of time was needed for the effectiveness of calcium to become
apparent. The effect was explained on the basis of the corrosion reactions
providing an environment at the surface conducive to the formation of CaC03,
even though the bulk water is below saturation with respect to CaC03.
Stumm measured corrosion rates of cast iron under relatively well char-
acterized conditions (97). Results are shown in Table 6. According to his
analysis, neither the CaC03 saturation equilibrium, nor the relative amount
of CaC03 deposited at the electrode surface are significant parameters of
4-13
-------
TABLE 6. CHEMICAL COMPOSITION AND CORROSION CHARACTERISTICS OF WATERS INVESTIGATED (97)
I
I—>
-ti.
Analysis
Water
Number
J
2
3
4
5
6
7
8
9
10
11
12
13
14
15b
16
17
Type3
Sy
Gr
Sy
Gr
Gr
Su
Su
Sy
Sy
Sy
Sy
Sy
Su
Su
Gr
Sy
Sy
Temperature
in degree
C
10
9.8
10
12
12
4.7
6.4
10
10
3-5
10
10
10
10-18
12
10
10
as mg per L
CaC03
PH
7.1
7.6
7.1
7.5
7.4
7.7
8.0
7.0
8.2
7.0
8.3
8.6
8.4
8.5
7.1
8.0
8.1
Ca+2
180
185
250
158
203
124
99
0
38
40
20
38
65
56
278
0
0
Mg+2
0
69
0
37
55
24
12
0
0
8
0
0
17
17
37
0
0
Alkalinity
250
245
250
180
228
136
103
250
83
12
100
209
47
47
285
250
180
cr
14
3
14
12
4
2
2
14
-
-
14
-
16
16
12
-
12
SO.,-2
20
20
20
-
-
16
-
20
20
-
20
20
32
32
-
20
20
1
mg per L ii
CaC03
)e position,
i mg per cm2
c
20
Si 0 Days
- 11.0
7 11. 0
- 11.0
- 2-4
- 9
2 11.5
2.5 12.2
- 11.0
- 11
- 12.5
- 11.0
- 11.0
- 11
- 3-11
- 5-6
- 11
- 11
0.35
3'. 5
0.24
-
-
2.6
6.5
0
2.4
-
2.5
6.1
0.03
2.6
-
0
0
50
Days
0.43
3.6
0.31
-
0.8
2.7
9.2
0
4.6
-
4.8
12.0
0.10
5.3
-
0
0
Saturation
Index, in
pH units
-0.40
+0.07
-0.25
-0.20
-0.03
-0.40
-0.20
-
-0.40
-2.5
-0.50
+0.30
-1.6
±0.0
-0.15
-
-
Buffer
Capacity,
mg per
PH
1.92
0.074
1.92
0.81
1.01
0.33
0.14
2.30
0.08
0.11
0.00
0.22
0.524
0.045
2.20
0.28
0.17
Corrosion
Ini-
tial
182
165
182
106
150
80
100
125
111
77
83.2
156
105
105
142
208
102
5
da; s
37.6
45
5
-------
corrosion inhibition. He used ground cast iron samples in a number of natur-
al and synthetic waters and exposures over 50 days. The deposition of CaC03
is primarily controlled by the electrochemical changes at the surface and
thus is associated with the corrosion reactions and accompanying pH changes.
He also speculates that the buffer capacity of the solution exerts a consid-
erable influence (greater buffer capacity, i.e.. alkalinity, being less cor-
rosive) and that the anode/cathode relative area is important and pH depen-
dent. The relative size of the local anode areas supposedly increases with
increasing pH. Deposition of CaC03 is stimulated by elevated pH of local
cathode areas but acts to reduce the anode area fraction (97). These consid-
erations make CaC03 deposition more effective at a pH of about 7 than at
higher pH values, and also more effectively applied to well buffered waters.
Patterson contends that effective CaC03 protection can only be provided
when the water contains an alkalinity of at least 50 mg/L (as CaC03), and an
equal amount of calcium (expressed as equivalent CaC03) (75). Using these
minimum values, the pH required to maintain the CaC03 coating is much higher
than the pH calculated using most saturation indices. The CaC03 layer depos-
ited at a high pH has often been found to be less effective than that formed
at moderate pH. Excessively high pH values may promote pitting and
tuberculation.
Recent work by Feigenbaum and co-workers stresses the structure of
natural calcium/iron scales (27). Fifteen natural scale layers formed in
potable water systems carrying waters of various compositions were examined
by scanning electron microscopy, x-ray diffraction, and microanalysis. The
specimens studied showed a distinct outer zone (adjacent to the scale/water
interface) and inner zone (adjacent to the metal/scale interface). The outer
zone is relatively compact and consists of contiguous crystals mainly of
calcite (CaC03). The inner zone is considerably more porous and comprised
of geatlite [aFeO(OH)], siderite (FeC03), and magnetite (Fe30i+) that favor a
needle-like and granular porous structure. A steep gradient in Fe and Ca
concentrations was found in the bulk scale. Depth of the gradient in the
scale varied from scale to scale and appeared to play a role in protective-
ness (27). In a later study, these workers proposed a model based on the
structure and porosity of the scales they had observed and made AC impedance
measurements on scaled specimens to associate with the diffusion resistances
used in the model (23). Correlations were developed between the individual
impedances of the 15 natural scales and their crystalline phase composition
and water composition. A new criterion for the tendency of protective scale
deposition was proposed and compared to others. Results of the correlation
of scale impedance (spatial compactness) and water quality factors are shown
in Table 7. Further comparison of scale resistance with long-term corrosion
experience indicated good correlation with the y value. According to this
criterion, provided sufficient temporary hardness exists, the presence of
chlorides and sulfates can improve the protective properties of scale (28).
4-15
-------
TABLE 7. RESULTS OF CORRELATION ANALYSIS (28)
Correlation Standard
Number Combinations Coefficient Deviation
rp,++~i Turn 1
i |.ia J L"iU3 J (-, 71 co
[C02]
2 Langelier index 0.34 70
3 [Alkalinity] Q 4g 223
[CT] + [S0^=]
4 Y = AH + B ([CL~] + [S04=])exp (-1/AH) + C 0.92 32
-H- — 9
where A = 3.5 x 10~\ B = 0.34, C = 19.0, and H = -^—^ ^HC03 J
[C02]
Effect of Flow Rate and Temperature—
Examples of the diverse and often opposing effects of solution flow rate
on corrosion of iron have been noted in the previous sections of this discus-
sion. The extremes of flow rate can produce serious corrosion: stagnant
situations promoting pitting and tuberculation, and very high flow rates
causing widespread metal losses due to erosion-corrosion. In the interme-
diate range, the effect of flow rate on corrosion rate has been modeled
(apparently for conditions where velocity dependent CaC03 deposition or high
oxygen passivation do not occur) (€6). The equations are based on a double
resistance model in which one resistance is significantly time dependent. An
adequate representation of new data obtained at 150°F and available litera-
ture data was obtained using the semi-empirical correlation and as a function
of Re number and a dimensionless diffusion group (66).
The effect of temperature on corrosion of iron in natural water is also
highly complex. It has received very little independent study. Temperature
changes can affect all of the aqueous equilbria, diffusion rates, deposition
rates and electrochemical reaction rates. In relatively simple systems such
as when the iron corrosion rate is controlled by diffusion of oxygen through
the reaction product film, the rate increases as the increase in oxygen
diffusion rate increases with temperature. In this case, the corrosion
rate doubles with every 30°C rise in temperature up to about 80°C. Above
80°C, in open systems, the corrosion rate decreases sharply due to the marked
decrease in solubility of oxygen with increasing temperature (107).
4-16
-------
Effects of Other Species in Solution--
This section gives a brief discussion of the effects of free chlorine,
chloramine, nitrate, humic acids, and sulfide on the corrosion of iron in
natural waters. Variation of species such as sodium ion, potassium ion, or
magnesium ion is not expected to have appreciable effects on corrosion rates.
The effect of free C12 concentration ( mg/L) is shown in Figure 9 where
they are superimposed on' data obtained with no C12 present (60). These
results were obtained for mild steel in aerated water of about 120 to 135
mg/L alkalinity, about 30 mg/L NaCl, at pH 7 and 8 and at low flow rates.
It can be seen that the corrosion rate is increased in the presence of free
chlorine concentrations greater than 0.4 mg/L. As shown, chloramine actually
acts as a mild inhibitor at low concentrations. The threshold concentration
of free chlorine for accelerated corrosion may be a function of the chloride
to alkalinity ratio, but this was not investigated. Chlorine can act as an
oxidizing agent which is "stronger" than oxygen in neutral solutions.
Chloramine (0.4 to 3.6 mg/1 as C12)
0.4
0.5
0.6
0.7
0.8
Equivalent Ratio Cl'/HCOs
Figure 9. Relative corrosion rates of mild steel at particular
chloride-bicarbonate ratios with and without chlorine (60)
Nitrate ion can be reduced on iron and play a role similar to that of
oxygen as a "cathodic depolarizer." The thermodynamic driving force is not
as high as for oxygen, but there are no solubility limits on nitrate and it
can be present under anaerobic conditions. A case has been described in
which severe corrosion of a 2.5 mile steel main carrying anaerobic well water
was caused primarily by 4-7 ppm (as N) nitrate (12). A detectable decrease
in nitrate concentration and corresponding increase in nitrite, ammonia and
hydroxyl ion (products of nitrate reduction) and dissolved iron was found as
4-17
-------
water passed through the main. Increasing the pH from 6.4 to 8.0 completely
arrested the corrosion both in the presence and absence of chlorine. Nitrate
can under some conditions act as a passivating agent for iron, but this is an
undependable type of inhibition.
The effect of humic acids on the corrosion of black steel pipes in
natural waters has recently been reported (86). These compounds were found
to inhibit corrosion for a range of hardness, flow rate, and chloride values.
The authors interpret this as being due to the inhibition by the humic mater-
ial of the oxidation of the siderite (FeC03) product layer. They attribute
considerable protective properties to siderite layers. It also seems possi-
ble that large organic molecules such as these could also act as direct
adsorption inhibitors or lead to the formation of reaction product layers
whose structure is more protective, regardless of composition.
Hydrogen sulfide or other sulfide species should not be present in any
properly maintained water system. In spite of this, cases do arise where
water containing sulfides is conveyed to consumers usually from small water
suppliers using underground sources (Ilia). The presence of sulfides is al-
most always objectionable to the consumer. In addition, sulfide waters can
be quite corrosive, attacking iron and steel to form "black water" and also
attacking copper, copper alloys, and galvanized piping, even in the absence
of oxygen. The mode of attack by sulfide is often complex and its effects
may either begin immediately or not be apparent for months only to become
suddenly severe. Much of the corrosive action of sulfide may be due to the
partial replacement of oxide or hydroxide films on iron or copper by metal
sulfide films which either disrupt the normal protective nature of the film
or initiate galvanic corrosion. Wells has discussed methods for removal of
hydrogen sulfide and sulfides from water in detail (Ilia).
Comparison of Cast Iron and Mild Steel--
Cast irons are ferrous alloys containing more than 1.7 percent carbon.
Gray fracture due to the presence of free graphite is seen in normal
slowly-cooled cast form. This graphite causes the brittleness of cast iron
and is the important metallurgical difference from mild steel. From a corro-
sion standpoint, the main differences are:
a surface skin of iron oxide, silicates, and alumina
silicates which is formed on cast iron during production.
the existence of graphite sites which occur at 0.04 mm
intervals on ground cast iron surfaces (57).
4-18
-------
graphitic corrosion of cast iron is possible.
The exterior skin can increase corrosion resistance of cast iron relative to
mild steel, but this layer is often partially removed by grinding, especially
prior to the application of linings. Grinding exposes the graphite sites,
and these can stimulate corrosion relative to steel during initial exposure
by galvanic attack. There seems to be little difference between corrosion
rates of ground cast iron and steel at long durations. Under some conditions
a selective leaching of iron (due to the galvanic cell formed by graphite and
iron) can occur ultimately leaving a porous mass consisting of graphite,
voids, and rust. This is usually a slow process.
Corrosion of Galvanized Iron
Galvanized (zinc coated) steel is an example of a coating used as a
cathodic protection device. The zinc coating is put on the steel not because
it is corrosion resistant, but because it is not. The zinc corrodes prefer-
entially and protects the steel, acting as a sacrificial anode. Electro-
deposited zinc coatings are more ductile than hot-dipped coatings and usually
quite pure. Hot-dipped coatings form a brittle alloy layer of zinc and iron
at the coating interface. Corrosion rates of the two coatings are comparable
except that hot-dipped coatings, compared to rolled zinc and probably elec-
trodeposited Zn, tend to pit less in hot or cold water. This difference
suggests that either specific potentials of the intermetallic compounds favor
uniform corrosion, or that the incidental iron content of hot-dipped zinc is
beneficial. In this connection, it is reported that Zn alloyed with either
5 or 8 percent Fe pits less than pure Zn in water (107). Zinc used for hot-
dip galvanizing may contain 0.01 to 0.1% cadmium and up to 1% lead as
impurities (73).
Effect of Water Quality Parameters--
In aqueous environments at room temperature the overall corrosion rate
of zinc in short-term tests is lowest within the pH range 7 to 12. In acid
or very alkaline environments, major attack is by H2 evolution. Above about
pH 12.5, zinc reacts rapidly to form soluble zincates by the following
reaction.
Zn + OH' + H20 -> HZn02- + H2
In general, both zinc and cadmium react readily with nonoxidizing acids to
release hydrogen and give divalent ions. Cadmium, however, is relatively
stable in bases since cadmiate ions, if formed, are much less stable than
zincate ions. The effect of pH on corrosion of Cd is shown in Figure 10.
In the intermediate pH range of main interest here, the main cathodic reac-
tion in aerated waters is probably reduction of oxygen. The corrosion rate
of zinc in distilled water increases with oxygen concentration and with C02
from air saturation (105). Nonuniform aeration of the surface can cause
differential concentration cells and localized corrosion of zinc. The corro-
sion rate of zinc increases with temperature as discussed below. In general,
corrosion in actual use is greater in soft waters than hard waters (52,108 ).
Chlorine additions, in the amounts normally used for health protection of
water supplies, do not increase the corrosion of zinc in water (2).
4-19
-------
.zoo
.180 -
.160
.140 -
.120 -
.100 -
.080 -
.060 -
.040-
.020-
- 1600
- 1400
- 1200
- 1000
- 800 e
- 600
- 400
- 200
| 2 3 4 5 6 7. 8 9 10 II 12 13 14
pH
Figure 10. Corrosion of Cadmium vs. pH in continuously
flowing, uniformly agitated and aerated solutions of HC1
or NaOH (108).
Material: 5 x 10 x 0.63 cm (2 x 4 x 1/4") cast.
cadmium.
Temperature: 24 ± 0.5°C (74 ± 1°F).
Time: 7 days for pH below 2; 41 days for pH above 2.
4-20
-------
Wagner has summarized results from field and laboratory tests on the
effect of water quality parameters on corrosion of galvanized steel tubes
(109). He shows a definite correlation between corrosion rate and pH, at
least for the zinc phase of the coating and with steady flow of water (at
0.5 m/s). These results, shown in Figure 11, indicate that corrosion rate
increases rapidly with a decrease in pH in the pH range 7 to 8. This effect
is said to exist in spite of other water quality parameters. According to
Wagner, there is negligible effect of buffer capacity and saturation index
on the corrosion rate of galvanized steel tubes, although the composition of
the deposits are altered. Corrosion rate does vary with time, first decreas-
ing as zinc corrosion products grow. Once formed, the coating gives a con-
stant (but pH-dependent) rate as long as the metallic zinc phase is present.
Once the Zn/Fe alloy phase is reached, the rate decreases again but reaches
another constant value which is also pH dependent. Effects of additives and
organic acids are also discussed (109).
T3
IN
E
^-»
cn
c
10,0
5,0
2,0
I 1.0
l/l
o
u
8 °-5
o
c
0,2
70
• Rotenb»rg
o Bobtingsn
• Mannhaim
a Witten
o
o
• 00
7.5
8,0
PH
Figure 11. Effect of pH on corrosion on galvanized steel tubes
4-21
-------
One of the important environmental factors for galvanized steel or iron
is the dissolved copper content of the water. A corrosion study of galvan-
ized steel and galvanized wrought iron pipes in 25 selected domestic waters
for two years has been reported (70). Computer correlation of corrosion
grades with a large number of factors such as chloride, pH, saturation index,
alkalinity, hardness, flow rate, etc. was attempted. The only definite
correlation for corrosion of galvanized pipe with water quality was for
dissolved copper concentration. Chloride concentration was a possible accel-
erating factor. In general, the remainder of the results were difficult to
interpret. Attack was observed only on the zinc, not iron. There was no
evidence that high alkalinity (above 100 ppm) or silica had inhibitory
effects. Several case histories illustrating the copper effect are given by
Cruse (19). His results, showing the correlation of copper found in corro-
sion products with maximum pitting rates in potable water, are shown in
Figure 12.
01
c
100
80
0) —
> — 60
i- E
1/1 i
.a
o D
40.
E fl3
I * =0
X - Cold Systems
o- Hoi Sy.lems lAJjuslcd lo V'O Fl
* Heavy S'lxa Deposit
-Heavy CalC'Mrr. CatboiMlc Deitn-ns
20
30
10
Copper Found in Corrosion
Products - mq/dm2
Figure 12. Corrosion of galvanized pipe in potable water systems (19).
Stagnant Conditions--
The increase in concentration zinc, along with cadmium and lead, from
hot-galvanized pipes in domestic drinking water systems has recently been
discussed (10). Under stagnant conditions, concentrations of 5 to 10 ppm
zinc may be obtained after 8 to 40 hour exposure to new galvanized pipes.
This behavior is almost independent of water composition. Most of the corro-
sion products are not dissolved but present as an "oxidic colloid" or finely
divided solid. The content of corrosion products in stagnant water could be
decreased by treating the water with phosphates or silicates. Their research
results indicate stagnation water in new pipes can also contain cadmium and
lead products but in rather lower proportion to zinc than their concentration
in the galvanizing layer. The authors are apparently referring to tests
using piping corresponding to a German standard in which the hot-galvanized
coating consists of about 97 percent Zn and 3 percent Fe with maximum limits
of 0.8 percent Pb and 0.01 percent Cd.
4-22
-------
A field investigation dating from 1973 was also undertaken for new
plumbing in a three story building (10). All samples were taken from one tap
on the top floor. The water composition is shown in Table 8; initial lead
and cadmium contents probably coming from "old pipework." Zinc, cadmium, and
lead contents in stagnation waters are shown in Table 9. The scatter in
individual measurements is attributed to the corrosion products being mainly
present in the form of suspended solids. Taking the initial values of Cd
and Pb to be 2 yg/L Cd and 25 yg/L Pb, the concentration increases due to
building plumbing and relative amounts shown in Table 10 were obtained.
Random samples from the pipe materials in this plumbing system showed a com-
position of 97 percent Zn, 0.7 percent Pb, and 0.06 percent Cd for the gal-
vanizing layer. This gives ratios of Zn/Cd = 1600 and Zn/Pb = 140 which,
according to the authors, indicate some retention of Pb and Cd in the pipe
covering layers (10).
TABLE 8. WATER COMPOSITION (10)
Parameter
Content (millimole/L,
except as noted)
PH
Alkalinity (to pH 4.3)
Acidity (to pH 8.2)
Chloride
Sulfate
Nitrate
Cu2+
Cd
Pb
02
7.0
3.6
0.85
1.6
0.93
0.2
1
20
3
-7.3 pH Units
- 4.5
- 1.3
- 2.3
-1.33
- 0.3
< 0.01 mg/L
- 3 ng/L
- 30 yg/L
- 7 mg/L
TABLE 9. ZINC, CADMIUM AND LEAD CONTENTS IN STAGNATION WATER (1-3 DAYS) (10)
TIME OF INVESTIGATION:
Cadmium Contents in yg/L
Number of Determinations
Range of Variation
Mean Values
Lead Contents in yg/L
Number of Determinations
Range of Variation
Mean Values
Zinc Contents, mg/L
Number of Determinations
Range of Variation
Mean Values
1973
12
2.1/9.9
6.9
1974
17
2/8
4.6
12
20/120
56
23
1.9/9.4
6.7
1975
9
5/10
6.0
9
69/90
80
9
5.6/7.1
6.3
1976
10
1/11.
5.4
10
47/96
67
10
3.7/6.7
5.6
1977
3
4.8/10
7.6
3
25/50
39
3
2.1/4.4
3.5
4-23
-------
TABLE 10. CADMIUM AND LEAD CONTENT OF STAGNANT WATER (1-3 DAYS) (10)
Period
1974
1975
1976
Mean Value
of Cd (yg/L)
2.6
4.6
3.1
Mean Value
of Pb (yg/L)
31
55
45
Zn/Cd Zn/Pb
(mass ratio)
2567
1370
2100
216
115
142
In flowing water the zinc content remained about 8 to 15 ppm for each
year from 1973 through 1977, apparently due to extraction from solid zinc
compound covering layers. No increase in Pb or Cd (within experimental pre-
cision) was noted in flowing water (10).
Comparative tests of galvanized iron in waters having a pH value between
7.5 and 9.5 and containing calcium bicarbonate, but having a very low content
of sulphates, chlorides, and nitrates, show that the attack on zinc was soon
mediated. But in waters low in calcium bicarbonate or containing appreciable
quantities of sulphates, chlorides, or nitrates, zinc suffers pitting attack.
Zinc tends to form an insoluble basic zinc carbonate layer on the attacked
areas which limits their size. Other findings were that zinc electrochemic-
ally protects iron when it is clean, but when coated with a resistant layer
of corrosion products the maximum distance at which it still gives protection
decreases. Zinc protects the alloy layer of a galvanized coating where it
has become laid bare, but the alloy layer will not afford electrochemical
protection to bared iron. The calcium carbonate which is deposited on alloy
or on iron in the course of electrochemical protection is itself protective,
and the ultimate success of the electrochemical protection of any exposed
area appears to depend on the building up of the protective effect of this
chalk layer at a greater rate than that at which protection is lost by disso-
lution of zinc from the area adjacent. The thickness of the zinc layer, the,.
calcium bicarbonate content, pH value, and conductivity of the water appeared
to be the deciding factors (48).
Hot Water Corrosion--
A number of studies have been performed on the corrosion of zinc in hot
water tanks (39,47 ). Some of the findings may be summarized as follows:
1. That low alkalinity water containing all or nearly all calcium
and magnesium normal carbonates, at a temperature of 150°F,
is more corrosive to galvanized metal than water containing no
normal carbonates, even though dissolved oxygen is appreciable
and some free C02 is present.
2. That normal calcium and magnesium carbonates deposit an uneven
scale, from 1/32 to 1/16 of an inch in thickness, which is some-
what adherent, whereas the bicarbonates of calcium and magnesium
4-24
-------
produce a much thinner scale, probably basic carbonate of zinc,
which adheres much better to the metal.
3. That corrosion of zinc depends more upon the alkalinity and pH
of the water than upon the dissolved oxygen present.
4. That the attack of normal carbonate water on the galvanized
tank, which brings about failure, starts at the pin-holes or
the weak spots in the galvanizing. The water soon penetrates
through such pin-holes and then attacks the iron. These pits
keep spreading and growing deeper until the tank fails.
5. That the products of corrosion fill the pin-holes, thus acting
as a protective film.
The effect of temperature on weight losses of zinc was established by
using distilled water with aeration and agitation. Weight losses at 150°F
were found to be six times greater than at 50°F. This increase in corrosion
was found to be caused by a change in the nature of the film from an adher-
ent gelatinous state to a nonadherent granular state. The nonadherent film
was found to exist between 131° to 167°F, which is within the range of use in
water tanks (39).
In many aerated hot waters, reversal of polarity between Zn and Fe
occurs at temperatures of about 60°C (140°F) or above (.107). This leads to
Zn having the characteristics of a noble coating instead of a sacrificial
coating, and hence a galvanized coating under these circumstances induces
pitting of the base steel. A 15-year service test on piping carrying Balti-
more water at a mean temperature of 46°C (115°F) and maximum of 80°C (176°F)
confirmed that pitting of galvanized pipe was 1.2 to 2 times deeper than in
black iron pipe (ungalvanized) of the same type, corresponding to shorter
life of the galvanized pipe. In cold water, however, pits in galvanized pipe
were only 0.4 to 0.7 times as deep as those in black iron pipe, indicating in
this case a beneficial effect of galvanizing. It was found that waters high in
carbonates and nitrates favor the reversal in polarity, whereas those high in
chlorides and sulfates decrease the reversal tendency (107).
The cause of this reversal is apparently related to the formation of
porous Zn (OH)2 or basic Zn salts, which are insulators, under those condi-
tions for which Zn is anodic to Fe, but to formation of ZnO, instead, under
conditions where the reverse polarity occurs. The latter compound conducts
electronically, being a semiconductor. It can therefore perform in aerated
waters as an 02 electrode whose potential, like mill scale on steel, is noble
to both Zn and Fe. Accordingly, in deaerated hot or cold waters in which an
02 electrode does not function because 02 is absent, Zn is always anodic to
Fe, but this is not necessarily true in aerated waters. Apparently, the
presence of HC03~ and N03" aided by elevated temperature stimulates formation
of ZnO, whereas Cl~ and SO^" favor formation of hydrated reaction products
instead.
At room temperature, in water or dilute NaCl, the current output of zinc
as anode decreases gradually because of insulating corrosion products which
4-25
-------
form on its surface. In one series of tests, the current between a couple
of Zn and Fe decreased to zero after 60 to 80 days and a slight reversal of
polarity was reported. This trend is less pronounced with high purity zinc
on which insulating coatings have less tendency to form.
Stainless Steels
According to Reedy, stainless steels used by the water works industry
can be divided into the three groups given below (32)-
1. Chromium steels (containing 12 percent or more Cr) designa-
ted by the American Iron and Steel Institute (AISI) as
Type 400 series (best known examples are Types 410 and 430).
2. The 18-8 chromium-nickel stainless steels, designated AISI
Type 300 series, of which Types 301, 302, and 304 are fre-
quently used.
3. AISI Type 316 stainless steel with a nominal composition of
18% Cr, 12% Ni, and 2.5% Mo.
Stainless steels are frequently used where protective coatings are not satis-
factory or cannot be used such as in corrosive areas of water treatment
plants and in components such as pumps, valves, meters, Venturis and pressure
regulators. Extensive use of stainless steel as a cladding material for up-
take and downtake shafts and for control and distribution chambers in a New
York City water tunnel has been reported (36). This project will use seven
million pounds of stainless steel clad.
The resistance of stainless steels to uniform corrosion is generally
good, but overall corrosion stability depends on maintenance of a passive
state. Passivity in the .sense used here is described below. Maintenance of
the corrosion resistant state depends on both the particular type of stain-
less steel employed and on the environmental conditions.
Passivity--
The corrosion resistance of stainless steels depends primarily on the
ability of the material to achieve and maintain a passive state in an aggres-
sive environment. The ability to achieve a passive state determines the
reduction in uniform corrosion rate from that of an active metal. The abili-
ty to maintain the integrity of the passive state determines the resistance
of the material to local attack caused by the chemical environment, eroding
particles, or stress. This passive state is characterized by a certain elec-
trode potential or limited potential range over which the dissolution rate
changes from a relatively high to a relatively low value. The electrode
potentials associated with the onset of passivity and the maximum current
density needed to reach the passive state from the active state are a func-
tion of the metal composition, chemical composition of the environment, and
temperature.
The detailed mechanism of formation and exact nature of the passive sur-
face components have been a subject of controversy for many years. It is
4-26
-------
probable that primary protection is offered by a very thin (on the order of
10 to 100 angstrom, amorphous, pore free metal oxide film. The film should
be an electronic conductor. Thicker, more porous crystalline layers formed
on top of the thin passive layer may offer substantial barriers for chemical
or mechanical attack on the basic film. However, the primary corrosion re-
sistance is dependent on the initial establishment of the passive layer and
its continued stability. The type of passivity defined here should be dis-
tinguished from that attributed only to the formation of sparingly soluble
reaction products on a metal surface. The latter process, which can be. pre-
dicted from Pourbaix diagrams, does not necessarily imply an established
passive state and the absence of corrosion. It is essential that the reac-
tion products be formed directly on the metal surface and as a direct conse-
quence of the anodic reation; complex processes such as chemisorption may
also be intimately involved.
Localized attack occurs when momentary passive film breakdown, in the
presence of chloride, exposes a small part of the metal surface which is
surrounded by a large film-covered cathodic area. The cathodic area can
drive anodic metal dissolution at the bare spot at a high rate (high current
density). Once started, hydrolysis of metal ions from the anodic reaction
causes the pH in the incipient pit or crevice to decrease which in turn dis-
courages film repair and augments attack. Growth involves migration of
chloride into the area which can also augment corrosion. Much of the greater
resistance of the more resistant materials is probably due to their greater
rate of film repair so that pit initiation is much less likely.
Type of Corrosion and Effect of Alloy Composition--
Passivity in relatively mild environments is obtained by alloying iron
with at least 12 percent chromium. The stainless steels formed with more
than 12 percent Cr as the only major added component, typified by the Type
400 materials, generally show relatively low uniform corrosion rates in
typical potable waters. They are highly susceptible, however, to severe
localized corrosion (pitting, crevice, and weld related attack) in waters
containing chloride and oxygen (37, 82, 96)• Additional alloying is needed
to obtain resistance to localized corrosion, but even the more resistant
alloys are occasionally susceptible to local breakdown of passivity. Type
400 materials are not recommended for submerged service in potable water
systems (32)- The stainless steel most favorably used for such service is
probably Type 304, which with 18 percent Cr and 8 percent Ni, has much better
localized corrosion resistance. Addition of about 2 percent molybdenum (and
more nickel) to Type 316 generally gives higher resistance to chloride
induced pitting and crevice corrosion than Type 304. While uniform corrosion
rates are low for Type 300 materials in potable water environments (36, 37,
96), their application must be well understood to avoid localized attack.
Good welding procedures and design to eliminate crevices are particularly
important. More resistant types than Type 316 are known, but they are mainly
used in extremely corrosive process industries.
Environmental Effects on Corrosion of Stainless Steels--
The two most important chemical factors in potable water systems are
chloride and oxygen. Oxygen plays a conflicting role in the corrosion of
stainless steels. Oxygen is necessary for the maintenance of the passive
4-27
-------
state for many stainless steels, but it also provides the driving force for
local disruption of the passive film in the presence of chloride to give
localized corrosion. The susceptibility of a given material to localized
attack increases with chloride concentration. Sulfate and perhaps bicarbo-
nate appear to act as pitting inhibitors and their concentration relative to
chloride is a relevant factor (61). The solution pH in the neutral region
has relatively Httle effect on pitting of 300 series stainless steels
(M. J. Johnson in Reference 3).
Temperature can have a major effect on the localized corrosion of stain-
less steels. Susceptibility to pitting increases with increasing temperature
for both Types 304 and 316 and the change is fairly marked for moderate
(15-30°C) temperature increases above 25°C. Scale and other deposits
decrease the corrosion resistance of stainless steels since they provide an
opportunity for establishment of differential aeration cells and for
crevice corrosion. Stagnant areas can be deleterious for similar reasons.
High flow rates can usually be tolerated, in the absence of eroding debris or
entrained solids, in the liquid phase. Careful design is needed to ensure
satisfactory performance of these materials.
Results in Potable Water--
While stainless steels have been tested under a variety of conditions in
the chemical process industry and laboratory, relatively few direct tests in
potable water systems have been reported. Results have been reported for
Type 410 and Type 316 in Southern California waters including treated Colora-
do River water (96). Type 410 was stable in four aqueduct and well waters,
but was very severely corroded by pitting in both treated and untreated Colo-
rado River water. Since pH and dissolved 02 were nearly the same in most
cases, the reason for increased corrosivity of the Colorado River water may
have been its high chloride content (85 ppm) relative to the other waters
(16-28 ppm). Type 316 was completely undamaged in any of the waters (96).
Extensive tests of a variety of stainless steels in New York City
Reservoir water have been reported (36,37 ). A typical average analysis of
this water is shown in Table 11. Results obtained during exposure in a semi-
stagnant area in a shaft 14 feet above the main flow are shown in Table 12.
Type 416 was not submerged. The pitting rates for the 300 series materials
and 17-4 PH (a cast stainless Steel) are all quite small. Weight loss corro-
sion rates did not exceed 0.01 mpy (36). Similar results were later reported
at longer exposures (up to 15 years) for samples submerged in jars containing
periodically renewed samples of the same water (37). A submerged Type 416
sample suffered severe crevice corrosion in these tests. Other results were
essentially the same; no attack was noted on 300 series or 17-4 PH materials.
CORROSION OF COPPER IN POTABLE WATER SYSTEMS
Copper is a highly regarded material for use in potable water service
lines. It is flexible, easy to join and install, has a low resistance to
flow, and is considered to be fairly resistant to the corrosive action of
most waters. Corrosion can occur under certain conditions, however, causing
a number of water quality problems. Most of these problems have been
4-28
-------
TABLE 11. TYPICAL AVERAGE ANALYSIS OF NEW YORK CITY RESERVOIR WATER
(Catskill-Delaware System—1969) (37)
Concentration (mg/L)
Component (except as noted)
Alkalinity (as CaC03)
Calcium
C02
Chloride
Copper
02
Hardness (as CaC03)
Sulfate
Silica (as Si02)
Total Solids
5
3
1.8
4.0
0.01
13.7
19
6.0
0.5
40
- 14
- 6.3
- 4.0
- 7.0
- 0.10
- 14.3
- 23
- 14.0
- 2.5
- 54
pH 6.5 _ 7.5
Specific Conductivity (ymho) 60 - 73
TABLE 12. STAINLESS STEELS—PITTING IN RESERVOIR WATER (36)
Test Duration Maximum Pit Depth
Alloy (years) (mpy)
304
304L
316
31 6L
317
17-4 PH
4161
2
9
9
9
9
2
4
0.25
0.23
0.11
0.11
0.34
0.05
7.20
Exposure at 100 percent humidity and 23°C.
NOTE: Each item represents average of 3 or more specimens
4-29
-------
delineated by an AWWA task group report and are summarized as follows (16).
Small concentrations of copper cause the formation of blue or blue-green
stains on procelain fixtures. Concentrations greater than 1.0 ppm react with
soap to produce insoluble green "curds." A bitter, unpalatable taste results
from copper when it is present in concentrations greater than 1.0-1.5 ppm.
Traces of copper can accelerate the corrosion of galvanized hot-water tanks
and cause pitting of aluminum pots and pans. Traces of copper are objection-
able in many industries, i.e., in those involved in the canning of foods and
in those using metallic-plating baths. Small amounts of copper in irrigation
water are toxic to sugar beets and barley grown in nutrient solution. A con-
centration of 2 ppm or more is believed to be toxic to tomatoes. Copper is
toxic to fish in concentrations of 0.25-1.0 ppm. Pitting corrosion may lead
to the failure of copper pipe as a result of pin-hole leaks.
Corrosion of copper in potable waters has been the subject of numerous
studies and papers. There are many factors which influence the corrosion
process. The interdependence and frequent lack of independent control over
these factors has led to a rather chaotic literature. However, favorable
conditions for the corrosion stability of copper must predominate in most
cases of potable water use based on its overall record. From World War II
to 1972 over 6 million miles of copper plumbing tube was put into service
(15). Relatively few cases of actual failure have occurred.
General Considerations
Although many modes of corrosion of copper have been distinguished, this
discussion will classify the types of corrosion of interest here into only
two types: general or uniform corrosion and localized corrosion. Velocity-
related corrosion will be discussed under localized corrosion. These modes
encompass the most prevalent ones from a water quality standpoint. Their
basis is interrelated to a certain degree. It often appears that conditions
which suppress uniform corrosion can give way to localized corrosion.
In the total absence of oxygen, copper is thermodynamically incapable of
corrosion in potable water environemnts. The usual presence of at least some
oxygen is the driving force for corrosion of copper in these environments,
but it is not usually the determining factor. It is doubtful that total
exclusion or removal of oxygen from potable water systems is desirable or
economically feasible. Such efforts could in fact be deleterious due to the
formation of sulfide, ammonia, or related compounds by bacterial action which
could be harmful to copper or other materials, especially if they are even-
tually exposed to waters containing oxygen or other oxidants.
Given the presence of oxygen and possibility of corrosion, the actual
occurrence of copper corrosion is governed by the presence and stability of
inorganic compound films on copper. Probably the most noteworthy of these
from a corrosion standpoint is cuprous oxide (Cu20) formed by initial corro-
sion of copper. As will be noted, the mode of formation is very important.
Films of cuprous oxide formed by annealing or other manufacturing steps may
be detrimental to the stability of the basic metal.
4-30
-------
The protective, metal-solution-grown layer of cuprous oxide may be very
thin and subject to further attack by solution species and physical erosion.
The cuprous oxide layer is often formed slowly and is tenuous in comparison
to the type of passive films which offer much of the corrosion protection to
metals such as iron, nickel, cobalt, and chromium. As such, it is fortunate
that other, thicker, films of insoluble copper compounds are often formed
over the primary film and can offer mechanical protection from erosion and an
additional barrier to diffusion by aggressive chemical species. Films formed
by deposition of calcium carbonate and similar compounds can offer similar,
but still somewhat unnatural, forms of protection. But the most basic pro-
tection is offered by the delicate, and often slowly grown, thin layers of
cuprous oxide. For this reason, conditions occurring during the initial ser-
vice exposure of copper are very important. Moderate flow rate and chemi-
cally mild environments favor formation of small grained, compact and protec-
tive layers of Cu20. Rapid deposition of disordered Cu20 under harsher
environments may be deleterious since the first layer is not very protective.
At high flow rates in aggressive environments, growth may be delayed or very
slow. Growth of localized layers of materials such as copper sulfides may
be very harmful due to their galvanic influence on adjacent areas of metal-
The dynamic nature of these processes must also be kept in mind.
Materials in service are often subjected to stresses which demand more or
less continuous maintenance of a protection mechanism.
Uniform Corrosion of Copper--
From experience, the uniform corrosion rate of copper is usually quite
low in potable water systems. This observation is based on the long life of
the majority of tubing in service. From a health standpoint, however, low
corrosion rates over a large uniform surface area can add a significant
amount of impurity to a relatively small volume of water. Therefore,
although low uniform corrosion rates may be acceptable based on corrosion
lifetimes, they may be significant from a water quality viewpoint.
Selected literature results on uniform corrosion of copper in potable
waters are given below. Emphasis is placed on references which are illustra-
tive of the literature as a whole. It is noteworthy that the maximum repor-
ted copper concentration in standing water of reasonable pH is about 5 ppm
and that this value is reported quite often. This could indicate that
copper concentration is ultimately limited by solubility of a reaction
product.
Effect of 02
The simplest overall electrochemical uniform corrosion mechanism for
copper, requiring simultaneous oxidation and reduction at the material/
solution interface, can be represented by the parallel reactions Eqs. 5 and
6.
Cu -> Cu2+ + 2e- ( 5)
1/202 + 2e- + 2H+ + H20 ( 6)
4-31
-------
There is evidence that, at least in acidic solution, 02 is also involved in
reaction with intermediate Cu+ as in the following more elementary steps
including the "chemical" step (Eq. 8).
Cu * Cu+ + e~ (7)
Cu+ + 02 + 2H+ + Cu2+ + H202 ( 8)
As previously mentioned, in the total absence of oxygen copper is
'thermodynamically incapable of corrosion in potable water environments. This
was substantiated by Schafer's study of corrosion of copper pipes in potable
water service in New Zealand (89) where a negligible amount of corrosion was
reported in the absence of oxygen. However, the amount of oxygen needed for
corrosion to occur may be small. Tronstad and Veimo (105, 106) determined
that varying the oxygen content of the tap water between 2.7 ppm and
25 ppm had a relatively small effect on the final dissolved copper concentra-
tion. With 100 ppm added NaHC03 (pH 7.0) the copper cencentration increased
about 50% for the 10-fold oxygen increases, while with 500 ppm NaHC02
(pH 7.2) the copper content increased by about a factor of 3 over the same
oxygen range. The authors attributed this to oxidation of dissolved cuprous
compounds.
Effects of pH
The effects of pH on copper corrosion have been studied in direct and
indirect tests. In their work with water standing in copper tubes, Tronstad
and Veimo Cl05 , 106J measured the copper concentrations in water after 24
hour exposure of tubing to tap water alone, and with various additions of
oxygen, carbon dioxide, sodium bicarbonate, sodium hydroxide, and calcium
oxide. The tap water composition and pertinent experimental details are
given in Table 13. All experiments were done at 18°C in the closed tubes.
This tap water can be classified as soft, having a fairly low pH, moder-
ate amount of aggressive C02, low alkalinity, and moderately high oxygen
content. In the tap water alone, the copper concentration reached a maximum
of 0.6 ppm after about 15 hours and then slowly declined. Addition of sodium
bicarbonate raised the pH but also consistently raised the amount of copper
dissolved over a 24-hour period. For example, addition of 0.4 g/£ NaHC03
produced a pH of 7.2 and dissolved copper of 1.9 ppm. Addition of calcium
bicarbonate had a similar effect. Addition of carbon dioxide lowered the pH
and increased both the copper content and rate at which a maximum was
reached. For example, 9 ppm C02 gave a pH of 5.65 and copper content of 3.9
ppm after 24 hours; with 45 ppm C02, a copper content of 4 ppm was reached
after only 2 hours.
4-32
-------
Addition of sodium hydroxide or lime to the tap water caused a decrease
in the copper concentration observed after 24 hours until a pH between 8 and
9 was reached. Above a pH of about 10, the copper concentration increased
once again. The lime additions resulted in somewhat lower copper concentra-
tions, at the same pH, than sodium hydroxide additions.
TABLE 13. CONDITIONS OF COPPER DISSOLUTION EXPERIMENTS
OF TRONSTAD & VEIMO (105)
Tap water composition:
Total solids
Ash
Chlorides as C£~
Iron
Combined C02
Free (aggressive) C02
PH
KMn04 consumed
NH3, NOX, phosphates
47.
28,
9,
ppm
ppm
ppm
0.07 ppm
8.0 ppm
4.0 ppm
6.3
65 m£/£ of 0.01 N KMnOi
Not detected
Properties of copper tubing:
Purity: 99.8 to 99.9%
Treatment: annealed, bent into loops, descaled in 2 N
NH4OH, washed with distilled water
Specimen Size:
Volume: 99 cm3
Internal surface area: 785 cm2
Internal diameter: 0.5 cm
Length: 500 cm
4-33
-------
Experimental studies under simulated domestic use conditions were done
to determine the effect of pH on copper corrosion and were reported by an
AWWA Task Group (16). The water composition was not specified, but appar-
ently it was aerated and contained C02. In the first set of experiments,
water at an adjusted pH value was allowed to flow through 60 feet of new
0.75 inch copper tubing at a rate of 0.067 gpm (0.05 fps). Samples were
collected at the end of I hour and again after 1.25 hours, the copper concen-
tration determined, and the two values averaged. The observed copper concen-
trations as a function of pH are shown in Figure 13.
Figure 13. Effect of pH on corrosion of copper (16),
In the second type of experiment, water at the desired pH was passed
through.new copper tubing of the same dimensions at a flow rate of 0.5 gpm
(0.37 fps) for 1 hour. The flow was stopped and the water allowed to stand
in the tubing for 16 hours (to simulate overnight conditions). The flow was
then started again with water at the initial pH and rate. Water samples were
collected immediately and at various time intervals and analyzed for copper.
Results as a function of time and pH are shown in Figure 14. The exponential
decay suggests a simple rinsing effect of the dissolved copper solution
formed during the stand.
The Task Group concluded from these two sets of tests and others that
the carbon dioxide content of a water (indirectly measured by its pH) has a
very significant effect on the corrosion solubility of copper (16). In
addition, raising the pH to a value above 7.0 "greatly minimizes" this
action.
4-34
-------
10 20 30 40 50 60
T i me - m i n
Figure 14. Effect of time and pH on corrosion of copper (16)
Effect of Free C07—
The free C02 or carbonic acid content of natural waters has often been
reported to play a significant role in the uniform corrosion of copper.
This could be due to its acting as a source of hydrogen ions at the surface
since the free C02 concentration can be many times the bulk hydrogen ion con-
centration. Free carbon dioxide may also act to dissolve or thin the protec-
tive films on copper, thus enabling higher corrosion rates. Both mechanisms
could occur in practical systems. The amount of C02 necessary to cause in-
creased corrosion in potable waters probably depends on many factors. A
concentration of 10 ppm is often quoted as a rule-of-thumb "threshold" con-
centration for anticipating or diagnosing C02 corrosion problems.
A review of early results on corrosion of copper and brass pipe and the
relations to health is given by Hale (40). His general conclusions are that:
Soft waters containing very little free carbonic acid will
show little corrosion of copper or brass pipe.
Soft waters containing considerable free carbonic acid
(10 ppm and up) may corrode copper and dezincify yellow
pipes.
4-35
-------
Several examples were given for soft, high C02 waters in which standing
copper concentrations of 5 ppm were obtained and up to 3.9 ppm with water
running continuously. Examples are also given in which treatment to remove
the C02 by pH adjustment and/or aeration were effective in controlling the
corrosion of copper. Aeration of water containing little natural oxygen is,
of course, discouraged.
Effects of Temperature
In their previously mentioned studies (105, 106} Tronstad and Veimo also
investigated the effect of elevated temperature. Copper concentration measure-
ments were made at different times at 46°C and 70°C and the data were corrected
for the change in solubility of C02 with temperature. Although the initial
rate of copper dissolution increased with temperature, the maximum concentra-
tion in solution did not change much, apparently due to solubility limita-
tions and consumption of the available oxygen.
Schafer (39) obtained copper corrosion data by sectioning and measuring
pipes that had been in service up to 30 years. He found that corrosion in
hot or previously heated water was usually less than in the same water before
heating, presumably due to removal of dissolved C02 and 02 at elevated temp-
eratures. Most corrosion appeared to occur in the first few years after
installation.
Effects of Miscellaneous Parameters
Low concentrations of iron (0.05 to 0.5 ppm Fe2+) have been found to
inhibit the corrosion of copper in service such as seawater desalination.
This concentration of iron could be dependent on the previous contact of
water with steel or cast iron pipe. The effect should be considered when
evaluating or designing tests of copper corrosion rates. It also raises the
possibility that much copper pipe is "protected" by upstream iron pipe.
The distinction between surface and artesian waters may also be of in-
terest in connection with reports of a natural inhibitor of localized cor-
rosion of copper. This unidentified, probably organic inhibitor, is said to
occur in most surface waters but not in underground waters (13). However,
the distinction between surface and artesian waters noted by Schafer with
regard to copper corrosion rates in cold (normally 10-20°C) potable waters
was that:
Surface waters usually showed a corrosion rate below 0.5 mpy
(to a first approximation) irrespective of pH, hardness,
chloride, and other water properties.
In low pH, high carbon dioxide artesian waters containing
moderate amounts of oxygen, corrosion was more rapid than
water and rates up to 15 mpy werp observed.
4-36
-------
A report of high concentrations of copper from 12-14 month old copper
service lines in one district in England has been given (8). The chemical
composition of the initial potable water was not reported. Initial morning
water samples from houses in towns 12 miles apart both showed maximum copper
concentrations of 5.5 ppm. These were serviced by 65 foot lengths of 0.5
inch copper tube. Four consecutive-day samples from one of these houses had
copper concentrations within the narrow range of 4.8 to 5.5 ppm. Other re-
sults given in the paper do not indicate any correlation of copper concentration
with length of service pipe. Based on the limited data given, the copper
concentration does appear to vary inversely with pipe diameter, which could
be due to the change in surface to volume ratio. Results are shown in Table
14.
TABLE 14. COPPER IN INITIAL MORNING SAMPLE OF
WATER FROM COPPER PIPES (8)
Copper
Diameter Length Concentration
Sample Type of Pipe (inch) (feet) (ppm)
A
B
copper tube, welded joints 1.25 6000 0.75
soft copper tube, mechanical 0.50 65 5.5
joints
C copper tube, mechanical 1.00 1740 1.2
joints
D hard (straight) copper tube 0.50 15 3.2
E soft (underground) copper 0.50 60 4.8 - 5.5
tube
NOTES:
1. Samples A, B, C - all from same main
2. Samples D, E - from 2 houses in a town 12 miles away
3. All pipe said to be to British Standard Specifications,
in service for 12 - 14 months.
4-37
-------
The same report states that copper content at any house during normal
day-time use did not exceed 0.5 ppm (8). Equipment was installed at one
house to dispense a "slowly soluble" hexametaphosphate to the water. This
caused the copper content of initial morning water samples to drop from a
consistent average 5 ppm to an average of 2 ±0.2 ppm (samples from 8 differ-
ent days). This relieved complaints of "blue water" and morning vomiting.
Fluoride in the concentrations added to many domestic water supplies
(up to 1 ppm) has no effect on the amount and rate of corrosion in cooper
distribution systems (62, 63). Sulfide is discussed in the section on iron.
Localized Corrosion of Copper—
In this report, localized corrosion of copper will include pitting,
impingement, and under-deposit forms of attack. Pitting is probably the
most common of these and will be discussed at some length, followed by a
shorter discussion of impingement. Under-deposit attack, a form of crevice
corrosion, has been observed for copper, but its occurrence should be rela-
tively rare for copper in potable water systems. However, when it does occur
in hot water systems, attack can be rapid (89).
The pitting of copper is a complex phenomenon which probably has several
different causes and a number of different contributing factors. The problem
is apparently not widespread, but it has recieved considerable attention
since failure of practically new pipe or tubing can occur in a short time.
Even in a given water distribution system, the failures often occur at
random locations.
The effects of copper pitting on potable water quality are difficult to
assess. The actual area of material that is affected is usually small and
the pits themselves are often capped with corrosion product which prevents
significant loss of soluble copper to solution. On the other hand, while
active pitting is occurring, the rate of attack per unit surface area is much
higher than the usual rate of uniform corrosion. Also, pits may rupture
unpredictably from time to time, releasing small amounts of concentrated
copper ion solutions. Copper in solution can promote the corrosion of other
metals, as discussed elsewhere in this report.
Causes of Pitting--
Much of the difficulty in characterizing copper pitting is in disting-
uishing between factors which initiate pitting and those which are necessary
to sustain it. Possible synergism between these factors also complicates
matters.
There are several elaborate theories on the growth and structure of
corrosion pits on copper (17, 26, 64, 80). These models do not appear to be
particularly helpful in determining either the probability of pit initiation
or the tendency of a given water to support pitting.
4-38
-------
Campbell has distinguished between two forms of copper pitting (13).
The first form, called "soft-water pitting," occurs only in certain soft
waters containing manganese. It usually is restricted to the hottest parts
of hot water systems and is associated with formation of a scale of manganese
oxides which forms an unfavorable galvanic couple with any exposed copper
surface. The second type, hard-water pitting, occurs in hard or moderately
hard waters and is nearly always restricted to cold water pipes. Campbell
states that this type of pitting is usually associated with the presence of
either a carbon film or a "glassy" copper oxide scale formed when certain
abnormal conditions prevail during the annealing process. These films are
cathodic to the copper surface and either one can provide a galvanic driv-
ing force which induces pitting.
The second form of pitting is prevalent in England and has promoted
considerable study (17, 81). Its occurrence is apparently prevented by a
naturally occurring organic inhibitor which is present in many surface
waters, but not in well waters. The inhibitor has not been isolated, but it
is removed by activated charcoal and has been further characterized (13).
The presence of carbon films on copper pipes manufactured in the United
States is said to be rare (43, 111). At least one possible case of pitting
due to a carbon film has been reported (111). Evidence has been presented
for pitting due to a "glassy" copper oxide scale in U.S. Service lines by
Cruse and Pomeroy (20). These authors examined over 65 pipe specimens.
Although carbon films were present, the correlation with tendency to pit was
less for carbon content than for the presence of glassy cuprite scales. They
conclude that well water containing dissolved oxygen, relatively high miner-
alization, and a pH below 7.5 is conducive to the rapid pitting of copper,
but that such pitting occurs only where the copper surface is sensitized by
relatively heavy glassy cuprite scales, carbon residues, or perhaps other
deposits or scratches.
From a survey conducted in the United States, pitting of copper tubing
has almost invariably been associated with cold, hard well waters, according
to Coher and Lynam of the Copper Development Association (15). They state
that a typical aggressive well water contains greater than 5 ppm dissolved
C02, dissolved oxygen up to 10 to 12 ppm (which may come from storage and
handling of the well water), and chlorides and sulfates. Their statistical
survey also shows that pitting failures are almost evenly distributed between
soft or annealed temper and hard drawn tube. They state that this refutes
hypotheses based on pitting due to surface conditions of the tube. They also
state that pitting can be prevented by treatment of the water to neutralize
the dissolved carbon dioxide. A case history discussing the effects of 02 and
C02 is presented in Section 6.
4-39
-------
Impingement Attack and Flow Rate Effects—
Copper is more susceptible than most engineering metals to a flow-
velocity dependent type of corrosion generally termed impingement attack
(42)- This is usually a localized attack caused by excessive liquid flow
velocities and aggravated by the presence of entrained solids or gas bubbles.
The resulting metal pits are undercut on the downstream ends and are fre-
quently horseshoe shaped.
The rate of attack depends to some extent on water composition. The
rate increases with increasing oxygen content and chloride content. Impinge-
ment attack rate increases with decreasing pH. Prevention can be obtained
by limiting flow velocity to 5 feet per second for most municipal waters and
to considerably lower values if entrained solids or air bubbles are present
(43).
Copper Alloys—
A wide range of copper alloys are used in potable water systems, partic-
ularly as valve parts or other components where their mechanical properties
are desirable. These alloys are divided here into the general classes of
brasses, bronzes, and copper-nickel alloys and each is discussed below. The
discussion of copper corrosion can be used as a basis for the behavior of the
alloys, but their corrosion resistance also depends on alloy composition.
One major difference compared to pure copper is that selective leaching is a
predominant mode of corrosion for some common copper alloys. This is a cor-
rosion process whereby one consituent of an alloy is removed from the metal,
leaving an altered residual structure. A common form of selective leaching
is dezincification of brasses. The fundamental mechanisms of selective
leaching are a subject of disagreement (44). One view is that the entire
alloy dissolves and then one of its components is replated from solution.
Another group contends that one component is selectively dissolved from the
alloy leaving the porous residue of the more noble species. Others believe
that both modes of corrosion occur.
Corrosion of Brasses--
The common brasses are alloys of copper with 10 to 50 percent zinc. A
number of other elements may be added either singly or in combination.
These elements are listed in Table 15, but not all of these are commonly
added to commercial Cu-Zn alloys (]QSI- Hundreds of modifications of the
brasses are known. Zinc dissolves in copper up to 39 percent to give a
single phase alloy, a brass. Another single phase alloy, 3 brass, is formed
with 47-50 percent zinc. At intermediate zinc levels, the alloy contains
both phases, a and 3 brass (11).
4-40
-------
TABLE 15. RANGES OF COMPOSITION OF Cu-Zn ALLOYS (l08)
Percentage
Lead
Aluminum
Tin
Nickel
Iron
Silicon
Manganese
Phosphorus
Arsenic
Antimony
Gold
Bismuth
Vanadium
Tungsten
Chromium
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.1
.1
.5
.5
.1
.1
.05
.01
.01
.01
.5
.1
.1
.1
.05
to
to
to
to
to
to
to
to
to
to
to
to
to
to
to
12
3.
6.
10
2.
2.
5
0.
1.
0.
1.
3.
0.
2
0.
0
0
(sometimes
0
0
(sometimes
10
0
1
0
0
5
5
up to 30)
up to 25)
Many brasses corrode in major part by dezincification. Two general
types of dezincification are recognized, the layer type and plug type. Layer
type attack occurs fairly uniformly along the surface while plug type attack
is localized and penetration occurs perpendicularly into the metal. Brasses
generally resist impingement attack better than pure copper (107). They are
susceptible to stress corrosion cracking under certain conditions but this is
not discussed here. Pitting can occur under some conditions but this is gen-
erally similar to copper and less important than dezincification. Generally,
the brasses show very good resistance to most types of unpolluted waters,
with corrosion rates averaging 0.1 to 1 mpy, in the absence of dezincifica-
tion (108). Soft waters containing high C02 can cause higher rates, often
accompanied by dezincification.
Dezincification is also generally favored by the following conditions
(62, 79, 107):
elevated temperatures,
stagnant solutions, especially if acid,
porous inorganic scale formation and crevices,
4-41
-------
residual stresses and local deformation, and
chlorides and copper ion buildup.
The composition of the alloy is also important. Brasses containing 85 per-
cent copper or more (red brasses) are generally resistant to dezincification.
Additions of iron or manganese to brass tend to accelerate dezincification
while additions of low concentrations of arsenic, antimony, phosphorus, bis-
muih, and tin have been used to reduce the rate of dezincification (62).
Thus, for example, Muntz metal, common yellow brass, and noninhibited alumi-
num brass are considerably less resistant than arsenical Admiralty brass and
arsenical aluminum brass (108). Layer-type dezincification tends to occur
more frequently for high zinc brasses and acidic environments. Plug type
attack seems to occur more often for low zinc brasses and neutral, alkaline,
or slightly acidic conditions (32). Many exceptions to these general state-
ments can occur, however.
The influence of dezincification on water quality was demonstrated by
extensive early work of Clark (14). Unfortunately, the pipe material was
only designated as "brass"; it was probably yellow brass ( 67-33 Cu-Zn) or
similar high zinc alloy. Water from several northeastern water supplies was
allowed to stand in, or was passed through lengths of new pipe. In general,
about the same amount of copper was dissolved from either pure copper or
brass pipe. Brass pipes, however, yield much more dissolved zinc than copper
(14).
Detailed observations of the corrosion of yellow brass and Muntz metal
pipes in domestic hot water systems using several municipal water supplies
(78). After 20 to 25 years service, the rate of uniform corrosion was low,
but most samples exhibited local corrosion. Corrosion proceeded in two
distinct phases, dezincification and then final corrosion of the copper
formed. The lag between the two phases varied considerably. In a/g brass
pipes the 3 phase was always attacked first. Initiation of the localized
corrosion was often associated with signs of residual stresses. Evidence
was found that the copper deposits were residual rather than formed by the
redeposition mechanism (78).
A recent study of valve stem brass corrosion in hot and cold potable
water at eleven cities in the United States has been reported (18). Commonly
used silicon red brasses (Cu-Zn-Si two phase alloys) were tested for one
year, primarily in cities where corrosion had been noted. Dezincification of
the uninhibited materials was severe and widespread. Addition of 0.03 to
0.06 percent arsenic prevented dezincification in both hot and cold water
while similar phosphorous levels were effective only in cold water. No
correlation could be found between water composition and dezincification for
the cities studied. Temperature was a most important external variable,
greatest attack occurring in hot water services (18).
Corrosion of Bronzes--
Originally, the term bronze was used for alloys of tin in copper (here
called tin bronzes), but it is now generally applied to casting alloys based
on copper whether or not tin is present. The tin bronzes are essentially
4-42
-------
solid solutions of tin in copper. The most common wrought forms contain 1-10
percent tin. Alloys with more than 8-10 percent tin are usually used in the
Cu- u 5rmi There are a number of modifications of Cu-Sn alloys, most of
which deal with variations in the concentrations of tin, zinc, lead, and
phosphorous. About twelve other elements have been added singly or in com-
bination, although seldom more than four are added at one time. A summary
of compositions is given in Table 16 (IQS). Additions of iron, antimony and
Dismuth are said to be dangerous and are tolerated only up to 0.2 to 0.5 per-
cent Ul). Aluminum bronzes generally contain up to 9-10 percent aluminum
as the important minor constituents in copper and sometimes small additions
of manganese and copper. Silicon bronzes contain up to 4.5 percent silicon
and minor additions of manganese, zinc, iron, or tin (62).
TABLE 16. RANGES OF VARIOUS COMPOSITIONS OF Cu-Sn ALLOYS (.108)
Copper
Tin
Zinc
Lead
Phosphorus
Cadmium
Nickel
Iron
Silicon
Aluminum
Arsenic
Antimony
Cobalt
Platinum
Tungsten
Manganese
Bismuth
Percentage
60.0 - 99.5
0.5 - 35.0
0.5 - 15.0 (sometimes up to 30)
•0.5 - 15.0
0.01 - 3.0
0.5 - 1.0
0.10 - 15.0
0.05 - 4.0
0.05 - 2.0
0.5 - 2.5
0.5 - 2.0
0.1 - 8.5
5.0~
10.0
10.0 > not common
3.0
0.5
4-43
-------
There is relatively little information on the corrosion behavior of the
bronzes in domestic fresh waters. Corrosion considerations for the tin
bronzes are generally similar to those for copper. The attack by water
depends on oxygen, carbon dioxide, and dissolved salt content and the forma-
tion of protective layers. Selective leaching of tin (destannification) has
been noted, but apparently occurs only under relatively extreme conditions
such as in superheated steam or for pump impellers handling hot feed water
(11). Alloys containing more than 5 percent tin are especially resistant
to corrosion by impingement (107).
Aluminum bronzes are said to show generally superior corrosion resis-
tance to other common copper alloys in sea water service (11). However,
selective leaching of aluminum can occur under ill-defined conditions. These
alloys are normally more resistant than copper and most brasses to erosion
corrosion or impingement attack. Several aluminum bronze samples exposed
for 12 years to stagnant New York City Reservoir water showed rather exten-
sive pitting with some cracking (37). Several tin bronzes exposed under the
same conditions showed essentially no pitting and average corrosion rates
less than 0.02 mpy. Extensive attack was noted for K manganese "bronze"
(65% Cu 22% Zn 2% Fe 6% Al 4% Mn). Additional results for a number of copper
alloys are given in References 36 and 37.
Silicon bronzes are used where high strength is required along with
corrosion resistance comparable to copper. In many environments the corro-
sion rates are about the same as pure copper (62). The silicon bronzes are
more resistant to acid attack and corrosion resistance increases with silicon
content.
Other Copper Alloys--
Alloys based on copper and containing 5 to 40 percent nickel have gener-
ally excellent corrosion resistance in sea water, brackish water, and fresh
water. These copper-nickel alloys are often used in heat exchangers for sea
water or for other brackish water applications where ordinary copper alloys
do not perform well. The 70/30 Cu/Ni and 90/10 Cu/Ni alloys are most often
found, with small additions of iron and manganese to improve corrosion resis-
tance under flow conditions (11, 62). Limited reference has been found to
the use of these materials in domestic potable water systems. Extended test-
ing of 90/10 and 70/30 cupronickels with 1 percent iron in New York City
Reservoir water showed superior resistance to pitting corrosion. Weight loss
general corrosion rates were reported 0.2 to 0.4 mpy (36) and 0.01 mpy (37).
Several nickel-copper alloy (Monels) specimens showed significant pitting
under the same conditions. Selective leading of nickel (denickelification)
form copper-nickel alloys can occur under special circumstances, but it is
relatively rare (11).
CORROSION OF LEAD IN THE WATER WORKS INDUSTRY
Because of its favorable structural characteristics, lead has been used
for transporting and distributing potable waters in the U.S. since the
nineteenth century. From an engineering standpoint, lead is highly resistant
to corrosion and attack by natural waters. It is also, however, an active
4-44
-------
accumulative toxicant, and the distribution of corrosive water through lead
pipe can constitute a serious health hazard to the consumer. The current
standard for lead, as established by the EPA, is 0.05 mg/£ in drinking water.
There are indications, however, that this standard may not provide a suf-
ficient margin of safety for the fetus and children under three years old.
The National Academy of Sciences has concluded that a lead standard greater
than 0.025 mg/a cannot assure a no-observed-adverse-health-effect status (50)-
Lead corrosion and high lead levels in potable waters have been identified
in several municipalities throughout the nation from systematic surveys and
sampling procedures. However, because of the expense and difficulty in im-
plementing a comprehensive sampling program, possible lead contamination in
many other communities has yet to be identified.
The most common use of lead in the water works industry is for lead
service and residential pipes and for lead-based solders. Lead is also widely
used as a pipe lining for zinc galvanized iron pipe to enhance durability and
extend the useful life of the pipe. Lead is used for "goosenecks" in smaller
piping systems to prevent undue stress on water mains, but some utilities
have discontinued this use in favor of less potentially harmful materials.
Lead and lead-lined galvanized pipe have a useful life expentancy of
35 to 50 years and longer. Consequently, many of the lead service lines
installed are still in operation. Many of the older municipalities in the
nation have a large number of lead service lines ranging in length from 30
to 100 feet, which connects the street main to the household plumbing.
The most commonly used lead-based solders are composed of 50 percent tin
and 50 percent lead or 60 percent tin and 40 percent lead. During installa-
tion, this solder may flow inside the pipes at the joint, thereby providing
a lead-based surface area exposed to potentially corrosive water.
A secondary use of lead in the water works industry is for lead gaskets
used as flanges for joining large valves and pipes in water treatment plants
and on water mains. However, exposed lead surface-areas are relatively small
and water contact time is usually short minimizing the potential for addi-
tional lead contamination.
Lead is also used for the production of brass and bronze. Brass is a
copper-zinc alloy which contains up to 12 percent lead, and bronze is a
copper-tin alloy which contains up to 15 percent lead. These materials are
corrosion resistant and are not suspected of contributing significantly to
lead contamination of potable waters. However, test results indicate some
dezincification in some brass, and it is reasonable to conclude that lead may
also be leached into the water system.
4-45
-------
Little information has been compiled which identifies or qualifies the
extent to which lead is used in the water works industry throughout the U.S.
The most comprehensive survey was completed by Donaldson and the results were
published in 1924 (24). His study included a survey of more than 500 water
distribution systems in 41 states and concluded that approximately 48 per-
cent contained lead lines. In the Boston metropolitan area, Worth estimated
that approximately 60 percent of the residences were serviced with lead pipes
(114). In a report published by EPA, it is estimated that approximately
1300 of 3300 service lines in Bennington, Vermont consist of lead (83)-
Bennington, Vermont is considered a typical New England community. Patterson
is currently conducting a survey of the quality of potable water at 1000
consumer's taps across the nation, and because the lead content of tap water
is usually proportional to the amount of lead in the plumbing system, this
study may indicate the extent of lead use in potable waters (76, 92).
Specific areas within a given city may contain greater percentages of lead
service lines, and these areas may be identified by a historical review of
plumbing codes and pipe sales correlated to the time of development of the
area.
In some municipalities where lead corrosion and potable water contamina-
tion have been documented, such as in Boston, Massachusetts, local or state
plumbing codes restrict the use of lead for soldering only. However, several
widely adopted codes, including the Uniform Plumbing Code and the Building
Officials Conference of America (BOCA) Code, currently allow the use of lead
pipe for transporting and distributing potable waters. Additionally, some
municipalities assume responsibility only for water mains and reserve the
selection and installation of water service lines to the home owners or
builders. Consequently, lead service lines are still being installed and
lead solders are widely used.
The primary water quality parameters related to lead corrosion and cor-
rosion rates are hardness, alkalinity, pH, total dissolved solids, dissolved
oxygen, and carbon dioxide. At least one investigator also attempted to
identify the effects of chlorine content on the corrosive nature of potable
waters. Although much research has been presented relating these parameters
to lead corrosion, other investigators maintain that the actual contribu-
ting factors are not "hardness" but hardness in association with tne corres-
ponding anionic component; not alkalinity but the "inorganic carbonate con-
centration"; and not TDS but possibly specific components or the effect of the
ionic strength of the water (34). Much of the literature reported here is
written in terms of the traditional parameters (hardness, alkalinity, TDS),
however, the factors noted above should be kept in mind when reviewing the
corrosion literature. Nevertheless, research should continue to determine
the specific mechanisms that contribute to the reported correlations and con-
tradictions. Physical characteristics of the water system such as water
velocity also influence lead corrosion. The length of time the lead pipe or
material has been in service will also affect the corrosion rate. New lead
pipe or materials are more susceptible to corrosion than older or used
materials.
4-46
-------
The occurrence of lead corrosion in potable water supplies is most prev-
alent among utilities which distribute corrosive or "aggressive" waters
through lead pipes. In general, soft waters containing dissolved oxygen,
carbon dioxide, and organic acids are corrosive to lead. However, many
investigations completed to date have addressed the corrosive effects of a
combination of parameters rather than a single parameter because of the
interactive effects. Thus, in reviewing the literature, the reader should
reflect on (1) how the parameter being correlated to corrosion reflects the
actual corrosive mechanism, and (2) whether this mechanism, as reflected by
the parameter, stands alone or is influenced by other factors inherent in
the investigation.
Effect of Flow Rate and Volume of Hater Flushed—
Because corrosion is a rate process, lead concentrations in water
exposed to lead surfaces will generally reach higher levels in standing water
than in running water and, consequently, a range of concentrations can be
expected from a given sampling point. In a survey of homes in Worcester,
Massachusetts completed in 1975 by O'Brien, nine pairs of water samples were
collected to compare lead concentration in standing water and running water
(74). The results of that survey are provided in Table 17 and show that
lead concentrations up to 1.90 mg/£ were observed in standing water while no
lead concentrations were observed in running samples.
TABLE 17. RESULTS OF WORCESTER LEAD SAMPLING ANALYSIS PROGRAM
(Lead Concentration, mg/a) (74)
LOCATION STANDING
A 0.05
B 0.08
C 0.10
D 0.06
E 0.04
F 0.00
G 0.06
H 0.17
I 1.90
RUNNING
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Wong and Berrang investigated potential sources of lead contamination
from water supply facilities using lead service pipes and lead-based solder-
ing to join copper pipes (113). In their investigation, they developed cor-
relations between volume of water flushed through the sytem and observed lead
concentrations for water supply facilities which were used only on occasion.
The results of their study are shown in Tables 18 and 19. This study
showed a decrease in lead concentration with increasing volume of water
flushed. It is important to note that very high lead concentrations were
observed from the facilities which were not used for an extended period of
time.
4-47
-------
TABLE 18. LEAD IN WATER TAKEN FROM AN OCCASIONALLY
USED TAP 24 HOURS AFTER LAST FLUSHING (113 )
Volume of Water Flushed
Through Tap
Lead Concentration
(ppb)
0.005
0.020
0.060
0.125
0.310
0.615
1.220
2.425
300.000
130
130
240
410
330
60
16
17
7
TABLE 19. LEAD IN WATER TAKEN FROM A TAP NOT USED
FOR ABOUT SIX MONTHS (113)
Volume of Water Flushed
Through Tap
Lead Concentration
(ppb)
0.005
0.025
0.055
0.105
0.210
0.415
0.920
1.920
6.920
200.000
300
3,000
2,300
2,100
2,500
490
190
64
15
12
4-48
-------
Effects of Dissolved Oxygen
Slunder and Boyd (95) discuss results reported by other investigators
concerning the influence of dissolved oxygen on lead corrosion. They in-
dicate that while some authors maintain that lead is insoluble in air-free
pure water, others claim that lead is noticeably soluble in pure water free
from gases. Despite these discrepancies, there is general agreement that
dissolved oxygen increases the corrosion of lead. Slunder and Boyd present
the results of Burns' work (Figure 15) showing that lead corrosion in distilled
water is directly proportional to the partial pressure of the oxygen in the
atmosphere above the water. An oxygen-nitrogen mixture was used, the test
waters were saturated with the mixture, and an adequate pressure was main-
tained over the water surface during the test.
Slunder and Boyd, also stated that lead in a carbon dioxide-free water
is strongly corroded because a protective film cannot form on the lead sur-
face. In the reported experiment, the lead surface became coated with small
crystals of lead oxide and hydroxide, the water became turbid, and lead con-
centrations rose to as much as 100 mg/£. Specifics of this investigation
were not presented.
00
cr
in
to
in
O
I
60
:0 20 40
oxygen in air above water,
I
80
100
Figure 15. Effect of Oxygen on Corrosion of
Lead Submerged in Distilled Water
at 75°F (95).
4-49
-------
Effect of Hardness—
M. R. Moore of the University of Glasgow attempted to develop a rela-
tionship between the rate of lead corrosion and water hardness (68). In his
investigation, he synthesized hard water by the addition of calcium chloride
to distilled water and measured the rate of dissolution. The results of this
study are shown in Figure 16. As is shown, he determined that the dissolu-
tion rate decreased exponentially with hardness. However, the specifics of
this evaluation were not presented, and it has been suggested that the in-
creased Ca2+ and Cl" contents might have altered the C02 absorption tenden-
cy, the pH, and hence the lead solvency (34).
1200
§ 1000,
en
c 800
600-
400-
200
OJ
c
o
c
o
TD
O)
o7iiToTO100
2+ 1
Ca Concentration (mg/£ )
Figure 16. Effects of synthetically hard water on lead corrosion (68).
Others imply that the anions associated with the hardness component are
the important factor. In a summary report, SI under and Boyd explain that
most natural waters contain some hardness components which will react with
lead to form adherent films, such as calcium carbonate, on lead surfaces
which will be protective and prevent further corrosion (95.) According to
Slunder and Boyd, a water hardness of 125 ppm as CaC03 is sufficient to form
the protective film and prevent corrosion.
To minimize possible anionic interferences, Naylor and Dague (71)
used calcium nitrate and magnesium chloride to produce desired hardness
levels. They determined that, in general, variations in the amount of hard-
ness cations had only small effects on lead solubility at the experimentalIv
maintained pH of 10.5. y
4-50
-------
In soft, aerated waters, corrosion and corrosion rates are dependent on
water softness and dissolved oxygen. In general, the softer the water and
the higher the dissolved oxygen concentration the greater the corrosion.
Additionally, the presence of organic acids whose lead salts are soluble
promotes corrosion. Water containing carbonic acids that dissolve calcium
deposits will encourage corrosion by forming soluble calcium bicarbonate
according to the reaction (95).
CaC0
H2C03 Z
Ca(HC03)
3J2
Effects of pH—
In another experiment, Moore investigated the effects of pH on lead dis-
solution from a lead pipe (68). In this experiment, water was allowed to
stand in^a lead pipe section for one hour with pH, adjusted by HC1 or NaOH
addition, being measured both before and after this time period.
The results of this study are shown in Figure 17 and indicate that the
rate of dissolution in distilled water increases considerably on both sides
of the pH range from six to eight with a minimum of approximately 1000 pg
Pb/1 Her/hour near pH 6.5.
s_ 4000 -
3
O
£. 3000
- 2000 -
c
o
b 1000
o
2 46 8 10 12
Figure 17. pH effects on lead corrosion (68),
4-51
-------
Naylor and Dague (71) indicate that in a solution of pH 8 or less,
lead ions will predominate. In the pH range of 8-11, lead precipitates as
the lead oxide:
Pb2+ + 20H" = PbO I +H20
However at a pH > 11, this oxide will dissolve according to:
PbO + 2H20 = Pb(OH)3" + H+
Their experiments on lead control by conventional lime and lime-soda ash water
treatment methods produced the lead solubility curve presented in Figure 18
Between pH 9.2 and 10.4, the lead levels were generally < 0.05 mg/n although
lead had been added at a rate of 2 mg/£ prior to pH control.
When reporting on the occurrence of lead in river systems, Hem and Durum
* V* P??duced So1uble lead-pH diagrams with respect to several concentrations
of total dissolved carbon dioxide species. Their data indicated that the
solubility of lead should be lower than 10 ug/£ above pH 8.0, regardless of
the alkalinity of the water. However, at a pH near 6.5, and in water with
rnNlH ™n1Jy °?nS tf]an 3° mg/£ as HC°3~) the soluble lead concentration
could range from 40 ug/a up to several hundred micrograms per liter
1.2
^ 1-0
<=<
J! 0.8
0.6
§ 0.4H
0.2.1
O)
0.0
8.0
10.0
DH
12.0
Figure 18. Effect of pH on Lead
Solubility (71)
4-52
-------
Effects of pH and Hardness--
Several studies have been completed which correlate lead corrosion with
soft acidic waters. Karalekas et al monitored lead concentrations in deliv-
ered potable waters at five municipalities and correlated the results with
raw and finished water quality (50). Table 20 is a summary of the results of
their study. Lead concentrations were monitored for water standing in the
interior household plumbing overnight, for water standing in the service
line, and for water from the main. These samples are identified in Table 20
as Samples 1, 2, and 3, respectively. Lead concentrations in finished water
were below the detectable limit of 0.005 mg/a for all municipalities
surveyed.
As can be observed, the water systems at Bridgeport, Connecticut and
Providence, Rhode Island which have higher pH values and hardness concentra-
tions experienced the least corrosion. Although the water supply at Mew Bed-
ford, Maine has a pH comparable to Bridgeport, the hardness concentration is
lower and the average lead concentrations are considerably higher. In the
case at Providence, Rhode Island, indications of lead corrosion are nearly
eliminated by maintaining a pH of 10.1 and a hardness concentration of 40
mg/£. In this study, no differentiation was made between particulate lead
(detached from the pipe or an un-adhering fresh precipitate) and lead truly
in solution (34).
In another study by Karalekas et al, the effects of water quality,
primarily hardness, on lead corrosion were investigated for the cities of
Boston, Cambridge, and Somerville, Massachusetts (49). These cities were
selected for the investigation because many of the homes in these three
cities are known to have lead or lead-lined water service pipes. Boston and
Somerville obtain water from the same source. In this investigation, both
running and standing water samples were collected. Characteristics of the
finished water supplied are shown in Table 21.
TABLE 21. CHARACTERISTICS OF FINISHED WATER SUPPLIED TO
CAMBRIDGE, BOSTON, AND SOMERVILLE (49)
Parameter
pH
Total Dissolved Solids (mg/£)
Chloride (mg/£)
Hardness
Cambridge
6.9-8.0
170
50
56
Boston &
Somerville
6.0-7.0
30
7
14
Water supplied to residences in Cambridge is higher in pH and hardness
concentrations than water supplied to Boston and Somerville. As expected,
the percentage of samples exceeding the lead standard of 0.05 mq/a was higher
in Boston and Somerville than in Cambridge as shown in Figure 19. Addition-
ally, the percentage of households having detectable levels of lead was
higher in Boston and Somerville than in Cambridge as shown in Figure 20.
4-53
-------
TABLE 20. RESULTS OF INVESTIGATION OF WATER QUALITY ON LEAD CORROSION (50)
en
Finished Water Quality
Municipality
Bridgeport, CT
Marlborough, MA
Chatham, MA
New Bedford, MA
Providence, RI
Hardness
pH (mg/£)
7.1 48
6.5 14
6.3 20
7.3 12
10.1 40
Alkalinity
(mg/£)
18
6
3
24
20
Average Lead
Concentration Observed*
mg/Jl
0.010
0.014
0.017
0.076
<0.005
mg/£
0.011
0.037
0.018
0.090
0.006
mg/£
<0.005
0.010
0.015
0.013
<0.005
Highest Lead
Concentration
Observed mg/£
0.04
0.250
0.098
0.260
0.050
*SAMPLING INSTRUCTIONS PROVIDED:
After 11:00 p.m., do not use the kitchen cold water faucet until collecting the water samples
the next morning. Using the following directions, in the morning, collect the water samples
at that faucet before using any faucet or flushing any toilets in the house. Fill the provided
containers to one inch below the top and put the caps on tightly.
SAMPLE 1. Open the cold water faucet and immediately fill bottle #1 and turn off the
water, recap this bottle.
SAMPLE 2. Turn the faucet on and place your hand under the running water, and immediately
upon noticing that the water turns colder, fill bottle #2. Cap bottle #2.
SAMPLE 3. Allow the water to run for three additional minutes and then fill bottle #3.
Cap bottle #3.
-------
19.0
17.8
7.6
Cambridge Boston Somerville
Figure 19. Percentage of samples exceeding lead standard (49)
30.1
25.5
14.5
Cambridge Boston
Somerville
Figure 20. Percentage of households exceeding
lead standard in one or more samples (49).
4-55
-------
Effects of Alkalinity--
Moderate carbonate alkalinity concentrations have been found to be bene-
ficial in controlling lead corrosion. The presence of this alkalinity will
encourage the formation of a very insoluble lead carbonate salt film on the
corroding lead surface. This film will adhere to the lead surface and form
a protection from the corrosive environment as well as limit lead solubility
(77).
In a sampling program to determine the extent of lead in potable water
in Boston, O'Brien found that approximately 29.6 percent of the samples ana-
lyzed had lead concentrations in excess of 0.05 mg/£ (74). O'Brien concludes
that the naturally low alkalinity of approximately 4.0 mg/£ as CaC03 is
responsible for this high lead concentration occurrence. He also points out
that the alkalinity is further depressed by an additional 2.6 mg/£ with
the addition of hydrofluorosilicic acid to provide 1.0 mg/a of fluoride ion.
However, a review of the analysis of water resource characteristics for this
study indicate that the water is very soft with a calcium and magnesium
concentration of only 2.8 and 1.0 mq/i, respectively. Additionally, the
water is slightly acidic with a pH of approximately 6.3. Therefore, this
source of water is highly corrosive and it is difficult to specifically iden-
tify low alkalinity as the primary cause of high lead levels in Boston's
potable water.
The chemical constituency of the protective film is a function of the
dissolved species, including hydroxyl, concentrations. Computer generated
Eh-pH diagrams presented by Schock indicate that a lead-hydroxy-carbonate,
Pb(OH)2(C03)2, may be the predominant solid (91). The tendency to form this
hydroxyl-carbonate decreases as the dissolved Pb(II) species concentrations
decrease from 10~6-0M to 10~6-62M or as the total carbonate concentration in-
creases from CT = 10-3-7M to 10~2-7M. Solid PbC03 is reported to form only
under low pH, high [Pb++] concentrations.
In a recent investigation, Thibeau et al, used Raman and infrared spec-
troscopy to analyze the surface film composition of potentiostatically oxi-
dized lead samples in 0.1 M sulfate solutions (loo). The results of these
analyses were then compared with the composition predicted by the calculated
Eh-pH diagram. Only partial agreement existed between the actual observa-
tions and the predictions based on thermodynamic equilibria. The major dis-
crepancies were that PbO, although not predicted by the diagram, was found
in neutral and basic solutions, and PbO-PbSOit, the compound predicted to be
stable under such conditions, was not. Thus, while Eh-pH diagrams may be an
aid in interpreting free energy data, their application to the real world
problem of corrosion is limited by (1) their assumption of equilibrium condi-
tions, and (2) their disregard for the effects of other species present
For instance, the presence of ions that form soluble lead salts, such as
nitrates, will interfere v
-------
While moderate levels of carbonate alkalinity have been shown to help
control lead corrosion, excessive concentrations may, in fact, enhance corro-
sion. Schaut attempted to duplicate actual distribution conditions and
study the corrosive action of various water quality parameters on lead pipe.
In one test, he exposed new lead pipe to waters with various alkalinity con-
centrations over various contact periods. Results of this test indicated
that for short contact periods, i ,e., less than three hours, alkalinity con-
centrations have little or no effect on lead corrosion. However, for contact
periods of 24 to 48 hours using cool water ranging from 35 to 40°F, he deter-
mined that lead concentrations in the water were almost doubled as alkalinity
concentrations were doubled. Using water at a higher temperature, the
effects of increased alkalinity concentrations were not as pronounced. Tests
on old lead pipes, which were approximately 35 years old, gave slightly
different results and the effects of increased alkalinity were not as obvi-
ous. However, increased corrosion rates were observed with increased alka-
linity concentrations. Schaut explains that this difference is due to the
formation of a basic lead carbonate which is relatively passive to the range
of alkalinity used or observed in drinking water. Schaut did not report the
values of alkalinity concentrations which he used (90).
In the investigations of lead removal by conventional water treatment
methods Naylor and Dague (71) maintained a 200 mg/£ alkalinity (as CaC03)
and varied the pH of their water. As expected, lead concentrations of < 0.1
mg/SL were found up to a pH of 8.6 in solutions initially prepared with
2 mg/£ lead. At higher pH, and contrary to expectations resulting from
their previous work on pH effects (see Figure 18), the lead levels rose to
0.9-1.1 mg/£. Because their investigations included various water treatment
unit operations, Naylor and Dague felt that physical, rather than chemical,
parameters were responsible for this increase in lead solubility.
Current research at the EPA, Municipal Environmental Research Labora-
toreis (MERL) in Cincinnati shows similar results with respect to alkalinity
concentrations (35). Studies are being conducted to evaluate the effects
several water quality parameters have on lead corrosion. Dr. Marvin Gardels
is currently examining changes in alkalinity and changes in pH for water
treatment to minimize corrosion. Results to date indicate that corrosion
may be enhanced with increasing alkalinity by carbonate addition above
approximately 40 to 50 mg/£ as CaC03. Gardels has found that increased pH
with a depressed carbonate concentration produces the least aggressive
waters. From his research, it appears that an optimum combination of pH and
carbonate concentration probably exists for maximum protection against corro-
sion. Initial results indicate, however, that the optimum pH values may be
slightly above 9.0.
4-57
-------
On the other hand, Schock presented computer generated activity ratio
diagrams (constant Eh) for the total lead (II) soluble species in relation
to pH, and dissolved lead (II) versus total carbonate concentrations for
several pH values. The results from his studies are shown in Figures 21 and
22. He substantiated the activity ratio diagrams with laboratory data and
concluded that the possibility of lead control by alkalinity-pH adjustment
was not as great as previously believed, and that in the pH range of 8-9.5,
there is little advantage to increasing the carbonate level above 30-40 mg/a
as CaC03 (91).
8r
0)
o 4
.a
a.
t-j
i
9
PH
i
11
I
Ol
O
8
13
Figure 21. Revised model (1=0.1, 25°C) for log C. = -3.6 ( );
-2.7 (— — );-!.!( ). Shown are data from precipitation
experiments of Patterson et al (1977) at initial log C+ values of -3.6
( A ); -2.7 ( o ); -1.1 ( o ) (91 ). *
4-58
-------
I
Ol
10
40
100
140 200 300
Total Carbonate, ing/1 as CaC03
400
Figure 22. Plot of total dissolved Pb(II) versus C. at fixed pH (1=0.005, 25°C).
Shown are pH = 7.0 ( n ); 8.0 ( • ); 8.5 ( A J; 9.0 ( A ); 9.5 ( o );
and 10.0 ( • ) (91).
-------
Effects of Temperature--
Results of several investigations have been reported which correlate
water temperature with lead corrosion. Using distilled water, Moore deter-
mined that the corrosion rate of lead increased exponentially with increas-
ing temperature and developed the relationship:
Pb corrosion rate = 101 exp(O.OlT)
where T is temperature (68).
in Figure 23.
Results of his experiment are shown graphically
3
o
O)
Q_
Cl
C/l
c
i.
1500-
1000 -I
20 40 60
Temperature, °C
Figure 23. Temperature effects on lead corrosion (68)
Effects of Chlorination--
Schaut attempted to describe the corrosive action of chlorination on
lead pipe. From his tests using municipal water with a pH range of 6.8 to
7.2 and an alkalinity of 35-50 mg/£ as CaC03, he concluded that in new lead
pipes, the rate of chlorine dissipation is dependent primarily on tempera-
ture. Additionally, he observed that when the chlorine was exhausted with
increasing temperatures, the water acquired a lead content which approxima-
ted the formation of PbCl2. Under all temperature conditions the lead con-
centration value for new lead pipe was 0.38 ppm. Duplicating his experiment
using old lead pipe, Schaut found that the lead concentration in the exposed
water did not reach a value of 0.38 ppm even with up to three days contact
(90). From the results of his data, Schaut concluded that chlorine contri-
butes its lead equivalent on a percentage basis about equally in old and new
lead pipe at maximum potable water temperature, at least for his experimental
eight-hour contact period. Additionally, he concluded that the time it takes
for water to acquire 0.1 mg/a lead in new lead pipes is approximately 1/4
4-60
-------
hour using warm water with a chlorine residual of 0.12 ppm. With cooler
water the time required is increased to approximately 1/2 hour (90)-
Schaut also investigated the combined effects of water temperature and
chlorine concentration and concluded that the combination of chlorine and
warm water is more corrosive than warm water alone. In his experiment,
Schaut held alkalinity and residual chlorine concentrations constant and
varied temperature. Corrosion measurements were made after an eight-hour
contact period. Results from this test showed a linear rather than exponen-
tial relationship with a doubling of the corrosion rate correlated with a
doubling of temperature. Again, Schaut did not provide numerical values.
Effects of Carbon Dioxide--
In a summary report by SI under and Boyd, results of previous research
on the corrosive effects of carbon dioxide content are discussed. Water con-
taining C02 in the absence of oxygen has little effect on the corrosion of
lead. The extent of corrosion when both C02 and oxygen, are present is con-
trolled primarily by the concentration of C02. Figure 24 is a graphical
summary of results reported by several investigators and prepared by SI under
and Boyd which shows the effects of C02 content on the corrosion of lead
(95). Unfortunately, the specifics of how the C02 was added to the water or
how the pH was maintained are not presented. As can be seen, when less than
2 mg/£ of carbon dioxide is present, corrosion proceeds linearly at an appre-
ciable rate, but at a C02 concentration of 60 mg/£, the corrosion rate is
much lower. During the tests used to develop this data, a white deposit,
probably a basic lead carbonate, was formed with the water having the higher
C02 concentration (95).
600-1
500-
400-
Dl
300
O)
-------
Lead Release from Solder Joints--
A1 though the primary source of lead in potable waters is thought to
occur from lead service lines and lead-lined galvanized pipe used in house-
hold plumbing, several studies have been completed to quantify the rate of
lead corrosion and water contamination from lead-based solders. Lyon and
Lenihan measured the magnitude of lead released from solder joints of copper
pipes and found the results to be much higher than expected (55). Their
laboratory experiment consisted of a running water test and a static water
test using deionized water. In the running water test, water was circulated
through a loop constructed of copper tubing with lead-based capillary joints.
From the results of the tests, they calculated a mean lead release of 322
yg/fitting/16 hours for the running water test and a mean lead release of
216 yg/fitting/16 hours for the static test. After four to five weeks,
a mean lead release of approximately 20 yg/fitting/16 hours was observed from
the static tests. These results were favorably compared to measurements
taken from capillary joints obtained from a five-year old building which
showed a mean lead release of 22 yg/fitting/16 hour period. Lyon and Lenihan
concluded that an initial release ranging from 200 to 300 yg/fitting/16 hours
can be expected after the first four to five weeks of operation, but that
this release will decrease to approximately 20 yg/fitting/16 hours and will
be maintained for a long period. It should be noted that these release rates
were developed using deionized water. Lyon and Lenihan also noted that the
magnitude of the pick-up experienced in the copper tubing was unexpectedly
high considering the relatively small surface area of solder exposed. From
this observation, it was concluded, through further experiments, that the
mechanism for the corrosion process results from galvanic action.
Wong and Berrang also attempted to determine the corrosion or lead pick-
up rate of lead-based solders used for joining copper pipes (ns ). In their
experiment, they simulated household copper tubing using 50 feet of one-half
inch diameter copper tubing soldered together with 20 soldered joints using
50/50, 60/40, and 95/5 (tin/lead) solder, as well as silver solder. The
results of this test for various volumes of water flushed are shown in
Table 22. From their experiments, they concluded that an average dissolution
rate of 0.4 yg/joint/hour can be expected after one year of service. This
value was favorably compared to results obtained from measurements taken from
an existing system in a one-year old house. In other experiments, they
determined that old lead service pipe will experience a dissolution rate of
30-240 yg/hour and that new lead service pipes will experience a dissolution
rate of 480 yg/hour.
TABLE 22. LEAD CONCENTRATIONS (PPB) IN WATER STAGNANT FOR
ONE HOUR IN A NEW SIMULATED HOUSEHOLD COPPER PLUMBING SYSTEM
(50 Feet Copper Tubings Joined by 20 Soldered Joints) (
Volume of Water Flushed Through the System (a)
Solder 80 1,200 12,000 25,000 150,000
50/50
60/40
95/5
Silver
Copper Only
1200
1100
3
2
1
150
130
2
2
2
96
49
—
--
--
34
25
1
1
1
9
7
—
--
--
4-62
-------
In his work on heavy metals release from residential plumbling, Rossum
( 84 ') determined that the typical lead "sweat" fitting provided a clearance
of 0.002 to 0.005 inches between the outside of the pipe wall and the inside
of the fitting. Thus a £ inch nominal size type L water tube would have an
average of 2.2 square milimeters exposed solder area. Furthermore, the solder
alloy was anodic to copper by 0.3 to 0.4 volts in tap water. Rossum reported
that lead was released into tap water, regardless of the water quality, from
new household plumbing but the length of release varied. Where a calcium
carbonate film was able to deposit on the pipe, the current established by
the solder-copper galvanic cell was reduced to the extent that within a few
weeks the lead release was undetectable. When the film deposition did not
occur, the water continued to pick up lead for longer than a year. Rossum
also reported that the corrosion inhibition by formation of the film is more
effective in flowing than in standing water situations. He also noted that
lead pick up may occur from lead impurities used in zinc galvanizing or from
brass faucets (typically composed of 6% lead) that may display an exterior
of chrome but are seldom plated on the interior walls. The possibility of
lead-related health disorders caused by the use of lead solder is documented
and problems of excessive lead concentrations occurring in portions of
Carroll County, Maryland are presented as a case history in section 6.4.
CORROSION OF ALUMINUM IN THE WATER WORKS INDUSTRY
The use of aluminum is relatively new to the water works industry so its
application is presently limited. However, because its corrosive behavior
is generally good, aluminum is currently being considered for more extensive
use. Typical applications of aluminum in the water works industry include
wier gates, storage tanks, reservoir roofs and supports, hot water systems,
and water pipelines (9).
Alloys used to manufacture aluminum materials for handling fresh waters
include copper, magnesium, silicon, iron, manganese, chromium, zinc, and
titanium. Their composition is shown in Table 23. These alloys are some-
times clad with a sacrificial alloy to provide cathodic protection to the
core metal. Corrosion induced penetration of an aluminum alloy cladding
layer, anodic to the core, will spread laterally after reaching the core.
However, as the area of the core is exposed, the resistance of the electro-
lytic path is increased and penetration of the core may proceed (6).
The corrosion behavior of aluminum is generally good owing to protection
afforded by oxide or hydrated oxide films formed on the metal surface. How-
ever, corrosion of aluminum in fresh waters can be severe depending on the
composition of the water and on the conditions of service.
4-63
-------
Several investigators have reported the results of their studies of the
effects of fresh waters on the corrosion of aluminum. In general, as alumi-
num corrodes, most of the surface is usually unaffected while the attack
takes the form of small scattered pits. General corrosion or gradual uniform
thinning does not occur (38). It has also been observed that, initially,
corrosion occurs rapidly through the formation of a large number of pits, but
slows considerably in a short period'. Porter and Hadden have studied the
corrosion of aluminum and have characterized this primary typn of •-'Vision
as "nodular pitting" (79). They observed that a mound of insoluble aluminum
hydroxide forms on the surface of the exposed aluminum while an acidic liquid
builds up underneath, causing the initiation of pits. The rate of corrosion
is initially rapid but stabilizes after about two weeks as a cathodic scale
forms on the surface of the pit (79).
TABLE 23. COMPOSITIONS OF TYPICAL ALLOYS USED IN FRESH WAFERS (6)
Alloy
Cu %
Mg
Si
Fe
Mn
Zn
Ti
1C
N3
N4
N5
N6
N8
H30
H19
H20
0.04
0.10
0.04
0.04
0.03
0.04
0.03
0.03
*
*
2.07
3.60
4.94
4.25
0.72
0.65
1.01
0.23
0.18
0.27
0.17
0.19
0.14
0.87
1.04
0.57
0.25
0.32
0.27
0.40
0.15
0.24
0.27
0.27
0.39
1.18
0.36
0.30
0.31
0.70
0.70
0.02
0.12
*
*
*
*
*
0.12
*
*
0.21
*
*
*
*
*
*
*
*
0.04
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.05
'•'Not normally added but may be present at an impurity level.
The predominance of nodular pitting was later confirmed by Davies, and
Rowe and Walker (21, 85). Davies investigated the effects of various water
quality parameters on exposed aluminum specimens. During his experiments,
he observed that bubbles formed on test specimens within 2 to 4 hours after
being submerged. Next, white corrosion products developed around the bubble
and enveloped it with pit formation beginning within 18 to 24 hours following
immersion (21).
Rowe and Walker, investigating the effects of various water quality
characteristics on corrosion of aluminum, made similar observations. They
also concluded that pitting corrosion was the predominant form of corrosion
of aluminum. They observed that the gas bubbles collected on the test speci-
men soon after being submerged in the test solutions were indicative of the
pit sites. At these sites, corrosion mounds began to develop and grew in
size with time as gas bubbles continued to rise form the center of the mound.
After a longer period, the growth rate of the mounds began to decline.
Mechanically removing the corrosion product mound revealed the formation of
a pit (86).
4-64
-------
Two approaches have been utilized in aluminum corrosion investigations.
Porter and Madden used a qualitative assessment by visual examination and a
quantitative assessment by measuring loss of weight, density of pitting, and
depth of pitting with all pits on every specimen being measured (79). This
basic methodology, with slight modification, was used by others in more
recent studies. Rowe and Walker, however, studied the effects of mineral
impurities in water on the corrosion of aluminum using an electrical conduc-
tance method by passing a known quantity of current through a specimen and
measuring the voltage drop. This method does not produce an "absol'fto"
corrosion rate measurement as pits may perforate the specimen. However, a
loss of metal by pitting is reflected in the measurement and the method is
useful for comparative studies (85).
The effects of various water quality characteristics on the corrosion
and pitting of aluminum have been investigated extensively. Characteristics
which have been identified as influencing corrosion include pH and total
hardness as well as the presence of chlorides, dissolved oxygen, and metal
ions. Conditions of service which influence corrosion of aluminum primarily
include water velocity, temperature, and time of contact.
Effects of Velocity
Most investigators agree that the corrosion of aluminum occurs more
readily in slow moving or stagnant waters than in fast moving waters. In
corrosion tests completed by Wright and Godard, it was shown that, in gener-
al, as velocity increased corrosion decreased and, from actual field observa-
tions, no pitting occurred on aluminum which was exposed to a water velocity
of 7 feet/second (lib). Other laboratory tests by Godard showed similar
results (38). Aluminum specimens immersed in a stagnant water pitted
normally while specimens exposed to the same water but at a water velocity
of 8 feet/minute showed no pitting. The results of tests by Wright and
Godard using Kingston, Ontario tap water are shown in Table 24 (115).
TABLE 24. WATER VELOCITY EFFECTS ON PITTING OF ALUMINUM (115).
Control Panels in Still Water Test Panels in Moving Water
Water Velocity Avg. # of Pits Avg. Max. Pit Avg. # of Pits Avg. Max. Pit
(fpm) Per Panel Depths (y) Per Panel Depths (p)
1
2
3
4
5
6
7
8
10
95
360
126
174
119
142
347
59
85
206
156
176
227
236
156
133
155
188
244
145
26
58
26
15
50
0
0
148
107
79
90
50
35
29
0
0
4-65
-------
Excessively high velocities may, however, enhance corrosion. From field
studies conducted by Godard, it was observed that at water velocities of
approximately 20 feet/second, turbulence occurred, especially at fittings,
resulting in pitting (38).
Effects of Temperature
Godard also investigated the effects of water temperature on the inci-
dent and growth of pits on aluminum (38). The results of his investigation
are shown in Figure 25. From Figure 25 it is shown that as the temperature
rises the probability of pitting increases and the pitting rate decreases.
In other experiments, Godard measured the current flow from machined pit
specimens exposed to various water temperatures (38). The results of that
experiment are shown in Figure 26. Godard found that the current flow
reached a maximum at around 40°C and dropped off quickly as the temperature
increased. At 70°C corrosion, as indicated from the current flow, was below
that observed at room temperature. In other experiments, Godard found that
the current flow decreased linearly over the entire temperature range and
no maximum was observed. From his experiments, he concluded that at above
40°C, the service life of aluminum equipment would increase with increasing
temperature (38). Consequently, aluminum is well suited for domestic hot
water system applications.
Godard also determined that in the pitting of aluminum, the rate of
penetration follows a rapidly decreasing rate curve that approximates a cube
root function. From examination of laboratory pitting data, he concluded
that the maximum pit depths (d) were proportional to the cube root of time
(t) and he described the rate of penetration by the expression
d = Kt /3
where K is a function of the alloy and water characteristics. Actual time is
measured from the initiation of the pit. This expression has been verified
using actual field observations (33).
Water Quality Effects
As early as 1920, Seligman and Williams investigated the effects of
various combinations of chlorides, sulfates, carbonates, and bicarbonates on
the corrosive behavior of aluminum. From the results of their investigation
they concluded that pitting of aluminum is due to the simultaneous presence
of chloride and bicarbonate in water provided there is free access of oxvoen
to the system (79). J-
4-66
-------
o. 400-
4=»
-J
d>
Q.
(U
200-
I
10
I
20
I I
30 40
Temperature °C
I
50
60
70
S-
O)
.a
-100
-50
Figure 25. Temperature effects on pitting of aluminum (38).
-------
-pi
o>
00
10-
9-
8-
^T 7-
£ 6-
t
3 5H
•i->
21 4-
3-
2-
1-
0
Temperature
Figure 26. Temperature effects on pit current (38),
-------
In later experiments, Porter and Madden isolated several water quality
characteristics to determine their singular effect on corrosion of aluminum.
These characteristics included copper ion, dissolved oxygen, and hardness.
In general, they determined that in the absence of copper ions, dissolved
oxygen, and hardness nodular type pitting is prevented. They concluded that
the characteristics which are necessary for the initiation of pitting include
temporary hardness, chlorides, copper, and dissolved oxygen. It was also
found that the water composition is more influenced on the corrosion of
aluminum than is the composition of the aluminum specimen (79).
Porter and Madden also investigated the effects of these parameters on
the maintenance of pits. These tests were preformed by transferring speci-
mens to other controlled aqueous environments after pitting was initiated.
It was found that dissolved oxygen was essential for maintenance of pitting
as pitting ceased in de-aerated waters. The removal or absence of copper
ion, however, did not prevent the maintenance of pitting, but the rate of
pitting was slowed.
In an effort to establish typical pitting curves, Godard compared the
constituents of seventeen fresh waters with the resulting corrosion (38).
He concluded that no simple correlation exists and, because of the wide var-
iation in the composition of waters, it would be difficult to establish the
aggressiveness of waters to aluminum from tables alone. He did conclude from
his data, however, that hard waters are generally more aggressive to aluminum
than soft waters. The partial composition and pitting data for fresh waters
compiled by Godard are shown in Table 25 (38).
TABLE 25. PARTIAL COMPOSITIONS AND PITTING DATA FOR SEVENTEEN FRESH WATERS
(In Increasing Order of Pitting Corrosivity) (38)
Order
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Test
No.
8
7
12
6
6
5
4
16
14
1
13
15
22
24
2
9
10
Weeks to
40 Mils. Location
953
453
207
205
175
147
83
46
25
23
17
8
6
6
4.4
2.6
Shawinigan South, Que.
Shawinigan, Que.
Crofton, B.C.
Hamilton Bay, Ont.
Kingston, Ont.
Credit Valley, Ont.
Columbia River, B.C.
Canyon Meadows, Alta.
Regal Golf Course, Calgary, Alta.
N. Sask. River, Drayton Vly. , Alta.
Peterborough, Ont.
R.G.May Golf Course, Calgary, Alta.
Billingham Beck, England
Jasper, Alta.
Lethbridge, Alta.
South Saskatchewan River, Sask.
Mossbank, Sask.
Hardness
pH p. p.m.
7.1
7.4
6.7
7.1
7.9
7.1
7.5
7.9
8.1
8.1
7.5
7.9
8.7
8.2
7.9
7.6
7.8
73
18
27
205
160
0
72
169
331
267
86
218
443
195
228
206
555
Copper
p. p.m.
0.04
0.04
0.024
0.003
0.005
0.028
0.017
0.005
0.007
0.11
0.012
0.002
0.011
0.007
0.017
0.04
0.005
4-69
-------
Davies investigated the effects of sodium chloride, calcium carbonate,
and dissolved copper on the pitting of aluminum under controlled conditions
using water which he composed in the laboratory. All testing was performed
under static conditions (21).
In general, Davies observed that pitting occurs more readily in waters
containing calcium bicarbonate, chloride, dissolved oxygen, and copper salts.
To further characterize the effects of these parameters on the corrosion of
aluminum, Davies investigated both the singular effects and the combined
effects of two and three constituents.
For the one constituent test, Davies prepared solutions of 10, 30, and
50 ppm of chloride ion in the absence of other ions, and solutions of 10, 80,
and 150 ppm calcium ion, as calcium bicarbonate, in the absence of other
ions. In each solution prepared, Davies exposed aluminum specimens and ob-
served the pitting or corrosion characteristics. For waters containing
chloride ions only, a negligible attack was observed. Even after six months
of exposure, the appearance of the test speicmen had not sufficiently
changed.
In tests with water containing calcium bicarbonate only, little corro-
sion was visible. However, specimens showed a slight tarnish which became
more pronounced with increasing calcium ion concentration.
For experiments containing two constituents, Davies prepared test
solutions by combining chloride and copper ions, calcium and copper ions, and
chloride and calcium ions. He observed a slight weight loss in specimens in
tests with waters containing both chloride and copper ions. Also, he obser-
ved the formation of a few nodular type shallow pits which increased in num-
ber and depth with increasing copper content.
Results of tests using solutions containing both calcium bicarbonate and
copper ions showed a negligible weight loss in test specimens. The aluminum
surface was essentially unchanged in appearance except for a slight dulling.
When the chloride and calcium bicarbonate ions were present in the
absence of copper, Davies observed a slight weight loss in the test speci-
mens. Additionally, only slight changes in the appearance of the specimens
were observed.
In experiments where all three constituents were present, a very pro-
nounced corrosion effect in the form of nodular pitting was observed. The
results of Davies experiments are shown graphically in Figures 27 through
29. in waters where the chloride content was equal to or in excess of the
calcium ion content, there was both a general attack as well as a localized
attack. The general attack was in the form of a brown stain which was
noticeable after two weeks and became more pronounced with time.
Davies compared his results using the laboratory solutions with tests
using tap water to determine the effects of the presence of copper ions. The
results of these tests are shown in Figure 30. No weight loss was observed
with tap water which did not contain copper ions. However, when copper ions
4-70
-------
0123 456
Time of Immersion, Months
A. CT = 30 ppm, Ca++ = 10 ppm, Cu++
B. CT = 30 pom, Cu++ = 0.2 ppm
C. Ca++= 10 ppm, Cu++ = 0.2 pom
= 0.2 ppm
Figure 27. Weight loss of aluminum in various water qualities (21),
20
16 '
12 •
01
in ,
-------
24
0123456
Time of Immersion, Months
A.
B.
C.
cr =
CT =
cr =
10
10
10
ppm,
ppm,
ppm,
Ca++ =
Ca++ =
Ca++ =
10
10
10
ppm,
ppm,
ppm
Cu++ =
Cu++ =
0
0
.2
.06
ppm
ppm
Figure .29. Effect of low calcium content on weight loss of aluminum (21),
10 1
0'
gj
1/1
C/l
c
6 -
4 -
2
0
^012 34
Time of Immersion, Months
A. Tan water + Cu++ 0 0.2 nnm
B. Tap water + Cu++ @ 0.06 oom
C. Tan water only
Figure 30. Weight loss of aluminum in tap water (21)
4-72
-------
were added to the tap water, similar corrosion results occurred as with the
laboratory test solution.
Because of the significant influence of the presence of copper ions on
the corrosion of the test specimens, Davies further studied the effects of
chloride and calcium ions by varying the chloride-calcium ion ratio and hold-
ing the copper ion constant at 0.2 ppm. The results of these experiments
after submerging the specimens for two weeks are shown in Figure 31. The
results indicate that with a two-week exposure period, a maximum weight loss
is observed with a calcium ion concentrate of approximately 50 ppm. Addi-
tionally, weight losses increased with increasing chloride ion concentra-
tions. For solutions containing less than 10 ppm calcium ion concentrations,
pits were very small and not of the nodular type. At calcium ion concentra-
tions above 150 ppm, almost no pitting was observed, but the specimens were
covered with a whitish deposit. It is important to note that the above
observations were made from experimental results obtained after exposure of
specimens for a two-week period only.
CM
-§
- 6
gj
CO
to
o
20 60 100 140 180 220 260
Ca++ Content, nnm
A. Cl~ = 50 nnm, Cu++ = 0.2 nnm
B. Cl- = 30 pnm, Cu++ = 0.2 nnm
C. CT = 10 ppm, Cu++ = 0.2 nnm
Figure 31. Effect of Ca++/Cl- ratio on weight loss (21)
Bell also investigated the effects of calcium carbonate on corrosion of
aluminum in waters containing chloride and copper. The results of his tests
did not provide any evidence of a maximum in the severity of corrosion at
any particular calcium or calcium carbonate concentration as previously
reported by Davies (5). Therefore, Bell conducted additional tests to iden-
tify the apparent discrepancy.
In his experiments, Bell exposed aluminum test specimens in water con-
taining various concentrations of calcium carbonate ranging from 50 to 600
ppm. The chloride ion content was held at 50 ppm and the dissolved copper
4-73
-------
ion content was held at 0.2 ppm. The pH ranged from 6.5 to 7.4, with the
higher values being recorded at the conclusion of the tests. The results
of his experiments for various lengths of exposure are shown in Figure 32,
and the maximum pit depths observed are shown in Table 26.
24 weeks
120
400
160
200
600 CaC03 npm
240 Ca++ nnm
Figure 32. Effect of calcium carbonate on weight loss of
aluminum specimens (solutions also contained 50 ppm chloride
and 0.2 ppm copper) (5).
TABLE 26. EFFECT OF CaC03 ON MAXIMUM PIT DEPTHS (mm)
(Solutions also Contained 50 ppm Cl and 0.2 ppm Cu) (5)
Time
Material
CaCOs Ca
p. p.m. p. p.m.
10 (4)
25 (10)
50 (20)
75 (30)
125 (50)
150 (60)
200 (80)
300 (120)
400 (160)
625 (250)
2
A
< 0.02
o.n
0.18
0.24
0.24
0.17
0.19
P--
0.12
0.11
Weeks
B
0.08
< 0.02
< 0.02
< 0.02
0.24
0.22
0.39
0.30
0.22
E
6 Weeks
B
0.15
0.19
0.28
0.32
0.37
0.42
0.32
0.55
0.42
0.31
12 Weeks
B
0.13
0.18
0.26
0.41
0.44
0.48
0.47
0.61
0.42
E
24
A
0.22
0.22
0.48
0.50
0.64
E
E
0.28
E
0.48
Weeks
B
0.27
0.28
0.25
0.37
0.55
0.51
0.71
0.52
0.67
0.61
E = attack on the edges only
4-74
-------
The results of Bell's studies indicate that a maximum weight loss is
dependent on the time of exposure for various calcium carbonate concentra-
tions. For example, for specimens immersed for two weeks, the maximum weight
loss occurred in the solution containing 75 ppm calcium carbonate. For
those specimens immersed for 12 to 24 weeks, the maximum weight loss was
least for waters containing 75 ppm calcium carbonate.
From his tests, Bell observed that when the calcium carbonate content
was less than the chloride ion content, corrosion proceeded slowly with a
slight general type of attack in the first two to six weeks. After six
weeks, a film formed on the surface of the specimen and the corrosion rate
increased sharply with the formation of numerous tiny mounds of corrosion
product each with a pin point underneath. These pits increased in size with
time of exposure. In tests where the calcium carbonate and chloride ions
were approximately equal in concentrations, the rate of weight loss decreased
with time as a result of film formation and very few pits were formed.
In waters containing more than 125 ppm calcium carbonate, nodular type
pitting developed. It was also observed that with increasing calcium carbo-
nate concentrations, the corrosion product mounds increased in size but
decreased in number. Additionally, it was observed that these pits tended .to
form on the edges of the specimen with increased calcium carbonate content.
For waters containing 125 to 150 ppm calcium carbonate, Bell observed
that a thick white amorphous film containing very little calcium carbonate
was deposited on the specimen. At calcium carbonate concentrations of 625
ppm, no white film was produced and deposits between the pits consisted only
of calcium carbonate (5).
From his experiments, Bell concluded that the relative loss of weight
of aluminum specimens is dependent on the period of immersion, and for
periods greater than 12 weeks, corrosion is least in waters with approxi-
mately equal concentrations of chloride and calcium carbonate. He also
states that corrosion of aluminum is also dependent on sulfate content and
pH and that measuring chloride, calcium carbonate, and copper concentrations
is not sufficient to adequately characterize the corrosion phenomena of
aluminum (5).
In their experiments, Rowe and Walker investigated the effects of
chloride, sulfate, bicarbonate, calcium, and copper on the corrosion of
aluminum. They observed that the corrosion rate was low in aerated distilled
water. Additionally, no substantial increase in corrosion rate was observed
when either chloride, sulfate, bicarbonate, or calcium were added to the
distilled water at concentrations as high as 300 ppm, or copper ion up to
2 ppm. The addition of a combination of any two of these constituents also
did not produce a substantial increase in corrosion rate. However, the com-
bination of chloride, bicarbonate, and copper ions in the presence of air did
produce a significant corrosion rate increase (85).
With these initial results, the additional testing by Rowe and Walker
focused on the combined effects of chloride, bicarbonate, and copper. They
concluded that a near maximum contribution to corrosion of aluminum occurs at
4-75
-------
an ion concentration of approximately 300 ppm for chloride and bicarbonate
and approximately 2 ppm for copper. The results of these tests are shown
in Figure 33 v85).
It should also be noted that while copper has been cited as the most
aggressive metal to aluminum, other metal ions such as tin and nickel (79)
and mercury (38) have been found to have detrimental effects on the corrosion
of aluminum.
Davies conducted corrosion tests on anodized aluminum specimens with
oxide films ranging in thickness from 0.05 to 1.35 y (21). These specimens
were immersed in a water containing 40 ppm chloride ion, 40 ppm calcium ion,
and 0.2 ppm copper ion for a period of two weeks. This water is known to be
aggressive to aluminum and gives rise to pit formation. Some specimens were
sealed while others were not.
Results of tests on the sealed specimens indicated that corrosion rate
is decreased even for specimens anodized for only a few seconds. One speci-
men with an oxide thickness of 0.4 y, which was formed after 50 seconds of
anodizing, was found to be immune to corrosion. However, Davies states that
if tested longer, pitting would probably occur. The same results were not
observed for unsealed specimens and pitting occurred. Davies concluded that
for anodizing to be effective, the specimen must be sealed (21).
Booth et al conducted studies to obtain data in an effort to predict
the service life of aluminum, primarily pipelines exposed to fresh waters
(6). They concluded that, in general, severe pitting of aluminum pipelines,
short of perforation, will not significantly affect their service life.
From pipe sections which were severely pitted, the materials tensile strength
was only slightly affected. The pipe bursting strength was affected, but
not to the point that would constitute a failure. Failure would occur first
by perforation. Booth et al also reported that for small to moderate diame-
ter aluminum water pipes, the hydraulic efficiency will be reduced approxi-
mately one percent per annum over the first ten years of service life (6).
ASBESTOS-CEMENT PIPE PERFORMANCE IN THE WATER WORKS INDUSTRY
Asbestos-cement pipe was first manufactured in Europe in 1913 and was
introduced in the U.S. in 1929 (98). Approximately one-third of all water
distribution pipe currently being sold in the U.S. is manufactured of
asbestos-cement. Since its introduction approximately 200,000 miles of
asbestos-cement pipe has been placed into service for transporting potable
waters (41).
Asbestos-cement pipe is composed of 15 to 20 percent asbestos fiber,
48 to 51 percent cement, and 32 to 34 percent silica. The cement portion is
either Portland cement, Portland blast furnace slag, cement, or Portland
pozzolena cement (98)-
4-76
-------
0.5
150 200
1.0 1.5
Concentrations, ppm
250
I
300
2.0
HCO;
CuH
Figure 33. Corrosion of aluminum after 22 hours of exposure in solution
containing chloride, bicarbonate, and copper (85).
-------
Asbestos is a generic term representing a number of fibrous silicate
minerals. These minerals vary in their metallic content, fiber diameter,
flexibility, tensile strength, and surface properties. Six asbestos minerals
have been identified for their commercial importance and are chrysotile,
amosite, crocidolite, anthophyllite, amphiboles, and actinolete. Chrysotile
referes to the serpentine [Mg6Silt010(OH)8] variety of asbestos and comprises
80 percent or more of the asbestos used in asbestos-cement pipe (41).
Chrysotile accounts for approximately 95 percent of the world's asbestos pro-
duction and is mined and quarried primarily in Quebec and Vermont (88). It
is identified physically as a tubular or hollow fibrous material. The
remaining type of asbestos used for manufacture of potable water pipe is the
crocidolite variety. This mineral is a fibrous blue or bluish green sili-
cate of iron and sodium.
Asbestos-cement pipe is available in sizes ranging from 4 to 36 inches
in diameter. Type II asbestos-cement pipe is autoclaved while Type I is
not. The interior base of the pipe is polished during production and, there-
fore, is very smooth and requires no interior coating. During installation,
asbestos-cement pipe lengths are joined by special couplings which are also
made of asbestos-cement. These couplings are machined to fit over the
machined ends of the pipe using two flexible 0-rings to ensure a watertight
seal. Asbestos-cement pipe can be easily drilled and tapped to provide
additional service when necessary. Advantages of asbestos-cement pipe
include an immunity to electrolysis due to its non-conductor status, and a
low hydraulic resistance due to its smooth interior.
Several investigations have been initiated to determine if asbestos
minerals are released from asbestos-cement pipe into potable water. In
general, these investigations have attempted to correlate various water qual-
ity conditions and pipe ages with asbestos fiber releases or occurrences in
potable waters after passing through a specified length of asbestos-cement
pipe. The results of these investigations will be discussed subsequently.
First, however, it is important to note the difficulty in conducting these
investigations and obtaining meaningful results. Therefore, a brief discus-
sion of the problems encountered in attempting to qualify or quantify the
potential release of asbestos fibers from asbestos-cement pipe into drinking
waters follows to provide an appreciation or understanding of the reported
results and their limitations.
Asbestos is ubiquitous and asbestos fibers of one variety or another
are present in soils throughout the U.S. The most frequent occurrences of
near surface asbestos fibers in soils are found in the western and Atlantic
seaboard states (51 ). These asbestos fibers are also present in our nation's
water supply sources as they are leached from the soils by runoff and re-
charge water. Natural wind erosion and earth disturbances will also act to
transport asbestos fibers into water ways. Therefore, it can be anticipated
that appreciable amounts of asbestos fiber may exist in potable water as it
enters a distribution system.
4-78
-------
To determine if asbestos fibers are released from asbestos-cement pipe
in potable water distribution systems, it is necessary to quantify incremen-
tal changes in asbestos concentrations or fiber counts as the water enters
and passes through the pipe. Observed incremental changes in asbestos con-
centrations or counts, however, do not necessarily indicate a release from
the asbestos-cement pipe. Increases in fibers or fiber concentrations may
result by contamination from the surrounding serpentine soil which remains
in the pipe following construction or repairs.
In one study, conducted by the Vermont Department of Health, Sargent
reported that asbestos fibers can appear in potable water distribution sys-
tems which do not use asbestos-cement pipe (88;- The objective of their
study was to compare various sources and to determine if asbestos fibers were
picked up in distribution systems. To eliminate the effect of the ubiqui-
tous nature of asbestos and its possible presence in source water, the inves-
tigators took samples of both the source and the distribution system and com-
pared the results of the analysis. The results showed that in 17 out of 23
systems initially sampled, the number of asbestos fibers increased from
source to distribution, while in six systems it actually decreased.
Incremental increases may also result from drilling and tapping opera-
tions when the interior surface of the pipe is disturbed. Although the
release of asbestos fibers via drilling and tapping is directly associated
with the use of asbestos-cement pipe, measured incremental increases observed
during field or laboratory studies should not be construed as normal release
of asbestos fibers from the smooth interior surface. Therefore, during such
investigations, any drilling and tapping or other pipe disturbances must be
identified for corrections.
The analytical procedures used to determine asbestos fiber release in
asbestos-cement pipes is presented in Section 5. However, it should be noted
that current techniques are based an microscopic quantification may be speci-
fic to a certain type of fiber and may not report fiber size. While the
effect of ingested asbestos fibers on health has not been determined, it is
assumed that the type and size may be important fiber parameters (51 ). Also,
these techniques are often imprecise and generally valid to within an order
of magnitude.
Causes of Asbestos Fiber Release
Several investigations have been conducted to ascertain that asbestos-
cement pipe does undergo deterioration resulting in the release of asbestos-
cement fibers in potable water systems. Other more recent investigations
have been initiated in an attempt to identify and quantify the various char-
acteristics which affect asbestos-cement pipe performance. The results of
most studies reported to date indicate that structural deterioration is usu-
ally negligible even with apparently high asbestos fiber counts, although
a measurable decrease in pipe thickness may occur. Primary characteristics
identified for examination include water quality, detention or pipe-water
exposure time, pipe age, and installation practices.
4-79
-------
Hallenbeck et al investigated the effects of pipe age on the release of
asbestos fibers (41). For this study, paired samples were taken from 15
public water systems in Northeast Illinois. The transmission electron micro-
scope analysis technique was used to detect and count chrysotile fibers.
Paired samples were collected for comparison and were representative of
before and after passing through asbestos-cement pipe. The field data col-
lected and the results of this study are shown in Table 27.
As can be observed, a wide variety of water quality characteristics
were investigated. Consequently, the authors performed statistical tests on
the before and after sample pairs. Although some increases occurred, the
authors concluded that no statistical significant release of chrysotile
fibers was observed. In some analyses, it was found that the fiber counts
increased. From fiber length measurements, it was determined that this
increase was probably due to breakage in fiber as the fibers were generally
shorter in the after exposure samples.
Tracy also investigated the effects of pipe aging on the release of
asbestos fibers from asbestos-cement pipe (104). In this investigation,
water quality samples for pH, hardness, and alkalinity were collected from
various locations in the distribution systems and changes were observed.
Asbestos-cement pipe sections of various ages were selected for this study
from three communities in Vermont which were Brattleboro, South Shaftsbury,
and South Burlington. Water quality observations were continued over nearly
a four-year period to identify any effects of pipe aging. Water quality
sampling and analytical results from this study are shown in Table 28.
In Table 28 it is shown that significant increases in pH, alkalinity,
and hardness were observed from samples collected from the Brattelboro and
South Shaftsbury distribution systems which were approximately five years
old. Samples collected from these same facilities after they had been in
service for approximately nine years showed less significant changes.
Additionally, Tracy observed from the results of samples taken from the nine-
year-old South Burlington system that only slight changes in water quality
occurred in portions of the distribution system where circulation was good
as compared to dead-end sections where circulation is minimal. From the re-
sults of this study, Tracy concluded the asbestos-cement pipe may stabilize
with age and become more resistant to water quality characteristics (104.).
Buelow et al investigated the behavior of asbestos-cement pipe under
various water quality conditions (9). The specific objective of their study
was to determine if asbestos-cement pipe would be attacked and asbestos
fibers released under the various conditions. Their approach was to select
ten water supply systems throughout the U.S. which utilized asbestos-cement
pipe and which had various water quality characteristics with respect to pH,
calcium hardness, and alkalinity. Pipe sections from most of the systems
were visually inspected and the samples were analyzed using the electron
microscope technique. The results of their study were reported as a corre-
lation between a water quality aggressive index, calculated from the value
of the water quality parameters listed above, and the incremental increase in
asbestos fibers observed from samples selected.
The aggressive index used for the Buelow et al study follows that proce-
dure identified by AWWA Standard C400-77 which establishes criteria for
4-80
-------
TABLE 27. FIELD DATA STUDY - DATA COLLECTED AND RESULTS (41)
I
00
WATER SYSTEM
Groundwater
Systems :
Westmont
Lisle
Hoffman
Estates
Rolling
Meadows
York
Center
Lake Michigan
Systems :
Bannockburn
Bradley Road
Zion-Benton
Waukegan
Zion
Midlothian
Blue Island
Brook Field
Glenview
Highland Park
PIPE AGE
(Years)
0.5
1.0
18.0
20.0
27.0
1.0
14.0
18.0
19.0
26.0
27.0
30.0
35.0
37.0
40.0
PIPE LENGTH
(Feet)
200
3,000
3,700
4,400
1,100
1,300
1,000
10,000
600
1,150
1,500
3,248
40,000
1,500
275
pH |
7.1
7.7
8.2
7.5
7.7
8.4
8.4
8.1
DATA
8.1
8.2
8.2
8.2
7.6
7.9
WATER QUALITY
Aggressive Index
11.2
12.5
12.7
12.2
12.8
12.4
12.4
12.1
NOT AVAILABLE
12.1
12.2
12.2
12.2
11.5
11.8
MASS OF CHRYSOTILE
FIBERS OBSERVED
10~5g/Grid Square
Before | After
0.15
0.13
0.28
1.06
0.03
0.53
5.97
0.41
17.21
0.28
1.27
29.90
0.92
0.39
0.24
0.92
0.44
0
0
0.27
3.78
0.41
1.25
4.71
0.34
1.05
6.63
0.28
0.82
0.10
-------
TABLE 28. EFFECTS OF PIPE AGING (104)
i
00
ro
BRATTLEBORO
(Installed in 1941)
Center
9/45
pH 7.2
Hardness 30.0
Total Alkalinity 30.0
of Vill
8/46
7.2
31.0
24.0
Reservoir
10/45
pH 6.6
Hardness 24.0
Total Alkalinity 12.0
(
City
11/45
pH 7.0
Hardness 54.0
Total Alkalinity 37.0
9/46
7.3
29.0
20.0
Install
Line
8/49
7.4
46.0
41.0
N . 1 . 5 mi on
age Cement -Asbestos Pipe
4/49 | 9/45 8/46 4/4<
N
9 4/45
7.5 9.2 9.1 8.2 9.9
46.0 48.0 48.0 40.0 57.0
24.0 41.0 36.0 27.0 55.0
SOUTH SHAFTSBURY
(Installed in 1940)
N. 0.6 mi on
Cement-Asbestos Pipe
'49 ave.| 10/45 9/46 5/4
N.W
9 10/45
7.2 6.8 7.8 8.3 9.6
32.0 26.0 29.0 24.0 45.0
15.0 22.0 12.0 35.0
SOUTH BURLINGTON
ed in 1936; extended in 1948)
E. 1.1 mi on
Cement-Asbestos Pipe
11/45 8/49
E . 1 . 9 mi
Dead End
11/45
7.2 7.4 8.2
54.0 50.0 54.0
41.0 42.0 44.0
. 2 mi at
Dead End
8/46
9.6
51.0
50.0
. 1 . 2 mi at
Dead End
9/46
8.4
50.0
41.0
4/49
8.6
42.0
30.0
5/49
9.5
40.0
26.0
E. 3.1 mi at
at New Dead End
After 1948
8/49 1
7.5
74.0
45.0
8/49
8.2
72.0
48.0
-------
determining the quality of water that can be transported through asbestos-
cement pipe without any adverse structural effects. Although this parameter
is often presented in asbestos-cement studies, it is not always accurate in
predicting a tendency to release fibers or to allow Ca(OH)2 leaching (34).
The aggressive index (AI) is calculated as:
Aggressive Index = pH + log [AH]
where,
pH = index of acidity or alkalinity in standard pH units
A = total alkalinity in mg/a as CaC03
H = calcium hardness in mgA as CaC03
Values greater than 12.0 identify non-aggressive water; values between 10.0
and 11.9 identify moderately aggressive water; and values less than 10.0
identify highly aggressive waters.
Three of the systems investigated had a water quality aggressive index
in excess of 12.0 and are, therefore, considered non-aggressive. Samples
collected from these systems were, in general, free of asbestos fibers. Only
two samples collected from the three systems which had passed through
asbestos-cement pipe had asbestos fiber counts which were statistically sig-
nificant. The highest value reported was 0.3 million fibers per liter (MFL).
In this analysis, a fiber count of 0.2 MFL was also indentified in the water
source or at the treatment facility.
Two of the water systems investigated had a water quality aggressive
index between 10.0 and 11.9 and are considered moderately aggressive. The
first system reported had an aggressive index of 11.56 and the second had an
aggressive index of 10.48. Only two samples collected from the first system
had fiber counts which were statistically significant. Both values were 0.2
MFL. A third sample taken from the well pump had an asbestos fiber count of
0.1 MFL.
In the second system which had a moderately aggressive water (aggressive
index = 10.43), changes in water quality with respect to pH, calcium hard-
ness, and alkalinity were also monitored at two sampling locations. It was
observed that pH and calcium concentrations increased as the water passed
through the asbestos-cement pipe. This increase indicates that calcium
hydroxide or other calcium products in the cement binder were being dissolved
resulting in an increase in pH and calcium concentrations in the water, and
demonstrates that water aggressive to asbestos-cement pipe will continue to
increase in pH and calcium with time of exposure as the water seeks its cal-
cium saturation level (9). In this system significant asbestos fiber counts
ranging up to 4.6 MFL were observed. However, because of the large fluctua-
tions in the number of fibers found in various samples, the authors explained
the high fiber counts as originating from pipe tapping in the sample collec-
tion area.
Five of the ten systems investigated had a water quality aggressive
index less than 10.0 and are considered highly aggressive to asbestos-cement
pipes. For these five systems surveyed, the aggressive index ranged from
5.34 to 9.51. From the results of this investigation, several important
4-83
-------
observations were made. In general, water samples taken from the system
showed that pH and the aggressive index increased as the aggressive water
passed through the asbestos-cement pipe indicating that the asbestos-cement
pipe serves a source of pH adjustment. With only one exception, high fiber
counts were measured in these water systems having highly aggressive waters
as was anticipated. In these tests pipe sections were removed for inspec-
tion and pipe deterioration and loosened fibers were apparent where high
fiber counts were observed. In one test where asbestos-cement pipe was ex-
posed to a water having an aggressive index of 8.74, the pipe inspection
showed that the cement binder had been dissolved to a depth of 1/8 inch.
In another test by Buelow et al, asbestos-cement pipe was exposed to
a water having an aggressive index of 6.0 to 7.5 and a pH ranging from 4.5
to 6.0. Although a high asbestos fiber count was expected, very few were
actually observed. Additionally, a visual inspection showed little deterio-
ration, but instead the presence of an iron rust-like coating. It is sus-
pected that this iron rust-like coating actually provides a protective
coating against pipe deterioration from aggressive water. Susequent labora-
tory testing confirmed this speculation (9). A summary of the results of the
field test completed by Buelow et al is shown in Table 29.
The Environmental Protection Agency Drinking Water Research Division
also conducted laboratory studies to investigate the performance of asbestos-
cement pipe under various water quality conditions (9). In the initial test-
ing, full lengths of four-inch and six-inch diameter pipes were used in an
effort to simulate actual conditions and minimize problems associated with
laboratory scale down. However, during the testing, water quality conditions
were difficult to maintain as a drift in pH and alkalinity concentrations
were observed owing to the exposure of the water supply source to carbon
dioxide in the atmosphere. Despite the problems encountered, some interes-
ting qualitative results were observed. For example, it was observed that
iron, dissolved in the water from some of the experimental equipment, preci-
pitated and provided a protective coating on the asbestos-cement pipe and
halted calcium leaching. From this initial experimental test it was also
verified that drilling and tapping of asbestos-cement pipe will generally
result in increased fiber counts in water and this increase can be
significant (9).
Because of the difficulties in controlling water quality conditions in
this initial experimental test, a laboratory scale coupon test experiment was
performed. The objective of this study was to investigate the effects of
controllable water quality conditions on asbestos-cement pipe deterioration.
This study included the use of chemical additives as a corrosion control
strategy. A summary of the water quality conditions used in the experiments
and general observations made are shown in Table 30.
A comparison between Tests 1 and 2 indicated that the addition of zinc
orthophosphate to a concentration of 0.3 to 0.5 mg/i provided protection for
the asbestos-cement pipe. It was observed that zinc was gradually depleted
but the phosphate was not.
Experimental Tests 3 and 4 were companion tests to further study the
potential of zinc orthophosphate for protection at a lower pH and lower
aggressive index. The results indicated that the use of zinc orthophosphate
at a lower pH or aggressive index was not as effective for preventing
4-84
-------
TABLE 29. SUMMARY OF FIELD DATA COLLECTED BY BUELOW ET AL (9)
Initial
Alkalinity
Aggressive
System
Index
pH
mg/«, as
CaC03
Calcium
Hardness
mg/£ as
CaC03
Pipe Wall
Consistently Deteriorated
Quantifiable as
Fibers
by
Determined
Inspection
Significant
Observations
CO
en
5.34 5.2
5.67 4.8
7.46 6.0
8.74 7.1
9.51 7.2
1.0
3.0
4.0
89.0
14.0
1.4
2.5
7.5
0.5
14.5
Yes
Yes
No
Yes
Yes
Yes Water pH and A.I. increased
as water passed through A/C
pipe; A/C pipe served as
source for pH adjustment.
Yes High fiber counts were obser-
ved in water samples; obser-
vation on pipe section remov-
ed confirmed pipe deteriora-
tion.
No Asbestos fibers were general-
ly absent from water samples;
observations of pipe section
suggested that an iron rust-
like coating provided protec-
tion from attack of this
highly aggressive water.
Yes High fiber counts were obser-
ved in water samples; obser-
vation on pipe section remov-
ed confirmed pipe deteriora-
tion.
Yes Water pH increased with expo-
sure time to A/C pipe.
Continued
-------
TABLE 29 (Continued)
Initial
Aggressive
System Index
PH
Calcium Pipe Wall
Alkalinity Hardness Consistently Deteriorated
mg/£ as mg/«, as Quantifiable as Determined
CaC03 CaC03 Fibers by Inspection
Significant Observations
(So
CTl
8
10.48 8.3
11.56 7.5
12.74 9.4
20.0
88.0
12.54 7.8 220.0
50.0
7.5
82.0
250.0
44.0
Yes
No
No
No
N.I.* Large fluctuations in water
sample fiber counts indica-
ted that pipe tapping may
be responsible for the pres-
ence of some asbestos fibers.
N.I. Water samples collected were
generally free of asbestos
fibers as is expected from
this moderately aggressive
water.
N.I. Water samples collected were
free of asbestos fibers.
N.I. Water samples collected were
free of asbestos fibers.
*N.I. = Not Inspected
-------
TABLE 30. WATER QUALITY CONDITIONS AND GENERAL OBSERVATIONS FOR SMALL SCALE EXPERIMENTS (9)
CO
Calcium
Experiment mg/£ as
No. pH CaC03
1 8.2 6
2 8.2 6
3 7.0 10
4 7.0 10
5 8.2 6
6 7.5 145
7 7.9 145
8 9.0 25
Total
Alkalinity
mg/£ as
CaC03
20
20
20
20
20
125
125
40
Aggressive
Index
10.28
10.28
9.30
9.30
10.28
11.76
12.16
12.00
Corrosion
Control
Method
None
Zinc
Orthophosphate
None
Zinc
Orthophosphate
Zinc
Chloride
None
Slightly
Positive
Langlier
Index
CaC03
Saturation
General Observations
Alkalinity and calcium concentra-
tions increased significantly
during experiment; coupon was
softened.
Alkalinity and calcium concentra-
tions increased slightly during
test; coupon retained hard surface;
light gray coating on the pipe
surface was observed.
Alkalinity and calcium concentra-
tions increased; coupon was soft-
ened.
Alkalinity and calcium concentra-
tions increased; coupon was soft-
ened.
Alkalinity and calcium concentra-
tions increased slightly; coupon
retained hard survace.
(Unsaturated with respect to CaCOs);
coupon was softened.
Coupon retained hard and clean
surface.
Coupon was slightly softened.
-------
asbestos-cement pipe deterioration. It does, however, appear to offer some
protection.
Experiment 5 was performed to determine if zinc alone, not phosphate,
was repsonsible for providing protection. Comparison of the results between
Experiments 2 and 5 verified that previous observation.
Experiments 6 and 7 were performed to demonstrate the performance of
CaC03 as a protection mechanism under conditions of saturation and unsatura-
tion. For these experiments, pH was used as the controlling variable for
CaC03 saturation. From Experiment 6, it was shown that the asbestos-cement
pipe was attacked by a water which was unsaturated or unstable with respect
to CaC03, although the aggressive index was high. Alternatively, Experiment
7 showed that a water which was saturated with respect to CaC03 did not
attack the asbestos-cement pipe.
Experiment 8 was a test of the aggressiveness of water at the point of
saturation. This condition is between the conditions tested in Experiment
6 and 7. Results of this test, as expected, showed a slight softenting of
the coupon.
Subsequent investigations have developed an asbestos-cement pipe protec-
tion model to alleviate problems of improper predictions based on the A.I.
by considering the overall water chemistry, and not just the CaC03
saturation (34).
Organic Release from Asbestos-Cement Pipe
The appearance of significant concentrations of tetrachloroethylene in
potable water has recently been associated with the use of lined asbestos-
cement pipe. In an investigation performed by the Environmental Protection
Agency, pipe sections of lined and unlined asbestos-cement pipe were immersed
in a beaker of water and water samples were analyzed at the start, one hour,
six hours, and 24 hours later. In these experiments no detectable level of
tetrachloroethylene was observed in samples taken from the unlined pipe
beaker. However, in the experiments using the lined asbestos-cement, the
following results were observed (55-|:
TETRACHLOROETHYLENE CONCENTRATION (yg/£)
Exposure Time Test 1 Test 2
0 hour Not Detectable Not Detectable
1 hour 8 14
6 hours 25 25
24 hours 41 20
Water quality samples have been collected from the field where lined
asbestos-cement pipe sections have been installed. Tetrachloroethylene con-
centration as high as 2508 yg/£ were observed from samples collected at
Brenton Point Park in Newport, Rhode Island, in October 1977 (55). Samples
collected from a new lined asbestos-cement service line in Newport showed a
4-88
-------
level of 56.7 yg/£ (1). Results showing levels in excess of 30 yg/Ji have
recently been reported in Vermont (55).
In an effort to identify the source of tetrachloroethylene, the Envir-
onmental Protection Agency has investigated the techniques used in fabrica-
tion and installation of asbestos-cement pipe. Tetrachloroethylene is used
to clean the internal surface of asbestos-cement pipe prior to application
of the liner. Therefore, it is concluded that the quantity or concentration
of tetrachloroethylene which is released to the water is at least paritally
dependent on the durability and integrity of the lining (55). It should be
noted that this process has been stopped, and no pipes manufactured with the
process are being sold.
CONCRETE PIPE
Concrete pipe was first used for transporting potable waters in 1910,
but widespread use of concrete pipe did not occur until after 1930. Concrete
pipe is composed of Portland cement, sand and gravel aggregates, water, and
reinforcing steel. Three types of concrete water pipe are available and
are classified in accordance with the method of reinforcement. These three
types are steel cylinder, not prestressed; steel cylinder, prestressed; and
noncylinder, not prestressed.
Concrete pipe for transporting potable waters can be either prefabrica-
ted at a central plant or manufactured on site. Concrete pipe can be con-
structed in any size, but pipe diameters generally range form 12 to 96
inches. Concrete pipe sizes up to 180 inches in diameter have been produced
for water systems.
Concrete pipes are usually coated or lined internally with a specified
mixture of mortar or concrete. If the pipe will be exposed to aggressive
water, an internal coating of cutback asphalt is sometimes spray applied.
Concrete pipe sections are joined with a modified bell and spigot joint, and
a gasket is used to ensure a watertight fit. The space between the pipe and
the two joining pipes is filled with mortar (98).
Concrete pipe has been used extensively for water distribution with pipe
being in service for 50 years or more in some locations. The suitability and
acceptance of concrete pipe for water mains is well established, but concrete
pipe can be attacked in some circumstances by aggressive waters or soil con-
ditions (94). Additional coatings are applied in such cases.
Although it is not strictly a concrete because aggregate is not present,
Portland cement coatings can be applied to protect cast iron or steel water
pipe on either the water or soil side or both. The cement protects the
underlying from corrosion by the aggressive environments. The coating which
may be applied by centrifugal casting, trowelling, or spraying ranges in
thickness from 0.25 to greater than one inch. The cement coatings are sub-
ject to the same types of attack as concrete pipe. A disadvantage of cement
coatings is the sensitivity to damage by mechanical or thermal shock.
4-89
-------
However, small cracks in cold-water pipes may be automatically plugged with
a reaction product of corrosion combining with alkaline products leached from
the cement.
A series of investigations during the 1950's were based on visual
inspection and surface layer analysis of cement lined or concrete pipe (29,
30). The samples were removed from various water supply service lines and
the following conclusions regarding their deterioration resulted:
1) Concrete pressure pipe is only slightly affected by even
aggressive water over service periods of 25 years or longer.
2) As seen in the cement-to-calcium oxide ratios shown in
Table 31, the removal of calcium oxide from concrete pipes
is limited to a surface layer less than 0.25 inches deep.
TABLE 31. CEMENT-TO-CALCIUM OXIDE RATIO
(With Respect to Depth from Pipe Surface) (29)
Depth
(inches)
City
Portland ME
(3 yrs/service)
Milton PA
(9 yrs/service)
St. Petersburg FL
(25 yrs/service)
Inside
0.075
1.77
1.76
2.24
Next
0.150
1.54
1.71
1.59
Next
0.150
1.53
1.59
1.50
Next
0.150
1.51
1.58
1.48
Next
0.150
1.54
1.63
1.48
Remaining
1.56
1.60
1.47
3) Reduction in CaO content is not the controlling factor in
determination of the service life of the pipes.
4) The limiting factor in leaching CaO from concrete pipe may
be the formation of a surface deposit of magnesium silicate
and calcium carbonate.
5) There appeared to be no difference in the amount of CaO
leached from either fine or coarse ground cement.
Dissolution of calcium compounds by aggressive waters are the primary
concern on the water side of concrete pipe, but attack by soil conditions is
also important, primarily to maintain structural integrity. Some soils will
react with the cement in the concrete or mortar. Alkali soils contain sul-
fate compounds that cause gradual deterioration of concrete made with stan-
dard Portland cement but there are formulations of sulfate-resistant cement
for use in these areas (4). Acid soils may contain sufficient acid to react
with concrete pipe or mortar. Cut-back asphalt, coal applied tar, or coal
4-90
-------
tar epoxy may be used to coat the exterior of the concrete pipe to protect
it from the acid content of the soil (4).
PLASTIC PIPE
Commercial plastic pipe was first introduced in 1930 in Germany and
later in 1940 in the United States. The first type of plastic pipe commer-
cially available was polyvinyl chloride (PVC). Large-scale production of
plastic pipe, however, did not begin until after 1948 with the production of
polyethylene (PE) for application in various water uses. Plastic pipe was
initially used in the water works industry for service lines and household
plumbing, and most pipe was two inches in diameter or smaller. However, with
continued development, a larger plastic pipe is now available and is used for
water distribution mains, service lines, and in-plant piping systems.
The use of plastic pipe and fittings is steadily increasing in potable
water systems as well as in other more corrosive environments. Several vari-
eties of plastics are used in making pipe. Characteristics and physical pro-
perties of plastics can vary within a chemical group as well as from one
group to another. The two major classifications of plastics are thermoplas-
tics and thermosets, and both are used in the manufacture of pipe. However,
thermoplastics are the material of choice for potable water systems. Thermo-
plastics soften with heating and reharden with cooling which allows them to
be extruded or molded into components for piping. Thermosets are permanently
shaped during the manufacture of an end product and cannot be softened or
changed by reheating.
Total use af thermoplastic piping in 1978 exceeded 3 billion pounds which
was approximately one-third of the footage of all piping (691. Approximately
two-thirds of the thermoplastic piping manufactured in the United States is
used for water supply and distribution, including community and municipal
systems and for drain, waste, and vent piping (lie). The principal thermo-
plastic materials in piping are as follows:
1) polyvinyl chloride including chlorinated polyvinyl chloride,
2) polyethylene,
3) acrylonitrile-butadiene-styrene,
4) polybutylene,
5) polypropylene,
6) cellulose acetate integrate, and
7) styrene-rubber plastics.
Other thermoplastics can also be made into piping for special applications.
The fist four plastics above account for approximately 95 percent of the
total plastic pipe and fittings produced (33). Polyvinyl chloride,
4-91
-------
polyethylene, and polybutylene are the plastics most often used for potable
water supplies. Short descriptions of the various plastics are given below.
Typical physical properites of the major thermoplastics are summarized in
Table 32.
Polyvinyl Chloride (PVC)
PVC is a good example of the variations that can occur within a chemical
group. The properties of the thermoplastic depend on the combinations of
PVC resins with various types of stabilizers, lubricants, fillers, pigments,
processing aids, and plasticizers. The PVC resin is the major portion of the
materials and determines the basic characteristics of the thermoplastic but
the amounts and types of additives influence such properties as rigidity,
flexibility, strength, chemical resistance, and temperature resistance.
Rigid PVC or Type I PVC are the strongest PVC materials because they
contain no plasticizers and the minimum of compounding materials. Type II
PVC materials are made by adding modifiers or other resins and are easier to
extrude or mold, have higher impact strengths, lower temperature resistance
and lower hydrostatic design stresses, and are less rigid and chemically
resistant. Chlorinated polyvinyl chloride (CPVC) is a Type IV PVC made by
the post chlorination of PVC. CPVC is similar to Type I PVC but has a higher
temperature resistance. Both Type I PVC and CPVC materials have a hydrosta-
tic design stress of 2000 psi at 75°F. Type I is useful up to 140°F while
CPVC is useful to 210°F.
The long-term strength and higher stiffness of PVC makes it the most
widely used thermoplastic for both pressure and non-pressure application.
PVC is used in water mains, water services, drain, waste, and vent, sewerage
and drainage, well casing, and communication ducts. The higher temperature
resistance of CPVC makes it applicable for hot/cold water and industrial
piping.
Polyethylene
Polyethylene is a polyolefin formed by the polymerization of the ethy-
lene. Polyethylene plastics are waxy materials that have a very high chemi-
cal resistance. The resistance of polyethylenes is such that piping struc-
tures must be joined by thermal or compression fittings rather than solvent
cements or adhesives. Carbon black may be added to polyethylene to screen
ultraviolet radiation.
Polyethylene compounds are classified by the density of the natural
resins. Type I materials are low density, relatively soft, flexible, and
have low heat resistance. Type I materials have a low hoop stress of 400 psi
with water at 73°F and are seldom used for pipe. When used for pipe, Type I
is used for low head piping or open-end piping; therefore, it is seldom used
in potable water systems. Type II polyethylenes are medium density com-
pounds. These materials are harder, more rigid, resistant to higher tempera-
tures, and more resistant to stress cracking. The high density polyethyl-
enes, Type III, have maximum hardness, rigidity, tensile strength, chemical
4-92
-------
TABLE 32. TYPICAL PHYSICAL PROPERTIES OF !1AJOR THERMOPLASTIC PIPING MATERIALS (69)
Property @ 75°F
Specific Gravity
Tensile Strength psi (103)
Tensile Modulus psi (105)
Impact Strength, Izod
ft-lbs/inch notch
*. Coeff. of Linear Expansion
10
9.0
2.9
III
0.95
3.2
1.3
>10
9.0
3.2
PB
0.92
4.2
0.55
>10
7.2
1.5
PP
0.92
5.0
2.0
2
4.3
1.2
PVDF
1.76
7.0
2.2
3.8
7.0
1.5
Specific Heat Btu/lb-F - 0.32 0.34 0.25 0.23 0.20 0.54 0.55 0.45 0.45 0.29
Approx. Operating Limit*
F,
F,
nonpressure
pressure
180
160
180
160
150
130
130
110
210
180
130
120
160
140
210
180
200
150
300
280
*Exact operating limit may vary for each particular commercial plastic material (consult manufacturer).
Effects of environment should also be considered.
-------
resistance, and temperature resistance. Their hydrostatic design stress for
water at 73°F is 630 psi.
Many water utilitites use polyethylene for cold water distribution and
service lines. The pipe most often used is two inches or less. The tough-
ness, low flexural modules, and chemical resistance are important considera-
tions in water service connections. It is most often used outside
buildings.
Polybutylene
Polybutylene is also a polyolefin. Its use in potable water systems
has been expanding considerably. Polybutylene is similar to low density
polyethylene in rigidity, but its strength is greater than that of high den-
sity polyethylene. However, its significant characteristic is its ability
to retain strength with increasing temperature. Polybutylene has a hydro-
static design stress of 1000 psi for water at 73°F and 500 psi for water at
180°F. Polybutylene is used for hot and cold water distribution, water dis-
tribution and service, gas distribution and services, and industrial piping.
The flexibility of polybutylene makes it useful for main-to-meter water
service tubing and well piping. It also protects against hot water backup
into cold water systems. Polybutylene is used inside buildings for hot and
cold water lines.
Acrylonitrile-Butadiene-Styrene (ABS)
ABS plastics are manufactured from the three monomers from which the
class name is derived. ABS piping materials are similar to Type II PVC but
vary according to the ratios of the component monomers. Acrylonitrile pro-
vides rigidity, strength, hardness, and chemical resistance. Butadiene
makes the plastic tougher. Styrene contributes gloss, rigidity, and easier
processing.
ABS plastic piping is relatively rigid with good impact shrength. The
hydrostatic design stresses for water at 73°F range from 800 to 1600 psi.
ABS plastic piping may be used up to 180°F in non-pressure applications.
ABS may be used to convey potable water but its most common use is for drain,
waste, and vent.
Polypropylene
Polypropylene is another polyolefin but it is not as widely used in
potable water systems as polyethylene or polybutylene. It is similar to
high density polyethylene, but it is more rigid and temperature resistant.
Its good chemical resistance makes it more useful in environments harsher
than potable water systems.
Deterioration and Releases from Plastic Piping
Very little direct information exists on the corrosion or, more appro-
priately, deterioration of thermoplastic materials in potable water systems.
One of the significant features of thermoplastics is the good chemical
4-94
-------
resistance of the compounds; this feature was responsible for many of the
early applications of thermoplastics in handling highly corrosive materials.
The ability of thermoplastics to withstand harsh chemical environments has
received most of the attention directed toward the corrosion of these mater-
ials. Most testing has concentrated on physical properties. Consequently,
little attention has been focused on thermoplastics in such relatively mild
environments as potable water systems. A recent study by the National Bureau
of Standards acknowledges the widespread acceptance of themoplastic piping
for residential plumbing and the absence of recent reports of failures due to
chemical attack or environmental stress cracking. This trend suggests that
these failures have ceased to be of significant concern in the use of thermo-
plastics in residential and related applications (116).
There are two general types of chemical attack on plastic pipe (33).
One is a solubility reaction where a chemical is removed from the plastic,
contaminating the fluid flowing in the pipe. The leached chemical may be
non-reacted components, reaction products, or impurities, but their leaching
should not significantly alter the physical properties of the pipe. From
steric considerations, the Teachable components probably lie close to the
pipe surface. The second type of chemical attack is where a polymer or base
resin molecule is altered by chain breakage, cross linkage, oxidation, or
substitution reactions. In these cases, the properties of the plastic may
be irreversibly altered, and the fluid flowing in the pipe may or may not
become contaminated. The chemical resistance of plastics may vary within
differenct grades of the same type as a result of minor chemical or process
differences. In general, a better chemical resistance exists when smaller
amounts of compounding additives are used. Most plastic pipe compounds
conforming to ASTM specifications use a minimum amount of compounding ingre-
dients, although CAB plastics may use chemically susceptible monomeric plas-
ticizers while PVC Type II uses chemically resistive impact modifiers.
Compared to metals and other construction materials, thermoplastics are
generally superior in resisting corrosion. Thermoplastics are not subject
to electrochemical corrosion because they are not conductors. Such electro-
chemical effects as galvanic corrosion do not occur with thermoplastics. As
examples, soils which are corrosive to metal pipes or in which stray currents
are present do not present problems for buried thermoplastic pipe. The
resistance of thermoplastics alleviate the need for such measures as cathodic
protection and special coatings.
Inorganics do not present significant threats to thermoplastics; most
are not affected by acid and alkaline salts. Thermoplastics are resistant
to polar active compounds such as acids, bases, and brines. The thermo-
plastics are resistant to chemical concentrations in normal household opera-
tions or potable systems. Although most plastics absorb water to a slight
extent, water does not produce corrosion or other types of deterioration.
Under some circumstances direct chemical attack by inorganic species such as
oxygen, chlorine, other strong oxidizers, very strong acids or alkalis, and
ultraviolet radiation may lead to deterioration of the plastics. Some
thermoplastics such as PVC have additives such as carbon black to protect
against ultraviolet rays which might otherwise degrade the long chain struc-
ture upon long duration exposure (72). However, it is unlikely that chemical
4-95
-------
attack by these types of species would be significant in potable water
systems because they would have to be present in such large concentrations
that hazards greater than thermoplastic deterioration would exist.
Since thermoplastics are organic materials, they are subject to deter-
ioration by reaction with some organic compounds, primarily via a solution
mechanism. The solvent cementing of plastic pipe is based on solution. The
effect of organic species on thermoplastics varies with the organic compounds
and plastics. For example, PVC is not affected by most esters and ketones
but cellulose acetate butyrate readily dissolves in most esters and ketones.
Aromatic species are the most likely class of compounds to attack thermo-
plastic piping. However, if organic compounds are present in sufficient
concentrations to deteriorate thermoplastic they present other more signifi-
cant and immediate problems from a water quality standpoint.
Environmental stress cracking is another form of degradation that may
affect thermoplastics in piping systems. The process is believed to occur
when a surface active agent such as an alcohol or detergent acts on surface
flows in a stressed or strained plastic (69). Some degree of stress concen-
tration, particularly at joints or fittings, might arise from 1) forced
alignment of pipes and fittings, 2) building settlement, 3) lumber shrinkage,
4) thermal expansion or contraction, or 5) long-term dimensional changes
(116). A chemical test for potable water pipe and fittings has been sugges-
ted by the Federal Construction Council of the Building Research Advisory
Board, but this test as well as others suffers from 1) uncertainties in the
representativeness of the conditions and 2) effects of exposure duration
(116%
Although the data are limited, there have been several studies of ther-
moplastic pipe deterioration in potable water and simulated environments
(72, 93, 99, 101, 102, 110.}. Tiedeman conducted studies to determine the
possible effects of plastic pipe on the safety, quality, and palatability of
water (101 , 102). He conducted extraction tests to determine the aggressive-
ness of several water systems on various thermoplastic pipes. The results
of the tests showed that no undesirable substances were extracted from the
plastic pipe, with the exception of three samples that were known to contain
substances which might be extractable. A typical set of results are shown
in Table 33. With a pH 9.6 in test waters, 0.34 ppm lead was extracted from
a plastic pipe in which a lead compound was used as a stabilizer. Lowering
the pH to 1.0 by adding hydrochloric acid extracted 2.0 ppm lead. However,
the results were obtained under extreme conditions of temperature, exposure
duration, and area of plastic exposed per unit volume of test waters (101 ).
Over the course of a three-year study, it was found that the most
aggressive potable water was a relatively soft water with the pH adjusted to
5 by adding carbon dioxide (102). This water extracted lead compounds from
specially prepared test plastics. However, the extraction results were
negative for all specimens of plastic pipe recommended for use with potable
water. A Soviet study of the extraction of lead from PVC pipe materials
also confirmed that lead stabilizing compounds could be leached from the PVC
in potable water supplies (93). However, these results were on Soviet pipes
which do not apply in this country. Changes in plastic pipe exposed to
4-96
-------
TABLE 33. TYPICAL EXTRACTION TEST RESULTS* (ltil)
-p.
10
ppi£fc ' (-'"lclr
Tni.il -Mili-lj Alkalinity
-Tur-
hMifv M.lnr T ,-TP*
No. ' •"'"" ] fpm ' \ Hf-iil- I>U- ! l>l,Hn..l. .- .
| 1
1
none
ct
120
160
170
none
ct
110
150
180
none
C|
200
6
6
8
6
6
0
0
3
5
5
7
6
5
5
0 i 5
0
0
0
(J
2
5
/
0
0
0
0
0
0
0
0
nicd
in
med
0
0 I 0
6.7 niL-d
0 ; nied
2
1
2
2
210 ! 0 < 17
21>l) 0 1
nicd
0
0
nicd
nicd
ined
t '
IM> MTlvr.1 • phthalrin ^'
ppm />/>"! rrm • '
152
168
164
17Z
176
176
176
184
188
184
144
160
168
156
172
164
176
172
176
192
184
176
182
144
128
140
168 ' 144
10
11
10
10
9
23
20
21
20
20
36
35
40
39
36
62
59
60
58
58
24 : 54
12
50
8 : 42
6 lf>
160 14S 24 (.«
;
i
i
„ Fc .VI NO, NOi i Cl *0i Hani- u.il i Ml
i"1 />/>«i r."1*1 : rrrn ~ rr** , /*f m /*/"•' Mri* *-i /*/*•"
9.70
9.70
9.65
9.65
9.60
9.90
9.70
9.85
9.85
9.75
9.50
9.45
9.22
9.20
9 50
0.1
0.1
0.1
0.25
0.1
0
0
0.2
trace
0
0.1
trace
0.1
0.1
trace
0.003
0.003
0.003
0.014
0.003
0
0
0
0
0
0
0
0
0
0.02
0.03
trace
0.01
0.04
0.01
0.0
0.1
0.1
0.2
0.002
0.004
0.004
0005
0 0 003
0.02
0.15
0.03
0.04
0:05
0.02
0.04
0.02
0.02
0.02
0.12
0.20
0.28
0.16
0.40
i
10.9
10.9
12.4
11.5
11.8
10.9
10.9
10.2
12.1
12.1
10.9
10.Q
10.3
10.7
10.7
53.9
54.7
54.2
56.2
55.2
64.3
62.9
61.3
71.0
63.4
48.2
41.3
49.4
3U
44.2
95
108
no
97
102
114
114
116
118
114
90
92
90
91
92
0.3
0.02
0.01
0.01
0.01
0.15
0.01
0
0.01
0
0.8
0
0
0
0.3
11.2
8.4
9.0
9.6
9.8
10.8
6.6
8.6
9.0
8.2
9.8
9.2
9.0
9.0
9.2
-------
outdoor conditions or buried in soil at pH 2.0 and held at 35°C were slight
after exposures of one year. Discoloration was the principal change in both
exposures (102).
One concern is the extraction or leaching of organic species from pipe
cements into water supplies. A recent study indicated that it is possible
to leach such solvents as 2-butanone (MEK) and tetrahydrofuran (THF) from
PVC pipe cement (110). Two sets of water samples were collected six and
eight months after PVC pipe installation and usage in a laboratory. About
40 gallons of water were used daily in the laboratory. The water temperature
was about 21°C. Seven water samples at different residence times in the PVC
pipe were taken for analysis. Results are summarized in Table 34. A com-
parison of the data from the two sets of samples indicates that concentration
of both MEK and THF in the second set were reduced to 1/2 of the concentra-
tion in the first set. About 2,400 gallons of water were used during the
period of samples taken between Set I and Set II. This water presumably
removed some of the MEK and THF from PVC pipe cement in the pipe.
TABLE 34. CONCENTRATION (PPM) OF MEK AND THF IN WATER
SAMPLES AT VARIOUS RESIDENCE TIMES IN THE PVC PIPE (no,)
Residence
Time (h)
0
4
8
16
24
48
64
72
96
Samples Taken 6 Months
After Pipe Installation
MEK
0
0.4
0.6
1.8
2.2
3.9
4.5
-
4.5
THF
0
1.0
1.7
5.8
8.9
12
13
-
13
Samples Taken 8 Months
After Pipe Installation
MEK
0
0.1
-
0.6
1.1
2.1
-
2.2
-
THF
0
0.7
-
2.4
3.7
6.3
-
7.5
-
Another series of tests, however, found that concentrations of MEK,
THF, cyclohexanone, and dimethylformamide (DMF) did not attain hazardous
levels in static water or usage simulation tests (103). An analysis based
on results of the tests stated that levels of the four solvents declined to
less than three parts per million in less than three weeks of static exposure
and that no significance in solvent leaching appears between poorly constructed
solvents cement joints and well constructed solvent cement joints. Testing
4-98
-------
was performed by a private consulting engineering firm while the analysis
presented was performed by representatives of the plastic resins, pipe,
fittings, and solvent manufacturers. Research in this area is currently
proceeding and should help to clarify the reported discrepencies concerning
release extents and possible health concerns from organic solvent leaching.
4-99
-------
REFERENCES
1. Adams, W. R., Jr., Regional Administrator, EPA, Region I, J. F.
Kennedy Federal Bldg., Boston, MA 02203, Letter to Dr = J. E. Cannon,
Director, Department of Health, Office of the Director, 75 Davis St.,
Providence, RI 02098, dated January 14, 1980.
2. Anderson, E. A., Reinhard, C. E., and W. D. Hammel, "The Corrosion of
Zinc in Various Waters," J. Am. Water Works Assoc., Vol. 26, No. 1*
1934, pp. 49-60.
3. ASTM Special Technical Publication 516, Localized Corrosion - Cause
of Metal Failure, American Society for Testing and Materials, 1916
Race Street, Philadelphia, PA 19103, 1972.
4. Bald, R. E., "Corrosion Resistance of Concrete Pipe," Hater Wastes
Engineering, Vol. 5, No. 11, 1968, pp. 50-52.
5. Bell, W. A., "Effects of Calcium Carbonate on Corrosion of Aluminum
in Waters Containing Chloride and Copper," J. Appl. Chem.. Vol. 12,
1962, p. 53.
6. Booth, F. F., Murray, G. A. W., and H. P- Godard, "Corrosion Behavior
of Aluminum in Fresh Waters with Special Reference to Pipeline,"
Br. Corros. J.. Vol. 1., No. 2, 1965, pp. 80-86.
7. Bopp, C. D., and S. A. Reed, "Stabilization of Product Water from
Sea Water Distillation Plants," U.S. Office of Saline Water Research
and Development Progress Report, No. 709, 1971.
8. Brighton, W. D., "Dissolved Copper form New Service Pipes," Water and
Water Engineering, Vol. 59, July, 1955, pp. 292-293.
9. Buelow, R. W., Millette, J., McFarren, E. and J. M. Symons, "The
Behavior of Asbestos-Cement Pipe under Various Water Quality Condi-
tions," A Progress Report, Presented at the American Water Works
Association, 1979 Annual Conference, San Francisco, June 27, 1979.
10. Burgmann, G., Friehe, W., and W. Schwenk, "Chemical Corrosion and
Hygienic Aspects of the Use of Hot-Galvanized Threaded Pipes in
Domestic Plumbing for Drinking Water," Pipes Pipelines Int., Vol. 23,
No. 2, 1978, pp. 11-15.
11. Butler, A. and H. C. K. Ison, Corrosion and Its Prevention in Waters,
Reinhold Publishing Corporation, New York, 1966.
12. Caldwell, D. H. and J. B. Ackerman, "Anaerobic Corrosion of Steel Pipe
Due to Nitrate," Journal-AWWA, Vol. 38, January 1946, pp. 61-64.
13. Campbell, H. S., "A Natural Inhibitor of Pitting Corrosion of Copper
in Tap-Waters," J. Appl. Chem.. Vol. 4, 1954, pp. 633-647.
4-100
-------
14. Clark, H. W., "The Effect of Pipes of Different Metals upon the
Quality of Water Supplies," Journal-New England Water Works Asso-
ciation. Vol. 41, 1927, pp. 31-51.
15. Cohen, A.,and W. S. Lyman, "Service Experience with Copper Plumbing
Tube," Materials Protection and Performance. Vol. 11, No.2, February
1972.
16. "Cold-Water Corrosion of Copper Tubing," Task Group Report, J.A.W.W.A.
Vol. 52, August 1960, pp. 1033-1040.
17. Cornwell, F. J., Wildsmith, G., and P. T. Gilbert, "Pitting Corrosion
in Copper Tubes in Cold Water Service," Br. Corros. J., Vol. 8, No. 5,
September 1973, pp. 202-209.
18. Costas, L. P., "Field Testing of Valve Stem Brasses for Potable Water
Service," Materials Performance, Vol. 16, No. 8, August 1977, pp. 9-16.
19. Cruse, H., "Dissolved-Copper Effect on Iron Pipe," Journal-AWWA.
Vol. 63, No. 2, 1971, pp. 79-81.
20. Cruse, H.,and R. D. Pomeroy, "Corrosion of Copper Pipes," Journal -
AWWA, Vol. 66, No. 8, August 1974, pp. 479-483.
21. Davies, D. E., "Pitting of Aluminium in Synthetic Waters," J. Applied
Chemistry, Vol. 9, December 1959, pp. 651-660.
22. Davies, D. H., and G. T. Burstein, "The Effects of Bicarbonate on the
Corrosion and Passivation of Iron," Corrosion-NACE, Vol. 36, No. 8,
August 1980, pp. 416-422.
23. De Waard, C., and D. E. Milliams, "Carbonic Acid Corrosion of Steel,"
Corrosion-NACE. Vol. 31, No. 5, May 1975, pp. 177-181.
24. Donaldson, W., "The Action of Water on Service Pipes," J. Am. Water
Works Assoc., Vol. 11, No. 3, 1924, p. 649.
25. Eliassen, R., Clemente, P., Romeo, A. J., and R. T. Skrinde, "Effects
of pH and Velocity on Corrosion of Steel Water Pipes," Journal-AWWA.
Vol. 48, August 1956, pp. 1005-1018.
26. Elzenga, C. H., and H. J. Boorsma, "Corrosion of Water Pipes in
Various Types of Drinking Water," International Water Supply Associa-
tion, 10th Congress. August 1974, pp. 1-7.
27. Feigenbaum, C., Gal-Or, L., and J. Yahalom, "Microstructure and
Chemical Composition of Natural Scale Layers," Corrosion, Vol. 34,
No. 2, February 1978, pp. 65-70.
28. Feigenbaum, C., Gal-Or, L., and J. Yahalom. "Scale Protection Criteria
in Natural Waters," Corrosion (Houston), Vol. 34, No. 4, 1978, pp. 133-
137.
4-101
-------
29. Flentje, M. E., and R. J. Sweitzer, "Further Study of Solution Effects
on Concrete and Cement Pipe," Journal-AWWA, Vol. 49, November 1975,
pp. 1441-1451.
30. Flentje, M. E., and R. J. Sweitzer, "Solution Effects of Water on
Cement and Concrete in Pipe," Journal-AWWA, Vol. 47, 1955, pp. 1173-
1194.
31. Foley, R. T., "Role of the Chloride Ion in Iron Corrosion," Corrosion-
NACE, Vol. 26, No. 2, February 1970, pp. 58-70.
32. Fontana, M. G., and N. D. Greene, Corrosion Engineering, McGraw-Hill
Book Company, New York, © 1978.
33. Ford, K. C. ed., Plastics Piping Manual. Plastics Pipe Institute, 355
Lexington Ave., New York, N. Y., 10017, pp. 26-28.
34. Gardels, M., and M. Schock, (E.P.A. Cincinnati Laboratory), Personal
Communication, via P- Lassovszky, January 12, 1981.
35. Garrels, R. M., Thompson, M. E., and R. Siever, "Control of Carbonate
Solubility by Carbonate Complexes," American Journal of Science,
Vol. 259, January 1961, pp. 24-45.
36. Geld, I., and C. McCaul, "Corrosion and Coatings Test Program of the
NYC Board of Water Supply," Materials Protection and Performance,
Vol. 11, No. 2, February 1972, pp. 41-44.
37. Geld, I., and C. McCaul, "Corrosion in Potable Water," JAWWA, Vol. 67,
No. 10, October 1975, pp. 549-552.
38. Godard, H. P., "The Corrosion Behavior of Aluminum in Natural Waters,"
The Canadian Journal of Chemical Engineering, Vol. 38, No. 5, October
1960, pp. 167-173.
39. Goetchins, D. R., "Porcelain Enamel as a Protective Coating for Hot
Water Tanks," J. Am. Ceramic Society, Vol. 25, 1942, pp. 164-168.
40. Hale, F. E., "Relation of Copper and Brass Pipe to Health," Water
Works Eng., Vol. 95, 1942, pp. 84-86, 156-159, 187-189, 207-"208~T
240-243, 264-265.
41. Hallenbeck, W. H., et al, "Is Chrysotile Asbestos Released from
Asbestos-Cement Pipe into Drinking Water?," Journal-American Water
Works Association, Vol. 70, No. 2, 1978, pp. 97-102.
42. Hatch, G. B., "Copper Corrosion, Quality Aspects of Water Distribution
Systems," Univ. 111. Eng. Expt. Sta. Cir. #81, 1963, pp. 32-40.
43. Hatch, G. B., "Unusual Cases of Copper Corrosion," Journal-AWWA,
Vol. 53, 1961, pp. 1417-1429.
4-102
-------
44. Heidersbach, R. H., and E. D. Verink, Jr., "The Dezincification of
Alpha and Beta Brasses," Corrosion-NACE, Vol. 28, No. 11, November
1972, pp. 397-418.
45. Hem, J. D., W. H. Durum, Solubility and Occurrence of Lead in
Surface Water, JAWWA, August 1973, pp. 562-568.
46. Hilbert, F., Mizoshi, Y., Eikhor, G., and W. J. Lorenz, "Correlations
Between the Kinetics of Electrolytic Dissolution of Iron, I Anodic
Dissolution of Iron," J. Electrochem. Soc.. Vol. 118, No. 12, 1971,
pp. 1919-1926.
47. Hoover, C. P., "The Corrosive Action of Various Types of Water on
Household Plumbing," Water Works and Sewerage, Vol. 83, 1936,
pp. 384-387.
48. Hubbard, D. J., and C. E. A. Shanahan, "Corrosion of Zinc and Steel
in Dilute Aqueous Solutions," British Corros. J., Vol. 8, No. 6,
November 1973, pp. 270-274.
49. Karalekas, P. C., Jr., Craun, G. F., Hammonds, A. R., Ryan, C. R.,
and D. J. Worth, M.D., "Lead and Other Trace Metals in Drinking Water
in the Boston Metropolitan Area," Journal-New England Water Works
Association. Vol. 90, No. 2, 1976, pp. 150-172.
50. Karalekas, P. C., Jr., Ryan, C. R., Larson, C. D., and F. B. Taylor,
"Alternative Methods for Controlling the Corrosion of Lead Pipe,"
J. New England Water Works Assoc., Vol. 92, No. 2, 1978, pp. 159-178.
51. Kuschner, M., et a'l, "Does the Use of Asbestos-Cement Pipe for
Potable Water Systems Constitute a Health Hazard?," Journal-AWWA,
Vol. 66, September 1974, pp. 4-22.
52. Lane, R. W., and C. H. Neff, "Materials Selection for Piping in
Chemically Treated Water Systems," Materials Protection, Vol. 8,
No. 2, February 1969, 27-30.
53. Langelier, W. F., "The Analytical Control of Anti-Corrosion Water
Treatment," Journal-AWWA, Vol. 28, No. 10, 1936, pp. 1500-1521.
54. Langelier, W. F., "Chemical Equilibria in Water Treatment," Journal-
AWWA, Vol. 38, No. 2, February 1946, pp. 169-178.
55. Larson, C. D., Chief, Technical Support Section, EPA Region I, J. F.
Kennedy Fed. Bldg., Boston MA 02203, Letter to J. Hagopian, Princi-
pal Sanitary Engineer, Rhode Island Dept. of Health, 75 Davis St.,
Health Bldg., Providence, RH 02908 dated January 31, 1980.
56. Larson, C. D., Chief, Technical Support Section, EPA Region I, J. F.
Kennedy Fed. Bldg., Boston MA 02203, Letter to J. Hagopian, Rhode
Island Dept. of Health, 75 Davis Street, Health Bldg., Providence
RH 02908 dated November 14, 1979.
57. Larson, T. E., "Corrosion by Domestic Waters," Illinois State Water
Survey, Urbana, Bulletin 59, 1975.
4-103
-------
58. Larson, T. E., and R. M. King, "Corrosion by Water at Low Flow
Velocity," Journal-AWWA, Vol. 46, No. 1, January 1954, pp. 1-9.
59. Larson, T. E., and R. V. Skold, "Current Research on Corrosion and
Tuberculation of Cast Iron," Journal-AWWA, Vol. 50, November 1958,
pp. 1429-1432.
60. Larson, T. E., and R. V. Skold, "Laboratory Studies Relating Mineral
Quality of Water to Corrosion of Steel and Cast Iron," Corrosion,
Vol. 14, June 1958, pp. 43-46.
61. Leckie, H. P., and H. H. Uhlig, "Environmental Factors Affecting the
Critical Potential for Pitting in 18-8 Stainless Steel," Journal of
the Electrochemical Society, Vol. 113, No. 12, December 1966,
pp. 1262-1267.
62. Leidheiser, H., Jr., The Corrosion of Copper, Tin, and Their Alloys,
John Wiley & Sons, Inc., New York, © 1971.
63. Levelton, B. H., and D. G. Kilburn, "Accelerated Corrosion Tests on
Copper Water Tubing," Materials Protection Journal, Vol. 5, No. 8,
August 1966, pp. 37-40.
64. Lucey, V. F., "Mechanism of Pitting Corrosion of Copper in Supply
Waters," Corrosion J., Vol. 2, No. 5, 1967, pp. 175-185.
65. Lyson, T. D. B., and J. M. A. Lenihan, "Corrosion in Solder Jointed
Copper Tubes Resulting in Lead Contamination of Drinking Water," Br.
Corros. J., Vol. 12, No. 1, 1977, pp. 41-45.
66. Hahato, B. K., Voora, S. K., and L. W. Shemilt, "Steel Pipe Corrosion
Under Flow Conditions - I. An Isothermal Correlation for a Mass
Transfer Model," Corrosion Science, Vol. 8, No. 3, 1968, p. 173.
67. McCauley, R. F,, and M. 0. Abdullah, "Carbonate Deposits for Pipe
Protection," Journal-AWWA, Vol. 50, 1958, pp. 1419-1428.
68. Moore, M. R., Plumbosolvency of Waters," Nature. Vol. 243, No. 5404,
May 25, 1973, pp. 222-223.
69. Mruk, S. A, "Thermoplastics Piping: A Review," Managing Corrosion
Problems with Plastics, Vol. 4, National Association of Corrosion
Engineers, P. 0. Box 218340, Houston TX 77218, © 1979, pp. 3-14.
70. NACE Technical Unit Committee T-7 on Potable Waters, Task Group
T-7B-2, Second Corrosion Study of Pipe Exposed to Domestic Waters,
NACE Publication 7B170, Materials Protection & Performance, Vol. 9,
No. 6, June 1970, pp. 34-37.
71. Naylor, Lewis M., Richard R. Dague, Simulation of Lead Removal
by Chemistry Treatment, JAWWA, October 1975, pp. 560-565.
72. Nesbitt, W. D., "PVC Pipe in Water Distribution: Reliability and
Durability," Journal-AWWA, Vol. 67, No. 10, October 1975, pp. 576-581.
4-104
-------
73. Nielsen, K., "Contamination of Drinking Water with Cadmium and Lead
from Brazed and Soldered Joints and from Other Metals in Plumbing
Systems," Second International Brazing and Soldering Conference,
London, Okt. 1975.
74. O'Brien, 0. E., "Lead in Boston Water: Its Cause and Prevention,"
Journal of the New England Hater Works Association, Vol. 90, No. 1,
January 1976, pp. 173-180.
75. Patterson, J. W., "Corrosion Inhibitors and Coatings," Proc.-AWWA
Seminar Controlling Corros. Water Syst., Paper No. 6 (4 pp.) 1978.
76. Patterson, J., Illinois Institute of Technology, Personal Communica-
tion, December 1979.
77. Patterson, J., and J. E. O'Brien, "Control of Lead Corrosion,"
Journal of the American Hater Works Association, Vol. 71., No. 5,
May 1979, pp. 264-271.
78. Polushkin, E. P., and H. L. Shuldener, "Corrosion of Yellow Brass
Pipes in Domestic Hot-Water Systems - A Metallographic Study,"
Corrosion, Vol. 2, No. 1, March 1946, pp. 1-19.
79. Porter, F. C., and S. E. Hadden, " Corrosion of Aluminium Alloys in
Supply Waters," J. Applied Chemistry, Vol. 3, September 1953,
pp. 385-409.
80. Pourbaix, M., "Recent Applications of Electrode Potential Measurements
in the Thermodynamics and Kinetics of Corrosion of Metals," Corrosion
(Houston), Vol. 25, No. 6, 1969, pp. 267-281.
81. Pourbaix, M., "Theoretical and Experimental Considerations in Corro-
sion Testing," Corrosion Science, Vol. 12, 1972, pp. 161-191.
82. Reedy, D. R., "Corrosion in the Water Works Industry," Materials
Protection, Vol. 5, No. 9, September 1966, pp. 55-59.
83. Report on Lead Contamination of Bennington, Vermont Drinking Water.
Report by USEPA Region I, Water Supply Branch, Boston MA, January 4,
1977.
84. Rossum, J. R., Pickup of Heavy Metals from Residential Plumbing,
Rough Draft prepared for the California Department of Health,
December 1, 1975.
4-105
-------
85. Rowe, L. C., and M. S. Walker, "Effect of Mineral Impurities in Water
on the Corrosion of Aluminum and Steel," Corrosion-National Assoc. of
Corrosion Engineers, Vol. 13, December 1961, pp. 105-113.
86. Rudek, R., Blankenhorn, R. and H. Sontheimer, "Verzogerung der
Eisenoxidation durch natiirliche organische Wasserinhaltsstoffe und
deren Auswirkung auf die Korrosion von schwarzen Stahlrohren,"
Vom Wasser, Vol. 53, 1971, pp. 133-146.
87. Russell, R. P., Chappell, E. L., and A. White, "The Effect of Velocity
on Corrosion of Steel under Water," Ind. Eng. Chem., Vol. 19, 1927,
pp. 65-68.
88. Sargent, H. E., "Asbestos in Drinking Water," Journal-New England
Water Works Association, Vol. 88., No. 1, 1974, pp. 44-57.
89. Schafer, G. J., "Corrosion of Copper and Copper Alloys in New Zealand
Potable Waters," New Zealand Journal of Science, Vol. 5, December
1962, pp. 475-484.
90. Schaut, G. G., "The Action of a Chlorinated Water Supply Upon Lead
Pipe," American Journal of Pharmacy, Vol. 114, 1942, pp. 241-249.
91. Schock, M. R., "Response of Lead Solubility to Dissolved Carbonate in
Drinking Water," Pre-Publication Copy, Physical and Chemical Contami-
nants Removal Branch, Drinking Water Research Diviison, U.S. Environ-
mental Protection Agency, 26 W. St. Clair Street, Cincinnati OH
45268, October 1979
92. Scholefield, R. J., "Metal Corrosion Products in Municipal Drinking
Waters," Thesis for Master of Science in Environmental Engineering at
Illinois Institute of Technology, Chicago IL, August 1979.
93. Sheftel, V. 0., "The Lixiviation of Lead Stabilizers from Polyvinyl
Chloride Water Pipes," Hygiene and Sanitation, Vol. 29, 1964. pp. 121-
122.
94. Simmonds, M. A., "Effect of Aggressive Waters on Cement and Concrete
with Particular Reference to Cement-Lined Mains," The Journal of the
Institution of Engineers, Australia. Vol. 26, January-February 1954,
pp. 9-16.
95. Slunder, C. J., and W. K. Boyd, Summary Report on Lead - Its Corrosion
Behavior to ILZRO, Battelle Memorial Institute, Columbus OH.
96. Streicher, L, "Effects of Water Quality on Various Metals," Journal -
AWWA, Vol. 43, No. 3, March 1956, pp. 219-238.
97. Stumm, W., "Investigation on the Corrosive Behavior of Waters,"
Proceedings of the American Society of Civil Engineers, Vol. 86,
No. SA-6, November 1960, pp. 27-45.
4-106
-------
98. Symons, G. E., Ph.D, "Water Systems, Pipes and Piping, Part I/Piping
Systems Design," Water and Hastes Engineering/Manual of Practice
Number Two, Vol. 4, No. 5, May 1976, pp. M3-M50.
"• Thermoplastic Piping for Potable Water Distribution Systems, Federal
Construction Council, Building the Search Advisory Board, Division
of Engineering, National Research Council, Technical Report No. 61,
National Academy of Sciences, Washington, D.C., 1971.
100. Thibeau, R. J. , Brown, C. W., Goldfarb, A. Z. and R. H. Heidersbach,
"Raman and Infrared Spectroscopy of Aqueous Corrosion Films on Lead
in 0.1 M Sulfate Solutions," J. Electrochem Soc.. Vol. 127, No. 9,
September 1980, pp. 1913-1918"
101. Tiedeman, W. D., "Studies on Plastic Pipe for Potable Water Supplies,"
Journal-AWWA, Vol. 46, August 1954, pp. 775-785.
102. Tiedeman, W. D., and N. A. Mil one, "Effects of Plastic Pipe on Water
Quality," Journal-AWWA, Vol. 48, 1956, pp. 1019-1023.
103. Toxicological Task Force of the Plastic Pipe and Fittings
Association, Toxicological Analysis of James M. Montgomery,
Consulting Engineers, Inc., Plastic Pipes Study Draft Report,
Glen Ellyn, Illinois, September 1980.
104. Tracy, E. L., "Observations on Water Samples form Cement-Asbestos
Pipe Systems in Vermont," Journal-New England Water Works Association,
Vol. 64, 1950, pp. 164-178.
105. Tronstad, L. and R. Veimo, "The Action of Water on Copper Pipes,"
Water and Water Eng., Vol. 42, May 1940, pp. 189-191.
1Q6. Tronstad, L. and R. Veimo, "The Action of Water on Copper Pipes,"
Water and Water Eng.. Vol. 42, June 1940, pp. 225-228.
107. Uhlig, H. H., Corrosion and Corrosion Control, an Introduction to
Corrosion Science, John Wiley & Sons, Inc., New York, 1963.
108. Uhlig, H. H, ed., The Corrosion Handbook, John Wiley & Sons, Inc.,
New York, 1948.
109. Wagner, I., "Influence of Water Quality and Water Treatment on
Corrosion and Coatings in Steel and Galvanized Steel Tubes,"
EUROCOR '77, 6th European Congress on Metallic Corrosion, (Met. A.,
ong
3^4
7807-72 0184), 1977, pp. 413-419.
110. Wang, T. C., and J. L. Bricker, "2-Butanone and Tetrahydrofuran
Contamination in the Water Supply," Bull. Environm. Contam. Toxidol.,
Vol. 23, 1979. pp. 620-623.
4-107
-------
111. Waters, D. M., "Internal and External Copper Corrosion in Domestic
Water Services," Proc. - AWWA Annu. Conf. 97th, Vol. 1, 1977.
Ilia. Wells, S. W., "Hydrogen Sulfide Problems of Small Water Systems,"
Journal - AWWA, Vol. 46, Feb. 1954, pp. 160-170.
112. Whitman, G. W., Russell, R. P., and V. J. Altieri, "Effect of
Hydrogen-Ion Concentration on the Submerged Corrosion of Steel,"
Industrial and Engineering Chemistry, Vol. 16, No. 7, July 1924,
pp. 665-670.
113. Wong, C. S. and P. Berrang., "Contamination of Tap Water by Lead Pipe
and Solder," Bulletin of Environmental Contamination and Technology,
Vol. 15, No. 5, 1976, pp. 530-534.
114. Worth, D., "Relationship of Blood Lead Levels to Lead in Household
Drinking Water," Proceedings - New England Water Works Association,
Walthan, MA, December 18, 1975.
115.-. Wright, T. E. and H. P. Godard, "Laboratory Studies on the Pitting
of Aluminum in Aggressive Waters," Corrosion, Vol. 10, June 1954,
pp. 195-198.
116. Wyly, R. S-, Parker, W. J., Pierce, E. T., Rorrer, D. E., Shaver, J.
R., Sherlin, G. C., and M. Tyron, "Investigation of Standards,
Performance Characteristics, and Evaluation of Criteria for Thermo-
plastic Piping in Residential Plumbing Systems, NBS Building Science
Series 111, National Bureau of Standards, Washington, D.C., May 1978.
4-108
-------
SECTION 5
CORROSION MONITORING AND DETECTION
Detection of degradation and measurement of corrosion will be desirable
for assessing the corrosivity of a given water, determining the efficacy of
water treatment or inhibitor programs, and evaluating health effects of
water system corrosion. The procedures involved in corrosion testing are
deceptively simple in the sense that measurements can be obtained using
relatively simple procedures. The detailed preparation of specimens and
apparatus, however, is critical to obtaining reliable numbers. And the
design of the experiment and use of the results for prediction requires con-
sideration of many aspects of corrosion. This section describes the basic
test methods applicable to corrosion in potable waters and gives references
to more detailed procedures.
The following general methods are discussed in this section.
• specimen exposure for an extended duration followed
by examination and weight-loss determination,
• electrochemical measurement of "instantaneous"
corrosion rates, and
• chemical analysis for changes in concentration
of a chemical species resulting from corrosion.
As with all corrosion tests, the value and reliability of these methods will
depend on proper planning and .execution of the details involved in the pro-
cedures. The applicability of a given procedure will depend on the objectives
of the tests.
This discussion is intended to apply primarily to testing under field
conditions (in the water treatment plant or distribution system). Testing
under laboratory conditions requires careful preparation and control of the
corrosive environment in addition to the other precautions. As in the rest
of this report, external corrosion will not be considered.
5-1
-------
SPECIMEN EXPOSURE TESTING
Placement of a test specimen in the corrosive environment and examination
after some exposure duration is the oldest corrosion test method. While
fundamentally simple, there are a number of details which must be considered.
One of the most basic considerations is that the test specimen should "see"
the same environment as the equipment of interest. This environment includes
the chemical content of the fluid, the temperature, flow rate, galvanic
coupling, periodic environment fluctuations, entrained solids or gases, etc.
While the test specimens cannot be exposed to exactly the same environment
as a given material in a water supply system, placement should be chosen to
be representative of the application of that material. It is often necessary
to consider the effect of specimen placement on the properties of the en-
vironment such as flow patterns and chemical content. Because corrosion is
a function of electrochemical kinetics and surface phenomena, it is not
surprising that surface preparation of specimen and careful documentation
of metallurgical history are important procedural considerations. Planning
and evaluation of tests should be done after careful review of factors
affecting the known corrosion behavior of the materials in similar environments.
The general procedures used for corrosion testing can be delineated as
follows:
• Selection of materials and specimens. Care should be taken
that factors such as heat treatment and chemical composition
are known and representative of the actual pipe or equipment
of interest.
• Surface preparation. Actual equipment surfaces generally
cannot be duplicated, but efforts to approach them with a
reproducible preparation method must be made.
• Measuring and weighing. Both surface area and weight must
be accurately measured with care taken to avoid fouling the
surface.
• Exposure technique. Proper placement should be maintainable
for the entire test period.
• Duration. Exposure time and an examination program should
be carefully planned before starting the test period.
' Examination and cleaning of specimens after test. This step
is important where documentation and use of proper technique
is critical.
• Interpretation of results.
5-2
-------
Details of these steps are discussed in large part by Fontana and
Greene (4). Procedures are also given in standards or recommended prac-
tices by the American Society for Testing and Materials (ASTM) and the
National Association of Corrosion Engineers (NACE). The main ASTM pub-
lication is the Standard Recommended Practice designated G4 on Conducting
Plant Corrosion Tests which gives general guidelines and information on
apparatus, test specimen preparation and placement, test duration, specimen
removal and examination, and reporting (2). The ASTM Standard Recommended
Practice Gl gives additional details on preparing, cleaning, and evaluating
corrosion test specimens fl). Another useful guide is the NACE Standard
TM-01-69 (1976 Revision) on Laboratory Corrosion Testing of Metals for the
Process Industries (12). Use of this guide in potable water corrosion con-
trol testing has been described by Mullen and Ritter (ll).
The size and shape of test specimens depends on several factors and
cannot be rigidly set. It is generally desireable to have a high ratio of
surface area to mass to obtain maximum corrosion loss. While the sample
should be as large as possible, it should not exceed the weight limitations
of the usual analytical balances (about 160 grams) or present problems in
placement in pipes or equipment. Thin sections can be used to satisfy several
of these requirements but the specimen should not be so thin as to be per-
forated by corrosion or to lack reasonable mechanical stability. The edges
of specimens should be finished by polishing or machining to eliminate cold-
worked metal. Specimens with sheared edges should not be used. Any dirt
or heat-treated scale should be removed and the specimens should be freed
from water breaks by suitable cleaning. Metal specimens should be abraded
to at least 120 grit surface finish. The specimen should be stamped for
identification, weighed to the nearest 0.1 mg on an analytical balance, and
their surface area accurately determined.
A number of methods can be used for supporting specimens for exposure.
The main considerations are that the corrosive media should have easy access
to the specimens, the supports should not fail during the tests, the specimens
should be insulated or electrically isolated unless the study of galvanic
effects is intended, and the desired degree of immersion should be obtainable.
Ready access to the specimens is also desireable. Apparatus for mounting
specimens is described in detail and with mechanical drawings in ASTM
G4-68 (2). They describe a spool rack in which specimens with a hole drilled
through their center are positioned on a metal support rod which is covered
with insulating plastic. Plastic tubing spacers also spooled on the center
rod keep the specimens separate and supported. Insulating end disks are pro-
vided and the assembly is completed by nuts which are tightened on either
end of the support rod. Other support methods are based on similar principles.
They should be tailored to fit the equipment and operating conditions at hand.
Misleading results may be obtained if exposure duration and number of
exposure periods are not carefully selected. It is often found that initial
corrosion rates are considerably higher than those obtained after some time.
However, in some cases pitting or crevice corrosion may not occur until after
5-3
-------
a certain incubation period. In general, tests run for long periods are con-
siderably more realistic than short term tests. For uniform corrosion, a
very rough guide for minimum exposure time suggested by both ASTM and NACE
is given by:
duration of test (hour) = corrj!^ rate (mpy)
This guideline is based on the general rule that the lower the corrosion rate,
the longer the test should be run. The guide can be used with an estimated
lower limit of corrosion rate or used to decide if tests should be repeated
for a longer period based on existing results.
Most sources recommend using the planned-interval test originally
proposed by Wachter and Treseder for setting up tests and evaluating results.
This procedure allows evaluation of the effect of time on corrosion of the
specimen and also on the corrosiveness of the environment. The procedure and
evaluation of results are given in Table 35 along with an example of its
application. This procedure is recommended by NACE TM-01-69 and also by
Fontana and Greene (4).
After removal from the test environment the appearance of the test
specimens and the rack should be noted. Specimens should be washed in
water to remove soluble materials from the surface. Color photographs of
the specimens should be made. The appearance and degree of adhesion of any
coatings or films or the surface should be noted. If possible, samples of
the corrosion product films should be preserved for future study. Specimens
are not generally weighed until corrosion products are totally removed, since
metal converted to corrosion product is structurally lost. But for potable
water studies, additional information on the addition of species to the water
stream might be obtained by also weighing the dried specimens at this point.
Following this, the corrosion layers should be removed by a method that does
not affect the base metal. The cleaning procedure is critical and will de-
pend on the base material as well as the nature of the corrosion products.
Procedures may include light mechanical cleaning (eg. rubbing with a rubber
stopper), electrolytic cleaning, and chemical cleaning. Detailed procedures
are given in ASTM Gl-72 and in Fontana and Greene (1, 4). The possibilty
of solid metal removal should be checked by applying the proposed method
to fresh and to already cleaned, dried, and weighed specimens to determine
any additional weight loss. After cleaning, the specimens should be dried
and weighed to the same accuracy as the initial pre-test weighing. Weight-
loss corrosion rates should be calculated for uniform corrosion cases. The
specimens should be carefully examined visually and any modes of degradation
such as pitting, crevice corrosion, dealloying, or other attacks noted.
Photographs of the specimens should again be made since cleaning will often
disclose more features of attack. If pitting occurs the maximum and average
pit depths should be measured and also the number, size, general
distribution, and shape of the pits should be noted. Distinction should be
made between pits which occur under insulating spacers and those on exposed
surfaces. The former is probably related to crevice corrosion. The depth
5-4
-------
TABLE 35. PLANNED INTERVAL TEST (4)
A7
o, ..
en
« U
Time
t+1
Identical specimens-all placed in the same corrosive fluid. Imposed
conditions of the test kept constant for entire time t + 1. Letters,
Aj, At, AI-H, B, represent corrosion damage experienced by each
test specimen. Aj is calculated by subtracting At from At+j.
o
I
Oi
Occurrences During Corrosion Test
Liquid corrosiveneu
Metal conodibility
Criteria
unchanged
decreased
increased
unchanged
decreased
increased
B
AI
Aj
B
= B
B Ai
AI
-------
of any crevice corrosion should be determined. Pitting rates are often
ambiguous and no extrapolations should be made. The actual pit depth and
length of exposure should be reported. Selective or localized attack can
be examined and recorded in greater detail by metallographic and microscopic
techniques.
Metals which may be susceptible to dealloying or stress cracking should
be bent after other examination and the development of any cracks should be
noted. These results should be compared to similar bend tests with unexposed
specimens (2).
ELECTROCHEMICAL TEST METHODS
Electrochemical methods for corrosion measurement are more complicated
than specimen exposure testing with respect to both equipment and inter-
pretation. Electrochemical methods are also relatively new and less estab-
lished than the conventional exposure tests. When properly applied, however,
the newer methods offer several advantages. The electrochemical methods of
main interest are very rapid and can be used for near-continuous monitoring
of corrosion rates under proper conditions. They are adaptable to measure-
ment of low corrosion rates which are most difficult to measure by weight
loss. Because the most often used electrochemical methods do not significantly
affect the specimen, time profiles of corrosion rates can be obtained. Also,
the effects of various water treatment methods can be monitored on a given
specimen.
The electrochemical measurements will be most reliable when the metal
of interest undergoes uniform corrosion in systems where scale formation is
minimal. The care required for sample preparation and placement is as impor-
tant for these methods as for the simple specimen exposure methods discussed
above. One important limitation on the use of electrochemical methods to
obtain rapid corrosion measurements is that the corroding behavior of a metal
often depends on the length of time it has been exposed to a given environment.
Electrochemical methods can provide "instantaneous" corrosion rates, but
variation of corrosion rates with time must also be considered.
The basis of these methods is the electrochemical nature of the corrosion
of metals in aqueous solutions. The methods are not applicable to nonmetallic
materials used in water supply systems. The rate of electrochemical reactions,
and thus the rate of corrosion reactions, can be expressed as an electrical
current. The driving force for obtaining the reaction giving rise to the
current is an electrical potential difference. In the general case, the
relation between current and potential difference for electrochemical re-
actions is non-ohmic (i.e., nonlinear); generally an exponential or complex
mixed relationship is seen. However, for small deviations in driving force
from some steady state, or open circuit corrosion potential, the current and
potential difference are approximately ohmic or linear. Use of such small
potential deviations (i.e., small polarization) methods forms the basis for
practically all commercial electrochemical corrosion rate instruments and
methods proposed for field or routine use. They are often referred to as
5-6
-------
"linear polarization" or "polarization resistance" methods. It is noteworthy
that the use of these small perturbation methods also causes less change in
the specimen surface and makes possible multiple measurements with the same
specimen. Even small perturbations are of some concern. However, tests of
repeatedly polarized cast iron specimens in a potable water environment gave
generally the same results as freely corroding samples (10). The possibility
of differences occurring with other metals should be considered.
The derivation of current-potential relationships for linear polarization
conditions has been described for various degrees of model sophistication
(4). A widely used form is given by the equation:
Q Q
AE ea ec
Ai dog 10)icorr(Ba + ec)
In this equation, AE is the deviation from the corrosion potential, Ai is the
current density (current per unit area of electrode specimen), icorr 1S the
corrosion current density which can be related to the corrosion rare, and
Bo and BC are parameters (so-called Tafel slopes) associated with the electro-
chemical kinetics of the individual anodic and cathodic corrosion reactions.
At least to a first approximation, Ba and ec are constants for a given corro-
sion system. For model systems, & values are such that the following simple
form is obtained.
AE _ RT
AT = T
corr
where RT/F consists of fundamental constants and the absolute temperature. The
value of RT/F is about 0.026 volts at 25°C. These equations are applicable for
AE values of 20 millivolts at most, and smaller AE values are probably prefer-
able. The simple form often gives useful approximations; although abso-
lute errors may be large in case of unequal Tafel slopes, relative measurements
may still be useful. The quantity 1 (microamperes) can be related to the
'corrosion rate by the following equation:
0 129 i A
Corrosion Rate (mpy) _ corr
nD
Here A is the atomic weight of the metal, n is the number of electrons lost
by the metal atom during corrosion (eg. n = 2 for iron going to Fe+2), and D
is the metal density (g/cm3). In some commerical instruments the conversion
factors are set by specimen size and electronic means to give direct read-
outs of corrosion rates.
There are a number of means of actually carrying out the linear polar-
ization measurement. There does not appear to be a definitive study to deter-
mine if any particular method is superior for use in potable water type systems.
Schematics of two types of electrode circuits are shown in Figure 34. The
three-electrode method is widely used and employs a test specimen electrode,
5-7
-------
3-Electrode
2-Electrode
Figure 34. Schematic of 3 and 2 electrode circuitry,
1000
100
1.0
low-Resistance,
Solutions
0.01
0.1
1.0
AE/cd (
ma/sq dm
10.0
Fiqur? 35.
Empirical relationship between initial slope
of polarization curve (resistance) and corrosion
rate determined by weight loss. (7)
5-8
-------
an auxiliary electrode and a reference electrode. External current flows
only through the test and auxiliary electrodes while the polarization of the
test electrode is measured by way of the referenced electrode. The degree
of polarization may be set at some constant value, 10 millivolts is often
used,or varied in the vicinity of the corrosion potential. It is also
possible to use two identical electrodes for test and auxiliary and measure
potential difference between these two along with the current, without using
a reference electrode. Comparison of three-electrode and two-electrode
methods have been given, along with a general analysis of errors in linear
polarization methods, by Bandy and Jones (3). This method is intended to
measure only the equivalent resistance of the corrosion reaction rate and,
therefore, care should be taken that other resistances such as that of the
solution or connectors are small and that abnormal scale or fouling of the
test electrode does not occur.
The applicability of the linear polarization method to certain potable
water systems was first demonstrated by Larson (7). The close correlation
found between corrosion rates obtained by weight loss measurement and the
linear polarization resistance is shown in Figure 35; These results were
obtained for cast iron and steel in potable water environments of various
composition. Several other evaluations of the linear polarization method
in potable waters have been reported (10,9).
Polarization curves can be recorded by the potentiodynamic method in
which the potential of the test specimen is electronically varied at a pre-
set linear rate and recorded along with the resulting current. For general
use the potential excursions can be limited to about ± 10 millivolts of the
open circuit voltage to decrease the chance of modifying the surface by
polarization. The potential scan rate should be low enough that current
due to double layer charging is negligible compared to the faradaic current.
The method has the advantages of providing a better measure of the current-
voltage slope which is the desired measurement and a readout by which any
departure from linearity of this plot can be obtained. This is obtained at
the cost of increased complexity in instrumentation and operation.
The linear polarization method has promise for rapid estimates of the
corrosivity of potable waters. It must be applied with care and time must
be taken to ensure that metal specimens have reached a steady-state corrosion
rate. With careful use, it should be possible to monitor the relative effects
of various treatment methods using these procedures. A number of probes of
the same material should be used and periodic inspection for localized cor-
rosion or scaling should be done. Integration of the instantaneous rates
and comparison with weight-loss specimens should be done for confidence in
absolute rates derived from linear polarization methods.
Linear polarization methods are applicable to uniform corrosion or at
best averaging small areas of pits with large uniform areas. Electrochemical
methods for determining localized corrosion susceptibility are being evaluated
under laboratory conditions, but have not been standardized. Potential
applications of such methods to field testing has recently been discussed
5-9
-------
by Martin (8). Additional work is needed to develop these methods for use
in potable water environments.
Another electrical, but not electrochemical, method for determining
corrosion rates is based on measuring the change in electronic resistance
of a metal specimen. As the relatively thin wire, tube or strip specimen
element corrodes, its electrical resistance increases due to the decrease
in cross sectional area. Measurements of resistance over a period of time
(several days or weeks) give an estimate of the change in specimen thickness
and, thus, a corrosion rate can be obtained. Results can be obtained more
quickly than with coupon tests, but are not "real time" in the sense of
linear polarization tests. For uniform corrosion behavior, the method is
relatively direct.
Reedy has described the field use of this type of resistance probe for
testing a number of metals in two natural water sources (13). A stabilization
period of 10-14 days immersion was required before consistent corrosion rates
were obtained. This result was probably due to the normal surface changes
that occur on exposure of a fresh metal specimen to a corrosive environment.
Long-term testing (50 days) confirmed that reproducible results could be
obtained after the initial period.
CHEMICAL ANALYSES FOR CORROSION PRODUCTS
Corrosion can be inferred from the increase in concentration of a metal
species in solution from one point in a water distribution system to another
point downstream of the first. Although this analysis is not a conventional
means of corrosion measurement, it provides a direct measure of the quantity
of interest and, in many cases, may be the only way of determining a health
hazard if construction materials are unknown or access to the distribution
system is difficult. Due to the uncertainties involved, this procedure will
probably serve best as an indicator of potential corrosivity of a water
system and the need for application of more quantitative corrosion rate
methods.
Sampling points and procedures will be dictated by the information
desired. To obtain baseline samples without any contamination by piping,
raw water samples should be dipped from surface sources, or for groundwater
supplies collected as close to the well as possible. In all cases samples
for trace metal analysis should be preserved with ultrapure nitric acid in
thoroughly cleaned sample bottles. Finished water samples can be obtained
after treatment at the treatment plant. Sampling in homes can be set up to
provide some differentiation between sources of metal contamination.
Karalekas and coworkers describe the following sampling program (5, 6).
The water samples were collected at the kitchen sink the first thing in the
morning before any water was used in the house.
5-10
-------
• Interior plumbing sample; this is collected immediately
upon opening the faucet and represents water that has
been standing overnight in the fixture and plumbing
serving the faucet,
• Service line sample; this is obtained after the sample
collector feels the water temperature change from warm
to cool, representing the water in the service line to
the house,
• Water main sample; this is collected after allowing the
water to run for several minutes and represents water
from the main which has had minimum contact time with
the service line and interior plumbing.
These samples provide a representation of the range of trace metal concentra-
tions the consumer is likely to experience as well as an indication of the
source of the contaminant.
Although the comparison of chemical analyses from several points seems
to provide a simple procedure, there are several complications. Because a
difference of two experimental numbers is required, the analytical methods
must be both accurate and precise. Surface water supplies may have trace
metal contents from a geochemical origin. Trace materials may also be in-
troduced by airborne or waterborne pollutants or as impurities in chemicals
added to the water during the treatment process. In analyzing any trace
component in water, care must be taken during sampling, sample storage, and
sample pretreatment to avoid large systematic errors due to such problems
as contamination by equipment, precipitation or adsorption of the measured
species, or contamination by impurities in reagents used for sample pre-
servation or pretreatment. The dependence of material concentration on
system flow history is well documented with regard to "stagnant" versus
"free-flowing". More complicated behavior may also occur in distribution
or household systems as corrosion products precipitate or undergo redox
chemistry or adsorption. Corrosion products may spall from walls or be re-
leased from pits at irregular intervals. The sorption of relatively large
quantities of lead and copper ion by hydrous ferric oxide has been studied
(14). Materials such as hydrous ferrous oxide could exist either as a layer
attached to iron pipes or in suspension. In view of these considerations,
reliable results may often require taking a large number of specimens or
continuous monitoring over a period of time.
If the materials in use in the water system are unknown or are not well
defined, the chemical analysis method for corrosion products is further
complicated. Some chemical analyses are expensive, and performing tests for
large number of possible contaminants may be cost prohibitive for small
public or private water systems. Additionally, specific corrosion products
can result or appear from the use of various materials.
5-11
-------
Analytical procedures used to quantify the existence of asbestos minerals
in water are severely limited and are subject to produce erroneous results.
Primary reasons for the extreme difficulty in determining asbestos fiber con-
centrations in water include 1) asbestos fiber concentration in potable water
is generally very low, 2) chemical analytical methods are not applicable be-
cause elements present are common to all rock-forming minerals, 3) asbestos
fibers cannot be concentrated or separated from other inorganic solids present
in the water, and 4) fiber sizes are often below the resolution of the optical
microscope (15). Also, a knowledge of field operations is necessary to deter-
mine the possibility of fiber release resulting from drilling and tapping as
opposed to regular deterioration.
One technique employed to analyze asbestos fiber concentrations or counts
in water utilizes the electron microscope. Using this technique, solids are
removed from the water sample by filtration on a membrance filter and the en-
tire sample is ashed to destroy the filter and any organic and oxidizable
inorganic solids that may be present. The inorganic residue is then rubbed
out in a dilute solution of nitrocellulose to reduce the particle sizes and
transferred to a standard electron-microscope grid. This sample is examined
under an electron microscope and the presence of chrysotile fibers is quanti-
fied by measuring the length and diameter of each fiber, calculating the total
mass and finally relating this mass to the original amount of water sampled.
This technique is specific to chrysotile fibers because of their recognizable
hollow-tube structure.
The accuracy and precision of this anlysis is very poor because only an
extremely small fraction of the sample can be examined and this sample generally
contains only a small amount of asbestos. Analysis of samples using the elec-
tron microscope technique cannot be duplicated by better than a factor of three,
and it is estimated that the measured value is accurate to within a factor of
ten of the true value. Therefore, measured values should be considered indi-
cators or indices of the relative amount of fiber present. Another limitation
of this technique is that the analysis is specific to the structurally recogniz-
able chrysotile fibers. Several other varieties of asbestos materials may be
present in water such as crocidolite which is also a component of most asbestos-
cement pipe (15).
Development of analytical1methods for trace material determination has
coincided with growing environmental concerns. A critical evaluation of the
many developing methods is outside the scope of this report. Three of the more
popular methods for the metals likely to be of interest are atomic absorption
spectroscopy (AAS), anodic stripping voltammetry (ASV), and neutron activation
analysis (NAA). Use of NAA requires access to a reactor. AAS often requires
considerable sample pretreatment as well as equipment usually restricted to a
permanent laboratory. Continuous monitoring of tap water for Pb, Cd, and Cu
has been demonstrated using ASV and related techniques by National Sanitation
Foundation personnel (9). The number of elements that can be analyzed by this
method is relatively small, but they correspond well to those of interest from
a corrosion and health standpoint. The instrumentation required for ASV is
somewhat complicated, but portable.
5-12
-------
REFERENCES
1. ASTM Standard Recommended Practice for Preparing, Cleaning, and
Evaluating Corrosion Test Specimens, Designation Gl-72, American
Society for Testing and Materials, Philadelphia, Pennsylvania
2. ASTM Standard Recommended Practice for Conducting Plant Corrosion
Tests, Designation G4-68, American Society for Testing and Materials,
Philadelphia, Pennsylvania
3. Bandy, R. and D. A. Jones, Analysis of Errors in Measuring Corrosion
Rates by Linear Polarization, Corrosion - NACE, Vol. 32, No. 4,
April 1976, pp. 126-134.
4. Fontana, Mars G. and Norbert D. Greene, Corrosion Engineering,
McGraw-Hill Book Company, New York, 1978.
5. Karalekas, Peter C., Gunther F. Craun, Arthur F. Hammonds,
Christopher R. Ryan, and Dorothy J. Worth, M. D., "Lead and
Other Trace Metals in Drinking Water in the Boston Metropolitan
Area", Journal—New England Water Works Association, Vol. 90,
No. 2, pp. 150-172, 1976.
6. Karalekas, Peter C., C. R. Ryan, C. D. Larson, and F. B.
Taylor, Alternative Methods for Controlling the Corrosion
of Lead Pipe, J. New England Water Works Assoc., Vol. 92,
No. 2, 1978, pp. 159-78.
7. Larson, T. E., Corrosion by Domestic Waters, Illinois State Water
Survey, Urbane, Bulletin 59, 1975.
8. Martin, R. L., Potentiodynamic Polarization Studies in the Field,
Materials Performance, Vol. 18, No. 3, March 1979, pp. 41-50.
9. McCelland, Nina I., and K. H. Mancy, Water Quality Monitoring in
Distribution Systems: A Progress Report, JAWWA, Vol. 64, No. 12,
1972, pp. 795-803.
10. McClanahan, Mark A., and K. H. Mancy, Comparison of Corrosion -
Rate Measurements on Fresh vs. Previously Polarized Samples,
JAWWA. Vol. 68, No. 8, August 1974, pp. 461-466.
11. Mullen, Edward D. and Joseph A. Ritter, Potable - Water Corrosion
Control, JAWWA, Vol. 66, No. 8, August 1974, pp. 473-479.
12. NACE Standard TM-01-69 (1976 Revision) Test Method: Laboratory
Corrosion Testing of Metals for the Process Industries, National
Association of Corrosion Engineers, Katy, Texas, approved March,
1969 (revised 1972, 1976).
5-13
-------
REFERENCES - SECTION 5 (continued)
13. Reedy, Donald R., Corrosive Effects of Southern California
Potable Waters, Materials Protection & Performance, Vol. 12,
No. 4, April 1973, pp. 43-48.
14. Swallow, Kathleen C., David N. Hume, and Francosi M. M. Morel,
Sorption of Copper and Lead by Hydrous Ferric Oxide, Environmental
Science & Technology. Vol. 14, No. 11, November 1980, pp. 1326-1331,
15. Wright, G. W. (Chairman), et al, Committee Report, "Does the Use of
Asbestos-Cement Pipe for Potable Water Systems Constitute a Health
Hazard?" JAWWA, September 1974, pp. 4-21.
5-14
-------
SECTION 6
CORROSION PREVENTION AND CONTROL
Corrosion control in potable water systems is most commonly attempted by
establishing a protective barrier between a corrodible material surface and a
corrosive water. This protective barrier can be applied mechanically as a
material coating or lining prior to installation of facilities or it can be
applied chemically by adjusting the water quality characteristics of the
potable water to precipitate a byproduct which forms on the surface of water
system materials. The most commonly used pipe coatings include coal tar
enamels, epoxy paint, and cement mortar. Water tank linings generally in-
clude coal tar enamels and paints, vinyl, and epoxy. Other coatings and
linings which are also used are hot and cold applied petroleum based waxes,
zinc paints, and asphalt. Any applied coating or lining must be environmen-
tally sound, including their application procedures, and must not impart
objectionable aesthetic or health effects into the water.
The addition of lime to induce calcium carbonate precipitation and form
a protective coating chemically is the most commonly used and accepted method
of corrosion control through water quality adjustment. Chemically applied
protective coatings are also formed by the addition of sodium silicate and
inorganic phosphates. These chemicals are generally referred to as corrosion
inhibitors.
Other corrosion control practices include adjusting the water quality
characteristics to render potable waters less aggressive or less corrosive.
As previously discussed, corrosive waters are generally characterized as
containing high dissolved oxygen and/or carbon dioxide, low alkalinity and
hardness, and low pH. Adjusting these parameters for optimum corrosion con-
trol of a specific material is sometimes practiced or attempted, but optimum
conditions are often difficult to determine.
Corrosion control alternatives currently being practiced and available
to the water works industry are discussed in this section. Also included in
this section is a brief description of the case history of the Seattle Water
Department attempts to correct or retard corrosion in its potable water supply
facilities. Their attempts emphasize the magnitude of the problems and dif-
ficulties in implementing a corrosion control program. A list of the applic-
ability of the mechanically applied linings and corrosion inhibitors discussed
is presented in Table 48 (Section 7):
6-1
-------
MECHANICALLY APPLIED PIPE LIMING AND COATINGS
As early as I860, attempts to solve corrosion problems in water pipes
resulted in dipping cast iron pipe in a bituminous coating material, and
cement lining was suggested by the French Academy of Sciences. By 1930,
the use of a cement mortar lining for pipes used in water distribution systems
was specified in eastern and southeastern areas of the United States, while
many areas in the Midwest specified coal tar coatings. Professional papers
in the 1930's and 40's ( 13, 14,21 ) extolled the virtures of cement
lined iron and cast iron pipes for removing "red water" problems, but incon-
sistencies between different systems prompted comparison investigations in
the 50's ( 3,50 ). These papers reported tubercule formation and
corrosion resistency for various cement, bituminous dips and enamels, pitch,
tar dip, and asphalt sealed cement linings on iron pipes. Water tank corro-
sion was controlled with paint coatings or cathodic protection which were
gradually replaced with coal tar enamel. More recent technologies have
produced vinyl, epoxy, chlorinated rubber, and other tank linings. The focus
of current research is to determine the extent of trace organics released and
subsequent public health affects from the use of bituminous based linings.
The three principal pipe linings currently used for corrosion protection
are coal enamel, epoxy, and cement mortar.
Hot Applied Coal Tar Enamel
Hot applied coal tar enamel is produced from coal tar pitch, a residue
in the fractional distillation of coal tar obtained from the destructive
distillation of bituminous coal. Coal tar used in potable water systems is
required to meet AWWA specifications for type of coal and production process.
Coal tar enamel is made by dispersing coal in a mixture of coal tar pitch and
coal tar gas oil. Talc is added to provide strength.
Hot applied coal tar enamel is used in steel pipes, and it is estimated
that between 50 and 80 percent of all steel pipe used for water distribution
is lined this way. The coating process involves spraying a primer coat,
usually a chlorinated rubber based resin, on the cleaned surface. After the
primer dries, the coal tar enamel is heated to a fluid consistency and poured
on the pipe surface from a trough extending lengthwise through the rotating
pipe. Fittings are manually lined. Thickness of the coating is 0.094 inches
(2.4mm). If connections are to be welded, field touch-up around welded joints
is necessary.
6-2
-------
Hot applied coal tar enamel was used to line steel water pipelines in
New York City as early as 1914. Coal tar enameled pipes, inspected after
many years of service in water supply systems, have shown no coating failures
when the lining was properly applied. Service life of this lining is rela-
tively long, often in excess of 50 years. Other advantages include good
erosion resistance to silt or small amounts of sand in the water and resis-
tance to biological attachments. Disadvantages include the need to re-apply
to welded areas and to use special care in handling pipes during weather ex-
tremes as cold causes brittleness and heat may initiate cracking. An in-
crease of trace organic compounds exists in water flowing through coal tar
lined pipe, but the potential effects of their release is presently unknown.
The extent of this release is currently under investigation at the EPA
laboratory in Cincinnati, Ohio.
Epoxy
Epoxy paint coatings have been used to line steel water pipes since the
mid 1960's and are becoming increasingly popular. Two types of epoxy systems
exist: single component and double component. The former consists of an
epoxy resin, pigments, drying oil, and a reactive resin. Drying results from
oxidation and polymerization. The latter consists of a base containing the
epoxy resin, pigment, and solvent and a polyamide or amine and solvent hardener.
The lining is usually applied in the field because welding will burn off
the epoxy and re-application can be avoided. The pipe is first cleaned and a
phosphate treatment or rust inhibiting zinc silicate primer is applied to the
surface. This step is not necessary, but it improves the epoxy adhesion and
abrasive resistance. Two coats of epoxy are then applied by spraying from a
"pig" pulled through the pipe.
A major advantage to using epoxy linings is the high smoothness coefficients
(Hazen-Williams coefficients of 138-172 (17 )) produced and the corresponding
reduced pumping costs. Also, epoxy paints can be formulated from components
approved by FDA for food contact surfaces. Principal disadvantages include
cost and a reduced resistance to abrasion, compared to coal tar enamel, which
limits service life to under 15 years.
Recent use of powdered epoxy coatings have reduced occupational health and
air pollution problems associated with solvents necessary for spray applications.
The powder is sprayed onto the heated surface and melts to a thin film that
fuses to the pipe. Because epoxy linings have been used to line steel water
pipes for fewer years than their anticipated service life, case histories are
currently just being developed. In 1964, however, the installation of 20,000
feet of epoxy lined 30-inch force main by an eastern Pennsylvania water company
resulted in lower construction costs and a more beneficial Hazen-Williams co-
efficient. As a result, the company installed an additional 28,000 feet of
epoxy lined 20- and 30-inch lines and, 10 years after the 1964 installation,
was planning further installations ( 17 ). It is also reported that three
pipeline companies use only epoxy to line their steel pipes ( 19 ).
6-3
-------
Cement Mortar
Although being used less to line steel pipes, cement mortar is the only
coating used to line cast iron and ductile iron pipes. Cement linings are
applied by introducing a slurry into a rotating pipe. Centrifugal action
distributes the mix to a uniform coating. An alternative process involves
using a rod with a revolving head. The rod is pushed through the pipe, and
the slurry is sprayed onto the surface of the pipe. Composition of the
applied slurry depends on the size of pipe being lined. Pipes under 20 cent-
imeters in diameter use a cement : sand : water weight ratio of 2 : 3 : 1,
whereas larger pipes use a ratio of 2 : 4 : 1 to control possible cracking.
Proper curing of the lined pipe is important. Steel pipes are cured by
sealing the ends, spraying water on the lining, or alternating applications
of steam and water. Mortar in cast iron and ductile iron pipes is coated
with a thin (0.001 inch, 0.0254mm) layer of an asphalt, mineral spirit, and
xylene sealant. This process provides a moisture barrier, constraining water
in the slurry, promoting a proper slow cure. This sealant has the added ad-
vantage of preventing decalcification of the lining in soft waters.
Advantages to using cement mortar lining include its cost and low sensi-
tivity to variations in the substrate quality or application procedures. It'
can be applied in situ on pipes whose lining has failed provided the loosely
coated areas are removed. Because the corrosion abatement mechanism is a
result of the calcium hydroxide released in addition to the physical barrier,
uncoated metal at pipe joints is protected. Also, surface fissures will heal
themselves when immersed in water.
The rigidity of the lining is a disadvantage because pipes subject to
deflections may experience lining cracking and sloughing. Also, the depth
of the coating reduces the cross sectional area of the pipe,-and hence its
carrying capacity significantly. For example, the area of a 20-inch I.D. pipe
with a cement mortar lining is reduced 6 percent. A coal tar enamel lining
reduces the area 2 percent, and an epoxy lining reduces the area by less than
0.2 percent. Also, decalcification in soft, calcium dissolving waters may
impart an objectionable taste to the water.
The first applications of cement mortar linings for steel water pipes was in
the late 1800's and some of these pipes have been in use since that time.
Cement lined sheet-iron pipes were also common at that time and, if designed
and installed properly, provided a 40 to 50 year service life. The first
cement lined cast iron pipe was installed in America in 1922, so the articles
published in the 1930's and 1940's citing the longevity of cement lined pipes
were in reference to the steel lined pipes. Published literature of the 1950's
( 2, 46 ) indicated the widespread use and acceptance of cement lined
cast iron pipe, but academic decisions were being deferred pending further use.
Today, cement mortar is the only lining used in cast iron and ductile iron pipe.
6-4
-------
TANK LININGS AND COATINGS
The principal types of water tank linings are coal tar enamels and paints,
vinyl, and epoxy. Other coatings exist and are mentioned, but they are either
no longer or not extensively used. Water tank linings should exhibit ease of
appl-ication, effective corrosion control, and good erosion resistance.
Coal Tar Based Coatings
Hot applied coal tar enamel is prepared and used as discussed in the
pipe lining section. This coating has a tendency to sag or ripple when applied
above the waterline when the tank walls are heated. Hot applied coal tar
enamel is the primary coal tar based coating used to line water tanks.
Coal tar paints are often used to reline existing water tanks. Cold
applied coal tar paint is prepared by adding coal tar solvents such as xylene
or naptha to coal tar enamel. It is brushed or sprayed on the surface to
form a relatively thick film resistant to sags or runs. Because it can impart
unpleasant tastes and odors to the water, its use is restricted to above water-
line surfaces. It is less durable than hot applied coal tar enamel and has
a service life of 5 to 10 years. A tasteless and odorless cold applied coal
tar paint is produced by combining coal tar enamel with solvents free of
phenols and other taste and odor constituents. This paint may be used below
water, but should not be exposed to sunlight or ice.
Coal tar epoxy paints used to line potable water tanks are generally two
component systems. One part consists of a coal tar pitch base with a poly-
amide resin, magnesium silicate, xylene, ethyl alcohol, a gelling agent, and a
catalyst added. The second part is a liquid epoxy resin. Some types of coal
tar epoxy paints sold do not conform to the Steel Structure Painting Council
specifications and may produce taste and odor problems in the water. Coal
tar epoxy paints are less resistant to abrasion than coal tar enamel. Expo-
sure to sunlight causes chalking to occur, but if this exposure is eliminated,
coal tar epoxy paints may provide a service life of over 20 years.
Coal tar urethane paints have been inconsistent in their service lives.
Some applications have provided a 25 year service life, but other applications
have failed within one year. This type of lining is not presently marketed.
Coal tar emulsion paint is a water based suspension of coal tar pitch,
magnesium silicate or other mineral filler, and a rust inhibitor. Although
it has good adhesive characteristics, is practically odorless, and resists
sunlight degradation, it is not as watertight as organic solvent coal tar
coatings. Consequently, its use below the waterline is limited.
6-5
-------
Investigations have shown that trace organics are released into water
stored in coal tar base lined tanks. The results of analytical tests on
water stored in a coal tar lined tank are shown in Table 36a. Possible health
affects associated with these organics in the concentrations observed are not
presently known. Research presently being done on the release of these or-
ganics, including rates, solubilities in water, and identification of decompo-
sition products will have to be combined with toxicological evaluations to
determine future uses or restrictions on coal tar based linings.
Vinyl paints are a mixture of a vinyl chloride-vinyl acetate copolymer
with a hydroxyl compound and/or a carboxy compound. These paints are applied
on steel in either a three-coat or a four-coat system, with different formu-
lations used in each system and within the five-coat system. Vinyl paints
are inert in water and produce a hard, smooth surface. Soft water conditions
may reduce the expected service life of 20 years. Recent vinyl failures in
California have been blamed on formula changes made to meet that state's air
pollution criteria. Another report stated that vinyl paints do not wear well
in soft waters, but one engineering consultant estimated (1977) that vinyl
paints are specified for 90 percent of their storage tank projects. Although
vinyls are one of the more popular water tank liners, their intricacies of
application prohibit their use in pipes.
Epoxy
Epoxy is produced and applied as previously discussed in The Pipe Lining
And Coating section. Like vinyl, epoxy produces a hard, smooth surface that
exhibits low water permeability and strong adhesion to steel when properly
formulated and applied. Reformulation necessary to conform to the California
air pollution control regulations has adversely affected their performance.
While most reports on epoxy linings are for pipe applications, where flow re-
sistance co-effecients are important concerns, tests by Bethlehem Steel have
indicated that epoxy wears as well as hot applied coal tar enamel in water im-
mersion, non-abrasive, testing ( 17 ).
Other Mechanically Applied Tank Linings
Hot and cold applied wax coatings are also used to line water tanks.
These coatings are blends of petroleum waxes and oil based corrosion inhibitors,
These coatings may be applied directly over old rust or paint, but commercial
blast cleaning of the surface prior to application is preferred. Application
either by stiff bristled brushes or spray equipment is followed by torch
flaming used to smooth and thoroughly close the surface. The major disadvan-
tage to using wax coatings is a relatively short service life of approximately
five years.
6-6
-------
TABLE 36a
ESTIMATED CONCENTRATIONS OF COMPOUNDS DETECTED IN THE WATER
IN THE BAYOU CASSOTTE GROUND STORAGE WATER TANK
USING GAS CHROMATOGRAPHY/MASS SPECTROMETRY (19)
Compound
naphthalene
methyl naphthalene
biphenyl
acenaphthene
dibenzofuran
fluorene
phenanthrene/ anthracene
carbazole
bromoform
C, alkylchlorobenzene
indene
C alkylbenzene
anthraquinone
methyl benzofuran
quinoline
methyl styrene/indan/ indene
methylene phenanthrene/methyl
phenathrene
pyrene
2,5-diethyltetrahydrofuran
dimethyl naphthalene
fluoranthene
Sample Date and Concentration
9/6/77
A B
5.4 6.7
0.75 1.4
0.21 0.40
2.8 4.6
3.1 5.0
3.4 5.1
8.7 9.3
0.70 1.3
< 10 < 10
<50 <50
< 10 < 10
< 10 <10
<10 <10
<10 <10
< 10. < 10
-------
Metallic sprayed zinc coating is.a relatively expensive process where
zinc wire is melted, atomized at a high pressure, and sprayed onto the sur-
face. The application requires special skills and equipment, but the coating
provides excellent rust inhibition and a service life of 50 years. Surface
preparation by blast cleaning with one of three specified grits is mandatory.
Zinc-rich paints, containing 80-95 weight percent zinc dust, will pro-
vide a hard, abrasion and rust resisting surface on steel. The cost is high,
and surface preparation involves near-white blast cleaning.
Chlorinated rubber paints may be used when the control of fumes from the
application of other linings is difficult or where their use is specified as
in Baltimore County. Chlorinated rubber compounds are formed by exposing
natural rubber to chlorine gas. Because the resultant material is brittle,
plasticizers or linseed oil are added, producing a more flexible and adherent
coating. Application requires a near-white, blast cleaned surface and a
zinc-rich primer. The chlorinated rubber paint is then spray applied in
several coats. The intercoat adhesion is strong enough to essentially pro-
duce a single unit. This benefit provides easy surface preparation and re-
application when the orginal coating nears the end of its service life.
Service lives have been estimated as 6 to 15 years.
Asphalt based linings are absorbed by water faster than coal tar based
linings, and their application is generally limited to reclining existing
asphalt lined tanks or sealing "green" cement mortar coating in pipes to pro-
vide a proper curing environment.
CORROSION INHIBITORS
Corrosion protection by formation of a film on the surface of a pipe may
be achieved by chemical as well as mechanical means. The importance of
calcium carbonate solubility in this regard was recognized in the early
twentieth century by the German chemist, Till mans, and in the U.S. by Baylis
( 29 ). Excess calcium and carbonates would form scale in pipes, reducing
carrying capacities and increasing pumping requirements. CaCO? deposition in
rust forming waters, however, could form an effective rust inhibitor. The
need to understand the deposition-solubilization tendency of CaC03 formed the
basis of the Langelier Saturation Index, a parameter formulated in 1936 but
still used in corrosion control. When silicates and polyphosphates were found
to inhibit corrosion, their theory and applications were investigated. Today,
orthophosphates are also being studied, but corrosion control in many public
water systems is still engineered from CaCO, solubility data. A summary sheet
of corrosion inhibitors is presented in Table 36b.
6-8
-------
TABLE 36b. DOSAGES AND PREFERRED WATER PARAMETERS FOR CORROSION INHIBITORS
(see text for discussion and references)
Dosages Velocity pH Alkalinity Hardness Additional
halogen + sulfate
< '
CaC03 M1 t 6.8-7.3 > 40 mg/L > 40 mg/L (
d considerately stagnant , . «
_ ,. No specific ..... < 8.4
Sodium _,_ _ Must be . n ,, , ,
_.,. ^ concentrations _n . A lower pH low very low
Silicate 1n o o /i flowing . . . ,
usually 2-8 mg/L is desired
nu u ,. no/, < 7-° n n
Phosphates 1-2 mg/L , , , low low
velocity lead pipe
-------
CaCCU Precipitation
Early observations that carbonate minerals in a' water supply tended to
inhibit corrosion in steel and cast iron pipes led to theories on calcium
carbonate deposition as an anti-corrosive mechanism as well as several indicies
of corrosion that are still used in contemporary water treatment. Effective
CaC03 protection depends on the presence of anodic (metal oxidation) and
cathodic (oxygen reduction) reaction products in the water. Reduced oxygen,
as the hydroxyl ion, reacts with bicarbonate to form water and carbonate ions.
The increased concentrations of carbonate ion along with the (oxidized) metal
ions (Fe, Cu, Zn, Pb) subsequently exceed the solubility product of and, con-
squently, form metal carbonates ( 1 ) and carbonate containing solids. If
this deposit satisfactorily adheres to both itself and the pipe surface, an
effective corrosion barrier forms.
Of the carbonates formed, zinc carbonate forms a less porous structure
than other metal carbonates. Calcium carbonate, however, forms a solid,
though soft, coating that has good adherent properties. The protective
ability of a CaC03 intermeshes with existing hydrous ferric oxides and fer-
rous carbonate (32 ). Thus, the most effective use of CaC03 depends on
maintaining an environment where CaC03 is slowly formed, along with corrosion
products, eventually producing a hard, impenetrable coating.
Corrosion waters are generally of low pH and low alkalinity, often with
high C02 concentrations. The introduction of a carbonate to this system will
affect all three parameters because of the relationships:
Turn ~~i
L - 3-J- = KI' (eq. a_) First dissociation constant of
[H2C03] carbonic acid.
! fCO =1
— - — -^- = K2' (eq. Ib) Second dissociation constant of
[HC03 ] carbonic acid.
[H ] [OH"] = K' w (eq. c_) Solubility product of water.
[Alk]+[H+] = 2[C03=]+[HC03>[OH-](eq. d) [Alk] is in equivalents/liter all
other concentrations are molal.
Also, the calcium concentration may control carbonate by the solubility
reaction: J
[Ca++] [C03=] = K' (eq. e)
6-10
-------
Empirical results have shown that optimum conditions for calcium car-
bonate scale formation and protection are ( 43 ):
1. Calcium carbonate oversaturation of 4-10 mg/1.
2. Calcium and alkalinity concentrations of at least 40 mg/1.
3. pH range of 6.8 to 7.3.
4. Halogen plus sulfate/alkalinity ratio of less than 0.2, on a
mi Hi equivalent basis.
By knowing the pH, alkalinity, and calcium concentration, the saturation
condition of the water may be determined by using Caldwell-Lawrence diagrams
( 35 ). Additions of calcium and carbonate alkalinity needed to meet the
optimum criteria are calculated from these diagrams. An alternate approach
( 34 ), if adequate alkalinity exists in the water, involves empirically
determining the saturation pH of the water and adding lime to achieve that
value. A third alternative is to adjust the water to achieve a zero or
positive number on the Langelier Saturation Index (SI) ( 29 ), defined in
section 2. Saturation index correlations with observed corrosion and red
water problems have been good but not completely consistent (12 ). Main-
tenance of the SI within one unit of zero is considered satisfactory to
eliminate extensive corrosion problems (16 ).
Methods used to increase calcium, alkalinity, and pH in corrosive waters
depends on existing water constituents and raw material procurement costs,
including transportation ( 5 ). One method used with desalinization waters
is to blend the process stream with a naturally hard water. Because the
distilled water is deaerated and of low pH, this water may need to be aerated
and limed to add necessary oxygen and raise the pH.
When carbon dioxide is inexpensive, it can be added to react with either
lime or limestone to produce a calcium bicarbonate rich water. A concentrated
solution is prepared in a split stream and blended into the entire flow.
Additional lime may be needed to raise the pH of the water to obtain a zero
Saturation Index. The use of pulverized limestone, ground to 80 mesh, is
generally insufficient to produce effective stabilization because of the re-
quired contact time. A pH above 6.5 is seldom achieved, but this method is
attractive if iron removal is also needed. Filtration through partially
calcined dolomite will raise the pH to 8.0 to 9.0, but the cost, compared to
lime, makes lime more attractive except in very small systems where simplicity
of operation is paramount. In areas where alkalinity is >25 mg/1 as CaC03,
the preferred method is lime addition. Lime readily dissolves in water and
is generally cheaper than limestone on the basis of equivalents of calcium
( 5 ).
6-11
-------
Although the use of calcium carbonate corrosion control is widespread,
best results are still based on empirical data. While the optimum guide-
lines have been set, variables of temperature, velocity, dissolved oxygen,
and other dissolved solids also affect the process. A temperature increase
will decrease the calcium carbonate solubility, causing a given water to be
closer to CaC03 saturation conditions. Also, reaction rates will increase.
In solid-liquid reactions, a 10°C rise will generally increase the reaction
speed by a factor of two ( 29 ). Saturated water heated to boiling will
dissociate bicarbonate ions to the corrosive free C02 and hydroxyl ions
(23 ). Water velocity affects the boundary layer and hence the transport
of corrosive or protective material to the surface. In general, velocities
greater than one fps are needed to deposit a protective barrier. Higher
velocities provide better coatings (33 ). Because the strength of the
CaC03 film depends on the presence of ferric oxides, the maintenance of 5 ppm
dissolved oxygen is recommended (5 ).
Corrosion in lead pipes can be controlled by calcium carbonate deposition
( 38 ). Again, the stability of CaC03 is regulated by adding calcium and/or
alkalinity or adjusting the pH. If pH adjustment alone is used, enough hard-
ness must be present to provide the necessary calcium carbonate. Patterson
and O'Brien ( 38 ) have suggested that pH adjustment without sufficient
carbonate and alkalinity may be detrimental due to the preferential formation
of Pb(OH)2. This compound is non adherent and is potentially toxic as are
all lead carbonate and hydroxide solids. Also, intermittent monitoring may
not be sufficient to detect its presence. In general, soft waters, which are
not amenable to pH adjustment only, occur on the eastern seaboard and in the
southeastern and northwestern portions of the United States.
Dissolved solids other than calcium and carbonate species influence
the rate of CaC03 formation. An increase in salinity causes an increase in
the solubility of calcium carbonate (CaC03 is about 500 times more soluble
in sea water than in fresh water) and explains why brackish waters are
unable to form protective pipe coatings ( 29 ). Divalent ions (Mg , SOO
also increase the solubility of CaC03, probably due to the increased ionic
strength where fewer calcium and carbonate ions would become available for
interaction ( 15 ). Because of the change in concentration of ionic species
after some deposition has occurred, it is not uncommon to experience excessive
deposition in and near the treatment plant with lessened protection in out-
lying areas. It should also be noted that increasing calcium and alkalinity
too far beyond the saturation point may produce excess deposition (scaling).
This is not uncommon when alkaline waters are lime treated without recarbona-
tion. Also, where chlorination is used as a disinfectant, a pH increase means
a decreased HOC! (the most effective chlorine form in water) content. Thus,
lime addition for corrosion control may necessitate increased chlorine doses
for effective disinfection.
Lime is available commercially as quicklime or hydrated lime Quicklime
is purchased in granular form and contains at least 90 percent CaO, magnesium
6-12
-------
oxide being the primary impurity. Hydrated lime is a powder containing
approximately 68 percent CaO. Both forms of lime are dissolved into a 5
percent slurry prior to addition to the water system. Alkalinity is often
applied in the form of soda ash (Na2C03), a greyish-white powder containing
about 98 percent sodium carbonate.
Lime addition requires the use of a slaker for quicklime while hydrated
lime is often prepared in a tank with a turbine mixer. The slaker uses a
gravimetric feeder to introduce the lime into a mixing chamber, where the
proper amount of water is added to produce the desired slurry concentration.
The ratio of water to lime is about 5 to 1, and pH controls that compensate
the water to lime ratio in relation to changes in water quality or lime
purity are available. Regulators protect against excessive temperatures
resulting from the chemical reaction:
CaO + H20 •* Ca(OH)2.
A grit remover is used to remove coarse material prior to the solution being
pumped to the slurry feeder. The feed rate into the water system is con-
trolled by an automatic pH and flow control system (52).
One carbonate process that is gaining popularity is the low-carbonate
method. Recently implemented in Bennington, Vermont, and soon to be used
in Seattle (see section on Case History), this method raises the pH of the
corrosive water to 8-8.3. Adding small amounts of alkalinity causes the
formation of insoluble carbonate salts on pipe surfaces. Effective cor-
rosion control is achieved without drastically altering the chemical make-
up of the water. Results in Bennington showed an 82 percent corrosion re-
duction in lead and an 80 percent reduction in copper pipes ( 39 )• This
method is also effective with galvanized surfaces.
Sodium Silicate
Sodium silicate, originally used to prevent metal solubilization in
lead pipes, has been used for rust inhibition for over 50 years ( 48 )•
Water conditioning by addition of sodium silicate is primarily applicable
to galvanized piping where it is used to curtail existing red water problems
and prevent corrosion, especially in hot water pipes. Sodium silicate is
non-toxic and has the ability to control pre-existing conditions.
Perhaps the first consideration for using sodium silicates for corro-
sion control in an entire utility was to prevent the solution of lead from
lead pipe in 1922. It was noted in this work that the corrosion of ferrous
materials was reduced ( 48 ). Since that time, the use of sodium silicates
has been attempted with some success over a wide range of applications ( 44 )
The mechanisms of corrosion control with the addition of sodium silicate
is thought to be the formation of a two-layered protective film between the
6-13
-------
material surface and the corrosive water. From results of laboratory studies,
Lehrman & Shuldener determined that in dilute sodium silicate solutions,
silica will exist in equilibrium between its ionic and collodial states
( 30 ). At low concentrations, such as those required for corrosion control
in potable water systems, this equilibrium is attained very rapidly, if not
instantaneously. The presence of silica in a zinc/iron system tends to make
the zinc anodic to iron and the zinc reacts with water causing the formation
of zinc hydroxide. The positively charged zinc hydroxide reacts with the
negatively charged collodial silica from the water and silica is absorbed
producing an amorphous silica (gel) deposit or layer capable of enmeshing
compounds of iron, calcium, magnesium, and organic matter above a layer of
metal corrosion products ( 30 ). In low alkalinity waters, this deposit
is basically hemimorphite [ZnItSi207(OH)2 H20] with some ferrous or calcium
silicate depending on whether the water is soft or hard (28 ), ( 11 ).
In waters of higher alkalinity, especially at higher temperatures, Shuldener
& Lehrman report that the zinc-iron potential can be reversed and the iron
becomes anodic to the zinc causing the formation of ferric hydroxide (45 ).
Although not as effective as zinc hydroxide, ferric hydroxide will also re-
move silica from water forming the amorphous silica layer.
Shuldener & Lehrman explain that when small amounts of silica are added
to systems containing bicarbonate, especially at higher temperatures, two
competing reactions can occur ( 45 ). The presence of silicate tends to
make the zinc anodic to iron and the bicarbonate tends to do the reverse.
The zinc corrodes to form a zinc hydroxide corrosion product which adsorbs
silicate. As the silicate is removed and the concentration lowered, its
effect is lowered and iron corrosion products are formed by the influence
of the bicarbonate concentrations. Because the iron is not as effective
at removing silica and because the silica concentration is now much lower,
some rust can appear.
Shuldener & Lehrman conducted laboratory studies to investigate temp-
erature, pH, and bicarbonate effects on the effectiveness of silicate for
corrosion control ( 45 ). They simulated a galvanized pipe carrying a
water at approximately 1.9 fps by rotating a pipe in water solution through
which C02-free air was bubbled. They determined that the corrosion rate,
as observed from rust formation, increased as the temperature increased from
72°F to 160°F, but that the addition of sodium silicate in the solution would
act as a corrosion inhibitor and decrease the corrosion rate. From the
laboratory studies on pH effects, Shuldener & Lehrman observed that in a
solution containing 8.5 mg/1 Si02 and 14 mg/1 bicarbonate, decreasing the
pH from 8.5 and 8.0 to 7.5 and 7.0, respectively, resulted in reducing the
corrosion rate. Similar results were obtained in waters containing 8.5 mg/1
Si02 and 54 mg/1 bicarbonate when the pH was decreased from 8.5 and 7.0 to
6.5. PH adjustment was with dilute h^SO^ but specifics of the pH measurements
in relation to the heating of the water or temperature corrections were not
presented.
6-14
-------
From their experiments it was shown that the presence of bicarbonate
has an accelerating effect on the rate of corrosion. As is noted, as the
pH increases in the range of 8 to 9, the concentration of bicarbonate de-
creases which implies that the effect of bicarbonate is closely associated
with the pH of the solution ( 45 ). From their studies they concluded that
bicarbonate has a very strong influence in causing reversal of potential with
a resulting rust formation at 160°F (45 ).
Lane et al5 conducted laboratory tests to investigate the effects of pH
at various silica levels on the corrosion of galvanized steel at a domestic
hot water temperature of 140°F. Their tests were conducted using a water with
a pH range of 7.2 to 9.0, a hardness concentration of 100 mg/1 (as CaC03) an
alkalinity concentration of 35-55 mg/1 and a velocity of 1.5 fps. Silica
levels were varied from 10.0 mg/1 to 20.0 mg/1. From results of their
experiments, they determined that initially, the higher concentration of 20
mg/1 silica was more effective than the lower concentration of 10 mg/1 silica
for controlling corrosion. However, after a sufficient period of time
(possibly 60 days), Lane et al concluded that the effects on corrosion control
is essentially the same at both levels of treatment indicating that there is
no advantage to using the higher application rates ( 28.). In both cases,
-•owever, corrosion control was most effective at the higher pH (8.5-9.0)
condition. From the results of their tests, Lane et al state that the
optimum silicate treatment appears to be a complement of pH and silica
dosage, i.e., lower pH requires higher Si02 dosage and vice-versa, and
is apparently influenced by water quality factors as calcium, magnesium,
alkalinity, chloride, sulfate, and pH ( 28 )• Their experiments did not
attempt to evaluate the effects of these water quality factors.
Sodium silicate dosages are independent of naturally occurring silica in
the water and there are no specific concentrations recommended for the various
conditions of water quality (51). As a general rule, however, an average con-
centration of 2 to 8 mg/1 and possibly up to 12 mg/1 Si02is sufficient to
maintain corrosion control in a system once a protective film is established.
This inhibitor has been found to be particularly useful in waters with very
low hardness, alkalinity, and pH < 8.4, and is more effective under higher
velocity flow conditions. The application of sodium silicate requires the
use of solution feeders, small positive displacement pumps that deliver a
specific volume of chemical solution for each piston stroke or impeller
rotation. The two general types of solution feeders are diaphragm and plunger
metering pumps, although some rotary pumps may be used as positive displace-
ment pumps as well. Both the diaphragm and plunger pumps may be controlled
manually by adjusting the stroke length or rate of reciprocation, or they may
use an automatic control unit that regulates stroking in proportion to water
flow (52).
6-15
-------
Inorganic Phosphates
Polyphosphates have been used to control calcium carbonate scale build-
up in water treatment plants and irrigation systems since the 1930's. Pre-
sently, both polyphosphates and orthophosphates may be used for scale
inhibition or corrosion control. The effectiveness of these chemicals is a
function of flow velocity, phosphate concentration, temperature, pH, and
calcium and carbonate levels.
Both ortho- and polyphosphates (especially tri-, pyro-, and higher
polyphosphate anions) are known to form complexes with a number of metal ions,
including calcium, iron, and lead ( 49 ). The presence of larger amounts
of polyphosphates has been blamed for assisting iron and lead uptake from
pipe surfaces in laboratory situations by upsetting local metal-ion equil-
ibrium through complexation. When concentrations generally less than 10 mg/1
of polyphosphates are introduced into a flowing water, however, they have a
tendency to form a thin film on the metal pipe surface that protects the
metal from further corrosion. This film will also adhere to calcite crystals,
preventing further growth. In this way, a water supersaturated with calcium
and carbonate can be prevented from scaling. The minimum concentration of
polyphosphates that will prevent crystal formation for a particular water is
called the threshold level. Subthreshold concentrations will produce dis-
torted calcite crystals ( 33 ).
The threshold concentrations of polyphosphates, usually added as sodium
hexametaphosphate, is pH dependent in calcium and carbonate containing waters
due to the pH dependency of CaC03 deposition. Successful red water control
has been achieved in waters that had unsuccessfully responded to lime teat-
ment. In Little Rock, Arkansas, Uniontown, Pennsylvania, and Newport, Rhode
Island, polyphosphate addition to lime treated waters with initially high
pH (>9.0) resulted in corrosion abatement at polyphosphate concentrations
of 1.0 to 2.0 ppm. However, the red water problems in Nitro, West Virginia,
persisted in the lime treated water even at 10, 5, and 1 ppm polyphosphate
concentrations.
Phosphates may be used to control corrosion in lead pipes, but because
corrosion control in lead pipes is primarily interested in preventing lead
being carried into the water as opposed to the pipe deterioration, the
phosphate concentration must be low enough not to form soluble lead complexes
( 21 ). Corrosion protection by addition of sodium hexametaphosphate is
best a pH <_ 7.0 ( 9 ). Above this pH, and in the absence of phosphates, a
lead-carbonate film will deposit on the surface and protect the pipe. If
polyphosphates are present, the formation of this film is retarded and more
lead may enter the water than would occur if either the addition of poly-
phosphates at low pH or the existence of a high pH alone were maintained.
Effective corrosion control in lead pipes at pH below 7.0 has been achieved
6-16
-------
with sodium hexametaphosphate concentrations at or below 2 mg/1 ( 21 ).
These experiments also showed that lead levels at pH 8.9 were equivalent
with or without polyphosphate addition and lower than those achieved at pH
7.0 or below with polyphosphate addition at low alkalinity.
Other pipe surfaces respond differently to polyphosphate corrosion con-
trol ( 9 ). The staining of plumbing fixtures due to the dissolution of
copper or the components in brass has been lowered but not stopped by poly-
phosphate addition.
Aluminum protection has been obtained in laboratory tests, especially
at low pH values. Zinc corrosion has also been controlled by 2 ppm poly-
phosphate addition to the water. At least two separate cases have been
reported where the protection of galvanizing was excellent. Specifics on
these tests, however, were not given ( 9 ).
One report of copper corrosion control by hexametaphosphate addition
( 6 ) reported that early morning copper concentrations were reduced from
an average 4.8 to 5.5 ppm to an average of 2 ± 0.2 ppm copper. The pipe
was 12 to 14 month old house lines in a district in England. Initial com-
position of the water was not reported. Hexametaphosphate addition relieved
complaints of blue water and emetic reactions to use of the water that had
stood in the pipes overnight.
A process involving orthophosphate addition has been developed (36,
27 ). The process involves the formation of a zinc orthophosphate, Zn3
(P0i»}2, in the pipeline and its subsequent precipitation as an insoluble
zinc orthophosphate film. This reaction involves adding zinc sulfate,
sulfamic acid, and monosodium orthophosphate to the water:
3ZnSO^ + 2HNH2S03 + 2NaH2PO,t -»- Zn3(P002 + 2NaNH2S03 + 3H2S04
Because zinc solubility, even in the presence of POiT, is pH dependent, an
initial application of zinc is empirically maintained at 2 to 3 ppm.
After the coating is formed, in about 3 weeks, the zinc is reduced to
1 ppm.- In highly alkaline waters, between 0.5 and 1 ppm sodium hexameta-
phosphate may have to be added to prevent CaC03 precipitation. This is be-
cause CaC03 weakens the Zn3(POit)2 film, decreasing its anti-corrosive ef-
ficiency. The results of one test using this procedure is presented in
Tables 37 and 33.
6-17
-------
TABLE 37. ANALYSIS OF A NEW JERSEY WATER
hardness, ppm 96
sulfate (SOf), ppm 44
pH 7.0
M.O.alkalinity, ppm 52
TDS, ppm 150
Langelier Index 1.1
TABLE 38. 63-DAY CORROSION TEST ON CAST IRON
USING TREATED NEW JERSEY WATER
Coupon
Control
Treated
Control
Treated
Exposure
(days)
31
31
63
63
Corrosion Rate
(mpy)
10.20
8.24
10.77
5.24
% Corrosion
Reduction
19
51
Source = (27),
Other recent tests ( 26, 38 ) on the ability of zinc orthophosphate
to prevent corrosion in lead pipes has presented negative results. Karalekas
et al added zinc orthophosphate to Boston water (low alkalinity and hardness,
pH < 7.0) for a six-month period and could not reduce the lead levels to the
0.05 mg/1 standard. Patterson and O'Brien immersed lead coupons in stagnant
water and determined that lead corrosion could increase from the use of zinc
orthophosphate. The pH conditions during their short term test varied from
7.24 to 9.13. The long term test pH was held between 6.52 and 6.82 and
showed a 60% increase in soluble lead content in the treated water. Murry
( 36 ) reported that zinc orthophosphate is insoluble in water so his pro-
cess was the in situ formation of zjnc orthophosphate during flowing con-
ditions. However a Ksp of 1 x 10" has been reported for zinc orthophosphate
( 4 ) and others ( 18 ) have reported that orthophosphate can substantially
reduce lead levels in waters of pH 7-8.2 with low alkalinities and carbonate
levels. Because of these inconsistent findings, more researcn is needed on
this corrosion control mechanism.
6-18
-------
Other organic and inorganic phosphate-phosphonate corrosion inhibitors
are used for industrial scale control but are not applicable to potable water
systems. Also, the possibility of eutrophication in lakes and streams that
ultimately receive the water has discouraged at least one city from using
inorganic phosphates for corrosion control. The critical level for algal
growth has been established as somewhere near 0.01 mg/1 ("42 ), and it was
believed that the wastewater treatment plants might not be able to reduce
the phosphorous content to this extent.
In summation, several important factors must be addressed prior to the
use of inorganic phosphates for corrosion control in any specific water
system. The effectiveness of polyphosphates, usually added as sodium hexa-
metaphosphate, is greater with increasingly turbulent velocities, and flow
velocities of 2-5 fps and above are not continually experienced in all parts
of any distribution system (51). Under stagnant or nearly stagnant condi-
tions, such as service lines, polyphosphates will not be effective. In waters
saturated with calcium carbonate, polyphosphates will prevent CaC03 deposi-
tion, and larger amounts of polyphosphates may assist in iron and lead concen-
trations through complexation. Polyphosphates have been shown to alleviate
"red water" in many, but not all, situations where the iron content was
originally in the water and where the iron was a result of pipe corrosion.
Polyphosphates can also prevent lead release into the water by forming a thin
film of complexation compounds across the lead surface. Effective polyphos-
phate concentrations reported are usually ^ 2 ppm, but an initial concentra-
tion of up to 10 mg/1 for up to three weeks has been used to produce an
initial film. Orthophosphates have produced both positive and negative
results for lead control and more clarifying research is needed.
Polyphosphate addition is by use of solution feeders as discussed in the
silicate section. Concentrated solutions are mildly corrosive, so stainless
steel is often used for the dissolving tanks, baskets, and the liquid end of
the chosen pump. The positive displacement pump(s) are generally selected
from accuracy, durability, capacity, corrosion resistance, and pressure
capability requirements, and most manufacturers have designs to accommodate
the specific need.
Miscellaneous Methods
Other corrosion control methods are applicable to more specific sit-
uations. In instances where high sulfate concentrations are producing cor-
rosive sulfides, which tend to accelerate the anodic dissolution of iron
( 43 ), the sulfate reducing bacteria responsible for the sulfides may be
eliminated by maintaining proper chlorine residuals in the water system.
This problem should not occur under normal conditions where disinfection
procedures and natural aerobic conditions occur.
Dissolved oxygen removal will curtail corrosion. Deaeration may be
accomplished by applying a pressure vacuum, heating and degasification, or
6-19
-------
deactivation. Deaeration is too expensive to be considered a viable cor-
rosion control for municipal water systems.
Cathodic protection, previously a common corrosion inhibiting method
in water tanks, occurs when electrons are supplied to the metal surface being
protected, decreasing that metal's tendency to oxidize. Electrons may be
supplied by a direct current or by a sacrificial anode. The latter consists
of a metal higher in the electromotive series than the protected metal and
is, therefore, preferentially corroded. Examples of a cathodic protection
are magnesium rods in hot water heaters and the zinc used to galvanize
steel.
ECONOMICS
The economics of corrosion abatement is often simplified to a comparison
of annual costs incurred by implementing additional chemical treatment to the
water supply and/or the increased cost of pipe lining versus the annual costs
for replacing distribution pipes ( 46 ). Increased pumping costs due to re-
duction in the hydraulic efficiency of corroded piping and increased costs
due to shortened pipe or appliance life are sometimes, but not always, in-
cluded in cost analysis. However, health and aesthetic effects and possible
increased cost due to excessive deposition (scale) are often ignored.
Benefit/Cost Analysis
In Benefit/Cost analysis, a price is determined for each aspect of cor-
rosion control and the sum of the annual costs for imlementation of the abate-
ment method is compared to the sum of the annual "lack of costs" experienced
by the utility or by the user and the utility. Costs for implementation of the
corrosion control are presented in the next sections and reflect construction,
materials, chemicals, and labor cost as well as intangibles such as public
acceptance and relative economy of the area. Benefits are harder to quantify,
especially in terms of pricing the "lack of costs" to be incurred by the con-
sumer.
User benefits may be classified as directly economic, aesthetic, or in
terms of health. The directly economic savings results from extended service
lives of piping, fixtures, and appliances, as well as lessened damage from
water leaks, and are generally formulated by first determining the extent of
corrosion-based annual costs and subtracting annual costs for similar repairs
experienced in similar areas practicing corrosion control. Methods used to
determine the extent of problems caused by corrosive waters include a review
of customer complaint records, the use of questionnaires, corrosion rate
testing, examination of plumbing pipes, and water quality monitoring in re-
sidences. From data collected, a service life for waterheaters, pipes,
faucets, shower heads, appliances, etc. is determined, and annual cost for
maintenance and repair or replacement is figured. Unfortunately, the prices
determined are applicable only to the individual distribution system samples
6-20
-------
supplying the data as results from similar areas using the same water source
have produced drastically different costs ( 8 ). Also, attempts to cor-
relate water quality parameters with corrosive effects in terms of economics
have been poor ( 37, 8 ). One study ( 37 ) tried to correlate the cor-
rosive potential with the Saturation Index, the Ryznar Index, and the
chloride plus sulfate concentration and ended up finding the greatest cor-
relation (r = 0.33) between IDS and houses that had to replace interior
piping. No positive correlation was found between IDS and replacement of
faucets, shower heads, or toilet flushing mechanisms. Whereas that study
assessed a penalty cost of 18
-------
TABLE 40
(1978'Prices)
Source (25)
COMPARISON OF* ESTIMATE!) ANNUAL COSTS AND SAVINCS
OK COIWOSKIN KI.I.ATEI) DETERIORATION
WITH AND WITHOUT COHItOS [UN CONTROL TREATMENT
lU'i. lilenl i.i 1 Call-Bury
Slni'.lf E-'»'l ly & Duplex
A 1 1 tin 1 1 s
Units Imllt 1971-75
All Hulls
llnll.-i IMI! It I9/I-/5
MiiUI-.!'.;»'i.[ly
All Units
Unllu Inil It I97I-/5
All Units
Units bul It 197 \- 1 5
I'lnmMng
Typo
C.ilv.-inUcd
SlUL'l
Calv.-inl7.cd
Stcul
Cn|ipci
Ctl|)|llT
C.i Iv.mlzud
S lv ul
f.'il Ivilll Izcd
Stud
Copper
Cupper
Number uf
Units
62,051
166
l/i. IUJ
6b/i
1,0. 700
t, 10
16.812
1.720
SOUTH AUKA
CKDAK UIVER
list hll.HC'd
Annuul Cunt/
Unit W/0
Treat men I
$39.35
7.60
).'J5
7.05
25. JU
*;.
4.70
2.90
/i . 90
KxpiTicil Annual
Coal S.w In|',3/Ucil(
I'rov lth:il Uy
Truiilinunt
$IO./iU
6. J5
J.'JO
6.95
11.75
) . 90
2.B5
',.H5
Expetr t etl
Annun 1 Cost/
Unit W/
Tl eacinjuL
$28.95
1.25
0.05
0. 10
10.55
O.HO
O.U5
0.05
NORTH ARLA
TOUT RIVKK
Number ul
Units
52,551
176
15.151
HUO
I2.9H8
152
8.720
760
Eut limited
Annual Coat/
Unit W/0
Treatment
$48.50
17.10
5.10
/. 10
32.90
12.45
J.20
4.80
Expected Animal
Cost SdVlngs/Unli
Provided By
Treatment
$7.20
7.55
4. U)
6.70
6.45
6.45
3.15
4.60
Expected
Animal Cost/
Unit U/
Treatment
$41.30
9.55
0.50
0.40
26.45
6.00
0.50
0.20
-------
Health and aesthetic problems resulting from corrosive waters are de-
finable but hard to quantify economically. Aesthetically, iron concentrations
may present turbidity and color (red water) problems; the drinking water limit
for zinc is based on taste tests; dissolved copper may cause a blue green
stain on porcelain where soap residues accumulate; and manganese may cause
brown-black staining. In terms of health, lead and cadmium may be toxic in
low levels, and other heavy metals may be leached from pipes and solders.
Furthermore, causes of the correlation between increased cardiovascular
disease and soft waters have not been positively identified but may result
from either the role of the bulk mineral content (Ca, Mg, Na) of hard waters
or the metal constituents in soft waters resulting from its corrosive nature
( 24 ). While cost figures have not been placed on the aesthetic problems,
it is estimated that the difference in cardiovascular mortality, between
corrosive and non-corrosive waters is 52 deaths/100,000 population annually,
which translates to 47,000 avoidable deaths attributable to a lack 'of cor-
rosion control, and an income loss to the U.S. of $3 billion annually ( 24 >.
Trends and Costs of Mechanically Applied Linings and Coatings
Cast iron and ductile iron pipe, composing over 76% of the pipe used
in U.S. water supply distribution systems serving over 2,500 people, is ex-
clusively lined with cement mortar. Steel pipe, at 6%, is increasingly being
lined with epoxy. Although epoxy is more expensive than coal tar, the benefit
of a high smoothness coefficient, a lack of having to control coal tar en-
amel application fumes, and possible adverse health concerns arising from
the use of coal tar based linings, have caused this increased epoxy usage.
There are two manufacturers of hot applied coal tar enamel for potable water
system lining in the United States with total 1976 coal tar enamel sales of
$1,500,000. One company estimated that half of the steel pipes installed by
the water supply industry were coal tar enamel lined.
The extent of use of various linings and coatings for water tanks is not
well defined. The extremes of material use estimates reported by various
vendors and manufacturers is shown in Table 41. As can be seen, a wide
range of discrepancy exists.
6-23
-------
TABLE 41.
EXTREMES OF VARIOUS ESTIMATES OF
USE OF TANK LININGS AND COATINGS
EXTREMES OF ESTIMATIONS
(10 FIRMS REPORTING)
OF % OF NEW
WATER TANKS LINED WITH A
COATING
Vinyl
Epoxy
Coal Tar
Other
GIVEN COATING
47.5 -
6
0
0
NATION-WIDE
90
50
90*
17
* 0-19% reported on all estimates except one firm indicated 90% (19)
Besides manufacturer's preferences, regional preferences exist in re-
lation to water tank linings as seen in Table 42.
TABLE 42.
REGIONAL PREFERENCE FOR WATER TANK LININGS
(compiled from (19) )
PREFERRED LINING/SPECIFICATION
CITY/STATE FOR STEEL' TANKS
Seattle ^100% Coal Tar Enamel
Atlanta Coal Tar Enamel
Houston Vinyls and Epoxys Used
Virginia Coatings containing coal tar,
vinyl or bituminous material not
approved.
Baltimore Chlorinated Rubber
California No longer uses coal tar enamel due
to possible health affects
6-24
-------
The cost of applying linings to water tanks varies widely and is a
function of the application method, surface preparation, cost of labor, and
intangibles such as regional competition and local materials preference.
Table 43 presents the extremes of cost estimates for the linings discussed
in the text.
Costs of Corrosion Control by Chemical Applications
After the chemical constituents of a corrosive water have been identified,
and potential control mechanisms are listed, the cost of treatment for each
method must be determined. Because local construction, materials, chemical,
and labor costs, along with intangibles such as the local economy and water
quality demands, influence cost estimating, this section should be regarded
as a general overview. Where possible, weight units are used because prices
quoted are apt to reflect temporal and regional values that may or may not
differ from current prices. Brand names used are those reported from the
articles cited and do not constitute an endorsement of the product.
Sodium silicate has been effective in reducing plumbing repair costs
when used to treat water used in groups of buildings or housing developments
( 44 ). This additive is more cost effective when used in very soft, low
alkaline waters, especially when compared to the use of lime, sodium hydroxide,
and/or soda ash. To obtain the 8 ppm silica concentration regarded as the
minimum level needed for corrosion control, a 28 ppm sodium silicate solution,
equivalent to 232 pounds of solution per million gallons of water must be used
( 48 ). The cost of equipment needed is that of a solution feeder. Prices
vary depending on the size, type (diaphragm or plunger metering pumps), and
manufacturer; and selection is usually based on accuracy, durability, cap-
acity, corrosion resistance, and pressure capablity.
The cost of using polyphosphates is seen in a 1966 study by the city of
Richmond, Virginia. They compared the use of lime, (Jalgon brand metaphosphate
with lime, and TG-10(a Calgon phosphate composition) with lime to stabilize
their corrosive waters (7 ). Empirical tests coupled with manufacturer's
data indicated that corrosion would be controlled using either 2.0 ppm Calgon
or 1.25 ppm TG-10, each with a reduced lime consumption. An additional cap-
ital cost ($3000) would be required for the polyphosphate addition and the
1966 costs for the use of TG-10 plus lime, Calgon plus lime, and lime would
have been $3.08, $2.35, and $1.00 per MG respectively. The report recommended
the use of TG-10 based on lower pipe line maintenance costs, reduced valve
and meter repair costs, and lower pumping costs from increased pipe diameters
expected to result from reduced scale formation.
The additional equipment required for the TG-10 and Calgon systems in-
cluded a i HP stainless steel Simplex pump (0.0-0.5 gal/min), two 400 gallon
stainless steel tanks, two 5-cubic foot stainless steel dissolving buckets,
two agitators (i HP motor, 3-three blade propellers), one Simplex pump
assembly, two electric timers, and one switch. Materials, based on a 40
MGD flow, were 2000 Ibs lime; 400 Ibs lime plus 717 Ibs Calgon; or 400 Ibs
lime plus 440 Ibs TG-10 to be used daily.
6-25
-------
CTl
I
CTl
TABLE 43
COST ESTIMATES (19)
SURFACE
PREPARATION
COATING
Coal Tar Enamel
(Hot)
3 Coat Vinyl
Metal 1 ized Zinc
li Coat Vinyl
Chlorinated Rubber
Coal Tar Paint
(Cold Appl ied
Coal Tar Epoxy
2-Component Epoxy
Asphal t
(Petropoxy)
Wax
1 -Component Epoxy
Asphal t
(Inertol #1)9)
Phenol ic
Zinc Rich Paint
Initial costs are as
$/Sq.
.1)0- .
.1)7-1.
.80-1 .
.21)-!.
.30- .
.30- .
.30- .
-30- .
.50-1.
.21)- .
.30- .
.30- .
.80-1.
reported and
Estimates were obtained from publ
from the 33rd edition
of Bu i 1 d i ng
Ft.
50
00
65
00
75
75
75
80
00
12
80
75
75
00
may
i shed
MATERIALS
$/Sq.
.10-.
.11-.
.09-.
.05-.
.13-.
.15-.
.15-.
.06-.
.10-.
.1 I-.
not i nc 1 ude f
1 i terature ,
Construction Cost
Ft.
15
22
50
20
17
25
20
50
21)
01)
25
1 1
13
25
i gures
pa int
Data,
APPLICATION
$/Sq.
-30-
.16-
1.15-1
.09-
.15-
.25-
.16-
.30-
.16-
.16-
.05-
Ft.
.60
.50
.75
.21
• 50
.30
-30
.70
.50
.29
.50
.30
-50
.15
reported in surface
supplies, engineering
1975, Robert
SERVICE
INITIAL COST* LIFE
$/Sq. Ft.
.911-2
.95-1
2.1)5-3
.55-1
.50-1
.73-1
-75-
.90-1
1.2ll-l
.35-
.1)7-1
.70-
.77-1
1.10-1
.00
.11
.90
.21
.11)
.30
.95
.95
.31)
.k5
.50
.91
.11
.1)0
preparation, material, or appl
consul tants ,
S. Means Company, and
YEARS
20-50
8-23
1)0-50
8-23
6-15
12-15
10-11)
10-15
15
k- 5
10
6
i)- 6
5
i cat ion
and painting contractor
the Estimating
Gu i de of
COST
EFFECTIVENESS
$/Sq. Ft./Yr.
.019-
.Qli8-
.Qii9-
.053-
.055-
.055-
.068-
.060-
.083-
.088-
.086-
-117-
.185-
columns .
.100
.139
.098
.I'll)
.190
.083
.075
.115
.089
.090
.115
.152
.193
.220
costs were calculated
the Painting and
Decorating Contractors of America^ 10th edition, 1977-To.
-------
Relative costs of calcium carbonate stabilization are presented in a
1971 report on the pacification of product water from a distillation plant
( 5 ). Costs of adding 40 ppm (as CaC03) of calcium bicarbonate to the
water are presented for the use of lime (CaO) and C02, limestone and C02,
calcined dolomite and C02, and limestone and H?SOi» for both 10 MGD and 50
MGD flows. The capital costs of the storage bins, dust filter, slaker unit,
mixers, slurry feed pump, recarbonator, aerator, instruments, etc. were
$110,000 and $357,000 for the 10 MGD and 50 MGD flows and salary related
costs were $27,000 and $68,000 per year, respectively. These costs were
essentially the same for all processes investigated. Chemical prices cited
were a reflection of shipping distances, especially in relation to the tran-
sportation charge of powered limestone, and the method of contracting for
liquid C02 that may vary the C02 price by a factor of two. The 1971 prices
given were:
Lime $30/ton
Limestone 15/ton
Calcined Dolomite 150/ton
Liquid C02 50/ton
H^ (93%) 23/ton
Chemical costs in 1971 dollars per 1000 gallons are presented in Table 44.
TABLE 44
CORROSION CONTROL BY CALCIUM CARBONATE STABILIZATION
COST CAPITAL AND
LABOR CHEMICALS TOTAL COST
ln Mnr rn Mrn
GAL 10 MGD 50 MGD 10 MDG 50 MGD
Lime and C02
Limestone and C02
Dolomite and C02
Limestone and H2SOi»
.89
.89
.89
.89
.45
.45
.45
.45
1.12
.82
.95
.98
2.01
1.71
1.84
1.87
1.57
1.27
1.40
1.43
from: (5) (1971)
6-27
-------
Although these costs reflect 1971 prices, it is significant to note that in
instances where limestone is half as expensive as lime and one tenth as ex-
pensive as calcined dolomite, and 93% H2S04 is half as expensive as liquid
C02, all on a weight basis, limestone and C02 would be the preferred process
( 5 ). Furthermore, a considerable savings would be realized if split
stream treatment were used on part of the flow followed by blending with the
remainder of the water.
Because of their applicability to calcium carbonate stabilization, cost
curves ( 20 ) for lime feed systems and recarbonation via liquid C02 systems
are included in this section. Construction costs for the lime feed system
(Figure 36 ) are based on hydrated lime use up to 50 Ib/hr and quicklime use
at higher rates. The hydrated lime arrives in 100 Ib bags, is introduced by
feeder to a dissolving tank, and gravity fed to the point of application.
The quicklime is stored in hoppers with a 30 day storage capacity (3 days if
recalcination is used) located over the slaker. The slurry is gravity fed
to the point of application. Operation and maintenace cost (Figure 37 ) do
not include the price of lime and are based on $0.03/kw hour and $10/hour
labor costs. Construction costs for liquid C02 system (Figure 38 ) include
a 10-day storage tank, C02 vaporizer, a solution-type C02 vaporizer, a sol-
ution-type C02 feeder, injector pump, main header, diffuser pipes, and an
automatic control system. Operation and maintenance (Figure 39 ) are based
on $0.03/kw-hr. and $10/hr. labor costs. Costs on these curves are based on
October, 1978 prices.
Estimates of the national cost of piping damage resulting from the dis-
tribution of corrosive waters range from $210 million annually (1975 dollars)
(47 ) to $375 million annually (1976 dollars) ( 24 ). Estimates for the
cost of implementing lime-addition corrosion control to stabilize these cor-
rosive waters are from $20 million ( 47 ) to $27 million ( 24 ) annually.
This simplified economic analysis does not account for the intangible benefits
of elimination of red water, possible decreases in CVD mortality, or lessened
pumping costs; yet, juding from the economics of pipe replacement alone, well-
run utilities with corrosive waters can ill-afford not to implement control
methods.
6-28
-------
9
8
7
6
5
4
1
2
!
6
5
4
3
2
I.OOO.C
9
s
6
5
4
-"»• 3
1
£ ?
v> e.
O
O
§ 100,0
•- 1
o ?
2 I
o: o
|- K
w 5
0
o ^
IO.OC
CO
00
^
/
0
J
^r
•*"
rf •
^ ^
i
— — ^
.-^^*^
0 ^"
WITHC
REGAL
Si«-
JT
CIN;
S2
L
T
"I
IV E
on
• ••
' 1^~~
UMT"LJ
WITH
RECA
T^-
i i
LI
uCI
«5S
A-C
VIC.
A
-• —
TJf
^ _
i
i
!
•*•*
10 2 3 4 5 6 789100 2 3456 7891000 2 3 456789
LIME FEED CAPACITY - Ib/ hr 10,000
10 100 1000
LIME FEED CAPACITY - kg/hr
Figure 36. Construction cost for lime feed systems (20).
6-29
-------
CO
o
o
3
7
6
5
4
3
2
100,
1
6
5
4
3
2
10,00 (
9
8
7
6
5
4
3
2
1000
9
8
7
6
5
4
3
: 1
6
5
4
3
2
000 10
9
8
7
6
5
4
3
2
3 10,0 (
9
8
7
6
5
4
3
W
£
IOOC
1 ft
- i 9
- DC 9
°> *
0
- _) 5
4
3
100
3,000
DO
X
s
)
^x
x^
X
^1
J^^
x^
^ X
. r
X
^
*
•r
•^
^
^
0*
+
X **
^^
^ " TO'
^^
, -^"^
<^
AL
^-^
LA
^
\
4*
01
*•
X*
:osr
^
o**
*
^
10 2 3 4 56783100 2 34 56789OOO
LIME FEED RATE- Ib/hr
+
3456 789
10,000
10
1
100
LIME FEED RATE-kg/hr
1000
Figure 37. Operation and maintenance requirements for
lime feed systems - labor and total cost.
(20)
6-30
-------
e
1
7
6
5
4
•s
2
6
5
4
3
2
1,000, OC
9
e
6
5
•*» 4
V)
O 2
Z
2 100,00
o S
o 5
o ^
10,000
)0
0
4
^
^ •*
_Mf^
r**^
T
>^^
,X
^
100 234 567891000 234 56789»OpOO 234 56789
100 OOO
INSTALLED CAPACITY - Ib C02/dy
1 • 1 1 1
160 1060 10,600
INSTALLED CAPACITY- kg C02/dy
Figure 38. Construction cost for recarbonation - liquid C02 source. (20)
6-31
-------
§
7
6
5
4
3
2
IOO.C
1
6
5
4
3
£ 2
-*»•
1
H- 10,00
§ I
_l 6
2 5
Z 4
2
IOOC
9
a
7
6
5
4
2
'9
8
7
4
T
2
00
6
5
3
2
D 100
9
8
7
— 5
- >s 2
_c
100
- ~ 9
" O 8
< 6
4
3
2
10
0
m
^.~*
.^^
-(ii^"^
^*
x*
J
X
X TC
*^LA
TA1
BOf
:os ••
'
100
34 567891000 234 56789IOpOO
FEED RATE- Ib C02/day
4 56789
100,000
100
1000
FEED RATE -kg C02/day
10,000
Figure 39. Operation and maintenance requirements for recarbon*tion-
liquid C02 as C02 source - labor and total cost. (20)
6-32
-------
CASE HISTORIES
Seattle
The Seattle Water Department has initiated a corrosion control program.
Although the system is not yet operational, a review of the engineering in-
volved in defining the problem and determining its solution demonstrates that
corrosion control in the water works industry is an intricate and multi-
dimensional process.
A slight increase in corrosion complaints occurred after Seattle started
to use the Tolt River waters in addition to the existing Cedar River supply
in 1964, but widespread complaints became common after 1970 when chlorination
doses were increased, ammoniation was stopped to increase the free chlorine
residual, and fluoridation with hydrofluosilicic acid began. Both water
supplies are surface waters from rain and snow runoff in mountains east of
the city. They are very soft, low TDS, low pH, highly oxygenated waters with
some organtcs, principally tannins and lignins. Addition of free chlorine
and fluorides caused this water to become excessively aggressive. Aestheti-
cally based consumer complaints were followed by a water metals survey.
Although metal increases in the distribution system mains were very low, lead,
iron, zinc, and copper were being appreciably leached in buildings and resi-
dences. Lead, iron, and copper limits often exceeded federal limits when
water was left standing in internal piping for as little as five hours. The
predominant pipes in these structures were galvanized iron or copper. Lead
increases were believed to be leached from soldering joints.
Because of adverse health and aesthetic reasons, as well as the expense
water customers had to assume in maintaining residential plumbing, a study
was initiated to determine the best applicable corrosion abatement process.
Alternatives included:
1. Alternative disinfectant chemicals to replace
gaseous chlorine.
2. Alternate fluoridation chemicals to replace
hydrofluosilicic acid.
3. Blending ground water with the surface water
to provide a less corrosive product.
4. Addition of corrosion inhibitors.
Alternate disinfection chemicals included calcium hypochlorite, sodium
hypochlorite, and ozone. The first two would provide a less corrosive water
but cost was two to six times greater than chlorine. Alternative fluoridation
chemicals, sodium silicofluoride and sodium fluoride, did not provide effective
corrosion control and were more expensive than the hydrofluosilic acid.
6-33
-------
Blending ground water was deemed unfeasible because the only available ground
water supply could be ecomonically used with only one of the two water systems.
Additional costs of well construction, pipelines, and pumping stations coupled
with the uncertainty of this means to affect a solution to the corrosion pro-
blem removed this alternative from consideration.
Corrosion inhibitors examined included lime, sodium carbonate, sodium
silicate, sodium bicarbonate, caustic soda, and various phosphate compounds.
The following are the results of the six alternatives selected for extensive
testing.
1. Low-carbonate system. Lime and sodium carbonate were used
to raise the pH to 9 and alkalinity to 10-30 mg/1. No
calcium carbonate scale formed on pipe surfaces, excellent
corrosion protection was provided for copper pipes and fair
protection for galvanized and black steel pipes. However,
the high pH attained raised questions about trihalomethane
formation as the surface waters contained organic precursors
and THM formation is known to be accelerated above pH 8.5.
Also, the high pH was unstable in contact with air, decreasing
by at least one pH unit within a day. Further testing with
NaHC03 eliminated both of these problems, yet provided the
same amount of corrosion protection.
2. Lime and zinc-orthophosphate. Although this method effec-
tively eliminated corrosion, it was felt that the potential
stimulant to algal growth from the phosphate and subsequent
taste, odor, plugging of fine filters, and additional chlorine
required did not warrant this as a desirable alternative.
3. Sodium silicate. Silica, added to a concentration of 10 mg/1,
gave corrosion protection, but evidence of increased pitting
rates coupled with an increased need to remove any silica
added to the water by industries that were using the existing
water without pretreatment made this alternative unattractive.
4.-6. These methods involved the carbonate method of corrosion
control based on CaC03 precipitation. Each of these methods
gave excellent protection for copper pipes and good protection
for galvanized and steel pipes.
4. High pH, balanced lime and alkalinity. Adding 15-20 mg/1 lime
and 15-20 mg/1 sodium bicarbonate produced a wa-ter with pH
8.8. Cost and THM formation possibilities made this alternative
unattractive.
6-34
-------
5. Moderate pH, balanced lime and alkalinity. By adding
35-40 mg/1 lime and 55-60 mg/1 C02 a water of about
75 mg/1 calcium and alkalinity (as CaC03) and pH 8.3
was achieved. Cost of this method is excessive and
the water would be classified as "moderately hard,"
imposing a much greater impact on industries that
pretreat the water for hardness removal.
6. High alkalinity, moderate pH. By adding 10-20 mg/1
lime and 65-130 mg/1 NaHC03, a very soft water with
pH 8-8.5 would result. Although attractive from a
corrosion control and water quality aspect, the cost
was excessive.
ESTIMATED
ANNUAL
METHOD COST ($)
Low carbonate 510,000
Lime and zinc orthophosphate 430,000
Sodium silica 630,000
High pH, balanced lime
and alkalinity 1,100,000
Moderate pH, balanced lime
and alkalinity 1,700,000
High alkalinity, moderate pH 1,400,000
The low carbonate method using sodium bicarbonate was chosen because of
its low cost and- minimal affect on existing water and environmental quality.
Additional sodium silicate will be added to the Tolt River system, bringing
the silica level up to that of the Cedar River (7-9 mg/1) and providing
additional corrosion protection. The mechanism of low carbonate control is
a result of reactions between pipe material and carbonate, forming insoluble
metal oxides that coat the pipe surface. Cost of the process and additional
chlorine needed due to the increased pH is expected to be $1.13 per thousand
gallons. Assuming a 100 gpcd water use rate, this amounts to $i.40/residence/
year. The computed savings to residences from expected lower piping repair
and service needs is about $7/year less than the present $35/year estimate.
In association with this process, city policy changes have been re-
commended. These include encouraging the use of less corrodible pipes, such
as cement lined steel and ductile iron, for distribution systems. The use
of asbestos cement and galvanized pipes is discouraged for municipal,
industrial, or commercial use. The use of copper or plastic pipe for re-
sidential piping is encouraged, as is the use of low-lead solder and glass
lined water heaters.
6-35
-------
Factors other than cost that influenced the decision to use this pro-
cess included the potential impact on industrial water use. Area firms are
adjusted to the relatively mineral free water presently supplied. Use of
the low carbonate process should only slightly affect ion-exchange units or
carbon adsorption beds, and few industries are expected to need extensive
adjustments. The waste water discharged into Puget Sound should have lower
metals concentrations, as will the waste water sludge. The corrosion abate-
ment process is presently being further tested and fine tuned, and a 21 month
plan, design, permit acquisition, and construction time period is anticipated.
Carroll County, Maryland
Carroll County, Maryland is a formerly rural area located about 30 miles
northwest of Baltimore that has experienced a large population increase in
the last two decades. The county currently has about 13,000 residences that
rely on individual wells for domestic water, and being built since the early
70's, many of these houses contain copper piping joined with 50% lead solder.
The major geologic formations supplying the well water include the Triassic
Sandstone, the Wakefield Marble, and the Pretty Boy Schist. The sandstone
waters are of variable pH, the marble waters are neutral or higher, but the
bedrock acidity of the schist formation produces low pH waters.
In November, 1975, a 45 unit apartment complex served by two wells was
found to have a 0.34 ppm lead concentration in a daytime water sample. At
the same time, several children living in the complex were found to have
moderately elevated blood lead levels. The complex consisted of relatively
new buildings (two to 14 years old), but the water system neutralzer was
poorly maintained and consequently ineffective. The neutralzer problem was
corrected, lead concentrations dropped to below the federal standard of
0.05 ppm, and the children's blood levels returned to normal over a period
of a few months.
A year later, in another part of the county, the physician of a family
experiencing a gastro-intestinal illness thought their drinking water might
be contaminated. A bacteriological analysis was negative, but the water's
pH of 5.5 coupled with an inspection of the type of plumbing used in the house,
suggested heavy metal contamination. Sampling indicated lead levels of 1.8
ppm and copper levels of 7.9 ppm in the tap water. Subsequent sampling of
14 new houses in this subdivision showed that 12 houses had lead concentrations
greater than 0.05 ppm and 10 houses had copper concentrations in excess of
1.0 ppm in early morning samples. A second sampling, after water had run
for five minutes, showed six houses with lead levels in excess of the fed-
eral standard. Water sampled directly from the well showed no lead or copper,
and homes with a neutralizer that adjusted the water to pH 7.2-7.5 also
showed no lead or copper. It was concluded that poor workmanship during the
copper piping installation and the use of lead solder had precipitated the
problem.
6-36
-------
This subdivision incident prompted a random sampling of the tap water
in homes in new subdivisions, and when 11 of 35 samples were found to con-
tain lead levels above 0.05 ppm in samples where the water had run for five
minutes, The Carroll County Health Department requested investigative
assistance from the Center for Disease Control. A subsequent random sampling
analyzed 350 homes in the county, determining pH, alkalinity, early morning
lead and copper levels, and lead and copper levels after running the water
for five minutes. Test results showed that low pH and low alkalinity were
associated with excessive lead and copper levels in the tap waters.
Consequences of this testing included the immediate recommendation that
the water, after standing overnight or for long periods, be allowed to run
for two minutes prior to use. The plumbing code in the county was revised
to permit the use of NSF-approved (grade 14) plastic tubing in new construc-
tion and repair work, and a neutralizer to raise water pH above 7.0 is pre-
sently recommended in existing homes with copper piping. The calcite filter
type of neutralizer is suggested as best for this problem. ( 31 )
Orange County, California
Between July, 1961 and March 1963, 206 houses were built by a land de-
velopment firm in an area of southern Orange County, California. All plumbing
was copper and water was supplied from wells by a local water utility district.
A typical water analysis showed calcium, 225 ppm; magnesium, 40 ppm; sodium,
160 ppm; sulfate, 525 ppm; chloride, 170 ppm; bicarbonate, 360 ppm; TDS,
1500 ppm; hardness, 735 ppm; pH 7.1; and free carbon dioxide, 36 ppm.
In January, 1963, a one million gallon reservoir and hydropneumatic tank
system began operation to provide better water service to the development's
growing population. Some houses had been occupied since the spring of 1962
and had used water pumped directly from the wells. However the problem of
perforation leaks in copper tubing throughout the development did not occur
until three months after the additional system began operation. In April, 15
leaks were reported, and the rate of about 15 leaks reported per month con-
tinued through October.
By May 1963, the builder had decided that the leaks were not the result
of faulty workmanship and sought the assistance of consultants, including two
corrosion experts, a testing laboratory, and a civil engineering firm. The
subsequent investigation determined that the corrosion was a form of severe
occasional pitting occurring on the inside surface of the copper tubes. The
pipe surface was covered with a thin green layer, assumed to be copper car-
bonates, and the pits formed conical penetrations into the tube material and
were covered by nodules up to 1/8 inch in height. The pitting occurred in a
random pattern and was not necessarily found near joints. The copper that
6-37
-------
had been used was both "Type K" and "Type L", had been supplied by three
different manufacturers, and was found to conform to ASTM Specification B88-
62 for that material. Copper staining of plumbing fixtures, that may result
from copper solubilization in soft, acidic waters, was not found. Leaks
appeared in occupied and unoccupied houses, and although electrical services
were grounded to the water systems, the unoccupied houses did not have their
power turned on. It was also noted that water from the same wells was used
in other parts of the utility's service area with few reports of errosion
failures.
The consultants decided that the probable cause of corrosion was the
combined presence of free carbon dioxide and dissolved oxygen, possibly en-
hanced by the high TDS concentration. Although the well water had minimal
dissolved oxygen, field measurements indicated that a dissolved oxygen con-
centration of about 3 ppm existed after the water passed through the reservoir
and hydropneumatic system. The consequent remedial action taken by the water
purveyor was to begin using water from two other wells. Because one of the
additional wells had high iron and manganese levels, a package treatment unit
consisting of potassium permanganate, caustic soda, and chlorine addition
with pressure-type sand filtration was installed. .Delivery of the treated
water began in September, and the rate of pipe failures dropped to nearly
zero in November.
In early 1964, copper tubing failures began to recur. The consultants
found that while the well providing water with high iron and manganese con-
centrations had initially shown low levels of free carbon dioxide, the level
had risen to about 25 ppm. Also, the addition of caustic soda that had
neutralized the free carbon dioxide, had been discontinued after less than
two months of operation. In April, the wells were taken out of service and
the delivery of imported water to the affected area began. After about a
month, the rate of tube failure again dropped to almost zero.
Damage resulting from the corrosion cost the builder substantial financial
losses, bad publicity, inability to sell houses, and the loss of FHA financing
commitments. He consequently sued the purveyor for negligence in not reducing
the carbon dioxide content of the water and for breach of warranty that es-
sentially says that in the absence of an explicit warranty, a seller warrants
his goods to be suitable for the purpose for which they are sold.
The court found that the water only was responsible for the corrosion and
that the addition of caustic soda was a simple and inexpensive remedy.
However, the water supplier was not found negligent for his failure to reduce
the free carbon dioxide concentration in the water. He was found guilty of
breach of warranty and the judgment was for the builder's claimed loss. The
court ruled that the water was defective in that the purveyor warranted that
the water would be reasonably fit for transmission through copper pipes and
would not damage or corrode the tubing during its normal life expectancy.
6-38
-------
Additional Corrosion Control Practices
The Catski11-Delaware water system supplying New York City is treated
with fluosilic acid for fluoridation that lowers the pH of this soft water
from 6.9 to 6.3. Caustic soda is then added to raise the pH to the 6.9-7.2
range. The effectiveness of this system has been judged on the basis of con-
sumer complaints, with an increase in the number of complaints of green
(copper) staining occurring when insufficient caustic soda was added. Samples
of water at consumer taps by the New York City Health Department indicated
that of 500 samples analyzed for lead, copper, and iron, only a few cases of
lead in excess of 0.05 ppm were found, contrary to the health department's
expectations ( 40 ).
In New London, Connecticut, severe corrosion problems experienced prior
to 1969 have been controlled by pH adjustment and corrosion inhibitor addition.
The raw water pH of 6.8 is adjusted to 7.2-7.5 and zinc orthophosphate is
added to produce a 0.5 mg/1 zinc concentration in the water. Tap water
sampling has shown this treatment to effectively control lead and copper con-
centrations.
Other large metropolitan areas with corrosion control methods are Sal em-
Beverly, Massachusetts (lime and zinc metaphosphate); Long Beach, California
(see October, 1970 issue of JAWWA); Middlesex, New Jersey (see August, 1974
issue of JAWWA); Waterbury, Connecticut; and Philadelphia's Schuylkill River
Plant.
6-39
-------
REFERENCES
1. Atkins, G. R.,"Soft Water Corrosion and Calcium Carbonate
Saturation',' The South African Industrial Chemist, Vol. 8,
June 1954, pp. 104-111.
2. Bay!is, John R., "Corrosion Studies,"J. New England Water
Works Assoc.. Vol. 67, 1953, pp. 38-73.
3. Baylis, John R., "Treatment of Water to Prevent Corrosion"
J. American Water Works Assoc., Vol. 27, No. 2, 1935,
pp. 220-234.
4. Bard, Allen J., Chemical Equilibrium, Harper & Row, New
York, 1966.
5. Bopp, C. D., and S. A. Reed, Stabilization nf Product
Water From Sea Water Distillation Plants, U.S. Office of
Saline Water Research & Development Progress Report No.
709, Oak Ridge National Laboratory, Tennessee, July 1971.
6. Brighton, William D.,"Dissolved Copper from New Service
Pipes,'Water and Water Engineering, Vol. 59, July 1955,
pp. 292-293.
7. Brown, Joseph E., Charles R. Pitts, Jr., Evaluation of
Three Different Agents for Stabilizing Wa^er, City of
Richmond, Department of Public Utilities, Richmond,
Virginia, April 1966.
8. California Department of Water Resources, Southern District,
Consumer Costs of Water Quality in Domestic Water Use Lompoc
Area, District Report, June 1978.
9. Committee Report,"The Value of Sodium Hexametaphosphate in
the Control of Difficulties Due to Corrosion in Water
Systems1; JAWWA, Vol. 34, No. 12, pp. 1807-1830.
10. Cornwell, F. J., G. Wildsmith and P. T. Gilbert,"Pitting
Corrosion in Copper Tubes in Cold Water Service,"Br.
Corrosion J., Vol. 8, No. 5, Sept. 1973, pp. 202-209.
11. Cox, Charles R.,"Corrosion Control by Water Treatment,"
Water Works Engineering, December 4, 1934, pp. 1514-1517.
12. DeMartini, F. E. ."Corrosion and the Langelier Calcium
Carbonate Saturation Index,"JAWWA, Vol. 30, No. 1, January
1938, pp. 85-111.
6-40
-------
REFERENCES (continued)
13. Esty, Roger W.,"Cement Lining of Pipe Corrects Bad Water
Troubles."The American City, April 1941, pp. 60-62.
14. Esty, Roger W.,"When Can Cement Lining of Pipe be Used
to Advantage."Water Works Engineering, September 10, 1930,
pp. 1363-4.
15. Feitler, Herbert,"Critical pH Scaling Indexes,"Materials
Protection and Performance, August, 1975, pp. 33-35.
16. Flentje, Martin E.,"Control of Red Water Due to Pipeline
Corrosion."JAWWA, December 1961, pp. 1461-1465.
17. Frye, S. C.,"Epoxy Lining for Steel Water Pipe,"jAWWA.
Vol. 66, No. 8, August 1974, pp. 498-501.
18. Gardels, M. and M. Schock, (EPA Cincinnati Laboratory),
Personnal Communication, via P. Lassovszky, January 12,
1981.
19. Goldfarb, Alan S., James Konz, and Pamela Walker, "Coal
Tar Based Materials and Their Alternatives," Interior
Coatings in Potable Water Tanks and Pipelines, The Mitre
Corp., Mitre Technical Report MTR-7803, U.S. EPA Contract
No. 68-01-4635, January 1979.
20. Gumerman, Robert C., Russell L. Gulp and Sigurd P. Hansen,
Estimating Water Treatment Costs, Vol. 2, Cost Curves
Applicable to 1 to 200 mqd Treatment Plants, EPA-600/2-79-
K2b, prepared for the Municipal Environmental Research
Laboratory Office of Research & Development, U.S. EPA by
Culp/Wesner/Culp Consulting Engineers, Santa Anna, California,
August 1979.
21. Hatch, G. B.."Inhibition of Lead Corrosion with Sodium
Hexametaphosphate,"JAWWA, Vol. 33, No. 7, pp. 1179-1187.
22. Heller, A., and K. C. Chang, and B. Mi Her, "Spectral
Response and Efficiency Relations in Semiconductor Liquid
Junction Solar Cells" J. Electrochem Soc., Vol. 124, No. 5,
May 1977, pp. 697-700.
23. Hopkins, Edward S.,"Basic Principles of Corrosion Control
by the Use of Lime" Paper Trade Journal. Vol. 127, No. 1,
July 1, 1948, pp. 61-63.
6-41
-------
REFERENCES (continued)
24. Hudson Jr., H. E. and F. W. Gilcreas,"Heath and Economic
Aspects of Water Hardness and Corrosiveness" JAWWA, Vol. 68,
1976, pp. 201-204.
25. Internal Corrosion Study, Summary Report, prepared for the
City of Seattle Water Department by Kennedy Engineers, 7708
Bridgeport Way W., Tacoma, Washington, 98467, February 17,
1978.
26. Karalekas, Peter C., Jr., C. R. Ryan, C. D. Larson, and F. B.
Taylor,"Alternative Methods for Controlling the Corrosion of
Lead Pipe."J. New England Water Works Assoc., Vol. 92, No. 2,
1978, pp. 159-78.
27. Kelly, T. E., M. A. Kise, and F. B. Steketee,"Zinc/Phosphate
Combinations Control Corrosion in Potable Water Distribution
Systems" Materials Protection and Performance, Vol. 12, No. 4,
April 1973, pp. 28-31.
28. Lane, R. W., T. E. Larson, S. W. Schelsky,'The Effect of pH on
the Silicate Treatment of Hot Water in Galvanized Piping" JAWWA,
August 1977, pp. 457-461.
29. Langelier, W. F.,"The Analytical Control of Anti-Corrosion
Water Treatment" JAWWA. Vol. 28, No. 10, 1936, pp. 1500-1521.
30. Lehrman, Leo, Henry L. Shuldener, "Action of Sodium Silicate
as a Corrosion Inhibitor in Water Piping," Industrial and
Engineering Chemistry, Vol. 44, No. 8, August 1952, pp. 1765-
1769.
31. Lovell, John, Richard Isaac, Ruth Singer, Control of Lead and
Copper in Private Water Supplies, Carroll County, Maryland,
Carroll County Health Department, October, 1978.
32. McCauley, Robert F. and Mahmond Omer Abdullah,"Carbonate
Deposits for Pipe Protecti on," JAWWA/Vol. 50, 1958, pp. 1419-1428.
33. McCauley, Robert F.,"Use of Polyphosphates for Developing
Protective Calcite Coatings,"JAWWA, January 1960, pp. 721-734.
34. McLaughlin, P. L./'Eliminating the Guess from Anti-Corrosion
Treatment" Water Works Engineering, 1937-
35. Merrill, Douglas T. and Robert L. Sanks,"Corrosion Control by
Deposition of CaC03 Films: Part 1, A Practical Approach for
Plant Operators,"JAWWA, Vol. 69, November 1977, pp. 592-599.
6-42
-------
REFERENCES (continued)
36. Murray|J( W. Bruce,"A Corrosion Inhibitor Process for Domestic
Water,"J. American Hater Works Assoc., Vol. 62, No. 10, Oct.
1970, pp. 659-662.
37. Orange County Water District, California, Water Quality and
Consumer Costs, Santa Ana, Calfornia, May 1972.
38. Patterson, James W. and Joseph E. O'Brien,"Control of Lead
Corrosion" J. of the American Water Works Assoc., Vol. 71,
No. 5, May 1979, p. 264-271.
39. Patterson, James W.."Corrosion Inhibitors and Coatings,"
JAWWA Conference, June 1978.
40. Report on Corrosion Control Practices, EPA Region 1, Water
Supply Branch, Division of Water Programs, Sept. 1975.
41. Sargent, Harold E.,"Asbestos in Drinking Water" J. New
England Water Works Assoc., Vol. 88, No. 1, 1974, pp. 44-57.
42. Sawyer, Clair N. and Perry L. McCarty, Chemistry for Sanitary
Engineers, 2nd Edition, McGraw-Hill Book Company, New York, 1967.
43. Scholefield, Ronald J., Metal Corrosion Products in Municipal
Drinking Waters, Thesis, Environmental Engineering, Illinois
Institute of Technology, August, 1979.
44. Shuldener, Henry L., Sidney Sussman,"Sodium Silicate - To Keep
Piping Young."Water Works Engineering. September 1960.
45. Shuldener, Henry L., Leo Lehrman,"Influence of Bicarbonate
Ion on Inhibition of Corrosion by Sodium Silicate in a Zinc-
Iron System,"JAWWA. November 1957, pp. 1432-1440.
46. Simmonds, M. A.,"Effect of Aggressive Waters on Cement and
Concrete, with Particular Reference to Cement-lined Mains,"
The J. of the Institution of Engineers, Australia, Vol. 26,
January-February 1954, pp. 9-16.
47. Singley, J. Edward, A. W. Hoadley, H. E. Hudson, Jr., Edna T.
Loehman, A Benefit/Cost Evaluation of Drinking Water Hygiene
Programs, U.S. EPA contract #68-01-1838, Univ. of Morida,
1969.
48. Stericker, William, "Sodium Silicates in Water to Prevent
Corrosion" Industrial and Engineering Chemistry. Vol. 30, #3,
March 1938, pp. 348-351.
6-43
-------
REFERENCES (continued)
49. Stumm, Werner and James J. Morgan, Aquatic^Chemistry, An
Introduction Emphasizing Chemical Equilibrium In Natural
Haters, Wiley-Interscience, New York, 1970.
50. Weir, Paul."Effects of Pipe & Tank Lining on Water Quality
at Altanta" JAWWA, Vol. 49, No. 1, January 1957, pp. 1-14.
6-44
-------
SECTION 7
CONSIDERATIONS FOR CORROSION CONTROL REGULATIONS
Detailed information is presented in this study which presents the
nature and magnitude of corrosion and corrosion control in the water works
industry. From the results of previous studies as presented in the litera-
ture, it is obvious that corrosion control is quite complex and development
of a responsible corrosion control strategy for the water works industry re-
quires a comprehensive approach. The objective of this Section is to sum-
marize the information presented in the preceeding Sections through the use
of charts and tables to provide the necessary basis for this comprehensive
approach. Information presented in these tables is developed directly from
data provided and discussed in this study. In as much as results of studies
presented in the literature are often in conflict and disagreement and the
fact that factors affecting corrosion and corrosion rates are most commonly
synergistic, the information and comments provided in these tables should be
considered as guidelines only. Additionally, in the interest of brevity to
present a comprehensive and readily usable overview, comments and statements
have been taken out of context and could be misleading to users lacking
sufficient technical understanding of the nature of corrosion in the water
works industry. Therefore, it is suggested that references be made to the
respective sections of this study, as needed, for effective use of the tables.
The tables and information presented in this Section were selected
and organized to follow a logical order for considering and developing a
corrosion control strategy for the water works industry. The various materi-
als and their specific use within the water works industry are first pre-
sented. Second, the general extent of each material's use as well as the
respective associated potential contaminants are presented. This information
is presented to help assess the significance of corrosion or deterioration
of the various materials and can be used to assign priority for corrosion
control strategies if desired. Once materials and their specific potential
contaminants have been identified, applicable corrosion monitoring and
detection techniques can be selected to determine the extent, if any, of
corrosion. The third table is perhaps the most comprehensive and provides
a brief summary of the various water quality and conditions of service para-
meters' effects on corrosion of each of the materials. This table can be
used most effectively to assess the potential for controlling all corrosion
in a system by comparing the preferred water quality and conditions of ser-
vice to minimize the corrosion of each material. Finally, a table is pre-
sented which identifies and summarizes the applicability of corrosion pre-
vention technologies for each material. The following is a brief description
of each of the tables included.
Table 45, Materials and Their Application in the Water Works Industry,
identifies specific uses of materials and attempts to provide a relative
quantification of occurrance. Specific uses include in-plant systems
(piping and appurtenances), transmission lines, storage, distribution mains,
7-1
-------
service lines, and household plumbing. Information in this table can be used
to assist in identifying the types of materials currently in place as well
as their relative quantity. This table does not, however, identify current
use patterns.
Table 461 Significance of Corrosion or Deterioration of Various
Materials Used in the Water Works Industry, presents a brief discussion of
the known extent of use of each material as well as the contaminants that
have been found to be associated with the use of each material. Concentra-
tions of contaminants released are provided where values are reported in the
literature. The importance of this table is its use for assessing corro-
sion significance of each material. It should be noted here, that it
appears from this table that the use of lead and asbestos-cement should be
of paramount concern as these two materials occur extensively within the
industry and have been associated -with releasing significant quantities or
concentration of contaminants, lead being potentially toxic and the effects
of asbestos fibers in drinking water yet to be determined. Additionally,
it should be noted that it is estimated that one-third of all water distri-
bution pipe currently being sold in the U.S. is manufactured of asbestos-
cement pipe. Alternatively, lead pipe is currently being used less exten-
sively but, owing to its relatively long life, many of the lead lines in-
stalled remain in service. Lead based solders continue to be used extensively.
Table 47, Preferred Water Quality and Conditions of Service to Minimize
Corrosion of Materials Used in the Water Works Industry, is the most compre-
hensive table and provides a brief overview of the factors influencing or
affecting corrosion of each material. Water quality parameters included are
pH, hardness, alkalinity, dissolved oxygen, carbon dioxide, total dissolved
solids, metal ions, and organic acids. Conditions of service included are
velocity and temperature. This table can be used to assess the potential
of controlling various parameters to control the corrosion within a water
system which contains a variety of materials. It should be noted that pH,
hardness, and alkalinity are the most controllable water quality parameters
using conventional water treatment practices. The other water quality
parameters are more difficult to control and/or maintain throughout a dis-
tribution system but are included herein to emphasize their synergistic
affects on corrosion of materials. The effects of water velocity can only
be controlled through design, and possibly, operation practices.
In general, water pH should be maintained near neutral or in the
slightly alkaline range. However, it is noted that localized corrosion
has been observed to peak in plain iron in the pH range of 6-9.
For most materials, the effect of hardness concentration is synergistic
with other parameters. In general, hard waters are considered less aggressive
then soft water to materials with the exception of aluminum in which the re-
verse is true. For iron-based materials, the presence of calcium ions has
been shown to control corrosion if sufficient alkalinity exists.
However, hardness concentrations do not appear to influence corrosion
7-2
-------
characteristics of stainless steel. In the case of copper, it has been found
that soft waters may not be corrosive if the carbon dioxide concentration
is low while pitting corrosion has been observed in cold hard water indicating
that hardness concentration does not act independently on the corrosion or
pitting of copper.
Corrosion and pitting appear to be inhibited in waters containing
higher alkalinity concentrations for most materials except copper. In copper
pipes, the addition of bicarbonate has been shown to actually enhance cor-
rosion under specific conditions. Desired alkalinity concentration ranges
of 20 mg/1 or higher for lead pipes ( 12 ) and 40 mg/1 or higher for con-
crete pipes ( 25 ) have been reported. For asbestos-cement materials, the
aggressiveness of water is defined as a function of the combined levels of
pH, hardness, and alkalinity concentrations. Although good studies on the
pH effect on asbestos-cement independent of the other factors is lacking
( 12 ), a less "aggressive" water should be produced by increasing any one
of tne three factors.
Dissolved oxygen concentrations cannot be controlled or maintained ef-
fectively throughout a water system. Nevertheless its presence or absence
can significantly affect corrosion or corrosion rates on various materials
and is therefore included in Table 47. In all cases with the metals, a low
concentration of dissolved oxygen is desirable to minimize corrosion. How-
ever, a higher dissolved oxygen concentration is generally desirable for the
formation of protective films. For the iron-based materials, corrosion ap-
pears to increase linearly with increased dissolved oxygen concentration.
For copper, however, the presence of dissolved oxygen is known to enhance
corrosion, but corrosion or the corrosion rate is not dependent on the dis-
solved oxygen concentration. Little information is provided in the litera-
ture identifying the effects of dissolved oxygen concentration on non-
metallic materials.
As with dissolved oxygen concentrations, low carbon dioxide concentra-
tions are desirable to minimize corrosion. For lead materials, it has been
stated that the presence of excess carbon dioxide concentration will tend to
dissolve protective carbonate films and assist corrosion.
Results of studies reported in the literature indicate that the effects
of total dissolved solids (IDS) concentrations on corrosion of the metals is
complex and no general characterization c'an be established. For plain
iron it has been reported that while the presence of IDS can decrease dis-
solved oxygen and carbon dioxide concentrations resulting in a reduction in
corrosion, the increased conductivity can, in fact, increase the range of
galvanic coupling or lead to the formation of less protective films. It is
also_stated in the literature that the presence of chloride (Cl~) and sulfate
(SO^ ) ions can increase corrosiveness of water while elsewhere ( H ) it is
stated that the presence of these two ions can improve the protectiveness
of scale.
7-3
-------
For galvanized iron, it is reported that chloride ion concentrations
observed in most potable water supplies do not increase corrosion or
corrosion rates while higher concentrations will tend to accelerate corrosion.
For the case of stainless steel, the literature reports that the presence of
sulfate ions will inhibit corrosion while the presence of chloride ions
can cause severe corrosion.
The corrosive effects of metal ion concentrations also varies with the
specific material as well as with the specific metal ion. The presence of
copper ions has been reported to be a major factor affecting and increasing
the corrosion of galvanized steels. Additionally, the presence of copper,
tin, nickel, and mercury have been shown to be detrimental to the corrosion
of aluminum. Conversely, low concentrations of iron (Fe++) have been re-
ported to inhibit corrosion in copper and to precipitate on asbestos-cement
to form a protective coating inhibiting calcium leaching ( 3 ).
Little information is presented in the literature describing the
corrosive effects of the presence of organic acids. For iron-based materials,
it is reported that the presence of humic acids can improve protective de-
posit formation and lead to reducing corrosion or the corrosion rate ( 30,
40 ). Alternatively for lead, the literature recommends that the occur-
rance of organic acids whose lead salts are soluble should be minimized to
prevent corrosion (32 ). It is also reported that the occurrence of an
unidentified high molecular organic acid found in surface but not in ground
water may inhibit corrosion in copper piping ( 5 ).
For most materials used in the water works industry, it is reported
that both stagnant waters and waters of high velocity will promote corrosion.
Where protective coatings may be formed, some flow is required. However,
excessive velocities can cause impingement attack and accelerate corrosion.
In general, but not in all cases, water velocities ranging between 2-7 fps
are desirable.
Temperature effects on the corrosion of materials are of most concern
at elevated temperatures above that observed from operations at a utility.
With the exception of aluminum, increased water temperatures will generally
increase corrosion of materials used in the water works industry. Water
temperatures in excess of 40°C are considered preferable for minimizing corro-
sion on aluminum. For plain iron, it is reported that the aggressiveness of
water increases with temperatures up to approximately 80°C and then decreases
at higher temperatures.
Table 48, Application of Corrosion Control Mechanisms, provides a
summary of the various corrosion control alternatives and their application
to the various materials. Corrosion control alternatives include both
coatings and linings and inhibitors. Coatings and linings are coal tar,
cement mortar, epoxy, vinyl, and miscellaneous non-coal tar paints. In-
hibitors include calcium carbonate, silicates, and phosphates. Cathodic
protection is also included as it is applied for protection of the inside
of steel tanks.
7-4
-------
TABLE 45
MATERIALS AND THEIR APPLICATION IN THE WATER WORKS INDUSTRY
MATERIAL
WROUGHT IRON
CAST/DUCTILE
STEEL
GALVANIZED
IRON
STAINLESS
STEEL
COPPER
LEAD
ALUMINUM
ASBESTOS-
CEMENT
CONCRETE
PLASTIC
IN PLANT
PIPING
/
//
/
/
/
/
/
/
SYSTEMS:
APPURTE-
NANCES
//
//
(brass)
/
(gaskets)
/
/
TRANSMISSION
LINES
//
/
/
//
//
STORAGE
//
//
J
DISTRIBUTION
MAINS
/
///
(cast iron)
/
/
/
//
/
/
SERVICE
LIMES
/
/
/
/
///
/
^
//
HOUSEHOLD
/
/
/
/
/
//
/
/
//
KEY: /// Used >50% for the particular service.
// Frequently used for the particular service.
/ Has been or is used for the particular service.
-------
TABLE 46
SIGNIFICANCE OF CORROSION OR DETERIORATION OF VARIOUS MATERIALS
USED IN THE WATER WORKS INDUSTRY
MATERIAL
EXTENT OF USE
ASSOCIATED CONTAMINANTS
IRON-BASED MATERIALS:
PLAIN IRON
Cast Iron is used in 75% of all major U.S. water supply
distribution systems (17). Also used in water appur-
tenances and treatment plants. Over 1/2 million steel
water storage tanks exist in the U.S.
Iron concentrations in excess of the 0.3 mg/1 approval limit
occur, resulting in ferric oxide (red water) complaints.
Generally limited to service lines, in plant systems, and
GALVANIZED IRON households. It requires threaded joints and gooseneck
connections and Is declining in usage.
Zinc concentrations will increase 5 to 10 mg/1 after 8 to 40
hour exposure to new galvanized pipe. Small amounts of Iron
will enter solution. Cadmium and lead (Impurities in galvanizing
process) concentrations will rise.
Seldom used for piping, but used where low maintenance and
STAINLESS STEEL reliable, continuous service is desired, such as pumps,
valves, meters, Venturis, and pressure regulators.
Besides iron, other metals used to manufacture stainless steel
that may enter the water through pitting or corrosion are chro-
mium, nickel and molybdenum.
COPPER
i
CD
Extensively used in household piping and service lines.
From World War II to 1972, over 6 million miles of copper
tubing was put into service. Bronze may be used for
appurtenances.
Copper, as well as iron, zinc, tin, and lead from associated
pipes and solder may be oxidized Into (solution. Cu concentrations
do not raise above about 5 mg/1. Impurities in brasses, such as
manganese, arsenic, antimony, phosphorus, bismuth, and tin may
also leak out.
LEAD
Little documentation available; according to Donaldson in
1924 ( 9 ) approximately 50% of water distribution systems
in the U.S. had lead lines, used primarily for service lines
and solders for copper pipes; =60% of residences 1n Boston
are serviced with lead lines.-
Lead concentrations ranging up to = 3.0 PPM from both lead lines
and lead based solders have been reported.
ALUMINUM
Use of aluminum is relatively limited; currently used for
weir gates, storage tands, reservoir roofs and supports,
hot water systems, and pipe lines.
No information available which identifies or quantifies potential
contaminants; could release traces of :opper, magnesium, silicon,
iron, manganese, chromium, zinc, or titanium as well as alumina
ions.
Approximately 1/3 of all water distribution pipe currently
ASBESTOS-CEMENT being sold in the U.S. is manufactured of asbestos-cement
pipe; approximately 200,000 miles has been placed into
service.
Asbestos-cement fibers counts in excess of 4.5 million fibers per
liter have been observed; tetrachlorethylene concentrations as
high as 2500 pg/1 ( 3 ) have been observed from lined asbestos-
cement pipes { 22 ).
CONCRETE PIPE
Extensively used in water distribution (and storage) systems
with service life of over 50 years in some locations. About
15% of new water tanks are concrete
Contaminant release is greatest when the pipe is first used and
decreases thereafter. Water hardness and pH initially increase.
Oxides of silicon, aluminum, iron, magnesium, and sulfur may
hydrolyze, releasing these elements.
PLASTIC PIPE
Currently growing in use for service lines and household
piping except in hot water systems; 1978 use was about 1/3
of all piping on a footage basis. Recent development has
produced larger pipes being used in distribution mains.
Lend stabilizing compounds may be leached from PVC pipes. Other
contaminants arise from the solvents used and include 2-butanone
(MEK) and tetrahydrofuran (THF).
-------
TABLE 47
PREFERRED HATER QUALITY AND CONDITIONS OF SERVICE TO MINIMIZE CORROSION
OF MATERIALS USED IN THE MATER WORKS INDUSTRY '
MATERIAL
IRON-BASED MATERIALS:
PLAIN IRON
GALVANIZED IRON
STAINLESS STEEL
COPPER
LEAD
ALUMINUM
ASBESTOS-CEMENT
CONCRETE PIPE
PLASTIC PIPE
pH
Long term: Little effect for pH 4-10 except localized
corrosion may peak at pH 6-9 range (39).
Short term: pH effects are a function of flow rate
and time (10).
1. Corrosion rate Increases inversely with pH.
2. Optimum pH range Is 7-12.
Little effect within range of water systems
1. A pH >7 will minimize uniform corrosion; also
uniform corrosion will decrease with increasing
pH ( 6).
2. Pitting corrosion will proceed at pH levels above
7.
1. A pH of 6-9 is pr»f»rr->H to minimize corrosion
(13, 20).
2. A pH of 6.5-7.0 is preferred to minimize corro-
sion (26).
Optimum pH is 7.0-7.5.
HARDNESS
Calcium (Ca'1"1') Inhibits corrosion In the presence of
sufficient alkalinity.
Hard waters are less aggressive than soft waters
(21, 37).
Not necessary for protection.
1. Soft waters are not corrosive if CO, 1s low
(18).
2. Pitting corrosion can occur 1n hard waters which
are cold (5).
1., A hardness of 10-100 PPM as CaC03 1s preferred
(26).
2. A hardness of 125 PPM as CaCO, 1s desirable
(32).
1. In general, soft waters are preferred.
2. The preferred concentration Is dependent on the
period of immersion.
3. CaC03 concentration should be approximately equal
to the chloride concentration (2).
ALKALINITY
1. Greater alkalinity produces less aggressive water
(34).
2. Anodic dissolution of iron is accelerated by
bicarbonate HC03~ through the localized formation
of Fe(C03r (7). -
Greater alkalinity produces less corrosive waters.
Increased bicarbonate alkalinity will, inhibit pitting.
Addition of bicarbonate may Increase corrosion (35, 36).
An alkalinity of 20 PPM 1s desireable to form a
protective film (12).
-
pH <• Log (Hardness x Alkalinity) should be >12.0 (3).
pH levels of 7.0 and greater are preferred to
inhibit leaching.
,
Hardness in excess of 16 PPM Ca++ is preferred to
inhibit leaching.
-
Alkalinity in excess of 40 PPM a CaCO, is preferred
to inhibit leaching.
-
-------
TABLE 47
PREFERRED HATER QUALITY AND CONDITIONS OF SERVICE TO MINIMIZE CORROSION
OF MATERIALS USED IN THE HATER WORKS INDUSTRY (Cont'd)
MATERIAL
IRON-BASED MATERIALS:
PLAIN IRON
GALVANIZED IRON
STAINLESS STEEL
COPPER
--J
LEAD
ALUMINUM
ASBESTOS-CEMENT
CONCRETE PIPE
PLASTIC PIPE
DISSOLVE OXYGEN
Corroslvity Increases linearly with dissolved oxygen
concentrations, however, better protective films
are formed at higher dissolved oxygen concentrations.
Corrosiveness of water increases directly with
Increased dissolved oxygen concentrations.
1. The presence of dissolved oxygen 1s necessary
for the formation of a protective film.
2. The presence of dissolved oxygen is aggres-
sive and encourages corrosion (14, 29, 33).
1. Corrosion is negligible in the absence of
dissolved oxygen (31).
2. The presence of dissolved oxygen will enhance
corrosion, but corrosion or the corrosion
rate is not dependent on the dissolved oxygen
concentration.
Lower concentrations favor Inhibition of corrosion.
Minimal dissolved oxygen 1s optimal (28).
-
-
-
C02
Carbonic acid Is aggressive to
iron ( 8).
Corrosiveness of water increases
with C02 concentration (38).
-
Dissolved C02 appears to enhance
corrosion.
Excess COj may dissolve protective
carbonate films and assist corro-
sion («)•
-
Low C02 concentrations are pre-
ferred.
Low C02 concentrations are preferred
-
TDS
1. Presence of TDS can decrease 02 and C02 content and, therefore, reduce
corrosion.
2. An increase in conductivity can increase the range of galvanic coupling or
lead to the formation of a less protective Fe(OH)2 film .
3. The presence of Cl~ and SGi," can Increase Corrosiveness (23).
4. The presence of Cl" and SOi," may improve protectlveness of scale (11).
1. At levels required for water treatment, Cl" does not increase corrosion
2. At levels above that required for water treatment, Cl" will accelerate
corrosion.
3. Corrosion 1s enhanced with higher neutral salt concentrations (39).
1. The presence of S0i,= will inhibit corrosion
2. The presence of Cl" can cause severe corrosion (severity depends on type
of stainless steel), (14, 29, 33).
Specific effects are difficult to identify.
1. Chlorides should be minimized (31).
2. Ions that form soluble lead salts should be minimized (32).
1. Chloride concentration should approximate the CaC03 concentration (2).
2. Preferred TDS concentrations are dependent on the period of immersion.
Possible inhibitory effect of dissolved solids.
-
-
-------
TABLE 47.
PREFERRED HflTER QUALITY AND CONDITIONS OF SERVICE TO MINIMIZE CORROSION
OF MATERIALS USED IN THE WATER WORKS INDUSTRY (Cont'd)
MATERIAL
IRON-BASED MATERIALS:
PLAIN IRON
GALVANIZED IRON
STAINLESS STEEL
COPPER
LEAD
ALUMINUM
ASBESTOS-CEMENT
CONCRETE PIPE
PLASTIC PIP-E
METAL IONS
-
Copper even at low concentrations will Increase zinc
corrosion; the presence of copper is often a main
factor 1n the corrosion of galvanized iron.
-
A low concentration of Fet+ (0.05-0.5 PPM) may
inhibit corrosion.
-
Metal ions, especially copper, tin, nickel, and
mercury should be minimized (15,28).
Fe++ can precipitate to form a protective coating
to inhibit calcium leaching (3).
Fe++ can precipitate to form a protective coating
to inhibit calcium leaching (3).
-
ORGANIC ACIDS
Presence of humic acids will inhibit corrosion
(30).
Presence of humic acids improves protective deposit
formation and reduces corrosion (40).
-
An unknown organic corrosion Inhibitor exists 1n sur-
face water but not groundwater; this natural Inhibitor
is probably of high molecular weight and may be an
organic acid (5).
Minimize the occurance of organic acids whose lead
salts are soluble (32).
-
-
-
-
VELOCITY
Stagnant waters can cause pitting and localized corrosion;
velocities >15 fps accelerates corrosion; optimum velocity
occurs between 1 fps and 8-15 fps.
Velocity differences produce little effect (4).
1. Stagnant waters are corrosive.
2. High velocity can be tolerated.
1. Some flow Is required to form protective Cu20 film.
2. Velocity In excess of 5 fps can cause imoingement
attack (]9).
Water should be running, not standing (27).
Prefer a minimum velocity of 8 fpm (15).
Water should be running, not standing (24).
Water should be running, not standing (24)
-
-------
TABLE 47.
PREFERRED WATER QUALITY AND CONDITIONS OF SERVICE TO MINIMIZE CORROSION
OF MATERIALS USED IN THE WATER WORKS INDUSTRY (Cont'd)
MATERIAL
IRON-BASED MATERIALS:
PLAIN IRON
GALVANIZED IRON
STAINLESS STEEL
COPPER
LEAD
ALUMINUM
ASBESTOS-CEMENT
CONCRETE PIPE
PLASTIC PIPE
TEMPERATURE
Aggressiveness of water Increases with temperature
up to = 80°C; at higher temperatures the aggres-
siveness decreases (37).
Increasing temperature will Increase corrosion
(16).
Increased water temperature above 25°C results 1n a
significant increase In pitting susceptibility.
Temperature effects are complex but usually not a
major factor.
Temperatures 20°C and less are preferred for corrosion
control (26).
Prefer higher temperatures of 40°C and up (15).
-
-
-
COMMENTS
Effects of any single variable are influenced by other
parameters, especially the Interrelation between pH,
temperature, dissolve oxygen, alkalinity, hardness,
TDS, and velocity.
Interrelation exists between pH, hardness, temperature,
alkalinity, TDS, plus organic acids or other stabiliz-
ing agents like phosphates or silicates.
Different types of Stainless Steel have different
corrosive tendencies. Cl" and Dissolve Oxygen are two
most important chemical factors in stainless steel
corrosion.
Copper concentration generally does not exceed E PPM —
may be limited by solubility of reaction product.
Aluminum corrosion is highly dependent on the period
of immersion.
-
Corrosion control is practiced by minimizing the disso-
lution of Ca++, often by regulating CaCOj stability
components.
1. Corrosion products that have been found are thought
to leach from solvents used for joints.
2. No variable cause-effect testing results are avail-
able
I
I—>
o
-------
TABLE 48. APPLICATIONS OF CORROSION CONTROL MECHANISMS
LININGS INHIBITORS !
Surface Material
Tanks:
Concrete
Steel
Pipes:
Iron
Steel
Asbestos-Cement
Reinforced Concrete
Lead
Copper
Plastic
Galvanized
Aluminum
4-
ra
\-
I™""
fO
o
o
^
J
TO
•M
4.
O
^ _>
c
o
E
0)
>7
/
»x
- — ...J
1
>^
X
o
LU
i
>'
/
1
,«
^.
C
•r-
>
^
^™
O
(13 O
o s-
CL.
/
..,_
Pi
O
o
n)
O
/
>x
/
/
/
/
/
10
ai
4«>
«3
O
•1—
r~~
tf~
00
y
/
/
! i
i
1 i
i
to
OJ
4_)
'O
^C^
CL
en
O
jr
a.
1
i
i
i
i
i
/
/
y
/
/
^
i
i
i
i
i
;
- -,-r
-------
REFERENCES
1. Anderson, E. A., C. E. Reinhard, and W. D. Hammel, "The Corrosion of
Zinc in Various Waters!' JAWWA. Vol. 26. No. 2, 1934, pp. 49-60.
2. Bell, Winifred,"Effect of Calcium Carbonate on Corrosion of Aluminum
in Waters Containing Chloride and Copper;1 Journal of Applied Chemistry,
12, February, 1962, pp. 53-55.
3. Buelow, R. W., J. R. Millette, E. F. McFarren, and J. M. Symons,"The
Behavior of Asbestos-Cement Pipe Under Various Water Quality Conditions"
Presentation-!979 AWWA Conference, San Francicso, June 27, 1979.
4. Burgmann, 6., W. Frieke, and W. S. Schwenk,"Chemical Corrosion and
Hygienic Aspects of The Use of Hot-Galvanized Threaded Pipes in Domestic
Plumbing for Drinking Water" Pipes & Pipelines Int., Vol. 23, No. 2,
1978, pp. 11-15.
5. Campbell, Hector S., A Natural Inhibition of Pitting Corrosion of
Copper in Tap-Waters^' J. Appl. Chem.. Vol. 4, 1954, pp. 633-647.
6. "Cold-Water Corrosion of Copper Tubing,'Task Group Report, JAWWA, Vol.
52, August, 1960, pp. 1033-1040.
7. Davies, D. H., and G. T. Burstein, "The Effects of Bicarbonate on The
Corrosion and Passivation of Iron" Corrosion-NACE, Vol. 36, No. 8,
August 1980, pp. 416-422.
8. DeWaard, C. and D. E. Milliams, "Carbonic Acid Corrosion of Steel",
Corrosion-NACE. Vol. 31. No. 5, May, 1975, pp. 177-181.
9. Donaldson, W., "The Action of Water on Service Pipes" JAWWA, Vol. 11,
No. 3, 1924, p. 649.
10. Eliassen, R., C. Pereda, A. J. Romeo and R. T. Skrinde,"Effects of pH
and Velocity on Corrosion of Steel Water Pipes" JAWWA, Vol. 48,
August, 1965, pp. 1005-1018.
11. Fergenbaum, C. L. Gabor and J. Yahalom, "Scale Protection Criteria in
Natural Waters" Corrosion (Houston), Vol. 34, No. 4, 1978, pp. 133-137.
12. Gardels, M. and Schock, E.P.A. Cincinnati Laboratory, Personnel
Communication via P. Lassovszky, Jan. 12, 1981.
13. Garrels, R. M., M. E. Thompson and R. Si ever, "Control of Carbonate
Solubility by Carbonate Complexes" American Journal of Science, Vol.
259, January, 1961, pp. 24-45.
14. Geld Isidore, and Colin McCaul, "Corrosion in Potable Water" JAWWA. Vol.
67, No. 10, October, 1975, pp. 549-552.
7-12
-------
15. Godard, H.P.,"The Corrosion Behavior of Aluminum in Natural Waters"
The Canadian Journal of Chemical Engineering, Vol. 38, No. 5, October,
I960, pp. 167-173.
16. Goetchins, D.R. ."Porcelain Enamel as a Protective Coating for Hot Water
Tanks1,1 J.Am. Ceramic Society, Vol. 25, 1942, pp. 164-168.
17. Goldfarb, A.S., J. Konz, and P. Wai ker,"Interior Coatings in Potable
Water Tanks and Pipelines" MITRE Corporation. Technical Report
MTR-7803, U.S. EPA Contract No. 68-01-4635, January, 1979.
18. Hale, F. E.,"Relation of Copper and Brass Pipe to Health" Water Works
Eng., Vol. 95, 1942.
19. Hatch, G.B.,"Unused Cases of Copper Corrosion,"JAWWA, Vol. 53, 1961,
pp. 1417-1429.
20. Karalekas, P. C., G. F. Craun, A. F. Hammonds, C. R. Ryan, and D. J.
Worth,"Lead and Other Trace Metals in Drinking Water in The Boston
Metropolitan Area,"J. New England Water Works Association, Vol. 90, No.
2, pp. 150-172, 1976.
21. Lane, R. W., anc' C. H. Neff/'Materials Selection for Piping in Chemically
Treated Water Systems."Materials Protection, Vol. 8, No. 2., February,
1969, pp. 27-30.
22. Larson, C. D., Chief Technical Support Section, EPA Region I, letter to
John Hagopian, Rhode Island Dept. of Health, November 14, 1979.
23. Larson, T. E., and R. V. Skold,"Laboratory Studies Relating Mineral
Quality of Water to Corrosion of Steel and Cast Iron" Corrosion,
Vol. 14, June, 1958, pp. 43-46.
24. McCauley, R. F., and M. 0. Abdul!ah,"Carbonate Deposits for Pipe
Protection,"JAWWA. Vol. 50, 1958, pp. 1419-1428.
25. Merrell, D. T. and R. L. Sanks,"Corrosion Control by Deposition of
CaCO- Films. Part 1, A Practiced Approach for Plant Operators."JAWWA,
Vol. 69, November, 1977, pp. 592-599.
26. Moore, M. R. ,"Plumbosolvency of Waters."Nature. Vol. 243, May 25, 1973,
pp. 222-223.
27. O'Brien, J. E."Lead in Boston Water: Its Cause and Prevention,"
Journal of The New England Water Works Association. Vol. 90, No. 1,
January, 1976, pp. 173-180.
28. Porter, F. C. and S. E. Hadden,"Corrosion of Aluminum Alloys in
Supply Waters,"J. Applied Chemistry. Vol. 3, September, 1953, pp. 385-409.
7-13
-------
29. Reedy, D. R.,"Corrosion in The Water Works Industry,"Materials
Protection. Vol. 5, No. 9, September, 1966, pp. 55-59.
30. Rudek, R., Blankenhorn, R., and H. Sontheimer,"Verzbgerung der
Eisenoxidation Durch Naturliche Organische Wasserinhaltsstoffe und
Duren Auswirkung auf die Korrosion Van Schwarzen Stah1rohren,"Von
Wasser, Vol. 53, 1979, pp. 133-146.
31. Schafer, G. T./'Corrosion of Copper and Copper Alloys in New Zealand
Potable Waters" New Zealand Journal of Science. Vol. 5, Dec. 1962,
pp. 475-484.
32. Blunder, C. J., and W. K. Boyd, Summary Report on Lead - Its Corrosion
Behavior to ILZRO, Battelle Memorial Institute, Columbus, Ohio.
33. Streicher, Lee,"Effects of Water Quality on Various Metals,"JAWWA. Vol.
48, No. 3, March, 1956, pp. 219-238.
34. Stumm, W.,"Investigation on The Corrosive Behavior of Waters,"
Proceedings of The American Society of Civil Engineers, Vol. 86, No.
5A-6, November, 1960, pp. 27-45.
35. Tronstad, L., and R. Veimo,"The Action of Water on Copper Pipes" Water
and Water Eng., Vol. 42, May, 1940, pp. 189-191.
36. Tronstad, L., and R. Veimo,"The Action of Water on Copper Pipes,"Water
and Water Eng., Vol. 42, June, 1940, pp. 225-228.
37. Uhleg, H. H., Corrosion and Corrosion Control as Introduction to
Corrosion Science, John Wiley & Sons, Inc., New York, 1963.
38. Uhleg, H. H., The Corrosion Handbook, John Wiley and Sons, Inc., New
York, 1948.
39. Wagner, Ivo,"Influence of Water Quality and Water Treatment on
Corrosion and Coatings in Steel and Galvanized Steel Tubes."EUROCOR
77, 6th Euorpean Congress on Metallic Corrosion, (MET. A., 7807-72
0184), 1977, pp. 413-419.
40. Waring, F. H.,"Prevention of Corrosion by The Application of
Inhibitors,"JAWWA, Vol. 30, No. 5, 1938, pp. 736-745.
7-14
-------
SECTION 8.0
RECOMMENDATIONS
The U.S. Environmental Protection Agency was mandated by the Safe Drinking
Water Act of 1974 (PL 93-523) to safeguard public drinking water supplies
and to protect the public health. Under the act, EPA is required to
establish and enforce National Drinking Water Regulations which include
corrosion by products in the water that pose' a threat to human health. In-
formation to form a basis for developing responsible and implementable cor-
rosion control regulations is partially available from results of historical
laboratory and field studies reported in the literature since the 1920's.
Many of these investigations were conducted to identify and quantify water
quality and conditions of service characteristics which influence corrosion
of the various materials used in the water works industry. To supplement
this data base, the EPA has initiated additional studies. These studies
have focused on more clearly defining and quantifying the various aspects
of potential corrosion control strategies which can assist in developing a
reasonable corrosion control program. Recent primary emphasis has been to
determine the magnitude of the problems and to search for both monitoring
and corrective actions which can be enforced. Results of these studies
have overwhelmingly concluded that the nature of corrosion and possible
corrosion control alternatives are extremely complex. Limiting the cor-
rosiveness of water by the use of a universal corrosion index or parameter
is not feasible at this time. Instead it appears that corrosion control can
only be accomplished through a comprehensively applied program on a community
water system case by case basis.
An example of the complexity and effort involved in administering a
responsible corrosion control program is provided by the Seattle Water
Department. Their approach includes an extensive monitoring program
coupled with an attempt to provide a wide range of treatment techniques
including the addition of various corrosion inhibitors. Monitoring is con-
tinued simultaneously with the application of treatment alternatives to
evaluate performance. In some cases, the results of the monitoring programs
are not in agreement with, or at least do not reflect the expectations of
th.e laboratory results. From this experience and from the results presented
in the literature it is evident that a complicated and potentially expensive
program is necessary to insure the success of corrosion control at each
community water system. The following procedures outline a comprehensive
program for corrosion control for specific water utilities.
8-1
-------
1) Identify and quantify the materials used in the
respective water works industry.
2) Characterize the delivered water quality with
respect to pH, and concentrations of alkalinity,
hardness, carbon dioxide, metal ions, total
dissolved solids, organic acids, and temperature
as a minimum.
3) Identify potential contaminants which would be
indicative of corrosion with respect to materials
used and water quality characteristics.
4) Develop and implement corrosion control treatment
alternatives which may be feasible for the particular
system.
5) Continue monitoring to determine effectiveness of
corrosion control strategies.
6) Develop guidelines for future construction practices
and local plumbing codes.
With an estimated 60,000 public water suppliers, it is evident that to
administer or enforce a program, with these requirements and recommendations,
on a nationwide scale would be both technically and economically prohibitive.
In many cases the treatment plant operators do not have the technical skills
nor do the small municipalities have the financial resources to implement
an extensive corrosion control program. Because some potential control
methods may enhance corrosion in specific instances, corrective action by
unskilled personnel may further aggravate an existing problem. It is also
important to note that relatively few water suppliers recognize a corrosion
problem or consider corrosion control a high priority.
It is recommended that corrosion control regulations be developed which
are both technically and economically reasonable and which can be enforced
effectively. This recommendation implies that regulations should be developed
which would first screen water suppliers to assess the potential for corrosion
in their respective systems. Only those systems suspected of having cor-
rosion problems should be required to initiate further investigations and
corrective actions.
8-2
-------
Specifically, water suppliers should be required to identify and quantify
materials and practices used in their respective water systems. For smaller
utilities this inventory can perhaps be easily compiled. For larger utilities,
an accurate inventory may be impossible to compile. In these cases, an in-
ventory estimate through use of recent records and historical plumbing codes
can be made. This approach implies that developed areas within a large
municipality can be sectioned with respect to historical growth, and applicable
plumbing codes, which were in affect during that development period, can be
superimposed on the respective area to provide some estimate. This procedure
will require that large municipalities provide an historical review of re-
spective local and state plumbing codes.
Water suppliers should also be required to conduct water quality in-
vestigations throughout their systems to assess corrosivity potential. As
addressed in this study, water quality parameters which have historically
been investigated to evaluate corrosion characteristics of specific materials
used in the water works industry are pH, and concentrations of hardness,
alkalinity, dissolved oxygen, carbon dioxide, transition metal ions, total
dissolved solids, organic acids, and temperature. It is recommended that
these water quality parameters be included as a minimum for a water quality
characterization portion of a corrosion control program.
The water quality sampling and analysis program must be designed to
effectively characterize the water quality conditions within the system.
Samples should be taken from both the water source and at locations following
treatment before it enters the distribution system. Samples should also be
taken at various representative points along the distribution system as well
as at consumer taps. It should be recognized, however, that water quality
changes may occur as the water passes through the distribution system and
these changes may produce erroneous results. For example, as previously
discussed, corrosive water parsing through asbestos-cement pipe will tend
to leach calcium from the pipe and become less aggressive. Corrosive water
samples taken from points long distances downstream of asbestos-cement pipe
sections may appear non-corrosive.
The water quality data collected, specifically pH, alkalinity, and
hardness can be used to develop corrosion indicators such as the Langelier
Index and/or the Aggressive Index. Although limited in use, these two
indices are apparently the most widely accepted indicators of corrosiveness
of water and should be used as appropriate. However, their limitations should
be recognized as previously discussed and corrosivity should not be assessed
exclusively by these parameters.
8-3
-------
Water suppliers which are identified as having conditions susceptible
to corrosion should be encouraged to initiate monitoring programs designed
in accordance with the results of the materials and practice inventory and
the water quality survey. Those utilities which are found to use lead or
materials which are sources of contaminants that adversely affect health
extensively should be given priority consideration.
Water systems suspect of having corrosion problems should be encouraged
to initiate a monitoring and detection program to determine the nature and
extent, if any, of corrosion. The design and extent of this monitoring
program is contingent on the potential contaminants which can be expected as
dictated by the specific materials used. Potential contaminants which are
associated with each material are listed in Table 49. It is suggested that
non-subjective (i.e. analytical as opposed to visual) monitoring techniques
be used and that samples be taken from consumer's taps and distribution mains.
Coupon testing and electrochemical techniques should currently be considered
only as indicators of corrosion. It is not necessary to monitor for every
associated potential contaminant but rather to monitor for only a few. The
potential magnitude of the corrosion problem and the staffing and financial
resources of specific water suppliers should dictate the extent of the mon-
itoring programs.
After identification of the specific corrosion related contaminants, an
applicable treatment technique relating to the material involved and water
quality parameters must be devised. Historically, attempts at corrosion
control by CaC03 deposition have been used most often, but corrosion in-
hibitors or modifications in calcium carbonate control as noted in chapter
6 should also be considered.
Local jurisdictions should respond to corrosion problems by adjusting
the local plumbing code to recommend some pipe materials and not allow others
to be used. Lead and unlined asbestos cement piping should be reviewed.
Plastic pipe is becoming more accepted and appears to resist corrosion, but
the possibility of small amounts of potentially harmful organics entering
the water should be further researched. Similarly, pipe coatings should be
reviewed in relation to the local waters. The low flow resistance of epoxy
coatings may be recommended pursuant to research on trace contaminant release.
8-4
-------
TABLE 49. MATERIALS AND THEIR ASSOCIATED CORROSION PRODUCTS
Material Potential Corrosion Products
Iron-tased materials Fe, Cd, Pb, Zn
Copper Cu, Fe, Zn, Sn, Pb, Mn, As,
Sb, P, Bi
Lead Pb
Aluminum Cu, Mg, Si, Fe, Mn, Cr,
Asbestos-Cement Asbestos fibers, and
tetrachlorethylene
Concrete Si, Al, Fe, Mg, S
Plastic Pb, components of various
solvents including 2-butanone
(MEK) and tetrahydrofuran
If several pipe materials are corroding, and the products are identified,
it may not be possible to engineer a single program to alleviate all of the
contaminants. In this case, where it is possible that the solution to one
facet of the problem may aggravate another, the primary concerns should be
control of those products deemed most detrimental in terms of health effects.
Specifically, this would include lead and asbestos fibers. These may not be
the easiest contaminants to justify in terms of economics or aesthetics, but
their control should be paramount.
Irrespective of corrosion control regulations, the EPA should review
local, state, and national plumbing codes and begin assessing the potential
for discouraging the use of potentially dangerous materials and practices.
Lead and lead-based solders should be seriously considered for discontinuing
their extensive use. Although a large quantity of lead pipe is currently in
service, its extensive use has begun to decline, but lead-based solders are
currently the most widely used solders for joining copper pipers.
3-5
-------
TECHNICAL REPORT DATA
(flette read Instructions on the rtvene be fort completing}'
. REPORT NO.
2.
3. RECIPIENT'S ACCESSION NO.
«. TITLE AND SUBTITLE
Corrosion in Potable Water Systems
6. REPORT DATE
February, 1982
I. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
David W. DeBerry, James R. Kidwell, David A. Malish
B. PERFORMING ORGANIZATION REPORT NO.
B. PERFORMING ORGANIZATION NAME AND ADDRESS
Sum X Corporation
P.O. Box 14864
1300 E. Braker Lane
Austin, Texas 78761
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Criteria and Standards Division
Office of Drinking Water
U.S. Environmental Protection Agency
Washington, D.C. 20460
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRAC
r The purpose of this investigation was to collect, review, evaluate, and
present existing information to determine whether a sufficient data base is
available to develop corrosion control regulations for the water works indus-
try as required by the Safe Drinking Water Act. To accomplish this objective,
an exhaustive literature search was completed which included a review of the
various materials used in the water works industry and their corrosion char-
acteristics. Results of laboratory and field research on each material as
related to corrosion in the water works industry are extensively reviewed and
data is presented as appropriate. Major emphasis is placed on assessing the
conditions of service and water quality characteristics in potable water
systems on the corrosion or deterioration of each material. A review of
corrosion monitoring and detection techniques is given which addresses the
various methodologies used to identify and evaluate corrosive waters. Avail-
able corrosion prevention and control techniques are also evaluated and pre-
sented. Additionally, case histories of corrosion control programs are
presented for examples. Finally, the information and data presented in these
reviews are compiled and presented in tabular form. These tables provide an
overall view of the nature of the corrosion problems in the water works indus-
try and can be used as a guide for the initial consideration of corrosion
control regulations.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
. COS AT i Field/Group
Corrosion, Water Supply Systems, Corrosion
Control, Corrosion Treatment, Drinking
Water Supply, Corrosion And Water Quality,
Corrosion Monitoring, Materials in Distri-
bution Systems.
Corrosion
Corrosion Control
Water Supply
Corrosion Measurement
8. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (Thit Report)
Unclassified
21. NO. OF PAGES
220
20. SECURITY CLASS (Thiipage)
Unclassified
22. PRICE
EPA P*«» '220-1 (R«». 4-77) Pncviouk COITION is OMOLKTE
------- |