University of Central Florida University of Central Florida

STARS STARS

Electronic Theses and Dissertations, 2004-2019

2008

Effects Of Orthophosphate Corrosion Inhibitor In Blended Water Effects Of Orthophosphate Corrosion Inhibitor In Blended Water

Quality Environments Quality Environments

Erica Stone University of Central Florida

Part of the Environmental Engineering Commons

Find similar works at: https://stars.library.ucf.edu/etd

University of Central Florida Libraries http://library.ucf.edu

This Doctoral Dissertation (Open Access) is brought to you for free and open access by STARS. It has been accepted

for inclusion in Electronic Theses and Dissertations, 2004-2019 by an authorized administrator of STARS. For more

information, please contact STARS@ucf.edu.

STARS Citation STARS Citation Stone, Erica, "Effects Of Orthophosphate Corrosion Inhibitor In Blended Water Quality Environments" (2008). Electronic Theses and Dissertations, 2004-2019. 3609. https://stars.library.ucf.edu/etd/3609

EFFECTS OF ORTHOPHOSPHATE CORROSION INHIBITOR IN BLENDED WATER QUALITY ENVIRONMENTS

by

ERICA DIANE STONE B.S. University of Central Florida, 2006

A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy

in the Department of Civil, Environmental, and Construction Engineering in the College of Engineering and Computer Science

at the University of Central Florida Orlando, Florida

Fall Term 2008

Major Professors: James S. Taylor and Steven J. Duranceau

ii

© 2008 Erica D. Stone

iii

ABSTRACT

This study evaluated the effects of orthophosphate (OP) inhibitor addition on iron,

copper, and lead corrosion on coupons exposed to different blends of groundwater, surface

water, and desalinated seawater. The effectiveness of OP inhibitor addition on iron, copper, and

lead release was analyzed by statistical comparison between OP treated and untreated pilot

distribution systems (PDS). Four different doses of OP inhibitor, ranging from zero (control) to

2 mg/L as P, were investigated and non-linear empirical models were developed to predict iron,

copper and lead release from the water quality and OP doses. Surface characterization

evaluations were conducted using X-ray Photoelectron Spectroscopy (XPS) analyses for each

iron, galvanized steel, copper, and lead/tin coupon tested. Also, a theoretical thermodynamic

model was developed and used to validate the controlling solid phases determined by XPS.

A comparison of the effects of phosphate-based corrosion inhibitor addition on iron,

copper, and lead release from the PDSs exposed to the different blends was also conducted.

Three phosphate-based corrosion inhibitors were employed; blended orthophosphate (BOP),

orthophosphate (OP), and zinc orthophosphate (ZOP). Non-linear empirical models were

developed to predict iron, copper, and lead release from each PDS treated with different doses of

inhibitor ranging from zero (control) to 2 mg/L as P. The predictive models were developed

using water quality parameters as well as the inhibitor dose. Using these empirical models,

simulation of the water quality of different blends with varying alkalinity and pH were used to

compare the inhibitors performance for remaining in compliance for iron, copper and lead

release.

iv

OP inhibitor addition was found to offer limited improvement of iron release for the OP

dosages evaluated for the water blends evaluated compared to pH adjustment alone. Empirical

models showed increased total phosphorus, pH and alkalinity reduced iron release while

increased silica, chloride, sulfate and temperature contributed to iron release. Thermodynamic

modeling suggested that FePO4 is the controlling solid that forms on iron and galvanized steel

surfaces, regardless of blend, when OP inhibitor is added for corrosion control. While FePO4

does not offer much control of the iron release from the cast iron surfaces, it does offer

protection of the galvanized steel surfaces reducing zinc release.

OP inhibitor addition was found to reduce copper release for the OP dosages evaluated

for the water blends evaluated compared to pH adjustment alone. Empirical models showed

increases in total phosphorus, silica and pH reduced copper release while increased alkalinity and

chloride contributed to copper release. Thermodynamic modeling suggested that

Cu3(PO4)2·2H2O is the controlling solid that forms on copper surfaces, regardless of blend, when

OP inhibitor is added for corrosion control.

OP inhibitor addition was found to reduce lead release for the OP dosages evaluated for

the water blends evaluated compared to pH adjustment alone. Empirical models showed

increased total phosphorus and pH reduced lead release while increased alkalinity, chloride, and

temperature contributed to lead release. Thermodynamic modeling suggested that

hydroxypyromorphite is the controlling solid that forms on lead surfaces, regardless of blend,

when OP inhibitor is added for corrosion control.

The comparison of phosphate-based inhibitors found increasing pH to reduce iron, copper

and lead metal release, while increasing alkalinity was shown to reduce iron release but increase

v

copper and lead release. The ZOP inhibitor was not predicted by the empirical models to

perform as well as BOP and OP at the low dose of 0.5 mg/L as P for iron control, and the OP

inhibitor was not predicted to perform as well as BOP and ZOP at the low dose of 0.5 mg/L as P

for lead control. The three inhibitors evaluated performed similarly for copper control.

Therefore, BOP inhibitor showed the lowest metal release at the low dose of 0.5 mg/L as P for

control of iron, copper and lead corrosion.

vi

ACKNOWLEDGMENTS

Special thanks to Dr. James Taylor and Dr. Steven Duranceau for serving as Chairs of my

committee and for all of their guidance and support. They have both been advisors and mentors

to me throughout my education and professional development. Thank you also to my other

committee members; Dr. Andrew Randall, Dr. Ni-bin Chang and Dr. Christian Clausen for

serving on my committee and providing valuable feedback for this work.

Thanks to Tampa Bay Water (TBW); Hillsborough County, Fla.; Pasco County, Fla.;

Pinellas County, Fla.; City of New Port Richey, Fla.; City of St. Petersburg, Fla.; and City of

Tampa, Fla., which are the Member Governments of TBW; and the American Water Works

Association Research Foundation (AwwaRF) for their support and funding of this project.

This work would not have been possible without all the hard work and dedication of the

faculty and students that worked on this project. Thanks to Dr. James Taylor, Dr. John Dietz,

Dr. Andrew Randall, Dr. Charles Norris, Maria Pia Real-Robert, Abdulrahman Alshehri, Jorge

Arevalo, Xiaotao Guan, Phillip Lintereur, David MacNevin, Raj Vaidya, Bingjie Zhao, Stephen

Glatthorn and Avinash Shekhar.

Finally, thanks to my family, especially my parents, for all the support, love and

encouragement they have given me over the years in furthering my education.

vii

TABLE OF CONTENTS

LIST OF FIGURES ....................................................................................................................... xi

LIST OF TABLES....................................................................................................................... xiv

CHAPTER 1 INTRODUCTION .................................................................................................... 1

Objectives ................................................................................................................................... 2

Theory ......................................................................................................................................... 2

Corrosion................................................................................................................................. 2

Phosphate Inhibitor ................................................................................................................. 5

Literature Review........................................................................................................................ 6

Water Blending ....................................................................................................................... 6

Metal Release.......................................................................................................................... 7

Modeling............................................................................................................................... 14

Surface Characterization....................................................................................................... 17

References................................................................................................................................. 19

CHAPTER 2 EFFECTS OF ORTHOPHOSPHATE CORROSION INHIBITOR ON IRON IN

BLENDED WATER QUALITY ENVIRONMENTS ................................................................. 24

Abstract ..................................................................................................................................... 24

Introduction............................................................................................................................... 25

Experimental Methods .............................................................................................................. 27

Experimental Design............................................................................................................. 27

Orthophosphate Inhibitor ...................................................................................................... 32

viii

Data Collection ..................................................................................................................... 33

Results and Discussion ............................................................................................................. 36

Dose Maintenance................................................................................................................. 36

Empirical Modeling .............................................................................................................. 37

Orthophosphate Inhibitor Performance................................................................................. 41

Surface Characterization....................................................................................................... 47

Thermodynamic Modeling.................................................................................................... 52

Conclusions............................................................................................................................... 56

References................................................................................................................................. 59

CHAPTER 3 EFFECTS OF ORTHOPHOSPHATE CORROSION INHIBITOR ON COPPER

IN BLENDED WATER QUALITY ENVIRONMENTS ............................................................ 61

Abstract ..................................................................................................................................... 61

Introduction............................................................................................................................... 62

Experimental Methods .............................................................................................................. 64

Experimental Design............................................................................................................. 64

Orthophosphate Inhibitor ...................................................................................................... 69

Data Collection ..................................................................................................................... 70

Results and Discussion ............................................................................................................. 73

Dose Maintenance................................................................................................................. 73

Empirical Modeling .............................................................................................................. 75

Orthophosphate Inhibitor Performance................................................................................. 78

Surface Characterization....................................................................................................... 83

ix

Thermodynamic Modeling.................................................................................................... 86

Copper Coupons.................................................................................................................... 90

Conclusions............................................................................................................................... 91

References................................................................................................................................. 93

CHAPTER 4 EFFECTS OF ORTHOPHOSPHATE CORROSION INHIBITOR ON LEAD IN

BLENDED WATER QUALITY ENVIRONMENTS ................................................................. 95

Abstract ..................................................................................................................................... 95

Introduction............................................................................................................................... 96

Experimental Methods .............................................................................................................. 99

Experimental Design............................................................................................................. 99

Orthophosphate Inhibitor .................................................................................................... 104

Data Collection ................................................................................................................... 104

Results and Discussion ........................................................................................................... 107

Dose Maintenance............................................................................................................... 107

Empirical Modeling ............................................................................................................ 109

Orthophosphate Inhibitor Performance............................................................................... 113

Surface Characterization..................................................................................................... 118

Thermodynamic Modeling.................................................................................................. 123

Conclusions............................................................................................................................. 129

References............................................................................................................................... 131

CHAPTER 5 COMPARISON OF PHOSPHATE INHIBITORS FOR IRON, COPPER AND

LEAD CORROSION CONTROL.............................................................................................. 133

x

Abstract ................................................................................................................................... 133

Introduction............................................................................................................................. 134

Experimental Methods ............................................................................................................ 136

Experimental Design........................................................................................................... 136

Phosphate Inhibitors............................................................................................................ 141

Data Collection ................................................................................................................... 142

Results and Discussion ........................................................................................................... 145

Iron...................................................................................................................................... 145

Copper................................................................................................................................. 158

Lead..................................................................................................................................... 171

Conclusions............................................................................................................................. 184

References............................................................................................................................... 186

xi

LIST OF FIGURES

Figure 2-1 Pipe materials (left) and parallel PDSs (right) ............................................................ 29

Figure 2-2 Orthophosphate inhibitor dosing................................................................................. 36

Figure 2-3 pH for OP and control PDSs ....................................................................................... 37

Figure 2-4 Predicted versus actual total iron concentrations using the empirical model ............. 40

Figure 2-5 Actual vs. predicted for empirical model by phase and PDS...................................... 41

Figure 2-6 Comparison of total Fe by phase and treatment.......................................................... 43

Figure 2-7 Comparison of influent and effluent iron and zinc ..................................................... 47

Figure 2-8 Distribution of iron compounds on iron coupons ....................................................... 49

Figure 2-9 Distribution of iron compounds on galvanized steel coupons .................................... 52

Figure 2-10 Pourbaix diagram for iron species with 1 mg/L P inhibitor addition........................ 54

Figure 3-1 Pipe materials (left) and parallel PDSs (right) ............................................................ 66

Figure 3-2 Corrosion shed (left) and corrosion loops (right)........................................................ 73

Figure 3-3 Orthophosphate inhibitor dosing................................................................................. 74

Figure 3-4 pH for OP and control PDSs ....................................................................................... 75

Figure 3-5 Predicted versus actual total copper concentrations using the empirical model......... 77

Figure 3-6 Actual vs. predicted for empirical model by phase and PDS, 90th percentiles ........... 78

Figure 3-7 Comparison of total Cu by phase and treatment ......................................................... 80

Figure 3-8 Distribution of copper compounds.............................................................................. 86

Figure 3-9 Scale on coupons incubating for 14 months ............................................................... 90

Figure 4-1 Pipe materials (left) and parallel PDSs (right) .......................................................... 100

xii

Figure 4-2 Corrosion shed (left) and corrosion loops (right)...................................................... 107

Figure 4-3 Orthophosphate inhibitor dosing............................................................................... 108

Figure 4-4 pH for OP PDSs and controls.................................................................................... 109

Figure 4-5 Predicted vs. actual for total lead empirical model ................................................... 111

Figure 4-6 Actual vs. predicted for empirical model by phase and PDS, 90th percentiles ......... 112

Figure 4-7 Comparison of total Pb by phase and treatment........................................................ 116

Figure 4-8 Distribution of lead compounds ................................................................................ 122

Figure 4-9 Pourbaix diagram for lead species with 1 mg/L P inhibitor addition ....................... 125

Figure 4-10 pC-pH diagram assuming hydroxypyromorphite as the controlling solid .............. 127

Figure 5-1 Pipe materials (left) and parallel PDSs (right) .......................................................... 138

Figure 5-2 Corrosion shed (left) and corrosion loops (right)...................................................... 145

Figure 5-3 Model predicted versus actual total iron concentration ............................................ 147

Figure 5-4 Comparison of inhibitor treatments on total iron release.......................................... 148

Figure 5-5 Phase I water quality iron simulation varying pH..................................................... 150

Figure 5-6 Phase II water quality iron simulation varying pH ................................................... 151

Figure 5-7 Phase III water quality iron simulation varying pH.................................................. 152

Figure 5-8 Phase IV water quality iron simulation varying pH.................................................. 153

Figure 5-9 Phase I water quality iron simulation varying alkalinity .......................................... 154

Figure 5-10 Phase II water quality iron simulation varying alkalinity ....................................... 155

Figure 5-11 Phase III water quality iron simulation varying alkalinity...................................... 156

Figure 5-12 Phase IV water quality iron simulation varying alkalinity...................................... 157

Figure 5-13 Model predicted versus actual total copper concentration...................................... 160

xiii

Figure 5-14 Comparison of inhibitor treatments on total copper release ................................... 161

Figure 5-15 Phase I water quality copper simulation varying pH .............................................. 163

Figure 5-16 Phase II water quality copper simulation varying pH............................................. 164

Figure 5-17 Phase III water quality copper simulation varying pH............................................ 165

Figure 5-18 Phase IV water quality copper simulation varying pH ........................................... 166

Figure 5-19 Phase I water quality copper simulation varying alkalinity .................................... 167

Figure 5-20 Phase II water quality copper simulation varying alkalinity................................... 168

Figure 5-21 Phase III water quality copper simulation varying alkalinity ................................. 169

Figure 5-22 Phase IV water quality copper simulation varying alkalinity ................................. 170

Figure 5-23 Model predicted versus actual total lead concentration .......................................... 173

Figure 5-24 Comparison of inhibitor treatments on total lead release........................................ 174

Figure 5-25 Phase I water quality lead simulation varying pH .................................................. 176

Figure 5-26 Phase II water quality lead simulation varying pH ................................................. 177

Figure 5-27 Phase III water quality lead simulation varying pH................................................ 178

Figure 5-28 Phase IV water quality lead simulation varying pH................................................ 179

Figure 5-29 Phase I water quality lead simulation varying alkalinity ........................................ 180

Figure 5-30 Phase II water quality lead simulation varying alkalinity....................................... 181

Figure 5-31 Phase III water quality lead simulation varying alkalinity...................................... 182

Figure 5-32 Phase IV water quality lead simulation varying alkalinity ..................................... 183

xiv

LIST OF TABLES

Table 2-1 Pipe materials in PDSs ................................................................................................. 28

Table 2-2 Finished source water descriptions............................................................................... 29

Table 2-3 Blend percentages for each phase................................................................................. 31

Table 2-4 Average water quality by phase ................................................................................... 31

Table 2-5 Orthophosphate inhibitor properties............................................................................. 33

Table 2-6 Water quality parameters and methods performed at University Laboratory .............. 34

Table 2-7 Water quality parameters and methods performed at Field Laboratory....................... 35

Table 2-8 Range of water quality in model development............................................................. 38

Table 2-9 Iron release summary ................................................................................................... 42

Table 2-10 Average influent water quality by PDS and phase..................................................... 44

Table 2-11 Elements found in XPS analysis of iron coupons ...................................................... 48

Table 2-12 Elements found in XPS analysis of galvanized steel coupons ................................... 50

Table 3-1 Pipe materials in PDSs ................................................................................................. 65

Table 3-2 Finished source water descriptions............................................................................... 66

Table 3-3 Blend percentages for each phase................................................................................. 68

Table 3-4 Average water quality by phase ................................................................................... 68

Table 3-5 Orthophosphate inhibitor properties............................................................................. 70

Table 3-6 Water quality parameters and methods performed at University Laboratory .............. 71

Table 3-7 Water quality parameters and methods performed at Field Laboratory....................... 72

Table 3-8 Range of water quality in model development............................................................. 76

xv

Table 3-9 Copper release summary .............................................................................................. 79

Table 3-10 Average influent water quality by PDS and phase..................................................... 81

Table 3-11 Elements detected in XPS analysis............................................................................. 84

Table 3-12 Thermodynamic modeling calculations for pH control PDSs.................................... 88

Table 3-13 Thermodynamic modeling calculations for OP PDSs................................................ 89

Table 4-1 Pipe materials in PDSs ................................................................................................. 99

Table 4-2 Finished source water descriptions............................................................................. 101

Table 4-3 Blend percentages for each phase............................................................................... 103

Table 4-4 Average water quality by phase ................................................................................. 103

Table 4-5 Orthophosphate inhibitor properties........................................................................... 104

Table 4-6 Water quality parameters and methods performed at University Laboratory ............ 105

Table 4-7 Water quality parameters and methods performed at Field Laboratory..................... 106

Table 4-8 Water quality range in model development................................................................ 110

Table 4-9 Lead release summary ................................................................................................ 113

Table 4-10 Observations below detection................................................................................... 114

Table 4-11 Average influent water quality by PDS and phase................................................... 115

Table 4-12 Elements found in XPS analysis............................................................................... 118

Table 5-1 Pipe materials in PDSs ............................................................................................... 137

Table 5-2 Finished source water descriptions............................................................................. 138

Table 5-3 Blend percentages for each phase............................................................................... 140

Table 5-4 Average water quality by phase ................................................................................. 140

Table 5-5 Inhibitor product properties........................................................................................ 142

xvi

Table 5-6 Water quality parameters and methods performed at University Laboratory ............ 143

Table 5-7 Water quality parameters and methods performed at Field Laboratory..................... 144

Table 5-8 Average water quality of each phase.......................................................................... 145

Table 5-9 Average water quality of each phase.......................................................................... 158

Table 5-10 Average water quality of each phase........................................................................ 171

1

CHAPTER 1 INTRODUCTION

This document is a Ph.D. dissertation entitled “Effects of Orthophosphate Corrosion

Inhibitor in Blended Water Quality Environments”. This dissertation is submitted in partial

fulfillment of the requirements for the Doctor of Philosophy in Environmental Engineering

degree in the Department of Civil and Environmental Engineering in the College of Engineering

and Computer Sciences at the University of Central Florida.

The Department of Civil and Environmental Engineering at the University of Central

Florida conducted a study to determine the impact of distribution system water quality in a

changing water quality environment using inhibitors. Corrosion inhibitors were utilized to offset

the adverse effects of changing finished water quality in the distribution system. The research

was conducted in an existing facility consisting of fourteen 80 foot pilot distribution systems

(PDSs). These were built from actual PVC, lined cast iron, unlined cast iron and galvanized

steel distribution system pipe. Three doses of each blended orthophosphate (BOP),

orthophosphate (OP) and zinc orthophosphate (ZOP) corrosion inhibitors, as well as pH control

strategies, were added to three blends of finished ground, surface and desalinated waters for

quarterly intervals over the course of a year. This allowed for the capability of corrosion

inhibitor to control distribution system water quality in a changing water quality environment to

be evaluated by process, pipe material, water quality and season.

This study evaluate the impacts of orthophosphate corrosion inhibitor on distribution

system water quality. Effects on metal release and insight into the mechanism by which

2

orthophosphate inhibits corrosion will be examined as well as a comparison to two other

phosphate based inhibitors; blended orthophosphate and zinc orthophosphate.

Objectives

The objectives of this research are:

• Determination of the effects of orthophosphate inhibitor addition on iron, copper and

lead release.

• Development of mathematical models to predict iron, copper and lead release from

distribution systems with orthophosphate inhibitor addition using water quality

parameters.

• Identify controlling solid phases on the surface of iron, galvanized steel, copper and

lead/tin coupons exposed to orthophosphate inhibitor to help determine the

mechanism of corrosion control.

• Compare phosphate-based corrosion inhibitors based on simulation of water quality

with developed mathematical models.

Theory

Corrosion

Corrosion is the oxidation process by which the native metal is converted to an oxidized

species. The metals used in distribution systems are generally not stable in water and will

oxidize, or corrode (Montgomery Watson Harza 2005). Slow corrosion will often convert the

3

metal directly to an oxide on the metal surface. Rapid corrosion will often cause the oxidized

metal to release into solution.

In order for corrosion to occur there must be an anode, a cathode, a conductor and a

conducting electrolyte. For example, with iron corrosion the iron metal will lose two electrons at

the anode and go into solution as ferrous iron. At the cathode, the hydrogen ions near the metal

surface accept the electrons generated by the iron and become reduced, and then combine to

form hydrogen molecules. The electrons will travel from one site to another through the

conductor, the pipe. The conducting solution, water, completes the electrical circuit.

Calcium carbonate saturation is thought to be a means of controlling corrosion in

distribution piping. The supersaturated solution deposits a protective layer of calcium carbonate

on the inside of the pipe. The Langelier saturation index is a measure of calcium carbonate

saturation that makes use of pHs, saturation pH, at which alkalinity and calcium hardness are in

equilibrium with each other as defined in Equation (1-1) through Equation (1-5).

][][][2][ 233

+−−− −++= HOHCOHCOAlk (1-1)

where alkalinity is defined with terms in molar quantities

]][[ 23

2 −+= COCaK s (1-2)

where Ks is the solubility constant for calcium carbonate. There are also related by the

equilibrium constant for the dissociation of bicarbonate.

4

+−− +↔ HCOHCO 233 (1-3)

][]][[

3

23

2 −

−+

=HCO

COHK

(1-4)

These simplify to calculate pHs.

][][ 22 AlkpCappKpKpH ss ++−= + (1-5)

The Langelier saturation index (LSI) is the difference in pHs and actual pH. Many

operate at neutral to slightly positive LSI in an effort to prevent corrosion.

In iron pipe, the formation of corrosion scale is the cause of red water problems. This

scale is made mostly of ferric oxides and hydroxides and can form tubercles over time.

Corroding iron will form ferric hydroxide, ferrous hydroxide and ferrous carbonate, with the

carbonate scale being the densest and most likely to help protect the pipe. Tubercles may form

where the elemental metal corrodes. As iron corrodes it becomes an electron donor and electrons

may travel through the metal or through magnetite (Fe2O3), which is a semiconductor. Oxygen

in the water becomes an electron acceptor and is reduced to water, consuming hydrogen ions.

This results in elevated pH which will cause localized calcium carbonate precipitation. As Fe2+

is generated, anions diffuse to maintain electroneutrality and anions like chloride can accelerate

the process.

5

Copper pipe corrosion occurs when copper is in water saturated with oxygen and an

adherent film of cuprous oxide forms. This film is a semiconductor in good electrical contact

with the metal itself. Oxygen is reduced to water on the outside of this layer. When this layer

reaches a certain thickness, a new oxide layer develops with a mixture of cuprous and cupric

salts. This outer later will form a porous scale layer due to the difference in coordination

requirements of the two layers.

Lead corrosion depends strongly on the pH and alkalinity of the water. At lower pHs,

lead is stable in the Pb2+ form. Lead carbonate is the stable form at neutral pHs. At higher pHs,

hydroxycarbonate and hydroxide forms of lead are favored. Lead becomes more soluble as pH

decreases and the Pb2+ and PbCO3 govern the distribution of oxidized forms. Since Pb2+ is a

dissolved ion and PbCO3 is a solid scale, lead release is greater at lower pHs.

Phosphate Inhibitor

Addition of phosphates to drinking water has long been used as a corrosion inhibitor.

The phosphate is thought to behave as an anodic inhibitor. This may be related to the inhibition

of the dissolution of the oxide layer or due to the formation of complexes on the surface that

compete with the ligands attempting to promote dissolution protective scale (Montgomery

Watson Harza 2005).

Orthophosphate inhibitor is added in the form of phosphoric acid (H3PO4). Phosphoric

acid dissociates to H2PO4-, HPO42-, and PO43- with ionization constants of pK1 equal to 2.1, pK2

equal to 7.2, and pK3 equal to 12.3. These are illustrated below in Equation (1-6) to Equation

(1-8) (Snoeyink and Jenkins 1980).

6

−+ +↔ 4243 POHHPOH 1.2

43

421 10

][]][[ −

−+

==POH

POHHK

(1-6)

−+− +↔ 2442 HPOHPOH 2.7

42

24

2 10][

]][[ −−

−+

==POHHPOHK

(1-7)

−+− +↔ 34

24 POHHPO 3.12

24

34

3 10][

]][[ −−

−+

==HPO

POHK (1-8)

Literature Review

Water Blending

Numerous studies indicate that blending different source waters can cause corrosion.

Corrosion can lead to red water problems as well as high lead and copper concentrations.

Corrosion also contributes to pipe failures, bitter tasting water, and health concerns (Dietrich et

al. 2004; Plottu-Pecheux et al. 2001).

A study at the Paris Water Utility found that blending of nanofiltered and surface finished

waters caused carbonate imbalance in the water, which resulted in excessive lead and copper

release. A one-year research project was conducted to select a model to calculate the suitable

dosage of soda ash (sodium carbonate) and contact time to obtain a stable pH. The Hallopeau-

Dubin model, which uses flow rate, temperature, conductivity, calcium hardness, alkalinity, and

pH as input parameters, led to very satisfactory results and could be generalized to other water

mixtures (Plottu-Pecheux et al. 2001).

Imran et al. (2006) found that varying water quality to minimize iron release could lead to

increased copper and lead release within the distribution system. It was found that increasing

7

alkalinity was beneficial in reducing the release of iron products from pipes but increased the

corrosion of copper and lead pipes. Increasing sulfates was found to reduce the release of lead

but increased the release of iron. These conflicting water quality requirements for iron, copper

and lead release require evaluation of the tradeoffs between water quality and the corrosion

response of all metals subject to corrosion in the distribution system. Schock and Clement

(1998) also discuss the need for tradeoffs, especially when using chemical addition for corrosion

control of different metals.

Metal Release

Iron

Cast iron has been used in drinking water distribution systems for 500 years (Gedge

1992). Under the right conditions, iron is released into the water causing water quality problems

and red water complaints (Burlingame, Lytle, and Snoeyink 2006). Phosphate inhibitors were

first added to drinking water to prevent excessive calcite precipitation (Hoover and Rice 1939).

The phosphate was thought to prevent calcite crystals from forming outside of the colloidal range

by sorption to the calcium carbonate nuclei (Hatch and Rice 1939). More recently, research

suggests mixed results for phosphate addition being used to prevent iron corrosion and red water

(McNeill and Edwards 2001). Research suggests orthophosphate addition forms a scale, while

polyphosphates sequester iron (Wagner 1992; Benjamin et al. 1990; Boffardi 1988; Wagner and

Kuch 1984; Huang 1980; Pryor and Cohen 1951).

Corrosion of iron pipes in a distribution system can cause three distinct problems. First,

pipe mass is lost through oxidation to soluble iron species or iron-bearing scale. Second, the

8

scale can accumulate as large tubercles that increase head loss and decrease water capacity.

Finally, the release of soluble or particulate iron corrosion by-products to the water leads to

consumer complaints of "red water" at the tap (McNeill and Edwards 2001).

It has been suggested that the presence of phosphate may interfere with Fe2+ oxidation

(Benjamin, Sontheimer, and Leroy 1996). It has also been suggested that phosphate inhibitors

can act as plugs, filling the pores of the corrosion scales (Sarin et al. 2004). This helps to prevent

diffusion of oxidants into the lower layers of the corrosion scale and further corrosion of the iron

pipe.

McNeill and Edwards (2000a) investigated phosphate inhibitors in new iron pipes in up

to 72 hours of stagnant conditions. Five different water qualities (variable pH and alkalinity)

were used, and dosed with orthophosphate, polyphosphate, or none. Examination of iron release

measurements showed that both ortho and polyphosphates either increased or had no effect on

iron concentrations. The only tested conditions that showed reduced iron release were at high

alkalinity (300 mg/L as CaCO3) and pH of 7.2. This observation of phosphate inhibitors

influence in stagnant pipes is opposite to common experiences in flowing water conditions. On

average, other pH and alkalinity conditions showed an increase in iron release during stagnation

when polyphosphate is used, compared to pipes receiving no inhibitor.

Another study tested iron corrosion in reactors, as well as the biological growth, with and

without phosphate inhibitor. Increasing phosphate doses were shown to help maintain a lower

corrosion rate. These increased doses are recommended for summer months because corrosion

rate was found to be highly related to seasonal variables, namely temperature. Lower phosphate

9

doses were recommended for the winter. The addition of phosphate did not increase biological

levels (Volk et al. 2000).

Another study involving temperature effects on iron corrosion was done with samples

held at a constant temperatures and samples that were exposed to a cycle of different

temperatures. They found samples of iron held at 5 °C had higher iron concentrations and

tuberculation than samples at 20 °C or 25 °C. However, large variation in the composition of the

scales was seen at each temperature, with temperature cycling not having as significant corrosion

as the constant temperature tests (McNeill and Edwards, 2000b).

Sontheimer (1981) suggests the formation of siderite (FeCO3) may help to control iron

corrosion, rather than just calcium carbonate formation alone. When siderite is not formed,

higher corrosion rates and non-uniform, thick scales were observed. It was suggested that

siderite competes with the oxidation of Fe2+ and then precipitates ferric oxides that form denser

and better protective layers than the direct oxidation and precipitation, resulting in lower

corrosion rates.

Copper

Copper levels in drinking water are regulated through the Lead and Copper Rule (LCR)

and limited to an action level of 1.3 mg/L at the 90th percentile of household kitchen taps

sampled by voluntary participant homeowners (Federal Register 1991). The source of copper in

drinking water comes primarily from corrosion of copper plumbing and is influenced by water

quality parameters like pH, alkalinity, chloride, nitrate, sulfate, sodium, calcium, and magnesium

(Edwards, Ferguson, and Reiber 1993). Addition of orthophosphates is believed to reduce

10

copper release by forming Cu3(PO4)2 or a similar scale on the surface of the pipe (Reiber 1989;

Schock et al. 1995). However, the benefits of using orthophosphate are thought to be limited to

cases of pH less than 8 (Reiber 1989; Schock, Lytle, and Clement 1995; Edwards, Jacobs, and

Dodrill 1999; Dodrill and Edwards 1995).

A study in the Greater Vancouver Regional District studied the effect of source water and

temperature on metal release. Copper release was found to be mostly from the household

plumbing system, rather than the distribution system or faucets. These concentrations were

found to be higher with more aggressive source waters and not influenced by temperature (Knox

et al. 2005).

Zhang et al. (2002) tested the corrosion of copper exposed to tap water with

monochloramine disinfectant for a period of 30 days. They found at a pH of 8, the copper

corrosion increased for six days and then became steady for the remaining days. Increasing ionic

strength, dissolved inorganic carbon and temperature promoted corrosion, resulting in thicker

oxide films

Pinto, McAnally, and Flora (1997) evaluated the addition of phosphates as well as pH

and alkalinity adjustment for corrosion control of copper in low hardness, low alkalinity waters.

Addition of phosphate corrosion inhibitor was shown to help reduce copper levels. Increasing

alkalinity was found to increase copper release. Phosphate addition was recommended because

it is effective at low doses.

Indian Hills Water Works in Ohio used elevated pH to treat high copper levels and zinc

orthophosphate for treatment of lead, but still had copper problems. To treat both, a study

showed an increase in orthophosphate inhibitor to a dose of 3 mg/L as PO4. This kept both lead

11

and copper within the action level and use of the zinc orthophosphate was discontinued (Schock

and Fox 2001).

A test of copper loops with stagnation and flow conditions similar to household plumbing

systems showed phosphate inhibitors to reduce copper concentrations. Stabilization of the

copper surface layer by building insoluble compounds of copper ions and phosphates was shown

to interfere with the reaction kinetics of the dissolution of the copper surface layer (Becker

2002).

Dodrill and Edwards (1995) conducted a survey of about 360 utilities to examine their

strategies in controlling lead and copper release, in response to the Lead and Copper Rule (LCR)

of 1991. For copper, the survey showed that at high pH, copper release is reduced with and

without inhibitors. At pH less than 7.8, copper release was high at high alkalinity, but inhibitor

use mitigated that release. However, at pH greater than 7.8, inhibitors had variable and adverse

effects on copper corrosion by-product release.

Edwards, McNeill, and Holm (2001) compared the benefits of orthophosphate versus

polyphosphate in controlling copper corrosion by-product release in aged copper pipes, at

variable pH and alkalinity values. Polyphosphate performed less favorably than orthophosphate

at comparable concentrations of 1 mg/L as P. It was believed that orthophosphate reduced

copper solubility by forming a cupric phosphate scale. While in the absence of any phosphate

inhibitors, an insoluble malachite scale formed over a period of years. Polyphosphate, however,

increased copper release in comparison to orthophosphate because it complexed copper,

increasing soluble copper release.

12

When inhibitors are not considered, copper release can be controlled best by raising pH

(Edwards, Hidmi, and Gladwell 2003). It was demonstrated that bicarbonates adversely affected

copper release, and that a pH increase (7.0-8.0) showed significant reduction in copper release.

CO2 stripping was the recommended method for raising the pH without raising alkalinity.

Lead

Lead levels in drinking water are regulated through the Lead and Copper Rule (LCR) and

limited to an action level of 0.015 mg/L at the 90th percentile of household kitchen taps sampled

by voluntary participant homeowners (Federal Register 1991). Lead in household tap water

originates from several sources including lead service lines, lead-tin solder, and brass fixtures

and faucets in bathrooms and kitchens (Singley et al. 1985; Lee, Becker, and Collins 1989;

Reiber 1991). Lead levels are a function of the water quality, plumbing materials, contact time,

pipe geometry, water temperature and age of materials (Boffardi 1995; Boffardi 1990).

Chemical treatment for control of lead corrosion includes pH adjustment, carbonate addition,

CaCO3 deposition and inhibitors (Boffardi 1995).

Numerous studies have found that orthophosphate inhibitor addition can reduce lead

concentrations in water under the correct pH conditions (Schock 1980; Hunt and Creasey 1980;

Sheiham and Jackson 1981; Gardels and Schock 1981; Gregory and Jackson 1983; Nriagu 1974).

Orthophosphate chemicals form passivating films on metallic surface anodic sites to suppress

electrochemical reactions. Zinc, lead, calcium or iron deposition can further enhance the

protection by forming films over cathodic sites (AwwaRF 1996). Orthophosphate can combine

13

with lead plumbing materials to form compounds that do not have a strong tendency to dissolve

into water, keeping lead concentrations low in drinking water (AwwaRF 2001, AwwaRF 1996).

A study in the Greater Vancouver Regional District studied the effect of source water and

temperature on metal release. Lead release was found to be from the plumbing as well as the

faucets of household piping rather than the distribution system. The source water and the

temperature were not found to significantly affect the lead concentrations (Knox et al. 2005).

Edwards and McNeill (2002) found dosing orthophosphate inhibitor to reduce lead

release by 70%, except with new pipes. New pipes had increased lead, suggested to be due to a

lower increase in pH during stagnation than seen without inhibitor. Particulate lead was the

dominant form observed and total lead levels decreased with aging.

Pinto, McAnally, and Flora (1997) showed addition of orthophosphate corrosion inhibitor

helped to reduce lead levels compared with zinc or blended orthophosphate inhibitors. Inhibitor

addition, as well as increasing the alkalinity in low alkalinity waters, reduced lead levels further

than alkalinity adjustment alone. Phosphate addition was recommended because of its

effectiveness at low doses. Alkalinity and pH adjustments are viewed as an alternative, but may

have adverse affects on the system scaling.

A study on the lead release seen in Washington D.C. with the conversion to

monochloromines as the disinfectant was done with corrosion inhibitor addition.

Orthophosphate was added to the distribution system and the scale formed on the lead service

lines was monitored. It was found that prior to orthophosphate treatment, lead scales on service

lines were mostly lead(IV) oxide. Use of orthophosphate took eight months to reduce lead levels

of 150 ppb down below the action level (Giani. Keefer, and Donnelly 2005).

14

Dodrill and Edwards (1995) conducted a survey of about 360 utilities to examine their

strategies in controlling lead and copper release, in response to the LCR. It was observed that

without phosphate inhibitors, higher alkalinity resulted in lower lead release. At low alkalinity,

using inhibitors reduced lead release compared to no inhibitors. The survey revealed that

utilities with the low alkalinity and pH below 7.4 benefited from using orthophosphate but not

polyphosphate. Moreover, polyphosphate increased lead release at higher alkalinity.

Hozalski, Esbri-Amador, and Chen (2005) studied the benefit of inhibitors in controlling

lead release from lead pipe. They used orthophosphate, polyphosphate, blended

ortho-polyphosphate (BOP), and stannous chloride (SnCl2). All inhibitors performed better than

the no-action alternative at reducing lead release with various degrees. The ranking of inhibitors

based on total lead concentrations, from lowest to highest, in the tested pipe loops was as

follows: ortho-P < SnCl2 < BOP < poly-P < control. The authors also observed that phosphate-

based inhibitors resulted in increased bio-growth compared to the stannous chloride and control

pipes.

Modeling

Taylor et al. (2005) developed statistical nonlinear relationships that described the metal

release in terms of source water quality. The iron, copper and lead release models that were

developed to analyze different blends of source waters and identify a feasible range of blends

that satisfy the water quality required for control of copper, lead and color release are shown

here.

15

Iron Release Model

It was determined that effluent values for apparent color correlate well with total iron

concentrations. Due to high coefficient of correlation (R2 = 0.82) between apparent color and

total iron concentrations, effluent values for apparent color can be used as a substitute

measurement for total iron. The correlation is presented in Equation (1-9).

lorApparentCoFe ×= 0132.0 (1-9)

Chlorides, sodium, sulfate, DO, temperature, HRT, and alkalinity were statistically

significant in a color release (iron) model that is shown in Equation (1-10). Increased chloride,

sodium, sulfate, dissolved oxygen, temperature and hydraulic residence time were shown to

contribute to increased apparent color, while alkalinity was shown to reduce apparent color.

912.0321.1

836.0813.0967.0118.024

561.0485.0

)(10)()()()()()(

AlkalinityHRTTDOSONaClC

−+−

=Δ (1-10)

where ΔC = increase in effluent values for apparent color in CPU Na, Cl, SO4

2- = sodium, sulfates and chlorides, respectively in mg/L Alk = alkalinity in mg/L as CaCO3 DO = dissolved oxygen content in mg/L T = temperature in oC HRT = hydraulic retention time in days

Copper Release Model

The model shown in Equation (1-11) was developed for copper release. The nonlinear

copper release model indicates alkalinity, temperature, sulfate, silica, and pH significantly affect

total copper release.

16

22.02

10.024

86.273.072.0 )(28.0 −−− ××××+= SiOSOpHAlkTempCu (1-11)where Cu = total copper release in mg/L

Temp = temperature in oC Alk = alkalinity in mg/L as CaCO3 SO4

2- = sulfate in mg/L SiO2 = silica in mg/L as SiO2

Both experimental data and the empirical model showed increasing alkalinity will

produce more total copper release than the other water quality parameters. Copper concentration

increased with increasing temperature and sulfate. Increasing pH and silica were observed to

decrease copper concentration for the conditions of this investigation.

Lead Release Model

A variety of regression functions combining ten major parameters; alkalinity, pH,

chloride, sulfate, DO, temperature, calcium, SiO2, conductivity, UV-254, and chloramine were

examined and evaluated to develop the lead release model. Temperature, alkalinity, pH,

chlorides and sulfate were found to be statistically significant in lead release model, as shown in

Equation (1-12).

228.02

4462.1726.2677.0)25( )()()()(027.1 −−−−= SOClpHAlkPb T (1-12)

where Pb = total lead release in mg/L Alk = alkalinity in mg/L as CaCO3 T = temperature in oC Cl = chloride in mg/L SO4

2- = sulfate in mg/L

The statistical model indicates that total lead release will increase as pH and/or sulfate

increases and decrease as alkalinity and/or chloride increases.

17

Surface Characterization

Vesecky, et al (1997) analyzed coupons of copper and lead exposed to similar water

qualities with different corrosion inhibitor chemicals for scale formation using X-ray

Photoelectron Spectroscopy (XPS) as a possible corrosion control mechanism. They found a

thick oxide layer is formed on lead coupons with only about 1.3 % phosphorus after treatment

with orthophosphate inhibitor. The copper coupons from this treatment showed a thin

incomplete oxide layer on the metal surface with only about 3.2% phosphorus.

Zhang et al. (2002) exposed copper coupons to tap water with monochloramine

disinfectant for 30 days. Using XPS, copper corrosion was found to form heavy cupric and

cuprous oxides after the first eight days of exposure. These oxide layers were suggested to offer

significant protection from further corrosion.

The nature of lead pipe corrosion in distribution systems was studied in water treated

with orthophosphate and pH adjusted. The water was characterized to have low alkalinity.

Analysis showed lead carbonate, lead oxide, and lead phosphate as the principle components of

lead corrosion (Davidson et al., 2004).

The use of phosphate inhibitors forms a layer on the surface of pipes, which protects

against corrosion (Benjamin et al. 1990). The surface speciation of orthophosphate ions on

goethite has been studied as a function of pH, time, total phosphate concentration and ionic

medium. Phosphate anions have a high affinity for corroded surface metals which leads to the

formation of a stable phosphate metal complex on the pipe surface and limits further corrosion

(Persson, Nilsson, and Sjoberg 1996).

18

Moriarty (1990) examined surface corrosion in cooling water systems when blended

orthophosphate was used as an inhibitor. The study showed two types of protective film formed

on the pipe surface: an inner thin monolayer film formed on the pipe surface from adsorption of

orthophosphate and an outer thicker layer composed of a weakly soluble compound, generally

porous, which was formed by the polyphosphate. The inner layer reduces iron dissolution while

the outer, protective layer serves as a physical barrier to diffusion and electron transfer.

Similar results were found in another analysis using Auger Electron Spectroscopy (AES)

in which corrosion inhibiting films formed on iron in phosphate-containing solutions were depth

profiled. In calcium free solutions, orthophosphate is not incorporated into the film in significant

amounts. When divalent calcium ion is added to the solution, the resulting film contains

significant amounts of phosphate, and the film composition depends on the applied potential

(Kamrath, Mrozek, and Wieckowski 1993).

The calcium phosphate deposition on iron in oxygen-containing neutral aqueous solutions

has been studied with respect to the corrosion inhibition properties of the deposited layer. Results

of this study show that calcium phosphate deposition on iron results from super saturation (pH

increase). Calcium carbonate co-deposition via heterogeneous nucleation accounts for up to

15% of inhibition. The apparent thickness of the calcium phosphate layer increases linearly with

temperature and bulk solution pH (Jovancicevic and Bauer 1989).

19

References

AwwaRF. 2001. Role of phosphate inhibitors in mitigating lead and copper corrosion. AwwaRF, Denver.

AwwaRF. 1996 (2nd ed). Internal Corrosion of Water Distribution Systems.. AwwaRF, Denver.

Becker, A. 2002. The effect of corrosion inhibitors in drinking water installation of copper. Materials and Corrosion, 53:560-567.

Benjamin, M.M., H. Sontheimer, and P. Leroy. 1996. Corrosion of Iron and Steel. In Internal Corrosion of Water Distribution Systems. 2nd ed. Denver, Colo.: American Water Works Association Research Foundation.

Benjamin, M.M., S.H. Reiber, J.F. Ferguson, E.A. Vanderwerff, and M.W. Miller. 1990. Chemistry of corrosion inhibitors in potable water. Report No 2P-3C-9056l-3/93-CM. Denver, Colo.:AWWARF.

Boffardi, B.P. 1995. Lead corrosion. Journal New England Water Works Association, 109(2):121-131.

Boffardi, B.P. 1990. Minimization of lead corrosion in drinking water. Materials Performance, 29(8):45-49.

Boffardi, B.P. 1988. Potable water treatment and monitoring for corrosion and scale control. Journal New England Water Works Associations, 102(2):111.

Burlingame, G.A., D.A. Lytle, and V.L. Snoeyink. 2006. Why red water? Understanding iron release in distribution systems. Opflow, 32(12):12-16.

Dietrich, A.M., D. Glindemann, F. Pizarro, V. Gidi, M. Olivares, M. Araya, A. Camper, S. Duncan, S. Dwyer, A.J. Whelton, T. Younos, S. Subramanian, G.A. Burlingame, D. Khiari, and M. Edwards. 2004. Health and aesthetic impacts of copper corrosion on drinking water. Water Science and Technology, 49(2):55-62.

Davidson, C.M, N.J. Peters, A. Britton, L. Brady, P.H.E. Gardiner, and B.D. Lewis. 2004. Surface analysis and depth profiling of corrosion products formed in lead pipes used to supply low alkalinity drinking water. Water Science and Technology, 49(2):49-54.

Dodrill, D. M. and M. Edwards. 1995. Corrosion control on the basis of utility experience. Journal American Water Works Association, 87(7):87-99.

20

Edwards, M., L. Hidmi, and D. Gladwell. 2003. Phosphate inhibition of soluble copper corrosion by-product release. Corrosion Science, 44:1057-1071.

Edwards, M., and L.S. McNeill. 2002. Effect of phosphate inhibitor on lead release from pipes. Journal American Water Works Association, 94 (1):79-90.

Edwards, M., L.S. McNeill, and T.R. Holm. 2001. Role of phosphate inhibitors in mitigating lead and copper corrosion. In Distribution Systems. Denver, Colo.: AWWARF.

Edwards, M., S. Jacobs, and D. Dodrill. 1999. Desktop guidance for mitigating Pb and Cu Corrosion By-Products. Journal American Water Works Association, 91(5):66.

Edwards, M., F.F. Ferguson, and S.H. Reiber. 1993. On the pitting corrosion of copper. Presented at the Annual Conference of AWWA, San Antonio, Texas.

Federal Register. 1991. Drinking Water Regulations; Lead and Copper Rule. 56 FR 26460

Gardels M.C. and M.R. Schock. 1981. Corrosion indices: invalid or invaluable? In Proceedings of the AWWA Water Quality Technology Conference. Dever, Colo.: American Water Works Association.

Gedge, G. 1992. Corrosion of cast iron in potable water service. Corrosion and related aspects of materials for potable water supplied. Proceedings Institute of Materials Conference, London.

Giani, R., W. Keefer, and M. Donnelly. 2005. Studying the effectiveness and stability of orthophosphate on Washington D.C.’s lead service lines. In Proc. of AWWA WQTC Conference. Quebec City, Quebec: AWWA.

Gregory, R. and P.J. Jackson. 1983. Reducing lead in drinking water. Report 219-S, Water Research Centre, Stevenage Laboratory, Elder Way, Stevenage, Herts.

Hatch, G.B. and O. Rice. 1939. Surface-active properties of hexametaphosphate. Industrial and Engineering Chemistry, 31(1):51.

Hoover, C.P. and O. Rice. 1939. Threshold treatment. Water Works and Sewage, 86:10.

Hozalski, R.M., E. Esbri-Amador, and C.F. Chen. 2005. Comparison of stannous chloride and phosphate for lead corrosion control. Journal American Water Works Association, 97(3):89-102.

Huange, D.J.-S. 1980. Polyphosphate for corrosion control in water distribution system. PhD dissertation, University of Missouri, Columbia.

21

Hunt, D.T.E and J.D. Creasey. 1980. The calculation of equilibrium trace metal speciation and solubility in aqueous systems by a computer method with particular reference to lead. Technical Report 151. Medmenham, England: Water Research Centre.

Imran, S.A., J. D. Dietz, M. Ginasiyo, and W. Xiao. 2006. Optimizing source water blends for corrosion and residual control in distribution systems. Journal American Water Works Association, 98(5):107-107.

Jovancicevic, V. and B. Bauer. 1989. Calcium phosphate deposition on iron in oxygen-containing neutral aqueous solutions: an electrochemical approach. Langmuir, 5(1):261-267.

Kamrath, M., P. Mrozek and A. Wieckowski. 1993. Composition depth profiles of potential-dependent orthophosphate film formation on iron using auger electron spectroscopy. Langmuir, 9(4):1016-1023.

Knox, G., D. Mavinic, J. Atwater, and D. MacQuarrie. 2005. Assessing the impact of corrosion control measures on tap drinking water of the Greater Vancouver Regional District. Canadian Journal of Civil Engineering, 32:948-956.

Lee, R.G., W.C. Becker, and D.W. Collins. 1989. Lead at the tap: sources and control. Journal American Water Works Association, 81(7):52-62.

McNeill, L.S., and M. Edwards. 2001. Iron pipe corrosion in distribution systems. Journal American Water Works Association, 93(7):88-100.

McNeill, L.S., and M. Edwards. 2000a. Phosphate inhibitors and red water in stagnant iron pipes. Journal of Environmental Engineering, 126(12):1096-1102.

McNeill, L.S., and M. Edwards. 2000b. Temperature effects on iron corrosion. Ph.D. diss., Virginia Polytechnic Institute, Blacksburg, Va.

Moriarty, B.E. 1990. Surface studies of corrosion inhibitors in cooling water systems. Materials Performance, 29(1):45-48.

Montgomery Watson Harza. 2005. 2nd ed. Water Treatment: Principles and Design. Hoboken: John Wiley & Sons.

Nriagu, J.O. 1974. Lead orthophosphates-IV. Formation and stability in the environment. Gerchim. Cosmochim. Acta, 38:887

Persson, P., N. Nilsson and S. Sjoberg. 1996. Structure and bonding of orthophosphate ions at the iron oxide-aqueous interface. Journal of Colloid and Interface Science, 177(1):263-275.

22

Pinto, J.A., A.S. McAnally, and J.R.V. Flora. 1997. Evaluation of lead and copper corrosion control techniques. Journal of Environmental Science and Health, A32(1):31-53.

Plottu-Pecheux, A., C. Democrate, B. Houssais, D. Gatel, and J. Cavard. 2001. Controlling the corrosiveness of blended waters. Desalination, 138:237-249.

Pryor, M.J. and M. Cohen. 1951. The mechanism of the inhibition of corrosion of iron by solutions of sodium orthophosphate. Journal Electrochemical Society, 98(7):263.

Reiber, S.H. 1991. Galvanic stimulation of corrosion on lead-tin solder-sweated joints. Journal American Water Works Association, 83(7):83-91.

Reiber, S. 1989. Copper plumbing surfaces: an electrochemical study. Journal American Water Works Association, 81(7):114-122.

Sarin, P., V.L. Snoeyink, D.A. Lytle, and W.M. Kriven. 2004. Iron Corrosion Scales: Model for Scale Growth, Iron Release, and Colored Water Formation. Journal of Environmental Engineering ASCE, 130(4):364-373.

Schock, M.R., and J.C. Fox. 2001. Solving copper corrosion problems while maintaining lead control in high alkalinity water using orthophosphate. Presented at the AWWA Annual Conference, Cleveland, Ohio, August 30, 2001.

Schock, M.R., and J.A. Clement. 1998. Lead and copper control with non-zinc orthophosphate. Journal New England Water Works Association, 112(1):20-42.

Schock, M.R., D.A. Lytle, and J.A. Clement. 1995. Effect of pH, DIC, Orthophosphate and Sulfate on Drinking Water Cuprosolvency. Risk Management Research Report EPA/600/R-95/085. Cincinnati, Ohio: U.S. EPA.

Schock, M.R. 1980. Response of lead solubility to dissolved carbonate in drinking water. Journal American Water Works Association, 72(12):695.

Seal, S. and T.L. Barr. 2001. Application of photoelectron spectroscopy in inorganic and organic material system. In Advances in Surface Science. Edited by H.S. Nalwa. San Deigo, Calif.: Academic Press.

Sheiham, I. and P.J. Jackson. 1981. The scientific basis for control of lead in drinking water by water treatment. Journal of the Institution of Water Engineers and Scientists, 35(6):491.

Singley, J.E. B.A. Beaudet, D.W. DeBerry, J.R. Kidwell, D.A. Malish, and P.H. Marker. 1985. Corrosion Prevention and Control in Water Treatment and Supply Systems, Pollution Technology Review No. 122, Noyes Publications, Park Ridge, NJ.

Snoeyink, V.L. and D. Jenkins. 1980. Water Chemistry. USA: John Wiley & Sons.

23

Sontheimer, H., W. Kolle, and V.L. Snoeyink. 1981. The siderite model of the formation of corrosion-resistant scales. Research and Technology, Journal AWWA, 73(11):572-579.

Taylor, J.S., J.D. Dietz, A.A. Randall, S.K. Hong, C.D. Norris, L.A. Mulford, J.M. Arevalo, S. Imran, M. Le Puil, S. Liu, I. Mutoti, J. Tang, W. Xiao, C. Cullen, R. Heaviside, A. Mehta, M. Patel, F. Vasquez, and D. Webb. 2005. Effects of blending on distribution system water quality. Denver, Colo.: AwwaRF and Tampa Bay Water.

Vesecky, S., J. Liu, R.M. Friedman, F. Pacholec, and J.B. Lechner. 1997. Comparison of film formation using phosphate inhibitor in systems with comparable water qualities. Journal New England Water Works Association, 111(3):258-284

Volk, C., E. Dundore, J. Schiermann, and M. Lechevallier. 2000. Practical evaluation of iron corrosion control in a drinking water distribution system. Water Research, 34(6):1967-1974.

Wagner, I. 1992. Influence of operating conditions on materials and water quality in drinking water distribution systems. Corrosion and related aspects of materials for potable water supplies. In Proceedings Institute of Materials Conference, London.

Wagner, I and A. Kuch. 1984. The influence of water parameters of corrosion rate, scale deposition and iron(III) uptake in unprotected iron pipes. Water Supply, 2(3/4):SS11.

Zhang, X., S.O. Pehkonen, N. Kocherginski, and G.A. Ellis. 2002. Copper corrosion in mildly alkaline water with the disinfectant monochloramine. Corrosion Science, 44:2507-2528.

24

CHAPTER 2 EFFECTS OF ORTHOPHOSPHATE CORROSION INHIBITOR ON IRON

IN BLENDED WATER QUALITY ENVIRONMENTS

Abstract

This study evaluated the effects of orthophosphate (OP) inhibitor addition on iron

corrosion on coupons exposed to different blends of groundwater, surface water, and desalinated

seawater. The effectiveness of OP inhibitor addition on iron release was analyzed by statistical

comparison between OP treated and untreated pilot distribution systems. Four different doses of

OP inhibitor, ranging from zero (control) to 2 mg/L as P, were investigated and non-linear

empirical models were developed to predict iron release from the water quality and OP doses.

Surface characterization evaluations were conducted using X-ray Photoelectron Spectroscopy

(XPS) analyses for each iron and galvanized steel coupon tested. A theoretical thermodynamic

model was developed and used to validate the controlling solid phases determined by XPS. OP

inhibitor addition was not found to offer much improvement on iron release for the OP dosages

evaluated for the water blends evaluated compared to pH adjustment alone. Empirical models

showed increased total phosphorus, pH and alkalinity reduced iron release while increased silica,

chloride, sulfate and temperature contributed to iron release. Thermodynamic modeling

suggested that FePO4 is the controlling solid that forms on iron and galvanized steel surfaces,

regardless of blend, when OP inhibitor is added for corrosion control. While FePO4 offers

limited control of the iron release from the cast iron surfaces, it offers significant protection of

the galvanized steel surfaces reducing zinc release.

25

Introduction

With increasing water demands and more stringent drinking water regulations, many

utilities are turning to desalinated sources to supplement their surface and groundwater supplies.

Tampa Bay Water (TBW) and the University of Central Florida (UCF) studied the effects of

blending multiple alternative source waters on distribution system water quality (Taylor et al.

2005). This study further evaluates the addition of orthophosphate (OP) corrosion inhibitor to

the blended source waters and the effects on iron corrosion.

Cast iron has been used in drinking water distribution systems for 500 years (Gedge

1992). Under the right conditions, iron is released into the water causing water quality problems

and red water complaints (Burlingame, Lytle, and Snoeyink 2006). Phosphate inhibitors were

first added to drinking water to prevent excessive calcite precipitation (Hoover and Rice 1939).

The phosphate was thought to prevent calcite crystals from forming outside of the colloidal range

by sorption to the calcium carbonate nuclei (Hatch and Rice 1939). More recently, research

suggests somewhat mixed results for phosphate addition used to prevent iron corrosion and red

water (McNeill and Edwards 2001). Research suggests orthophosphate addition forms a scale

while polyphosphates sequester iron (Wagner 1992; Benjamin et al. 1990; Boffardi 1988;

Wagner and Kuch 1984; Huang 1980; Pryor and Cohen 1951).

Corrosion of iron pipes in a distribution system can cause three distinct problems. First,

pipe mass is lost through oxidation to soluble iron species or iron-bearing scale. Second, the

scale can accumulate as large tubercles that increase head loss and decrease water capacity.

Finally, the release of soluble or particulate iron corrosion by-products to the water leads to

consumer complaints of "red water" at the tap (McNeill and Edwards 2001).

26

McNeill and Edwards (2000a) investigated phosphate inhibitors in new iron pipes in up

to 72 hours of stagnant conditions. Five different water qualities (variable pH and alkalinity)

were used, and dosed with orthophosphate, polyphosphate, or none. Examination of iron release

measurements showed that both ortho and polyphosphates either increased or had no effect on

iron concentrations. The only tested conditions that showed reduced iron release were at high

alkalinity (300 mg/L as CaCO3) and pH of 7.2. This observation of phosphate inhibitors

influence in stagnant pipes is opposite to common experiences in flowing water conditions. On

average, other pH and alkalinity conditions showed an increase in iron release during stagnation

when polyphosphate is used, compared to pipes receiving no inhibitor.

Another study tested iron corrosion in reactors, as well as the biological growth, with and

without phosphate inhibitor. Increasing phosphate doses were shown to maintain a lower

corrosion rate. These increased doses are recommended for summer months because the

corrosion rate was found to be highly related to seasonal variables, namely temperature. Lower

phosphate doses were recommended for the winter, and the addition of phosphate did not

increase biological growth (Volk et al. 2000).

Another study involving temperature effects on iron corrosion was done with samples

held at a constant temperatures and samples that were exposed to a cycle of different

temperatures. They found samples of iron held at 5 °C had higher iron concentration and

tuberculation than samples at 20 °C or 25 °C. However, large variation in the composition of the

scales was seen at each temperature, with temperature cycling not having as significant corrosion

as the constant temperature tests (McNeill and Edwards, 2000b).

27

Sontheimer (1981) suggests the formation of siderite (FeCO3) may help to control iron

corrosion, rather than just calcium carbonate formation alone. When siderite was not formed,

higher corrosion rates and non-uniform, thick scales were observed. It was suggested that

siderite competes with the oxidation of Fe2+ and then precipitates ferric oxides that form denser

and better protective layers than direct oxidation and precipitation, resulting in lower corrosion

rates.

It has been suggested that the presence of phosphate may interfere with Fe2+ oxidation

(Benjamin, Sontheimer, and Leroy 1996). It has also been suggested that phosphate inhibitors

can act as plugs, filling the pores of the corrosion scales (Sarin et al. 2004). This helps to prevent

diffusion of oxidants into the lower layers of the corrosion scale that further corrosion the iron

pipe.

This study evaluated the effects of orthophosphate inhibitor addition to blended treated

surface, ground, and seawater sources of varying blend percentages. The effects of water quality

were evaluated and a model predicting total iron release using water quality and total phosphorus

concentrations was developed. XPS analysis of cast iron and galvanized steel coupons was

evaluated for solid phase surfaces present on the coupon, and thermodynamic modeling was

performed to gain insight into inhibitor control of iron release.

Experimental Methods

Experimental Design

Experimentation was conducted with the use of pilot distribution systems (PDSs) built

from actual pipelines extracted from TBW member governments distribution systems

28

(Hillsborough County, Fla.; Pasco County, Fla.; Pinellas County, Fla.; City of New Port Richey,

Fla.; City of St. Petersburg, Fla.; and City of Tampa, Fla.). Details regarding the PDS and prior

study results are reported elsewhere (Taylor et al. 2005). Each PDS runs in parallel with

segments of PVC, lined cast iron, unlined cast iron, and galvanized steel pipes that were placed

sequentially to simulate actual distribution systems. The materials and their respective diameter

and length are shown in Table 2-1. Each PDS was fed blends of groundwater, surface water, and

desalinated seawater along with different types and doses of corrosion inhibitors.

Table 2-1 Pipe materials in PDSs

Pipe Material Length (ft)

Diameter (in)

PVC 20 6 Lined Cast Iron 20 6

Unlined Cast Iron 12 6 Galvanized Steel 40 2

Images of materials as well as the full PDSs used to represent full scale distribution

systems are shown in Figure 2-1. The pipe materials used, shown in the left image, are

galvanized steel, PVC, lined cast iron, and unlined cast iron from left to right. The image on the

right shows the PDSs with the materials connected in series from influent (closest to viewer) to

effluent (furthest from viewer) of the system. The pipe materials were connected in sequence of

PVC, lined cast iron, unlined cast iron, then galvanized steel.

29

Figure 2-1 Pipe materials (left) and parallel PDSs (right)

The PDSs were fed blends of conventionally treated groundwater (GW), enhanced

coagulation-sedimentation-filtration (CSF) treated surface water (SW), and desalinated seawater

by reverse osmosis (RO). A description of the three finished source waters is presented in Table

2-2.

Table 2-2 Finished source water descriptions

Water Source System Description

GW Groundwater Ground water source. Treatment by aeration, disinfection by free chlorine with a residual of 5 mg/L after a 5 minute contact time. 5.0 mg/L chloramine residual.

SW Surface water

TBW treatment plant: Treatment by ferric sulfate coagulation, flocculation, settling, filtration, disinfection by ozonation and chloramination. Project site: adjustment of chloramine residual to 5.0 mg/L chloramine residual.

RO Groundwater Treatment by membrane reverse osmosis, aeration, disinfection by free chlorine with a residual of 5 mg/L after a 5 minute contact time. 5.0 mg/L chloramine residual.

30

The GW unit used raw well water from the Cypress Creek well field owned by TBW.

The GW was treated with aeration, disinfection, and pH stabilization. Aeration was achieved in

the GW by pumping the raw water to the top of the finished water tank through a spray nozzle.

Sodium hypochlorite was used for primary disinfection and was dosed to provide a 5 mg/L

residual after a 5 minute contact time. Afterwards, ammonium chloride was added to produce a

5 mg/L monochloramine residual. Ammonia was added in the form of NH4Cl at a 5:1 ratio. The

Cl2: NH3 ratio was initially 4:1 to protect against DBP formation. This ratio was increased to 5:1

in after 6 months of operation to reduce free ammonia.

SW was treated at the TBW Regional Surface Water Treatment Facility by enhanced

coagulation, ozonation, biologically activated carbon (BAC) filtration, aeration, and

chloramination. The SW was hauled weekly to the field facility for use and temporarily stored in

two 7000 gallon storage tanks before being transferred to the SW finished water tank. In the SW

finished tank, the chloramine residual was adjusted to 5 mg/L as Cl2.

The RO pilot plant was housed in a trailer at the testing facility and utilized raw

groundwater for the feed stream. The RO treatment pilot system required the addition of TDS,

calcium, and alkalinity to the RO permeate to represent the finished water produced by the TBW

Regional Desalination Facility. RO pretreatment consisted of 2.7 mg/L antiscalant addition

(Hypersperse MDC700TM, GE Water, Minnetonka, Minn.) followed by 5-micron cartridge

filtration. The RO membrane unit was operated at 72-73% recovery, producing 9.3 gpm

permeate flow, which was aerated by a 10-inch diameter aeration tower filled with tripack plastic

packing. After aeration, 50 mg/L of sea salt was added to the aerated permeate stream to

simulate the TBW desalination process. Calcium chloride and sodium bicarbonate were also

31

added to meet the calcium and alkalinity specifications. The finished water was stabilized with

sodium hydroxide to 0.1-0.3 pH units above pHs.

The effects of water quality were evaluated by varying the blend quarterly, while

seasonal effects were evaluated by maintaining the same blend in the summer and winter. The

quarterly phases and percentages of the blends are shown below in Table 2-3. The average water

quality of each of the source waters in each phase is shown in Table 2-4. The effects of season

are seen in the temperature as well as rainy and dry season effects on the surface water between

Phases I and III. The blends with a high percentage of groundwater in Phases I and III are

characterized by high alkalinity and pH. Phase II had the highest percentage of surface water

and is characterized by high sulfate concentrations. Phase IV has average water quality

parameters due to the equal percentage of GW and SW.

Table 2-3 Blend percentages for each phase

Phase Quarter % GW % SW % RO I Feb-May 2006 62 27 11 II May-Aug 2006 27 62 11 III Aug-Nov 2006 62 27 11 IV Nov 2006-Feb 2007 40 40 20

Table 2-4 Average water quality by phase

Phase pH Alkalinity (mg/L CaCO3)

Chloride (mg/L)

Sulfate (mg/L)

Temperature (°C)

I 8.0 161 45 62 21 II 7.9 104 67 103 26 III 8.0 150 68 66 26 IV 7.9 123 59 76 21

32

The feed rate of the blend into each PDS was maintained to achieve a two-day hydraulic

residence time (HRT). Pumps maintained the blend flow as well as the inhibitor addition into

each PDS. The PDSs each were fed different inhibitor types and doses. The inhibitor was dosed

to the PDSs at three different levels, categorized as low dose, medium dose, and high dose.

Orthophosphate (OP) was maintained at a target dose of 0.5 mg/L as P for the low dose, 1.0

mg/L as P for the medium dose, and 2.0 mg/L as P for the high dose. Control PDSs were not fed

any chemical inhibitor; one was maintained at pHs and a second was treated with elevated pH,

maintained at pHs+0.3. The PDS at pHs+0.3 was maintained at a positive LSI to assess the affect of

elevated pH treatment as a means of iron release control.

Orthophosphate Inhibitor

The orthophosphate inhibitor used in this study is Inhibit-All WSF-36 (SPER Chemical

Corporation, Clearwater, Fla.). It is made of monosodium ortho-phosphate blended into 17

megohm purified water at a concentration of 36%. It is a clear, slightly viscous liquid with a

bulk density of 11.25 lbs/gal. The specific gravity is 1.35 the pH of 1% solution is 5.1 to 5.4.

Table 2-5 shows a summary of the properties of the OP inhibitor along with manufacture

recommendations for use.

33

Table 2-5 Orthophosphate inhibitor properties

Property Value Manufacturer SPER Chemical Corp. Product Inhibit-All WSF-36 Percent Orthophosphate 36% Bulk Density 11.25 lbs/gal Specific Gravity 1.35 pH 1% solution 5.1 – 5.4 Recommended dose (low-med-high) 1.25-2.75-4.25 mg/L as Product Used dose (low-med-high) 0.5-1.0-2.0 mg/L as P Recommended pH range 7.4 – 7.6 (optimum), 6.8 – 7.8

Data Collection

Water quality parameters were collected and analyzed weekly in the influent and effluent

of the PDSs. The water quality parameters monitored and methods used for samples transported

to the UCF laboratory for analysis are shown in Table 5-6. Table 5-7 shows the water quality

parameters monitored and methods used for the parameters analyzed in the field laboratory at the

testing facility. The method detection limit (MDL) is also shown. Methods noted as SM are

from APHA, AWWA, and WEF (1998).

34

Table 2-6 Water quality parameters and methods performed at University Laboratory

Parameter Method Reference Method Description MDL

Aluminum SM 3120B ICP Method 0.001 mg/L Bicarbonate SM 2320B Titration Method 5 mg/L Calcium SM 3120B ICP Method 0.1 mg/L

Chloride SM 4110 Ion Chromatography with Chemical Suppression of Eluent Conductivity 0.1 mg/L

Color SM 2120A Or Hach 8025 Cobalt-Platinate Method (with spec) 1 CPU

Conductivity SM 2510B Laboratory Method 1 μmho/cm Copper SM 3120B ICP Method 0.001 mg/L Iron SM 3120B ICP Method 0.001 mg/L Lead SM 3120B ICP Method 0.001 mg/L Magnesium SM 3120B ICP Method 0.1 mg/L Nitrogen (NH3,TKN) SM 4500-Norg Macro-Kjeldahl Method 0.1 mg/L

NPDOC SM 5310C Persulfate-UV Oxidation Method 0.1 mg C/L pH SM 4500-H+ B Electrometric Method ± 0.01 pH units Phosphorus SM 3120B ICP Method 0.001 mg/L Silica SM 3120B ICP Method 0.001 mg/L Sodium SM 3120B ICP Method 0.1 mg/L Solids (TDS) SM 1030E Estimation of TDS by major ion sum 1 mg/L

Sulfate SM 4110 Ion Chromatography with Chemical Suppression of Eluent Conductivity 0.1 mg/L

Turbidity SM 2130B Nephelometric Method 0.01 NTU UV-254 SM 5910 UV Absorption at 254 nm 0.0001 cm-1 Zinc SM 3120B ICP Method 0.001 mg/L

35

Table 2-7 Water quality parameters and methods performed at Field Laboratory

Parameter Method Reference Method Description MDL Alkalinity SM 2320 B Titration 5 ppm Ammonia-N SM 4500-NH3 C Membrane Probe Method 0.1 ppm Chloride SM 4500-Cl- B Argentometric Titration 1 mg/L Chlorine, free SM 4500-Cl G or Hach 8021 DPD colorimetric 0.1 ppm Chlorine, total SM 4500-Cl-G or Hach 8167 DPD colorimetric 0.1 ppm Color, apparent SM 2120 B Visual Comparison

(by spectrometer) 1 CPU

Conductivity SM 2510 B Conductivity Bridge 1 μmho/cm Hardness (total, calcium) SM 2340 C EDTA Titration 5 mg/L

Nitrate Hach 8192 Cadmium reduction 0.1 mg/L Nitrite Hach 8507 Diazotization 0.1 mg/L Oxygen, Dissolved (DO) SM 4500-O G Membrane probe 0.1 mg/L

pH SM 4500-H+ B Electrometric ± 0.01 pH units

Phosphate-P (Reactive) SM 4500-P E. or Hach 8048 Ascorbic Acid Method 0.1 mg/L

Silica, SiO2 (reactive) SM 4500-SiO2 or Hach 8185 Molybdosilicate Method 0.1 mg/L

as SiO2 Temperature SM 2550 B Direct reading 0 deg C Turbidity SM 2130 B Nephelometric 0.01 NTU UV254 SM 5910 A UV spectrometry 0.0001 cm-1

Iron and galvanized steel coupons were placed in cradles that received flow in parallel

with each PDS. The coupons were evaluated for surface characteristics after incubation during

each phase. X-ray Photoelectron Spectroscopy (XPS) analysis was then performed on each type

of coupon to identify chemical components on the outer layer of the corrosion surface. A survey

scan reveals the presence of elements, whereas a high resolution scan of those elements found on

the outer layer shows the chemical states, providing detailed surface characterization

information.

36

Results and Discussion

Dose Maintenance

Three of the PDSs were treated with orthophosphate corrosion inhibitor at a low,

medium, and high dose. These doses were targeted to maintain at 0.5 mg/L as P, 1.0 mg/L as P,

and 2.0 mg/L as P, respectively. The average orthophosphate inhibitor dose for the course of the

study in each PDS is shown in Figure 2-2. Error bars represent the minimum and maximum

observations. The low dose of orthophosphate averaged 0.51 mg/L as P, the medium dose

averaged 0.94 mg/L as P, and the high dose averaged 1.83 mg/L as P.

1.83

0.94

0.51

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Low Med High

OP Inhibitor Dose

Tota

l P (m

g/L

as P

)

Figure 2-2 Orthophosphate inhibitor dosing

37

The orthophosphate inhibitor is added as phosphoric acid, so it has an effect of lowering

the pH in those PDSs treated with inhibitor. The pH maintained in each of these PDS is shown

in Figure 2-3 with the error bars representing the minimum and maximum observations. The

difference in pH maintained in each of the PDSs is significant with the exception of the low

orthophosphate dosed PDS pH being the same as the elevated pH PDS at pHs+0.3.

7.94

7.71

7.867.907.92

6.8

7.0

7.2

7.4

7.6

7.8

8.0

8.2

8.4

8.6

OP-low OP-med OP-high pHs pHs+0.3

pH

Figure 2-3 pH for OP and control PDSs

Empirical Modeling

A non-linear empirical model was developed from the water quality and phosphorus dose

for the three PDSs with orthophosphate inhibitor and the PDSs maintained at pHs and pHs+0.3.

38

Water quality parameters found to be significant using ANOVA procedures at 95% confidence

were retained as variables in the model. All variables retained p-values less than 0.05. The

range of the water quality parameters included in the model are shown in Table 2-8.

Table 2-8 Range of water quality in model development

Parameter Minimum Maximum Total Phosphorus, mg/L as P 0.01 2.69 Silica, mg/L as SiO2 3.8 13.6 pH 7.4 8.4 Alkalinity, mg/L as CaCO3 84 170 Chloride 38 123 Sulfate, mg/L 53.5 116.4 Temperature, °C 10.4 29.7

Equation (2-1) presents a non-linear least squares regression model developed using

results of the study. The water quality parameters silica, pH, alkalinity, chloride, sulfate and

temperature remained significant as well as the total phosphorus, which represents the inhibitor

addition. Total phosphorus, pH and alkalinity helped mitigate iron release, as shown by the

negative exponent on the variable in the model. However, silica, chloride, sulfate and

temperature are shown to increase iron release for the conditions experienced during this testing.

The low values of the OP dose used and the near-zero value of the exponent on the total

phosphorus term yields predictions of total iron concentration that are not very sensitive to the

inhibitor dose.

39

25611.04

383.0649.0987.1603.02

031.0 012.1 −−−− ××××××= TSOClAlkpHSiOTPFeTotal (2-1)where: Total Fe = total iron, mg/L

TP = total phosphorus, mg/L as P SiO2 = silica, mg/L as SiO2 pH = -log[H+] Alk = alkalinity, mg/L as CaCO3 Cl = chloride, mg/L SO4 = sulfate, mg/L T = temperature, °C

The model fits the data with an R2 of 0.29. The predicted iron release is plotted against

the actual iron release in Figure 2-4 by PDS where the diagonal line represents an ideal

agreement between the observed and predicted values. The high values of iron release observed

in the pHs PDS tend to be under-predicted by the model. These observations also tend to be over

the iron secondary standard of 0.3 mg/L. Most of the observations exceeding the standard

suggest there is merit in elevating pH or adding corrosion inhibitor to reduce iron release. This is

consistent with the model presented previously that suggests increases in pH and total

phosphorus will help to reduce iron release.

40

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40Actual Total Fe (mg/L)

Pred

icte

d To

tal F

e (m

g/L)

OP-low OP-med OP-high pHs pHs+0.3

Figure 2-4 Predicted versus actual total iron concentrations using the empirical model

Figure 2-5 shows the average actual and predicted iron release for each of the OP treated

PDSs where the error bars represent the minimum and maximum. It is shown that no prediction

exceeds the 0.3 mg/L standard, however, there are some instances of actual observations that do

exceed this level. This again shows the model to under-predict observations in excess of the

secondary standard. Also shown in Figure 2-5, the iron release is not very sensitive to the

phosphorus dose in the model or in actual observations. The data and the empirical model are in

agreement that increasing the dose of OP inhibitor does not offer a trend of improved iron

control.

41

0.00

0.10

0.20

0.30

0.40

0.50

0.60

I-OPlow

I-OPmed

I-OPhigh

II-OPlow

II-OPmed

II-OPhigh

III-OPlow

III-OPmed

III-OPhigh

IV-OPlow

IV-OPmed

IV-OPhigh

Phase-Treatment

Ave

rage

Tot

al F

e (m

g/L)

Actual Predicted

Figure 2-5 Actual vs. predicted for empirical model by phase and PDS

Orthophosphate Inhibitor Performance

A summary of the iron release for each of the PDSs is presented in Table 2-9 by phase.

There is not an obvious improvement in mitigation of iron release with the increase in OP dose.

Phase I had the lowest iron release with no observations above the secondary standard of 0.3

mg/L. In Phase II, both the pH control PDSs and the medium dose OP PDS had at least one

observation exceed the secondary standard. This was similar to observations in Phase III except

the elevated pH treatment showed no observation exceeding the standard. In Phase IV, the pHs

PDS and the high OP dose PDS both had observations above 0.3 mg/L.

42

Also shown in Table 2-9, the dissolved iron concentrations are very low in comparison to

the total iron concentrations. This suggests the majority of the iron release observed is in the

particulate form. This is likely due to oxidized scale build-up breaking loose from the pipe wall.

This also explains why the release of iron is difficult to predict using water quality parameters

because the high percentage of particulate iron can be released somewhat randomly.

Table 2-9 Iron release summary

Dissolved Fe (mg/L) Total Fe (mg/L) Phase Treatment Avg Min Max Avg Min Max

I OP-low 0.030 0.001 0.111 0.123 0.062 0.203 OP-med 0.028 0.002 0.158 0.138 0.103 0.191 OP-high 0.031 0.001 0.130 0.114 0.060 0.288 pHs 0.030 0.001 0.106 0.151 0.098 0.201 pHs+0.3 0.036 0.003 0.181 0.113 0.070 0.206 II OP-low 0.004 0.001 0.015 0.159 0.083 0.217 OP-med 0.009 0.001 0.055 0.210 0.104 0.334 OP-high 0.006 0.001 0.028 0.174 0.089 0.238 pHs 0.025 0.001 0.116 0.277 0.171 0.371 pHs+0.3 0.030 0.001 0.134 0.188 0.116 0.332

III OP-low 0.008 0.001 0.026 0.144 0.099 0.214 OP-med 0.009 0.001 0.026 0.230 0.155 0.309 OP-high 0.007 0.001 0.027 0.144 0.080 0.179 pHs 0.011 0.001 0.041 0.268 0.157 0.371 pHs+0.3 0.009 0.001 0.037 0.140 0.094 0.183

IV OP-low 0.002 0.001 0.003 0.135 0.009 0.201 OP-med 0.002 0.001 0.009 0.134 0.006 0.197 OP-high 0.002 0.001 0.006 0.202 0.009 0.513 pHs 0.001 0.000 0.003 0.192 0.051 0.315 pHs+0.3 0.002 0.001 0.003 0.141 0.062 0.237

Figure 2-6 shows the total iron release data for each PDS for each phase with the

secondary standard of 0.3 mg/L indicated. The error bars represent the minimum and maximum

43

observations. Table 2-10 shows the water quality parameters found to be significant in the

empirical model presented previously for each phase and PDS.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Phase I Phase II Phase III Phase IV

Tota

l Fe

(mg/

L)

OP-low OP-med OP-high pHs pHs+0.3

Secondary Standard = 0.3 mg/L

Figure 2-6 Comparison of total Fe by phase and treatment

44

Table 2-10 Average influent water quality by PDS and phase

Total Phosphorus Silica pH Alkalinity Chloride Sulfate TemperatureTreatment Phase (mg/L as P) (mg/L as SiO2) (std. units) (mg/L as CaCO3) (mg/L) (mg/L) (deg. C)

OP-low I 0.49 10.8 8.0 164 43.2 61.8 21.0 II 0.55 5.0 7.8 106 67.2 104.2 25.6 III 0.54 10.4 8.0 152 63.6 65.1 25.6 IV 0.47 6.3 7.8 125 56.6 76.3 21.2 Average 0.51 8.2 7.9 143 54.7 73.8 22.9

OP-med I 1.04 10.8 8.0 164 43.1 62.0 21.0 II 0.90 4.9 7.8 107 67.4 104.3 26.0 III 0.99 10.4 8.0 152 62.8 65.1 25.6 IV 0.82 6.3 7.8 125 56.4 76.0 21.0 Average 0.94 8.2 7.9 143 54.5 73.8 23.0

OP-high I 1.84 10.7 8.0 163 43.2 61.5 20.9 II 1.91 5.0 7.8 107 67.8 104.8 26.0 III 1.87 10.4 7.9 151 65.0 65.7 25.8 IV 1.69 6.2 7.8 125 57.9 77.7 21.2 Average 1.83 8.2 7.9 142 55.4 74.1 23.0

pHs I 0.11 11.0 7.7 146 64.2 61.4 21.3 II 0.03 4.9 7.8 92 82.8 100.3 26.4 III 0.04 10.2 7.7 149 88.4 65.9 25.3 IV 0.05 6.3 7.6 119 68.1 75.8 21.0 Average 0.06 8.2 7.7 130 75.2 73.5 23.2

pHs+0.3 I 0.16 10.8 8.0 164 43.2 62.2 21.2 II 0.04 5.2 7.9 106 65.1 102.2 26.5 III 0.04 10.5 8.0 151 64.8 66.5 25.6 IV 0.06 6.3 7.9 125 57.9 76.0 21.3 Average 0.08 8.3 7.9 139 57.2 75.5 23.5

45

The low dose OP PDS showed fairly consistent iron release across all phases. No

observation ever exceeded the secondary standard of 0.3 mg/L. The addition of phosphate

appears to have dampened the effects of water quality, but the higher doses do not always offer

this control of iron release. The medium dose OP PDS showed significantly lower iron release

in Phases I and IV than in Phases II and III. Phases II and III each had one observation exceed

the standard, while Phases I and IV had none. This is likely due to the increased chlorides and

temperature observed in Phases II and III and decreased alkalinity. This is consistent with the

model presented previously and the model developed in the previous study without corrosion

inhibitor addition presented in Equation (2-2) (Taylor et al. 2005).

8360813096705610

485011809120047750....

...-

(HRT)re)(Temperatu(DO)×(Sodium)(Chloride)(Sulfate)y)(Alkalinit.)= Color(CPUΔApparent

(2-2)

where: Δ Apparent color = CPU

Alkalinity = mg/L as CaCO3 Sulfate = mg/L Chloride = mg/L Sodium = mg/L DO = dissolved oxygen, mg/L Temperature = oC HRT = hydraulic residence time, days

The medium dose OP PDS had higher iron release than the low dose OP PDS except in

Phase IV. Therefore the water quality affected iron release more at the medium dose than the

low dose of inhibitor. The high dose OP PDS had only one observation exceed the standard in

Phase IV, but no other phase. Phases II and IV had higher average iron release than Phases I and

III.

46

The iron release tends to be more influenced by water quality than by inhibitor addition.

Phase I and III were mostly groundwater, which is characterized by high alkalinity. Phase II had

the lowest percentage of groundwater and therefore the lowest alkalinity. The negative exponent

on the alkalinity term suggests reduction of iron release can be achieved by increasing the

alkalinity. This is confirmed with Figure 2-6, especially for the Phase I data that has the lowest

iron release and Phase II that has the highest iron release for both actual data and the predictions

by the model. The model is sensitive to the alkalinity of the blend due to the high concentration

of alkalinity and the larger exponent.

The model is also sensitive to chloride, but increases contribute to iron release, as

suggested by the positive exponent on the chloride term in the model. Phase I has the lowest

average chloride concentration. The reduced chloride concentration is another explanation for

the low levels of iron observed in Phase I.

Overall, the iron release is reduced by increases in the pH, and therefore alkalinity, from

the elevated pH treatment. OP inhibitor addition performed as well or sometimes worse than the

elevated pH treatment. Water quality effects had a greater impact on the iron release than the

inhibitor dose, suggesting inhibitor addition has limited control of iron release at the HRT of this

system. As suggested by the previous study (Taylor et al. 2005), iron release is best controlled

by reducing stagnation time and HRT in the system.

At the two-day HRT used in this study, iron release as well as zinc release were shown to

increase in the system from the unlined cast iron and galvanized steel pipe, respectively. This is

illustrated in Figure 2-7. While the inhibitor was not shown to offer much reduction in iron

release, an improvement with zinc release was observed. This is evidenced by the improved zinc

47

release with increasing inhibitor dose. This suggests the inhibitor offers protection of the

galvanized steel pipe surface better than the unlined cast iron pipe surface. The low OP dose did

not offer as much reduction in zinc release as the elevated pH treatment, but the medium and

high dose reduced zinc release further than elevated pH alone.

0.00

0.05

0.10

0.15

0.20

0.25

OP-low OP-med OP-high pHs pHs+0.3Treatment

Tota

l Fe

(mg/

L)

Influent Effluent

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

OP-low OP-med OP-high pHs pHs+0.3Treatment

Tota

l Zn

(mg/

L)

Influent Effluent

Figure 2-7 Comparison of influent and effluent iron and zinc

Surface Characterization

Iron Coupons

Surface characterization analysis by XPS was done on cast iron coupons that were

incubated for each phase. Each phase a new coupon was exposed to each treatment. The

coupons exposed to the inhibitor were exposed to the medium dose. A summary of the elements

detected on iron coupons is presented in Table 2-11.

48

Table 2-11 Elements found in XPS analysis of iron coupons

Element OP pHs pHs+0.3 (4 total) (4 total) (4 total)

Carbon 2 3 4 Calcium 2 3 4

Iron 3 4 4 Oxygen 4 4 4

Phosphorus 2 0 0 Silicon 3 3 3

Zinc 4 3 2

Carbon was detected on the OP treated coupons in Phases II and IV as carbonate.

Calcium was also detected in Phases II and IV, so the compound CaCO3 is suggested to be

present in these two phases. Iron corrosion products were found in all but Phase III coupons.

Phase IV was the only phase where FePO4 was detected. Phosphorus was detected in two of the

four coupons as PO43-. Silicon was detected in all phases but Phase II and zinc was detected in

all phases as ZnO and Zn2+.

For the pHs coupons, calcium and carbon in the carbonate form were detected for all

phases except Phase III suggesting calcium carbonate formed. For the pHs+0.3 coupons, calcium

carbonate was detected in all four phases. The most notable difference between the pH control

coupons and the OP treated coupons is there were no phosphorus forms detected on any of the

pH control coupons like there were for the OP treated coupons. This suggests the only source of

phosphorus is from the inhibitor.

Both of the pHs and pHs+0.3 coupons had iron corrosion products of Fe2O3, Fe3O4, and

FeOOH on all coupons in all phases. Oxygen was also detected on all coupons as O2-, CO32-,

and OH-. Silica was detected in all phases except Phase III for the pHs coupons and Phase II for

49

the pHs+0.3 coupons. ZnO was detected in all but Phase III for the pHs coupons and in Phases III

and IV for the pHs+0.3 coupons.

The distribution of compounds found on each of the cast iron coupons is presented in

Figure 2-8. FeOOH tends to be the dominant form found on coupons, with the percentage being

reduced when FePO4 forms. This may be an explanation for the slight, but inconsistent

improvement seen with inhibitor addition. The dominant corrosion products are similar for all

treatments, but FePO4 may form on the cast iron surface when orthophosphate inhibitor is

present.

0

20

40

60

80

100

OP pH

Inhibitor

% A

rea

of C

ompo

und

Fe2O3 Fe3O4 FePO4 FeOOH

Figure 2-8 Distribution of iron compounds on iron coupons

50

Galvanized Steel Coupons

A summary of the elements detected by XPS on the galvanized steel coupons incubated

for each phase and treatment is presented in Table 2-12. Iron corrosion products were detected

in all coupons exposed to OP inhibitor with FePO4 found on all coupons except Phase IV. This

corresponds to the phosphorus found as PO43- on three of the four coupons. Oxygen was

detected on all coupons as CO32- and O2- and on three of four coupons as OH- and PO4

3-.

Calcium was present as calcium carbonate for Phases I and III. Silica was detected in all but

Phase IV. Zinc was also detected in all four phases as ZnO and as Zn(OH)2 in Phases II and IV.

Table 2-12 Elements found in XPS analysis of galvanized steel coupons

Element OP pHs pHs+0.3 (4 total) (4 total) (4 total)

Carbon 4 4 4 Calcium 2 3 4

Iron 4 4 4 Oxygen 4 4 4

Phosphorus 3 0 0 Silicon 3 4 2

Zinc 4 4 3

The pHs and pHs+0.3 coupons showed oxygen in the form of CO32-, O2-, and OH- in all

phases for both treatments. Calcium was detected as calcium carbonate in all coupons for the

elevated pH PDS and three of four coupons for the pHs PDS. Silicon was observed on all four

coupons for pHs and in the Phases II and III for the pHs+0.3 coupons as a mixture of SiO and

SiOx. Zinc was detected as ZnO was found on all coupons for pHs and for three of four coupons

51

for pHs+0.3. Zinc hydroxide (Zn(OH)2) was found on the Phase IV coupon for pHs and on the

Phase II coupon for pHs+0.3.

The presence of phosphate on three of four of the OP coupons suggests the phosphate

may be inhibiting the release of zinc from the galvanized steel coupons since phosphate was not

detected on any of the pH control coupons. The phosphate was also present more frequently and

at a higher percentage than with the cast iron coupons, offering explanation for the better control

of zinc release over iron release.

The distribution of iron compounds identified on the galvanized steel coupons is shown

in Figure 2-9. For all treatments the compounds tend to be a mix with no dominant corrosion

product. As was seen with the cast iron coupons, the presence of orthophosphate inhibitor leads

to the formation of FePO4. The percentage of FeOOH is again reduced when FePO4 forms and

FePO4 is a higher percentage of the corrosion products than was seen on the cast iron coupons.

The formation FePO4 may help in reducing the release of zinc from the surface of the galvanized

coupons, as was seen in the OP PDSs.

52

0

10

20

30

40

50

60

70

80

90

100

OP pH

Inhibitor

% A

rea

of C

ompo

und

Fe2O3 Fe3O4 FePO4 FeOOH

Figure 2-9 Distribution of iron compounds on galvanized steel coupons

Thermodynamic Modeling

Pourbaix diagrams are constructed using the Nernst equation and the chemical

relationships between the different iron components of interest. These diagrams show the

products that are thermodynamically stable and dominant under varying pH and potential

conditions. The diagram constructed here also varies alkalinity and total dissolved iron

concentration. Without the addition of a phosphate-based inhibitor, carbonate, oxide, and

hydroxide scales tend to predominate within drinking water distribution systems.

53

Compounds were selected for construction of these diagrams based on past literature

(Snoeyink and Jenkins 1980; Benjamin, Sontheimer, and Leroy 1996). Solid products Fe(OH)3,

FePO4, Fe(OH)2, and FeCO3 were all considered and included where appropriate. Complexes

considered include Fe2+, Fe3+, FeOH+, FeOH2+, Fe(OH)2+, and Fe(OH)4

-. Dominant complexes

that were not taken over by regions of insoluble species appear in the Pourbaix diagrams.

The thermodynamic model is represented by a Pourbaix diagram varied across a range of

alkalinities that used the constituents previously discussed shown in Figure 2-10. The diagram

was constructed for the iron species at a temperature of 298 K. As shown in the figure, the

native iron pipe is only stable at very low potentials. As the potential increases the iron will

oxidize to Fe(II) species and then to Fe(III) species. Depending on the characteristics of the

layer of solids that form, these solids can form a passivating layer to help protect the iron metal

from corrosion. Formation of the different oxidation layers is due to the diffusion of oxidants

from the bulk water through the corrosion scales. Diffusion is higher at the surface of the scale

so the more oxidized Fe(III) products are observed near the upper layers. Fe(II) products form

on the layer above the pipe surface where diffusion is limited.

54

-20

-10

0

10

20

0 2 4 6 8 10 12 140

100

200300

pe

pH

Alkalinity (mg/L-C

aCO

3 )

Iron Speciesat T = 298 K

with 1 mg/L P addition

CT,Fe = 10-6 MCT,Fe = 10-5 MCT,Fe = 10-4 M

Fe2+

Fe3+

FePO4

FeCO3Fe

Fe(OH)2

Fe(OH)3

Fe(OH)4-

TypicalDrinkingWater

Conditions

Figure 2-10 Pourbaix diagram for iron species with 1 mg/L P inhibitor addition

55

The Pourbaix diagram shows the thermodynamically stable products that form. The

corrosion products that form in practice may not be thermodynamically stable, so would not

appear on the diagram but would be seen in an actual pipe. Also, the bulk water pH, alkalinity

and potential are known, but the characteristics of the conditions within corrosion scales may be

different and likely unknown (Benjamin, Sontheimer, and Leroy 1996).

The oxidation of iron pipe is limited by the ability of the oxidants like DO and

disinfectant residual to diffuse through the corrosion scale layers to the pipe surface. Factors like

the flow velocity of the bulk water and the porosity and thickness of the corrosion scale impact

the pipe corrosion. Therefore, thicker more dense corrosion scale acts to protect the pipe from

further corrosion. However, these scales can release into the bulk or release soluble iron

products in periods of stagnation.

The varying concentration of total dissolved iron, CT,Fe, is shown to broaden the regions

of the solid products with increasing concentration. For example, Figure 2-10 shows Fe(OH)3

appears in a pH range of 7.2 to 10.1 when CT,Fe is 10-6 M (0.056 mg/L) and is in the pH range of

7.2 to 12.1 when CT,Fe is 10-4 M (5.6 mg/L). This suggests the increased concentration has a

larger pH range to form a protective scale layer rather than be in a complex dissolved form

Also shown in Figure 2-10 are typical drinking water conditions, represented by the

shaded cube. This region corresponds to conditions that would be observed in the bulk water of

a typical distribution pipe having an alkalinity between 50 and 300 mg/L-CaCO3, a pH of 7.5 to

8.5, and a pe spans from 9.0 to 12.5. This pe is typical for a system maintaining a

monochloramine residual. The dominant form of iron in this range is Fe(OH)3. This is

consistent with observations seen in practice. The upper layer of scale formation is usually

56

Fe(III) forms of iron due to the proximity to the bulk water and oxidants like DO and disinfectant

residual. Also, it has been observed that occurrences of red water at consumers’ taps is in the

form of precipitated Fe(OH)3 flocs (Benjamin, Sontheimer, and Leroy 1996). The Fe(OH)3

forms when Fe2+ ions are either released through diffusion from the corrosion scale or the metal

pipe and oxidize and precipitate as Fe(OH)3.

Figure 2-10 shows a Pourbaix diagram constructed with consideration of iron-phosphate

products with 1 mg/L of phosphorus addition to the system. It can be seen that FePO4 forms an

insoluble region at lower pH than is typical of the bulk water in the distribution system. The

Fe(OH)3 is still the dominant form in the bulk water conditions. However, the boundary between

the Fe2+ ion and the FePO4 is lower than the boundary observed between the Fe2+ ion and the

Fe(OH)3. This may explain the limited iron release control observed when phosphate inhibitor

treatment is used.

Conclusions

• Empirical modeling of water quality to predict total iron release found the total

phosphorus, silica, pH, alkalinity, chloride, sulfate, and temperature to be significant

parameters. The addition of phosphorus and increases in pH and alkalinity were

found to decrease the total iron release while the silica, chloride, sulfate, and

temperature were found to contribute to total iron release. The model was not very

sensitive to inhibitor dose, suggesting water quality has a greater effect on iron

release than inhibitor addition.

57

• Observations found PDSs treated with orthophosphate inhibitor did not help to reduce

iron release any better than elevating the pH, and in some cases performed worse. On

average no treatment in this study exceeded the secondary standard of 0.3 mg/L, but

the PDS at pHs had the most observation points above 0.3 mg/L. The low OP dose

had the most consistent control of iron release.

• XPS analysis of cast iron coupons exposed to the medium OP dose showed evidence

of FePO4 formation in two of the four phases at an average of 7.4% of the iron

corrosion products. Both the PDSs at pHs and pHs+0.3 showed no evidence of

phosphate corrosion products, suggesting the phosphate is formed from inhibitor

addition. Analysis of the galvanized steel coupons also showed evidence of

formation of FePO4 on three of four coupons at an average of 28% of the iron

corrosion products. Both PDSs at pHs and pHs+0.3 had no phosphate products present

on the galvanized steel coupons. The greater FePO4 formation on the galvanized steel

coupons offers a good explanation for why zinc release was reduced from the

galvanized steel coupons but iron release from the cast iron coupons was

inconsistently controlled with the addition of orthophosphate inhibitor.

• Thermodynamic modeling of iron corrosion products shows FePO4 to be

thermodynamically stable when phosphate is present. The formation of solid FePO4

offers some control of iron oxidation at a lower pe than Fe(OH)3. However, this

control is thermodynamically stable at a lower pH then observed in the bulk water.

This offers explanation for the inconsistent control of iron release with

58

orthophosphate inhibitor addition and validation of formation observed in XPS

analysis.

59

References

APHA, AWWA, and WEF (American Public Health Association, American Water Works Association, and Water Environmental Association). 1998. Standard Methods for the Examination of Water and Wastewater. 20th ed. Washington, D.C.: APHA.

Benjamin, M.M., H. Sontheimer, and P. Leroy. 1996. Corrosion of Iron and Steel. In Internal Corrosion of Water Distribution Systems. 2nd ed. Denver, Colo.: American Water Works Association Research Foundation.

Benjamin, M.M., S.H. Reiber, J.F. Ferguson, E.A. Vanderwerff, and M.W. Miller. 1990. Chemistry of Corrosion Inhibitors in Potable Water. Denver, Colo.: AWWA Research Foundation and American Water Works Association.

Boffardi, B.P. 1988. Potable water treatment and monitoring for corrosion and scale control. Journal New England Water Works Associations, 102(2):111.

Burlingame, G.A., D.A. Lytle, and V.L. Snoeyink. 2006. Why red water? Understanding iron release in distribution systems. Opflow, 32(12):12-16.

Gedge, G. 1992. Corrosion of cast iron in potable water service. Corrosion and related aspects of materials for potable water supplied. Proceedings Institute of Materials Conference, London.

Hatch, G.B. and O. Rice. 1939. Surface-active properties of hexametaphosphate. Industrial and Engineering Chemistry, 31(1):51.

Hoover, C.P. and O. Rice. 1939. Threshold treatment. Water Works and Sewage, 86:10.

Huange, D.J.-S. 1980. Polyphosphate for corrosion control in water distribution system. PhD dissertation, University of Missouri, Columbia.

McNeill, L.S., and M. Edwards. 2001. Iron pipe corrosion in distribution systems. Journal American Water Works Association, 93(7):88-100.

McNeill, L.S., and M. Edwards. 2000a. Phosphate inhibitors and red water in stagnant iron pipes. Journal of Environmental Engineering, 126(12):1096-1102.

McNeill, L.S., and M. Edwards. 2000b. Temperature effects on iron corrosion. Ph.D. diss., Virginia Polytechnic Institute, Blacksburg, Va.

Pryor, M.J. and M. Cohen. 1951. The mechanism of the inhibition of corrosion of iron by solutions of sodium orthophosphate. Journal Electrochemical Society, 98(7):263.

60

Sarin, P., V.L. Snoeyink, D.A. Lytle, and W.M. Kriven. 2004. Iron Corrosion Scales: Model for Scale Growth, Iron Release, and Colored Water Formation. Journal of Environmental Engineering ASCE, 130(4):364-373.

Snoeyink, V.L., and D. Jenkins. 1980. Water Chemistry. New York: John Wiley & Sons, Inc.

Sontheimer, H., W. Kolle, and V.L. Snoeyink. 1981. The siderite model of the formation of corrosion-resistant scales. Research and Technology, Journal AWWA, 73(11):572-579.

Taylor, J.S., J.D. Dietz, A.A. Randall, S.K. Hong, C.D. Norris, L.A. Mulford, J.M. Arevalo, S. Imran, M. Le Puil, S. Liu, I. Mutoti, J. Tang, W. Xiao, C. Cullen, R. Heaviside, A. Mehta, M. Patel, F. Vasquez, and D. Webb. 2005. Effects of blending on distribution system water quality. Denver, Colo.: AwwaRF and Tampa Bay Water.

Volk, C., E. Dundore, J. Schiermann, and M. Lechevallier. 2000. Practical evaluation of iron corrosion control in a drinking water distribution system. Water Research, 34(6):1967-1974.

Wagner, I. 1992. Influence of operating conditions on materials and water quality in drinking water distribution systems. Corrosion and related aspects of materials for potable water supplies. In Proceedings Institute of Materials Conference, London.

Wagner, I and A. Kuch. 1984. The influence of water parameters of corrosion rate, scale deposition and iron(III) uptake in unprotected iron pipes. Water Supply, 2(3/4):SS11.

61

CHAPTER 3 EFFECTS OF ORTHOPHOSPHATE CORROSION INHIBITOR ON

COPPER IN BLENDED WATER QUALITY ENVIRONMENTS

Abstract

This study evaluated the effects of orthophosphate (OP) inhibitor addition on copper

corrosion on coupons exposed to different blends of groundwater, surface water, and desalinated

seawater. The effectiveness of OP inhibitor addition on copper release was analyzed by

statistical comparison between OP treated and untreated pilot distribution systems. Four

different doses of OP inhibitor, ranging from zero (control) to 2 mg/L as P, were investigated and

non-linear empirical models were developed to predict copper release from the water quality and

OP doses. Surface characterization evaluations were conducted using X-ray Photoelectron

Spectroscopy (XPS) analyses for each copper coupon tested. A theoretical thermodynamic

model was developed and used to validate the controlling solid phases determined by XPS. OP

inhibitor addition was found to reduce copper release for the OP dosages evaluated and the water

blends evaluated compared to pH adjustment alone. Empirical models showed increased total

phosphorus, silica and pH reduced copper release while increased alkalinity and chloride

contributed to copper release. Thermodynamic modeling suggested that Cu3(PO4)2·2H2O is the

controlling solid that forms on copper surfaces, regardless of blend, when OP inhibitor is added

for corrosion control.

62

Introduction

With increasing water demands and more stringent drinking water regulations, many

utilities are turning to desalinated sources to supplement their surface and groundwater supplies.

Tampa Bay Water (TBW) and the University of Central Florida (UCF) studied the effects of

blending multiple alternative source waters on distribution system water quality (Taylor et al.

2005). This study further evaluates the addition of orthophosphate (OP) corrosion inhibitor to

the blended source waters and the effects on copper corrosion.

Copper levels in drinking water are regulated through the Lead and Copper Rule (LCR)

and limited to an action level of 1.3 mg/L at the 90th percentile of household kitchen taps

sampled by voluntary participant homeowners (Federal Register 1991). The source of copper in

drinking water comes primarily from corrosion of copper plumbing and is influenced by water

quality parameters like pH, alkalinity, chloride, nitrate, sulfate, sodium, calcium, and magnesium

(Edwards, Ferguson, and Reiber 1993). Addition of orthophosphates is believed to reduce

copper release by forming Cu3(PO4)2 or a similar scale on the surface of the pipe (Reiber 1989;

Schock, Lytle, and Clement 1995). However, the benefits of using orthophosphate are thought to

be limited to cases of pH less than 8 (Reiber 1989; Schock, Lytle, and Reiber 1995; Edwards,

Jacobs, and Dodrill 1999; Dodrill and Edwards 1995).

Zhang et al. (2002) tested the corrosion of copper exposed to tap water with

monochloramine disinfectant for a period of 30 days. They found at a pH of 8, the copper

corrosion increased for six days and then became steady for the remaining days. Increasing ionic

strength, dissolved inorganic carbon and temperature promoted corrosion, resulting in thicker

oxide films

63

Pinto, McAnally, and Flora (1997) evaluated the addition of phosphates as well as pH

and alkalinity adjustment for corrosion control of copper in low hardness, low alkalinity waters.

Addition of phosphate corrosion inhibitor was shown to help reduce copper levels. Increasing

alkalinity was found to increase copper release. Phosphate addition was recommended because

it is effective at low doses.

Indian Hills Water Works in Ohio used elevated pH to treat high copper levels and zinc

orthophosphate for treatment of lead, but still had copper problems. To treat both, a study

showed an increase in orthophosphate inhibitor to a dose of 3 mg/L as PO4. This kept both lead

and copper within the action level and use of the zinc orthophosphate was discontinued (Schock

and Fox 2001).

A test of copper loops with stagnation and flow conditions similar to household plumbing

systems showed phosphate inhibitors to reduce copper concentrations. Stabilization of the

copper surface layer by building insoluble compounds of copper ions and phosphates was shown

to interfere in the reaction kinetics of the dissolution of the copper surface layer (Becker 2002).

Dodrill and Edwards (1995) conducted a survey of about 360 utilities to examine their

strategies in controlling lead and copper release, in response to the Lead and Copper Rule (LCR)

of 1991. For copper, the survey showed that at high pH, copper release is reduced with and

without inhibitors. At pH less than 7.8, copper release was high at high alkalinity, but inhibitor

use mitigated that release. However, at pH greater than 7.8, inhibitors had variable and adverse

effects on copper corrosion by-product release.

Edwards, McNeill, and Holm (2001) compared the benefits of orthophosphate versus

polyphosphate in controlling copper corrosion by-product release in aged copper pipes, at

64

variable pH and alkalinity values. Polyphosphate appeared to perform less favorably than

orthophosphate at comparable concentrations of 1 mg/L as P. It was believed that

orthophosphate reduced copper solubility by forming a cupric phosphate scale. In the absence of

any phosphate inhibitors, an insoluble malachite scale formed over a period of years.

Polyphosphate, however, increased copper release in comparison to orthophosphate because it

complexed copper, increasing soluble copper release.

When inhibitors are not considered, copper release can be controlled best by raising pH

(Edwards, Hidmi, and Gladwell 2003). It was demonstrated that bicarbonates adversely affected

copper release, and that a pH increase (7.0-8.0) showed significant reduction in copper release.

CO2 stripping was the recommended method for raising the pH without raising alkalinity.

This study evaluated the effects of orthophosphate inhibitor addition to blended treated

surface, ground, and seawater sources of varying percentages. The effects of water quality were

evaluated and a model predicting total copper release using water quality and total phosphorus

concentrations was developed. XPS analysis of copper coupons was evaluated for solid phase

surfaces present on the coupon, and thermodynamic modeling was performed to gain insight into

inhibitor control of copper release.

Experimental Methods

Experimental Design

Experimentation was conducted with the use of pilot distribution systems (PDSs) built

from actual pipelines extracted from TBW member governments distribution systems

(Hillsborough County, Fla.; Pasco County, Fla.; Pinellas County, Fla.; City of New Port Richey,

65

Fla.; City of St. Petersburg, Fla.; and City of Tampa, Fla.). Details regarding the PDS and prior

study results are reported elsewhere (Taylor et al. 2005). Each PDS runs in parallel with

segments of PVC, lined cast iron, unlined cast iron, and galvanized steel pipes that were placed

sequentially to simulate actual distribution systems. The materials and their respective diameter

and length are shown in Table 3-1. Each PDS was fed blends of groundwater, surface water, and

desalinated seawater along with different types and doses of corrosion inhibitor.

Table 3-1 Pipe materials in PDSs

Pipe Material Length (ft)

Diameter (in)

PVC 20 6 Lined Cast Iron 20 6

Unlined Cast Iron 12 6 Galvanized Steel 40 2

Images of materials as well as the full PDSs used to represent full scale distribution

systems are shown in Figure 3-1. The pipe materials used, shown in the left image, are

galvanized steel, PVC, lined cast iron, and unlined cast iron from left to right. The image on the

right shows the PDSs with the materials connected in series from influent (closest to viewer) to

effluent (furthest from viewer) of the system. The pipe materials were connected in sequence of

PVC, lined cast iron, unlined cast iron, then galvanized steel.

66

Figure 3-1 Pipe materials (left) and parallel PDSs (right)

The PDSs were fed blends of conventionally treated groundwater (GW), enhanced

coagulation-sedimentation-filtration (CSF) treated surface water (SW), and desalinated seawater

by reverse osmosis (RO). A description of the three finished source waters is presented in Table

3-2.

Table 3-2 Finished source water descriptions

Water Source System Description

GW Groundwater Ground water source. Treatment by aeration, disinfection by free chlorine with a residual of 5 mg/L after a 5 minute contact time. 5.0 mg/L chloramine residual.

SW Surface water

TBW treatment plant: Treatment by ferric sulfate coagulation, flocculation, settling, filtration, disinfection by ozonation and chloramination. Project site: adjustment of chloramine residual to 5.0 mg/L chloramine residual.

RO Groundwater Treatment by membrane reverse osmosis, aeration, disinfection by free chlorine with a residual of 5 mg/L after a 5 minute contact time. 5.0 mg/L chloramine residual.

67

The GW unit used raw well water from the Cypress Creek well field owned by TBW.

The GW was treated with aeration, disinfection, and pH stabilization. Aeration was achieved in

the GW by pumping the raw water to the top of the finished water tank through a spray nozzle.

Sodium hypochlorite was used for primary disinfection and was dosed to provide a 5 mg/L

residual after a 5 minute contact time. Afterwards, ammonium chloride was added to produce a

5 mg/L monochloramine residual. Ammonia was added in the form of NH4Cl at a 5:1 ratio. The

Cl2:NH3 ratio was initially 4:1 to protect against DBP formation. This ratio was increased to 5:1

in after 6 months of operation to reduce free ammonia.

SW was treated at the TBW Regional Surface Water Treatment Facility by enhanced

coagulation, ozonation, biologically activated carbon (BAC) filtration, aeration, and

chloramination. The SW was hauled weekly to the field facility for use and temporarily stored in

two 7000 gallon storage tanks before being transferred to the SW finished water tank. In the SW

finished tank, the chloramine residual was adjusted to 5 mg/L as Cl2.

The RO pilot plant was housed in a trailer at the testing facility and utilized raw

groundwater for the feed stream. The RO treatment pilot system required the addition of TDS,

calcium and alkalinity to the RO permeate to represent the finished water produced by the TBW

Regional Desalination Facility. RO pretreatment consisted of 2.7 mg/L antiscalant addition

(Hypersperse MDC700TM, GE Water, Minnetonka, Minn.) followed by 5-micron cartridge

filtration. The RO membrane unit was operated at 72-73% recovery, producing 9.3 gpm

permeate flow, which was aerated by a 10-inch diameter aeration tower filled with tripack plastic

packing. After aeration, 50 mg/L of sea salt was added to the aerated permeate stream to

simulate the TBW desalination process. Calcium chloride and sodium bicarbonate were also

68

added to meet the calcium and alkalinity specifications. The finished was stabilized with sodium

hydroxide to 0.1 to 0.3 pH units above pHs.

The effects of water quality were evaluated by varying the blend quarterly, while

seasonal effects were evaluated by maintaining the same blend in the summer and winter. The

quarterly phases and percentages of the blends are shown below in Table 3-3. The average water

quality of each of the source waters in each phase is shown in Table 3-4. The effects of season

are seen in the temperature as well as rainy and dry season effects on the surface water between

Phases I and III. The blends with a high percentage of groundwater in Phases I and III are

characterized by high alkalinity and pH. Phase II had the highest percentage of surface water

and is characterized by high sulfate concentrations. Phase IV has average water quality

parameters due to the equal percentage of GW and SW.

Table 3-3 Blend percentages for each phase

Phase Quarter % GW % SW % RO I Feb-May 2006 62 27 11 II May-Aug 2006 27 62 11 III Aug-Nov 2006 62 27 11 IV Nov 2006-Feb 2007 40 40 20

Table 3-4 Average water quality by phase

Phase pH Alkalinity (mg/L CaCO3)

Chloride (mg/L)

Sulfate (mg/L)

Temperature (°C)

I 8.0 161 45 62 21 II 7.9 104 67 103 26 III 8.0 150 68 66 26 IV 7.9 123 59 76 21

69

The feed rate of the blend into each PDS was maintained to achieve a two-day hydraulic

residence time (HRT). Pumps maintained the blend flow as well as the inhibitor addition into

each PDS. The PDSs each were fed different inhibitor types and doses. The inhibitors were

dosed to the PDSs at three different levels, categorized as low dose, medium dose, and high dose.

Orthophosphate (OP) was maintained at a target dose of 0.5 mg/L as P for the low dose, 1.0

mg/L as P for the medium dose and 2.0 mg/L as P for the high dose. Control PDSs were not fed

any chemical inhibitor; one was maintained at pHs and a second was treated with elevated pH,

maintained at pHs+0.3. The PDS at pHs+0.3 was maintained at a positive LSI to assess the affect of

elevated pH treatment as a means of copper release control.

Orthophosphate Inhibitor

The ortho-phosphate inhibitor used in this study is Inhibit-All WSF-36 (SPER Chemical

Corporation, Clearwater, Fla.). It is made of monosodium ortho-phosphate blended into 17

megohm purified water at a concentration of 36%. It is a clear, slightly viscous liquid with a

bulk density of 11.25 lbs/gal. The specific gravity is 1.35 the pH of 1% solution is 5.1 to 5.4.

Table 3-5 shows a summary of the properties of the OP inhibitor along with manufacture

recommendations for use.

70

Table 3-5 Orthophosphate inhibitor properties

Property Value Manufacturer SPER Chemical Corp. Product Inhibit-All WSF-36 Percent Orthophosphate 36% Bulk Density 11.25 lbs/gal Specific Gravity 1.35 pH 1% solution 5.1 – 5.4 Recommended dose (low-med-high) 1.25-2.75-4.25 mg/L as Product Used dose (low-med-high) 0.5-1.0-2.0 mg/L as P Recommended pH range 7.4 – 7.6 (optimum), 6.8 – 7.8

Data Collection

Water quality parameters were collected and analyzed weekly in the influent and effluent

of the PDSs. The water quality parameters monitored and methods used for samples transported

to the UCF laboratory for analysis are shown in Table 3-6. Table 3-7 shows the water quality

parameters monitored and methods used for the parameters analyzed in the field laboratory at the

testing facility. The method detection limit (MDL) is also shown. Methods noted as SM are

from APHA, AWWA, and WEF (1998).

71

Table 3-6 Water quality parameters and methods performed at University Laboratory

Parameter Method Reference Method Description MDL

Aluminum SM 3120B ICP Method 0.001 mg/L Bicarbonate SM 2320B Titration Method 5 mg/L Calcium SM 3120B ICP Method 0.1 mg/L

Chloride SM 4110 Ion Chromatography with Chemical Suppression of Eluent Conductivity 0.1 mg/L

Color SM 2120A Or Hach 8025 Cobalt-Platinate Method (with spec) 1 CPU

Conductivity SM 2510B Laboratory Method 1 μmho/cm Copper SM 3120B ICP Method 0.001 mg/L Iron SM 3120B ICP Method 0.001 mg/L Lead SM 3120B ICP Method 0.001 mg/L Magnesium SM 3120B ICP Method 0.1 mg/L Nitrogen (NH3,TKN) SM 4500-Norg Macro-Kjeldahl Method 0.1 mg/L

NPDOC SM 5310C Persulfate-UV Oxidation Method 0.1 mg C/L pH SM 4500-H+ B Electrometric Method ± 0.01 pH units Phosphorus SM 3120B ICP Method 0.001 mg/L Silica SM 3120B ICP Method 0.001 mg/L Sodium SM 3120B ICP Method 0.1 mg/L Solids (TDS) SM 1030E Estimation of TDS by major ion sum 1 mg/L

Sulfate SM 4110 Ion Chromatography with Chemical Suppression of Eluent Conductivity 0.1 mg/L

Turbidity SM 2130B Nephelometric Method 0.01 NTU UV-254 SM 5910 UV Absorption at 254 nm 0.0001 cm-1 Zinc SM 3120B ICP Method 0.001 mg/L

72

Table 3-7 Water quality parameters and methods performed at Field Laboratory

Parameter Method Reference Method Description MDL Alkalinity SM 2320 B Titration 5 ppm Ammonia-N SM 4500-NH3 C Membrane Probe Method 0.1 ppm Chloride SM 4500-Cl- B Argentometric Titration 1 mg/L Chlorine, free SM 4500-Cl G or Hach 8021 DPD colorimetric 0.1 ppm Chlorine, total SM 4500-Cl-G or Hach 8167 DPD colorimetric 0.1 ppm Color, apparent SM 2120 B Visual Comparison

(by spectrometer) 1 CPU

Conductivity SM 2510 B Conductivity Bridge 1 μmho/cm Hardness (total, calcium) SM 2340 C EDTA Titration 5 mg/L

Nitrate Hach 8192 Cadmium reduction 0.1 mg/L Nitrite Hach 8507 Diazotization 0.1 mg/L Oxygen, Dissolved (DO) SM 4500-O G Membrane probe 0.1 mg/L

pH SM 4500-H+ B Electrometric ± 0.01 pH units

Phosphate-P (Reactive) SM 4500-P E. or Hach 8048 Ascorbic Acid Method 0.1 mg/L

Silica, SiO2 (reactive) SM 4500-SiO2 or Hach 8185 Molybdosilicate Method 0.1 mg/L

as SiO2 Temperature SM 2550 B Direct reading 0 deg C Turbidity SM 2130 B Nephelometric 0.01 NTU UV254 SM 5910 A UV spectrometry 0.0001 cm-1

Portions of the flow from each PDS were fed to a corrosion loop consisting of 30 feet of

5/8 inch copper tubing with one lead/tin coupon to represent solder. Each loop holds

approximately 1.8 L of water. The copper tubes were flushed every morning with 2 gallons of

the PDS water. Weekly samples were collected after a six-hour stagnation period in order to

simulate tap monitoring as described in the Lead and Copper Rule. The corrosion shed and

housed corrosion loops are show in Figure 3-2.

73

Figure 3-2 Corrosion shed (left) and corrosion loops (right)

Copper coupons were placed in cradles that received flow in parallel with each PDS. The

coupons were evaluated for surface characteristics after incubation during each phase. X-ray

Photoelectron Spectroscopy (XPS) analysis was then performed on each type of coupon to

identify chemical components on the outer layer of the corrosion surface. A survey scan reveals

the presence of elements, whereas a high resolution scan of those elements found on the outer

layer shows the chemical states, providing detailed surface characterization information.

Results and Discussion

Dose Maintenance

Three of the PDSs were treated with orthophosphate corrosion inhibitor at a low,

medium, and high dose. These doses were targeted to maintain at 0.5 mg/L as P, 1.0 mg/L as P,

and 2.0 mg/L as P, respectively. The average orthophosphate inhibitor dose for the course of the

study in each PDS is shown in Figure 3-3. Error bars represent the minimum and maximum

74

observations. The low dose of orthophosphate averaged 0.51 mg/L as P, the medium dose

averaged 0.94 mg/L as P, and the high dose averaged 1.83 mg/L as P.

1.83

0.94

0.51

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Low Med High

OP Inhibitor Dose

Tota

l P (m

g/L

as P

)

Figure 3-3 Orthophosphate inhibitor dosing

The orthophosphate inhibitor is added as phosphoric acid, so it has an effect of lowering

the pH is those PDSs treated with inhibitor. The pH maintained in each of these PDS is shown in

Figure 3-4 with the error bars representing the minimum and maximum observations. The

difference in pH maintained in each of the PDSs is significantly different with the exception of

the low orthophosphate dosed PDS being the same as the elevated pH PDS at pHs+0.3.

75

7.94

7.71

7.867.907.92

6.8

7.0

7.2

7.4

7.6

7.8

8.0

8.2

8.4

8.6

OP-low OP-med OP-high pHs pHs+0.3

pH

Figure 3-4 pH for OP and control PDSs

Empirical Modeling

An empirical model for predicting total copper release was developed using the water

quality data collected from the PDSs with OP inhibitor addition as well as the pH control PDSs.

The range of the data used in development of this model is presented in Table 3-8. The inclusion

of water quality parameters was based on ANOVA procedures for parameters that were

statistically significant. Non-linear least squares regression was performed and independent

parameters not significant at a 95% confidence level were eliminated. The resulting model is

presented in Equation (3-1). All parameters shown in the model retained p-values less than 0.05.

76

Table 3-8 Range of water quality in model development

Parameter Minimum Maximum Total Phosphorus, mg/L as P 0.01 2.69 Silica, mg/L as SiO2 3.8 13.6 pH 7.4 8.4 Alkalinity, mg/L as CaCO3 84 170 Chloride 38 123

408.0459.1591.4281.02

280.0446.1 ClAlkpHSiOTPCuTotal ×××××= −−− (3-1)where: Total Cu = total copper, mg/L

TP = total phosphorus, mg/L as P SiO2 = silica, mg/L as SiO2 pH = -log[H+] Alk = alkalinity, mg/L as CaCO3 Cl = chloride, mg/L

This model suggests the addition of the orthophosphate inhibitor, as measured by the total

phosphorus concentration, mitigates copper release. This is shown by the negative exponent on

the total phosphorus term. Similarly, higher silica and pH reduce copper levels in the corrosion

loops. The pH term suggests the elevated pH treatment to pHs+0.3 is a valuable treatment.

However, increased alkalinity and chloride contribute to copper release.

The fit of the model to the data has an R2 value of 0.71 and is shown graphically in

Figure 3-5. The performance of each corrosion control treatment can also be seen in Figure 3-5.

The highest dose of orthophosphate inhibitor has the lowest copper concentration followed by

the medium and low doses, respectively. The elevated pH treatment at pHs+0.3 is next followed

by no treatment at pHs with the highest observed copper release. This agrees with the pH effect

discussed previously.

77

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0Actual Total Cu (mg/L)

Pred

icte

d To

tal C

u (m

g/L)

OP-low OP-med OP-high pHs pHs+0.3

Figure 3-5 Predicted versus actual total copper concentrations using the empirical model

For the data, but for the pHs PDS, the model under-predicts the copper concentration in

some cases. With respect to the action level of 1.3 mg/L of total copper, the model describes the

data below the action level better than above the action level. If exceeded, then the model

under-predicts very high concentrations. Therefore, the model is useful as a predictor of copper

release with OP inhibitor addition. All of the PDSs operating with OP inhibitor addition are

maintained below the action level.

Figure 3-6 shows the model predictions of the total copper release for each PDS by

phase. The bars represent the 90th percentile total copper concentrations observed or predicted

78

for comparison to the action level of 1.3 mg/L for the 90th percentile of samples. The trend of

decreasing concentration with increasing inhibitor dose is well defined by the model. It is also

shown that actual and predicted total copper concentration for all PDSs receiving the OP

inhibitor were maintained below the action level of 1.3 mg/L.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

I-OP lo

w

I-OP m

ed

I-OP hi

gh

II-OP lo

w

II-OP m

ed

II-OP hi

gh

III-O

P low

III-O

P med

III-O

P high

IV-O

P low

IV-O

P med

IV-O

P high

Phase-Treatment

Tota

l Cu

90th

per

cent

ile (m

g/L)

Actual Predicted

Figure 3-6 Actual vs. predicted for empirical model by phase and PDS, 90th percentiles

Orthophosphate Inhibitor Performance

Table 3-9 shows a summary of the copper release in each of the OP treated and pH

control PDSs. As shown with the predictive model presented previously, Table 3-9 shows

copper concentration to decrease with increasing dose of orthophosphate inhibitor. Also shown

79

in Table 3-9 is the total copper concentration was consistently 80 to 90 % in the dissolved form.

These values are also shown graphically in Figure 3-7 with the action level of 1.3 mg/L noted.

The error bars represent the 90th percentile observations for the action level and the average

values are displayed. Again it can be seen that no observation in the PDSs treated with

orthophosphate inhibitor exceeded the action level. This figure also shows higher copper

concentrations were observed in the pHs PDS than the pHs+0.3 PDS.

Table 3-9 Copper release summary

Dissolved Cu (mg/L) Total Cu (mg/L) Phase Treatment Avg Min Max Avg Min Max 90th percentile

I OP-low 0.46 0.41 0.57 0.50 0.44 0.59 0.56 OP-med 0.38 0.26 0.48 0.41 0.27 0.48 0.45 OP-high 0.26 0.16 0.38 0.30 0.16 0.60 0.33 pHs 0.95 0.69 1.41 1.04 0.75 1.52 1.22 pHs+0.3 0.89 0.29 1.29 1.00 0.76 1.37 1.16

II OP-low 0.33 0.13 0.49 0.39 0.28 0.52 0.42 OP-med 0.28 0.15 0.38 0.35 0.18 0.43 0.42 OP-high 0.21 0.10 0.28 0.24 0.12 0.31 0.27 pHs 1.04 0.73 1.53 1.16 0.67 1.71 1.62 pHs+0.3 0.73 0.13 1.24 0.82 0.21 1.34 0.95

III OP-low 0.36 0.29 0.44 0.40 0.25 0.50 0.47 OP-med 0.29 0.24 0.37 0.35 0.28 0.43 0.41 OP-high 0.19 0.14 0.26 0.21 0.16 0.28 0.24 pHs 1.32 0.84 1.84 1.48 0.97 2.08 1.91 pHs+0.3 0.79 0.62 1.13 0.85 0.69 1.14 1.02

IV OP-low 0.34 0.20 0.63 0.40 0.26 0.73 0.55 OP-med 0.32 0.19 0.56 0.36 0.24 0.64 0.51 OP-high 0.21 0.12 0.34 0.23 0.15 0.37 0.29 pHs 1.14 0.06 2.25 1.41 0.34 2.51 1.99 pHs+0.3 0.83 0.50 1.53 0.93 0.56 1.72 1.36

80

0.0

0.5

1.0

1.5

2.0

2.5

Phase I Phase II Phase III Phase IV

Tota

l Cu

(mg/

L)

OP-low OP-med OP-high pHs pHs+0.3

Action Level = 1.3 mg/L

Figure 3-7 Comparison of total Cu by phase and treatment

The elevated pH treated PDS had observations exceeding the standard in all but Phase III,

but the 90th percentile was only exceeded for Phase IV. However, the pHs PDS had observations

exceeding the standard in every phase and exceeded at the 90th percentile in all but Phase I.

The average water quality observed in each PDS in each phase is presented in Table 3-10.

The water quality parameters presented here are those that were significant in the empirical

model presented previously. As seen with the model, the total phosphorus concentration has

mitigating effects on copper release. Increased phosphorus concentration decreases copper

concentration. Also consistent with the model, the elevated pH treatment consistently has lower

81

copper concentrations than the PDS at pHs. The additional phosphate in the OP PDSs reduced

copper levels further than elevated pH treatment alone.

Table 3-10 Average influent water quality by PDS and phase

Total Phosphorus Silica pH Alkalinity Chloride Treatment Phase (mg/L as P) (mg/L as SiO2) (std. units) (mg/L as CaCO3) (mg/L)

OP-low I 0.49 10.8 8.0 164 43.2 II 0.55 5.0 7.8 106 67.2 III 0.54 10.4 8.0 152 63.6 IV 0.47 6.3 7.8 125 56.6 Average 0.51 8.2 7.9 143 54.7

OP-med I 1.04 10.8 8.0 164 43.1 II 0.90 4.9 7.8 107 67.4 III 0.99 10.4 8.0 152 62.8 IV 0.82 6.3 7.8 125 56.4 Average 0.94 8.2 7.9 143 54.5

OP-high I 1.84 10.7 8.0 163 43.2 II 1.91 5.0 7.8 107 67.8 III 1.87 10.4 7.9 151 65.0 IV 1.69 6.2 7.8 125 57.9 Average 1.83 8.2 7.9 142 55.4

pHs I 0.11 11.0 7.7 146 64.2 II 0.03 4.9 7.8 92 82.8 III 0.04 10.2 7.7 149 88.4 IV 0.05 6.3 7.6 119 68.1 Average 0.06 8.2 7.7 130 75.2

pHs+0.3 I 0.16 10.8 8.0 164 43.2 II 0.04 5.2 7.9 106 65.1 III 0.04 10.5 8.0 151 64.8 IV 0.06 6.3 7.9 125 57.9 Average 0.08 8.3 7.9 139 57.2

The effect of silica was shown to be beneficial for copper release in the empirical model

presented previously. This was also observed and significant in the empirical model developed

in the previous study without inhibitor addition shown in Equation (3-2) (Taylor et al. 2005).

82

22.02

10.024

86.273.072.0 )(28.0 −−− ××××+= SiOSOpHAlkTempCu (3-2) where: Cu = total copper concentration, mg/L

Temperature = oC Alkalinity = mg/L as CaCO3 pH = -Log [H+] Sulfate = mg/L Silica = mg/L as SiO2

The effects of silica tend to be confounded with the negative effects of alkalinity because

of the high levels of each associated with the blends enriched with groundwater. The high

alkalinity observed in Phase I tends to have a greater effect on the copper release than the high

silica concentration seen in Phase I. This is seen in Table 3-10 with the higher copper

concentrations observed during Phase I. Alkalinity was also high in Phase III due to the same

blend as Phase I. However, copper concentrations in Phase III tended to be lower than Phase I.

The lowest alkalinity was observed in Phase II with the lowest percentage of groundwater.

For the low OP dose PDS, Phase I copper release is significantly higher than the other

phases. However, Phases II, III, and IV all experienced about the same copper release. This

suggests the high alkalinity in Phase I was able to affect the copper release with a low dose of

inhibitor, but the copper release was maintained to be very stable in other phases. Even in Phase

I, the copper release did not exceed the action level for the low OP dose PDS.

The medium OP dose PDS also saw its highest copper release in Phase I, but not as

significant. The copper release was lower for the medium dose than with the low dose and was

better able to maintain a consistent copper concentration, regardless of water quality. Similar

results were seen with the high OP dose PDS, except even lower copper release was maintained.

83

The control PDSs had higher concentrations than any of the OP treated PDSs and the

PDS at pHs was higher than the PDS at pHs+0.3. The pHs+0.3 PDS saw its the highest

concentration in Phase I, like was seen with the OP PDSs. The pHs PDS saw the opposite, with

the lowest concentrations seen in Phase I. This is likely due to the lower pH and higher

alkalinity observed in Phase III and IV compared with Phase I.

Overall, the copper release was still able to be maintained below the action level with the

addition of orthophosphate inhibitor, even in the presence of high alkalinity and the slight

depression of pH. The elevated pH treatment could not always mitigate the copper

concentrations below the action level and in the presence of high alkalinity could lead to

decreased carrying capacity of pipes due to calcium carbonate scale build up.

Surface Characterization

Copper coupons were exposed to the orthophosphate corrosion inhibitor at the medium

dose and the pHs and pHs+0.3 treatment for the duration of each phase. XPS analysis on each of

the coupons was completed after incubation to determine the elements and their states that had

formed on the surface of the coupon. A survey scan revealed the elements present. A high

resolution scan of each element was then deconvoluted to determine the identity and percentage

of the compounds associated with that element.

XPS analysis revealed Cu(OH)2, CuO, Cu2O and Cu(II) scales on these coupons with

Cu(OH)2 and CuO being the most predominant. A summary of the elements found on the

coupons exposed to orthophosphate compared to the pHs and pHs+0.3 exposed coupons is

presented in Table 3-11. Orthophosphate was detected on three of the four coupons exposed to

84

orthophosphate inhibitor, and calcium carbonate was detected on two of those three. Calcium

phosphate was not detected on any of the coupons.

Table 3-11 Elements detected in XPS analysis

Element OP pHs pHs+0.3 (4 total) (4 total) (4 total)

Carbon 4 4 4 Calcium 2 0 1 Copper 4 4 4 Oxygen 4 4 4

Phosphorus 3 0 0

Carbon was detected on all of the four coupons exposed to OP inhibitor. It was always in

the carbonate form. The two calcium coupons were from Phase I and III, appearing in the

CaCO3 form. This is likely due to the high percentage of groundwater in these two phases in the

presence of OP. The OP addition may encourage the formation of CaCO3 films. The copper

detected on the four coupons was in the form Cu2O, CuO, Cu(OH)2, and Cu(II) salts. The Cu(II)

salts did not appear in Phase I. The salts may appear as carbonate compounds or phosphate

compounds since calcium was only found in Phase III. Phosphate may also be absorbed to the

surface as liquid-solid film. Oxygen was detected on all four coupons in the form of O2-, CO32-,

OH-, and PO43-. However, in Phase IV no phosphate was found on the copper coupon.

In comparison, the coupons exposed to pHs had carbon detected in all four phases in the

carbonate form, but calcium was never detected, supporting CaCO3 film formation does not

occur at pHs. Copper was detected on all four coupons, in the forms Cu2O, CuO, Cu(OH)2, and

Cu(II) salts. The salts may be in the carbonate form since carbonate was detected in each phase

85

without calcium. Oxygen appeared on all four coupons in the form of O2-, CO32-, and OH-.

None of the coupons had phosphorus compounds, suggesting the only source of phosphate is

from the inhibitor.

For the pHs+0.3 coupons, carbon appeared in all four phases as carbonate, but calcium

only appeared in Phase IV. Similar to the pHs coupons, copper was detected on all four coupons,

in the forms Cu2O, CuO, Cu(OH)2, and Cu(II) salts. The salts may be in the carbonate form

since carbonate was detected Phase I, II and III without calcium. Oxygen appeared on all four

coupons in the form of O2-, CO32-, and OH-. Again, none of the coupons had phosphorus

compounds, showing the only source of phosphate is from the inhibitor.

Figure 3-8 shows the percent area of the deconvoluted XPS high resolution scan over all

phases for the OP inhibitor coupons compared with the pH control coupons. Cu(OH)2 appears to

be the controlling solid phase for copper solubility in all PDSs so addition of orthophosphate

inhibitor does not appear to affect the controlling solid phase. However, the addition of inhibitor

was shown to reduce copper release and thermodynamic modeling shows Cu3(PO4)2·2H2O

provides a good prediction for dissolved copper release in the PDSs receiving ortho-phosphate

inhibitor addition, as discussed below.

86

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

OP pH

Inhibitor

% A

rea

of C

ompo

und

Cu(II) Cu(OH)2 Cu2O CuO

Figure 3-8 Distribution of copper compounds

Thermodynamic Modeling

Thermodynamic modeling was done to validate the copper controlling solid phase. As

previously discussed, the controlling solid phase for all PDSs appears to be Cu(OH)2 as found

from XPS analysis. However, phosphate was present on the coupons exposed to the

orthophosphate inhibitor. The equilibrium model, assuming Copper (II) species developed for

TBW I (Taylor et al., 2005), was expanded for phosphate, sulfate, chloride and ammonia

complexes. The model is presented in Equation (3-3). This model was used to calculate the total

dissolved copper concentration based on the assumption of a controlling solid phase.

87

][][])([])([

])([])([][][][

][])([])([])([

{][])([])([][][

02

253

243

233

223

23

04

04

422

3232

3

0333

02

2

CuClCuClNHCuNHCu

NHCuNHCuCuNHCuSOCuHPO

POCuHCOOHCuCOOHCuCOCu

CuCOCuHCOOHCuOHCuCuOHCuCuT

++++

+++++

++++

+++++=

+++

+++

+−−−

+−++

(3-3)

The empirical model developed previously suggests chloride and alkalinity contribute to

copper release, while pH and total phosphorus help to mitigate it. The equilibrium model

presented here has species to account for changes in pH (OH), alkalinity (CO3), phosphate (PO4),

sulfate (SO4), chloramine disinfectant (NH3), and chloride (Cl) observed in the system.

From the XPS analysis, it was determined that cupric hydroxide (Cu(OH)2) was the

predominant copper corrosion product present on the copper coupons. Tenorite (CuO), cuprous

oxide (Cu2O), and cupric salts (Cu(II)) were also present on the copper coupons. Cu(OH)2 has

been identified as a metastable intermediate that forms in young copper pipe, and has been used

as the basis for thermodynamic models of copper (Schock, Lytle, and Clement 1995; Xiao 2004).

Cupric hydroxide can age to its dehydrated form, tenorite, CuO, which is less soluble. In the

presence of alkalinity, old copper pipe tends to form the more stable, less soluble cupric

hydroxycarbonate Cu2(OH)2CO3, malachite (Schock, Lytle, and Clement 1995).

Both cupric hydroxide and tenorite were considered as the controlling solid phase during

thermodynamic modeling. Table 3-12 shows the cupric hydroxide thermodynamic model

predicted the dissolved copper concentrations for the pHs and pHs+0.3 PDSs. The tenorite

88

thermodynamic model under-predicted the dissolved copper concentrations significantly for both

control PDSs.

Table 3-12 Thermodynamic modeling calculations for pH control PDSs

Actual Copper Release (mg/L) Modeled Copper Release (mg/L) Phase Treatment Diss Cu Total Cu Cu(OH)2 CuO

I pHs 0.95 1.04 0.83 0.11 pHs+0.3 0.89 1.00 0.57 0.08

II pHs 1.04 1.16 1.02 0.05 pHs+0.3 0.73 0.82 1.10 0.05

III pHs 1.32 1.48 2.16 0.13 pHs+0.3 0.79 0.85 0.97 0.05

IV pHs 1.14 1.41 0.90 0.13 pHs+0.3 0.83 0.93 0.52 0.07

Average pHs 1.11 1.27 1.13 0.10 pHs+0.3 0.81 0.90 0.75 0.06

For the pHs+0.3 control, actual dissolved and total copper release data were compared to

predicted dissolved copper release from tenorite, CuO, and cupric hydroxide, Cu(OH)2, models.

The predicted copper release from the Cu(OH)2 model varied, appearing over and under than the

actual copper release observed for pHs+0.3. The tenorite model under-predicted the dissolved

copper concentrations by about 0.7 mg/L as Cu.

For the OP inhibitor PDSs, equilibrium calculations considered cupric hydroxide

(Cu(OH)2), tenorite (CuO), cupric phosphate dihydride (Cu3(PO4)2·2H2O) and cupric phosphate

(Cu3(PO4)2) as the controlling solid phase and are presented in Table 3-13. These compounds

are considered because cupric hydroxide and tenorite were the most predominant scales found on

the coupon. Additionally, thermodynamic data was available for cupric phosphate dryhydride

89

and cupric phosphate to account for the addition of orthophosphate inhibitor corresponding to the

reduction in copper release compared with the pH control PDSs.

Table 3-13 Thermodynamic modeling calculations for OP PDSs

Actual Copper Release (mg/L) Modeled Copper Release (mg/L) Phase OP dose Diss Cu Total Cu Cu(OH)2 CuO Cu3(PO4)2·2H2O Cu3(PO4)2

I low 0.46 0.50 0.51 0.07 0.90 3.96 med 0.38 0.41 0.54 0.08 0.53 2.34 high 0.26 0.30 0.58 0.09 0.35 1.54

II low 0.33 0.39 1.07 0.06 0.62 2.65 med 0.28 0.35 1.15 0.06 0.47 1.99 high 0.21 0.24 1.40 0.07 0.27 1.16

III low 0.36 0.40 1.05 0.06 1.03 4.42 med 0.29 0.35 1.07 0.06 0.68 2.93 high 0.19 0.21 1.23 0.07 0.44 1.87

IV low 0.34 0.40 0.59 0.08 0.64 2.78 med 0.32 0.36 0.61 0.09 0.43 1.86 high 0.21 0.23 0.68 0.10 0.26 1.14

Average low 0.38 0.42 0.74 0.07 0.80 3.48 med 0.32 0.37 0.78 0.07 0.53 2.30 high 0.21 0.25 0.88 0.08 0.33 1.43

It can be seen that cupric hydroxide and cupric phosphate dyhydride over-predict the

dissolved copper concentration observed in the PDSs. If the equilibrium constant for cupric

hydroxide were elevated to match the observed copper release in the pHs+0.3 PDS, then the cupric

phosphate dihydride model would explain the decrease in copper release with orthophosphate

inhibitor addition. This trend of decreasing copper with increasing OP dose can be seen in Table

3-13.

Cu3(PO4)2·2H2O models provided the best prediction of copper release, over predicting

by no more than about 0.3 mg/L Cu. The CuO model consistently under-predicted copper

release by about 0.25 mg/L Cu. In contrast, the values predicted by a cupric phosphate

90

dihydride, Cu3(PO4)2·2H2O would be a closer match to the data. A review of the literature

confirmed the lack of information identifying the form of copper orthophosphate solids in water

distribution systems. However, the forms Cu3(PO4)2·2H2O and Cu3(PO4)2, were modeled since

thermodynamic data was available for these copper forms (Shock, Lytle, and Clement 1995).

Copper Coupons

Copper coupons were left to incubate for 14 months after removal from the PDSs in

100 mL of distribution system water. These coupons were from Phase III of operations and

shown in Figure 3-9. As seen in the picture, a blue-green scale formed on the copper exposed to

the medium and high dose orthophosphate inhibitor, but not on the pHs+0.3 control PDS coupon.

The pHs coupon similarly showed no blue green scale. The scale also appears to be thicker on the

high dosed coupon.

Figure 3-9 Scale on coupons incubating for 14 months

91

The dissolved copper concentration was measured in the water these coupons had been

incubating in. Measurements showed the low, medium and high dosed coupons to have

dissolved copper concentrations of 0.41 mg/L, 0.26 mg/L, and 0.23 mg/L, respectively. These

still show decreasing concentration with increasing dose as was observed over the course of the

study and were also lower than the controls. The coupon exposed to pHs had a dissolved copper

concentration of 0.76 mg/L and the coupons exposed to pHs+0.3 had a concentration of 0.74

mg/L. The OP treated coupons had similar concentrations to those observed during the study

and have had time to reach near equilibrium. This suggests the controlling solid for the OP

treated coupons is different from the controlling solid that forms when inhibitor is not present.

Conclusions

• Empirical modeling of water quality resulted in a model with an R2 of 0.71. The

model showed total phosphorus, silica, and pH to reduce copper release while

alkalinity and chloride increase copper release. This suggests addition of

orthophosphate inhibitor will reduce copper release, as will elevating pH. However,

elevating pH could cause scaling problems in the distribution system.

• Addition of orthophosphate inhibitor reduced copper release and maintained levels

below the action level at the doses evaluated, regardless of blend. This was

improvement over pH elevation alone because copper remained low, even in the

presence of high alkalinity, compared to the operation at pHs.

92

• XPS analysis showed Cu(OH)2 to be the controlling solid phase for copper solubility

in all PDSs, regardless of blend or corrosion control strategy. Addition of

orthophosphate inhibitor does not appear to affect the controlling solid phase.

However, phosphate forms were detected on the copper coupons exposed to OP

inhibitor that were not detected in either of the pHs or pHs+0.3 treatments.

• Thermodynamic modeling suggested Cu3(PO4)2·2H2O as a controlling solid phase

and followed the trend of dissolved copper release decreasing with increasing dose

better than the Cu(OH)2 model. Also, visual inspection of equilibrated coupons

showed different scale formation for the OP treated coupons than the pH control

coupons, suggesting a different controlling solid for copper surfaces exposed to OP

inhibitor.

93

References

APHA, AWWA, and WEF (American Public Health Association, American Water Works Association, and Water Environmental Association). 1998. Standard Methods for the Examination of Water and Wastewater. 20th ed. Washington, D.C.: APHA.

Becker, A. 2002. The effect of corrosion inhibitors in drinking water installations of copper. Materials and Corrosion, 53:560-567.

Dodrill, D. M. and M. Edwards. 1995. Corrosion control on the basis of utility experience. Journal American Water Works Association, 87(7):87-99.

Edwards, M., L. Hidmi, and D. Gladwell. 2003. Phosphate inhibition of soluble copper corrosion by-product release. Corrosion Science, 44:1057-1071.

Edwards, M., L.S. McNeill, and T.R. Holm. 2001. Role of phosphate inhibitors in mitigating lead and copper corrosion. In Distribution Systems. Denver, Colo.: AWWARF.

Edwards, M., S. Jacobs, and D. Dodrill. 1999. Desktop guidance for mitigating Pb and Cu Corrosion By-Products. Journal American Water Works Association, 91(5):66.

Edwards, M., F.F. Ferguson, and S.H. Reiber. 1993. On the pitting corrosion of copper. Presented at the Annual Conference of AWWA, San Antonio, Texas.

Federal Register. 1991. Drinking Water Regulations; Lead and Copper Rule. 56 FR 26460

Pinto, J.A., A.S. McAnally, and J.R.V. Flora. 1997. Evaluation of lead and copper corrosion control techniques. Journal of Environmental Science and Health, A32(1):31-53.

Reiber, S. 1989. Copper plumbing surfaces: an electrochemical study. Journal American Water Works Association, 81(7):114-122.

Schock, M.R., and J.C. Fox. 2001. Solving copper corrosion problems while maintaining lead control in high alkalinity water using orthophosphate. Presented at the AWWA Annual Conference, Cleveland, Ohio, August 30, 2001.

Schock, M.R., D.A. Lytle, and J.A. Clement. 1995. Effect of pH, DIC, Orthophosphate and Sulfate on Drinking Water Cuprosolvency. Risk Management Research Report EPA/600/R-95/085. Cincinnati, Ohio: U.S. EPA.

Taylor, J.S., J.D. Dietz, A.A. Randall, S.K. Hong, C.D. Norris, L.A. Mulford, J.M. Arevalo, S. Imran, M. LePuil, S. Liu, I. Mutoti, J. Tang, W. Xiao, C. Cullen, R. Heaviside, A. Mehta,

94

M. Patel, F. Vasquez, and D. Webb. 2005. Effects of Blending on Distribution System Water Quality. Denver, Colo.: AwwaRF and Tampa Bay Water.

Xiao, W., 2004. Effect of Source Water Blending on Copper Release in Pipe Distribution System: Thermodynamic and Empirical Models. Ph.D. diss., University of Central Florida, Orlando, Fla.

Zhang, X., S.O. Pehkonen, N. Kocherginski, and G.A. Ellis. 2002. Copper corrosion in mildly alkaline water with the disinfectant monochloramine. Corrosion Science, 44:2507-2528.

95

CHAPTER 4 EFFECTS OF ORTHOPHOSPHATE CORROSION INHIBITOR ON LEAD

IN BLENDED WATER QUALITY ENVIRONMENTS

Abstract

This study evaluated the effects of orthophosphate (OP) inhibitor addition on lead

corrosion on coupons exposed to different blends of groundwater, surface water, and desalinated

seawater. The effectiveness of OP inhibitor addition on lead release was analyzed by statistical

comparison between OP treated and untreated pilot distribution systems. Four different doses of

OP inhibitor, ranging from zero (control) to 2 mg/L as P, were investigated and non-linear

empirical models were developed to predict lead release from the water quality and OP doses.

Surface characterization evaluations were conducted using X-ray Photoelectron Spectroscopy

(XPS) analyses for each lead coupon tested. A theoretical thermodynamic model was developed

to predict lead concentration and used to validate the controlling solid phases determined by

XPS. OP inhibitor addition was found to reduce lead release for the OP dosages evaluated for

the water blends evaluated compared to pH adjustment alone. Empirical models showed

increased total phosphorus and pH reduced lead release while increased alkalinity, chloride, and

temperature contributed to lead release. Thermodynamic modeling suggested that

hydroxypyromorphite is the controlling solid that forms on lead surfaces, regardless of blend,

when OP inhibitor is added for corrosion control.

96

Introduction

With increasing water demands and more stringent drinking water regulations, many

utilities are turning to desalinated sources to supplement their surface and groundwater supplies.

Tampa Bay Water (TBW) and the University of Central Florida (UCF) studied the effects of

blending multiple alternative source waters on distribution system water quality (Taylor et al.

2005). This study further evaluates the addition of orthophosphate (OP) corrosion inhibitor to

the blended source waters and the effects on lead corrosion.

Lead levels in drinking water are regulated through the Lead and Copper Rule (LCR) and

limited to an action level of 0.015 mg/L at the 90th percentile of household kitchen taps sampled

by voluntary participant homeowners (Federal Register 1991). Lead in household tap water

originates from several sources including lead service lines, lead-tin solder, and brass fixtures

and faucets in bathrooms and kitchens (Singley et al. 1985; Lee, Becker, and Collins 1989;

Reiber 1991). Lead levels are a function of the water quality, plumbing materials, contact time,

pipe geometry, water temperature, and age of materials (Boffardi 1995; Boffardi 1990).

Chemical treatment for control of lead corrosion includes pH adjustment, carbonate addition,

calcium carbonate deposition and inhibitors (Boffardi 1995).

Numerous studies have found that orthophosphate inhibitor addition can reduce lead

concentrations in water under the correct pH conditions (Schock 1980; Hunt and Creasey 1980;

Sheiham and Jackson 1981; Gardels and Schock 1981; Gregory and Jackson 1983; Nriagu 1974).

Orthophosphate chemicals form passivating films on metallic surface anodic sites to suppress

electrochemical reactions. Zinc, lead, calcium or iron deposition can further enhance the

protection by forming films over cathodic sites (AwwaRF 1996). Orthophosphate can combine

97

with lead plumbing materials to form compounds that do not have a strong tendency to dissolve

into water, so lead concentrations tend to remain low in drinking water (AwwaRF 2001;

AwwaRF 1996).

Edwards and McNeill (2002) found that dosing orthophosphate inhibitor reduces lead

release by 70%, with exception to new pipes. New pipes had increased lead, suggested to be a

result of a lower increase in pH during stagnation. Particulate lead was the dominant corrosion

product form observed. Total lead levels decreased with pipeline aging.

Pinto, McAnally, and Flora (1997) showed addition of orthophosphate corrosion inhibitor

reduced lead levels as compared with zinc or blended orthophosphate inhibitors. Inhibitor and

alkalinity addition in low alkalinity waters reduced lead levels further than alkalinity adjustment

alone. Phosphate addition was recommended because of its effectiveness at low doses.

Alkalinity and pH adjustments are viewed as an alternative, but may have adverse secondary

affects on water system scaling.

Washington D.C.’s conversion to monochloromines caused high levels of lead release in

the distribution system. A study was done with orthophosphate addition and the scale formed on

lead service lines was monitored. It was found that prior to orthophosphate treatment, lead(IV)

oxide was dissolving due to the lower ORP experienced with the monochloramine residual

compared with the free chlorine. Use of orthophosphate was found to be the optimal inhibitor

choice and took eight months to form a passivating scale and reduce lead levels of 150 ppb down

below the action level (Giani, Keefer, and Donnelly 2005).

Dodrill and Edwards (1995) conducted a survey of about 360 utilities to examine their

strategies in controlling lead and copper release in response to the LCR. It was observed that

98

without phosphate inhibitors, higher alkalinity resulted in lower lead release. (This is contrary to

observations in the system used in this study (Taylor et al. 2005)). At low alkalinity, using

inhibitors reduced lead release compared to no inhibitors. The survey revealed that utilities with

the low alkalinity and pH below 7.4 benefited from using orthophosphate but not polyphosphate.

Moreover, polyphosphate increased lead release at higher alkalinity levels.

Hozalski, Esbri-Amador, and Chen (2005) studied the benefit of inhibitors in controlling

lead release from lead pipe. They used orthophosphate, polyphosphate, blended

ortho-polyphosphate (BOP), and stannous chloride (SnCl2). All inhibitors performed better than

the no-action alternative at reducing lead release with various degrees. The ranking of inhibitors

based on total lead concentrations, from lowest to highest, in the tested pipe loops was as

follows: ortho-P < SnCl2 < BOP < poly-P < control. The authors also observed that phosphate-

based inhibitors resulted in increased biological growth compared to the stannous chloride and

control pipes.

This study evaluated the effects of orthophosphate inhibitor addition to blended treated

surface, ground, and seawater sources of varying blend percentages. The effects of water quality

were evaluated and a model predicting total lead release using water quality and total phosphorus

concentrations was developed. XPS analysis of 50/50 lead/tin coupons was evaluated for solid

phase surfaces present on the coupon, and thermodynamic modeling was performed using

collected experimental information to gain insight into the mechanism of inhibitor control of lead

release.

99

Experimental Methods

Experimental Design

Experimentation was conducted with the use of pilot distribution systems (PDSs) built

from actual pipelines extracted from TBW member governments distribution systesms

(Hillsborough County, Fla.; Pasco County, Fla.; Pinellas County, Fla.; City of New Port Richey,

Fla.; City of St. Petersburg, Fla.; and City of Tampa, Fla.). Details regarding the PDS and prior

study results are reported elsewhere (Taylor et al. 2005). Each PDS runs in parallel with

segments of PVC, lined cast iron, unlined cast iron, and galvanized steel pipes that were placed

sequentially to simulate actual distribution systems. The materials and their respective diameter

and length are shown in Table 4-1. Each PDS was fed blends of groundwater, surface water, and

desalinated seawater along with different types and doses of corrosion inhibitor.

Table 4-1 Pipe materials in PDSs

Pipe Material Length (ft)

Diameter (in)

PVC 20 6 Lined Cast Iron 20 6

Unlined Cast Iron 12 6 Galvanized Steel 40 2

Images of materials as well as the full PDSs used to represent full scale distribution

systems are shown in Figure 4-1. The pipe materials used, shown in the left image, are

galvanized steel, PVC, lined cast iron and unlined cast iron from left to right. The image on the

right shows the PDSs with the materials connected in series from influent (closest to viewer) to

100

effluent (furthest from viewer) of the system. The pipe materials were connected in sequence of

PVC, lined cast iron, unlined cast iron, then galvanized steel.

Figure 4-1 Pipe materials (left) and parallel PDSs (right)

The PDSs were fed blends of conventionally treated groundwater (GW), enhanced

coagulation-sedimentation-filtration (CSF) treated surface water (SW), and desalinated seawater

by reverse osmosis (RO). The description of the three finished source waters is presented in

Table 4-2.

101

Table 4-2 Finished source water descriptions

Water Source System Description

GW Groundwater Ground water source. Treatment by aeration, disinfection by free chlorine with a residual of 5 mg/L after a 5 minute contact time. 5.0 mg/L chloramine residual.

SW Surface water

TBW treatment plant: Treatment by ferric sulfate coagulation, flocculation, settling, filtration, disinfection by ozonation and chloramination. Project site: adjustment of chloramine residual to 5.0 mg/L chloramine residual.

RO Groundwater Treatment by membrane reverse osmosis, aeration, disinfection by free chlorine with a residual of 5 mg/L after a 5 minute contact time. 5.0 mg/L chloramine residual.

The GW unit used raw well water from the Cypress Creek well field owned by TBW.

The GW was treated with aeration, disinfection, and pH stabilization. Aeration was achieved in

the GW by pumping the raw water to the top of the finished water tank through a spray nozzle.

Sodium hypochlorite was used for primary disinfection and was dosed to provide a 5 mg/L

residual after a 5 minute contact time. Afterwards, ammonium chloride was added to produce a

5 mg/L monochloramine residual. Ammonia was added in the form of NH4Cl at a 5:1 ratio. The

Cl2:NH3 ratio was initially 4:1 to protect against DBP formation. This ratio was increased to 5:1

in after 6 months of operation to reduce free ammonia.

SW was treated at the TBW Regional Surface Water Treatment Facility by enhanced

coagulation, ozonation, biologically activated carbon (BAC) filtration, aeration, and

chloramination. The SW was hauled weekly to the field facility for use and temporarily stored in

two 7000 gallon storage tanks before being transferred to the SW finished water tank. In the SW

finished tank, the chloramine residual was adjusted to 5 mg/L as Cl2.

102

The RO pilot plant was housed in a trailer at the testing facility and utilized raw

groundwater for the feed stream. The RO treatment pilot system required the addition of TDS,

calcium and alkalinity to the RO permeate to represent the finished water produced by the TBW

Regional Desalination Facility. RO pretreatment consisted of 2.7 mg/L antiscalant addition

(Hypersperse MDC700TM, GE Water, Minnetonka, Minn.) followed by 5-micron cartridge

filtration. The RO membrane unit was operated at 72-73% recovery, producing 9.3 gpm

permeate flow, which was aerated by a 10-inch diameter aeration tower filled with tripack plastic

packing. After aeration, 50 mg/L of sea salt was added to the aerated permeate stream to

simulate the TBW desalination process. Calcium chloride and sodium bicarbonate were also

added to meet the calcium and alkalinity specifications. The finished was stabilized with sodium

hydroxide to 0.1 to 0.3 pH units above pHs.

The effects of water quality were evaluated by varying the blend quarterly, while

seasonal effects were evaluated by maintaining the same blend in the summer and winter. The

quarterly phases and percentages of the blends are shown below in Table 4-3. The average water

quality of each of the source waters in each phase is shown in Table 4-4. The effects of season

are seen in the temperature as well as rainy and dry season effects on the surface water between

Phases I and III. The blends with a high percentage of groundwater in Phases I and III are

characterized by high alkalinity and pH. Phase II had the highest percentage of surface water

and is characterized by high sulfate concentrations. Phase IV has average water quality

parameters due to the equal percentage of GW and SW.

103

Table 4-3 Blend percentages for each phase

Phase Quarter % GW % SW % RO I Feb-May 2006 62 27 11 II May-Aug 2006 27 62 11 III Aug-Nov 2006 62 27 11 IV Nov 2006-Feb 2007 40 40 20

Table 4-4 Average water quality by phase

Phase pH Alkalinity (mg/L CaCO3)

Chloride (mg/L)

Sulfate (mg/L)

Temperature (°C)

I 8.0 161 45 62 21 II 7.9 104 67 103 26 III 8.0 150 68 66 26 IV 7.9 123 59 76 21

The feed rate of the blend into each PDS was maintained to achieve a two-day hydraulic

residence time (HRT). Pumps maintained the blend flow as well as the inhibitor addition into

each PDS. The PDSs each were fed different inhibitor types and doses. The inhibitors were

dosed to the PDSs at three different levels, categorized as low dose, medium dose, and high dose.

Orthophosphate (OP) was maintained at a target dose of 0.5 mg/L as P for the low dose, 1.0

mg/L as P for the medium dose, and 2.0 mg/L as P for the high dose. Control PDSs were not fed

any chemical inhibitor; one was maintained at pHs and a second was treated with elevated pH,

maintained at pHs+0.3. The PDS at pHs+0.3 was maintained at a positive LSI to assess the

affect of elevated pH treatment as a means of lead release control.

104

Orthophosphate Inhibitor

The orthophosphate inhibitor used in this study is Inhibit-All WSF-36 (SPER Chemical

Corporation, Clearwater, Fla.). It is made of monosodium orthophosphate blended into 17

megohm purified water at a concentration of 36%. It is a clear, slightly viscous liquid with a

bulk density of 11.25 lbs/gal. The specific gravity is 1.35 the pH of 1% solution is 5.1 to 5.4.

Table 2-5 shows a summary of the properties of the OP inhibitor along with manufacture

recommendations for use.

Table 4-5 Orthophosphate inhibitor properties

Property Value Manufacturer SPER Chemical Corp. Product Inhibit-All WSF-36 Percent Orthophosphate 36% Bulk Density 11.25 lbs/gal Specific Gravity 1.35 pH 1% solution 5.1 – 5.4 Recommended dose (low-med-high) 1.25-2.75-4.25 mg/L as Product Used dose (low-med-high) 0.5-1.0-2.0 mg/L as P Recommended pH range 7.4 – 7.6 (optimum), 6.8 – 7.8

Data Collection

Water quality parameters were collected and analyzed weekly in the influent and effluent

of the PDSs. The water quality parameters monitored and methods used for samples transported

to the UCF laboratory for analysis are shown in Table 4-6. Table 4-7 shows the water quality

parameters monitored and methods used for the parameters analyzed in the field laboratory at the

testing facility. The method detection limit (MDL) is also shown. Methods noted as SM are

from APHA, AWWA, and WEF (1998).

105

Table 4-6 Water quality parameters and methods performed at University Laboratory

Parameter Method Reference Method Description MDL

Aluminum SM 3120B ICP Method 0.001 mg/L Bicarbonate SM 2320B Titration Method 5 mg/L Calcium SM 3120B ICP Method 0.1 mg/L

Chloride SM 4110 Ion Chromatography with Chemical Suppression of Eluent Conductivity 0.1 mg/L

Color SM 2120A Or Hach 8025 Cobalt-Platinate Method (with spec) 1 CPU

Conductivity SM 2510B Laboratory Method 1 μmho/cm Copper SM 3120B ICP Method 0.001 mg/L Iron SM 3120B ICP Method 0.001 mg/L Lead SM 3120B ICP Method 0.001 mg/L Magnesium SM 3120B ICP Method 0.1 mg/L Nitrogen (NH3,TKN) SM 4500-Norg Macro-Kjeldahl Method 0.1 mg/L

NPDOC SM 5310C Persulfate-UV Oxidation Method 0.1 mg C/L pH SM 4500-H+ B Electrometric Method ± 0.01 pH units Phosphorus SM 3120B ICP Method 0.001 mg/L Silica SM 3120B ICP Method 0.001 mg/L Sodium SM 3120B ICP Method 0.1 mg/L Solids (TDS) SM 1030E Estimation of TDS by major ion sum 1 mg/L

Sulfate SM 4110 Ion Chromatography with Chemical Suppression of Eluent Conductivity 0.1 mg/L

Turbidity SM 2130B Nephelometric Method 0.01 NTU UV-254 SM 5910 UV Absorption at 254 nm 0.0001 cm-1 Zinc SM 3120B ICP Method 0.001 mg/L

106

Table 4-7 Water quality parameters and methods performed at Field Laboratory

Parameter Method Reference Method Description MDL Alkalinity SM 2320 B Titration 5 ppm Ammonia-N SM 4500-NH3 C Membrane Probe Method 0.1 ppm Chloride SM 4500-Cl- B Argentometric Titration 1 mg/L Chlorine, free SM 4500-Cl G or Hach 8021 DPD colorimetric 0.1 ppm Chlorine, total SM 4500-Cl-G or Hach 8167 DPD colorimetric 0.1 ppm Color, apparent SM 2120 B Visual Comparison

(by spectrometer) 1 CPU

Conductivity SM 2510 B Conductivity Bridge 1 μmho/cm Hardness (total, calcium) SM 2340 C EDTA Titration 5 mg/L

Nitrate Hach 8192 Cadmium reduction 0.1 mg/L Nitrite Hach 8507 Diazotization 0.1 mg/L Oxygen, Dissolved (DO) SM 4500-O G Membrane probe 0.1 mg/L

pH SM 4500-H+ B Electrometric ± 0.01 pH units

Phosphate-P (Reactive) SM 4500-P E. or Hach 8048 Ascorbic Acid Method 0.1 mg/L

Silica, SiO2 (reactive) SM 4500-SiO2 or Hach 8185 Molybdosilicate Method 0.1 mg/L

as SiO2 Temperature SM 2550 B Direct reading 0 deg C Turbidity SM 2130 B Nephelometric 0.01 NTU UV254 SM 5910 A UV spectrometry 0.0001 cm-1

Portions of the flow from each PDS were fed to a corrosion loop consisting of 30 feet of

5/8 inch copper tubing with one lead/tin coupon to represent solder. Each loop holds

approximately 1.8 L of water. The copper tubes were flushed every morning with 2 gallons of

the PDS water. Weekly samples were collected after a six-hour stagnation period in order to

simulated tap monitoring as described in the Lead and Copper Rule. The corrosion shed and

housed corrosion loops are show in Figure 4-2.

107

Figure 4-2 Corrosion shed (left) and corrosion loops (right)

50/50 lead-tin coupons were placed in cradles that received flow in parallel with each

PDS. The coupons were evaluated for surface characteristics after incubation during each phase.

X-ray Photoelectron Spectroscopy (XPS) analysis was then performed on each type of coupon to

identify chemical components on the outer layer of the corrosion surface. A survey scan can

reveal the presence of elements, whereas a high resolution scan of those elements found on the

outer layer can show the chemical states, providing detailed surface characterization information.

Results and Discussion

Dose Maintenance

Three of the PDSs were treated with orthophosphate corrosion inhibitor at a low,

medium, and high dose. These doses were targeted to maintain at 0.5 mg/L as P, 1.0 mg/L as P,

and 2.0 mg/L as P, respectively. The average orthophosphate inhibitor dose for the course of the

study in each PDS is shown in Figure 4-3. Error bars represent the minimum and maximum

108

observations. The low dose of orthophosphate averaged 0.51 mg/L as P, the medium dose

averaged 0.94 mg/L as P, and the high dose averaged 1.83 mg/L as P.

1.83

0.94

0.51

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Low Med High

OP Inhibitor Dose

Tota

l P (m

g/L

as P

)

Figure 4-3 Orthophosphate inhibitor dosing

The orthophosphate inhibitor is added as phosphoric acid, so it has an effect of lowering

the pH is those PDSs treated with inhibitor. The pH maintained in each of these PDS is shown in

Figure 4-4 with the error bars representing the minimum and maximum observations. The

difference in pH maintained in each of the PDSs is significantly different with the exception of

the low orthophosphate dosed PDS being the same as the elevated pH PDS at pHs+0.3.

109

7.94

7.71

7.867.907.92

6.8

7.0

7.2

7.4

7.6

7.8

8.0

8.2

8.4

8.6

OP-low OP-med OP-high pHs pHs+0.3

pH

Figure 4-4 pH for OP PDSs and controls

Empirical Modeling

A non-linear empirical model was developed using the water quality data and phosphorus

dose for the three PDSs with orthophosphate inhibitor addition and the PDSs maintained at pHs

and pHs+0.3. Water quality parameters found to be significant using ANOVA procedures at 95%

confidence were retained as variables in the model. All variables retained p-values less than

0.05. The range of the water quality parameters included in the model are shown in Table 4-8.

110

Table 4-8 Water quality range in model development

Parameter Minimum Maximum Total Phosphorus, mg/L as P 0.01 2.69 pH 7.4 8.4 Alkalinity, mg/L as CaCO3 84 170 Chloride 38 123 Temperature, °C 10.4 29.7

Equation (4-1) presents a non-linear least squares regression model developed using

results of the study. The water quality parameters pH, alkalinity, chloride, and temperature

remained significant as well as the total phosphorus, which represents the inhibitor addition.

Total phosphorus and pH mitigated lead release, as shown by the negative exponent on the

variable in the model. However, alkalinity, chloride, and temperature are shown to increase lead

release for the conditions experienced during this testing.

25168.1853.2064.12435.0 144.1 −−− ××××= TClAlkpHTPPbTotal (4-1)where: Total Pb = total lead, mg/L

TP = total phosphorus, mg/L as P pH = -log[H+] Alk = alkalinity, mg/L as CaCO3 Cl = chloride, mg/L T = temperature, °C

The model fits the data reasonably well with an R2 of 0.62. The predicted lead release by

the model compared with the total lead release observations is shown graphically in Figure 4-5.

It can be seen that most observations from the PDSs treated with orthophosphate inhibitor were

below the detection limit for lead measurement of 0.001 mg/L. The observations at higher

111

concentrations are from the pHs and pHs+0.3 control PDSs. This further shows the benefits of

addition of the corrosion inhibitor to reduce lead release, often to levels below detection.

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07

Actual Total Pb (mg/L)

Pred

icte

d To

tal P

b (m

g/L)

OP-low OP-med OP-high pHs pHs+0.3

Figure 4-5 Predicted vs. actual for total lead empirical model

The model fits the data, but tends to under-predict the very high concentrations. All

observations for the OP treated PDSs were below the action level of 0.015 mg/L and are well

described with the model. Observations from the pHs PDS tend to be much higher than the

action level and are under-predicted by the model. Therefore the model is best suited for

predicting cases where treatment is used to maintain lead release near and below regulated

concentrations.

112

Figure 4-6 presents the 90th percentile of lead release at the inhibitor doses evaluated as

compared to the 90th percentile of the model prediction. The model predicts decreasing lead

release with increasing inhibitor dose. The lowest inhibitor dose has the highest lead release, but

the highest inhibitor dose appears to have the same lead release as the medium dose, due to so

many observations below detection. Paired t-tests show no significant difference between the

medium and high dose PDSs lead release at a 95% confidence level. The other PDSs have a

significant difference between lead release observations. Also, both the model predictions and

actual observations for the lead release are always below the 90th percentile action level of 0.015

mg/L for all OP treated PDSs.

0.000

0.002

0.004

0.006

0.008

0.010

0.012

I-OP lo

w

I-OP m

ed

I-OP hi

gh

II-OP lo

w

II-OP m

ed

II-OP hi

gh

III-O

P low

III-O

P med

III-O

P high

IV-O

P low

IV-O

P med

IV-O

P high

Phase-Treatment

Tota

l Pb

90th

per

cent

iles (

mg/

L)

Actual Predicted

Figure 4-6 Actual vs. predicted for empirical model by phase and PDS, 90th percentiles

113

Orthophosphate Inhibitor Performance

A summary of the lead release from the orthophosphate treated PDSs and the pH control

PDSs is shown in Table 4-9. The total lead release remained below the action level of 0.015

mg/L for the orthophosphate treated PDSs except one observation for each the low and high dose

in Phase IV. This one observation does not exceed the 90th percentile for action level. The

pHs+0.3 PDS exceeded 0.015 mg/L at the 90th percentile in Phase II and the pHs PDS exceeded

the action level in Phases II and III.

Table 4-9 Lead release summary

Dissolved Pb (mg/L) Total Pb (mg/L) Phase Treatment Avg Min Max Avg Min Max 90th percentile

I OP-low 0.001 0.001 0.002 0.004 0.001 0.015 0.010 OP-med 0.001 0.001 0.002 0.001 0.001 0.003 0.002 OP-high 0.001 0.001 0.001 0.002 0.001 0.005 0.004 pHs 0.003 0.001 0.008 0.006 0.001 0.016 0.015 pHs+0.3 0.001 0.001 0.006 0.004 0.001 0.015 0.011

II OP-low 0.001 0.001 0.001 0.005 0.001 0.009 0.009 OP-med 0.001 0.001 0.001 0.001 0.001 0.001 0.001 OP-high 0.001 0.001 0.001 0.001 0.001 0.004 0.001 pHs 0.005 0.001 0.010 0.012 0.001 0.024 0.021 pHs+0.3 0.004 0.001 0.011 0.008 0.001 0.023 0.018

III OP-low 0.001 0.001 0.002 0.004 0.001 0.009 0.007 OP-med 0.001 0.001 0.001 0.001 0.001 0.001 0.001 OP-high 0.001 0.001 0.001 0.001 0.001 0.001 0.001 pHs 0.017 0.003 0.035 0.032 0.005 0.064 0.058 pHs+0.3 0.005 0.001 0.022 0.010 0.001 0.036 0.015

IV OP-low 0.001 0.001 0.004 0.004 0.001 0.017 0.005 OP-med 0.001 0.001 0.001 0.001 0.001 0.001 0.001 OP-high 0.001 0.001 0.001 0.003 0.001 0.026 0.001 pHs 0.003 0.001 0.007 0.004 0.001 0.009 0.008 pHs+0.3 0.002 0.001 0.005 0.003 0.001 0.014 0.004

114

As a result of the low lead release levels observed in this study, the average values

calculated are skewed. This is due to many lead observations being below the measurement

detection limit of 0.001 mg/L. Table 4-10 shows the number of observations for each PDS and

phase that were below detection as well as the total observations for that phase. This suggests

the medium and high dose PDSs performed the best because of the high percentage of

observations below detection. The pHs+0.3 PDS performed about as well as the low

orthophosphate dose PDS, while the pHs PDS had very few observations below detection.

Table 4-10 Observations below detection

Phase OP-low OP-med OP-high pHs pHs+0.3 Total Obs. I 5 11 10 2 7 14 II 1 13 12 2 1 13 III 0 12 13 0 1 13 IV 2 12 10 3 2 12

Total 8 48 45 7 11 52

Table 4-10 shows the average water quality for each PDS and each phase for the water

quality parameters of importance in the empirical model presented previously. Phase I and III

are high in alkalinity due to the high percentage of groundwater. Phase I and IV have low

temperature while Phase II and III have higher temperature due to the season.

115

Table 4-11 Average influent water quality by PDS and phase

Total Phosphorus pH Alkalinity Chloride Temperature Treatment Phase (mg/L as P) (std. units) (mg/L as CaCO3) (mg/L) (deg. C)

OP-low I 0.49 8.0 164 43.2 21.0 II 0.55 7.8 106 67.2 25.6 III 0.54 8.0 152 63.6 25.6 IV 0.47 7.8 125 56.6 21.2 Average 0.51 7.9 143 54.7 22.9

OP-med I 1.04 8.0 164 43.1 21.0 II 0.90 7.8 107 67.4 26.0 III 0.99 8.0 152 62.8 25.6 IV 0.82 7.8 125 56.4 21.0 Average 0.94 7.9 143 54.5 23.0

OP-high I 1.84 8.0 163 43.2 20.9 II 1.91 7.8 107 67.8 26.0 III 1.87 7.9 151 65.0 25.8 IV 1.69 7.8 125 57.9 21.2 Average 1.83 7.9 142 55.4 23.0

pHs I 0.11 7.7 146 64.2 21.3 II 0.03 7.8 92 82.8 26.4 III 0.04 7.7 149 88.4 25.3 IV 0.05 7.6 119 68.1 21.0 Average 0.06 7.7 130 75.2 23.2

pHs+0.3 I 0.16 8.0 164 43.2 21.2 II 0.04 7.9 106 65.1 26.5 III 0.04 8.0 151 64.8 25.6 IV 0.06 7.9 125 57.9 21.3 Average 0.08 7.9 139 57.2 23.5

The average total lead release is shown in Figure 4-7 with the action level of 0.015 mg/L

noted. The error bars represent the 90th percentile concentrations. Lead release for the low dose

OP PDS is lower in Phase I and IV than in Phase II and III. This is believed to be a result of the

increased temperature observed in Phase II and III. The empirical model presented previously

also showed this trend of increasing lead release with increasing temperature. The temperature

effect is coupled with low chloride in Phases I and IV compared with Phases II and III. The

negative effect of increased chlorides was also observed in the empirical model. Trends of

116

increasing lead release with increasing temperature and chloride are consistent with the lead

release model developed for the previous study (Taylor et al. 2005) presented in Equation (4-2).

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

Phase I Phase II Phase III Phase IV

Tota

l Pb

(mg/

L)

OP-low OP-med OP-high pHs pHs+0.3

Action Level = 0.015 mg/L

Figure 4-7 Comparison of total Pb by phase and treatment

( ) 228.0-462.1726.2-677.025027.1 SulfateChloridepHAlkalinityPb T ××××= − (4-2)where: Pb = total lead concentration, μg/L

T = Temperature, °C Alkalinity = mg/L as CaCO3 pH = -Log [H+] Chloride = mg/L Sulfate = mg/L

117

The medium dose OP PDS and the high dose OP PDS have consistent lead release

between each other and between each phase. As previously discussed, most observations with

these two treatments were below detection. Consequently, any water quality effects in either of

these two PDSs are dampened and are not shown to have an effect on lead release.

In the pHs PDS, Phase I and IV had lower lead release than Phase II and III with Phase III

having the highest lead release. Again, Phase I and IV had lower temperature and chloride as

was seen with the low OP dose PDS discussed previously. However, the effects of alkalinity are

now seen between Phases II and III with similar temperatures and chloride concentrations. The

increased alkalinity in Phase III from the high percentage of groundwater caused higher lead

release in Phase III than Phase II. This is consistent with the negative effects of alkalinity

described by the empirical model presented previously in Equation (4-1) as well as the lead

release model developed in the previous study (Taylor et al. 2005) in Equation (4-2). This

alkalinity effect is not as apparent in the OP treated PDS due to the dampening the inhibitor

addition has on the water quality effects on lead release.

The pHs+0.3 PDS had similar trends as the low OP dose PDS. Lead release was lower in

Phases I and IV than in Phases II and III. The effects of water quality are somewhat reduced by

the elevated pH treatment as was seen with the low OP dose. This confirms the mitigating

effects of increased pH as seen in the empirical model presented previously as well as the water

quality model developed in the previous study shown in Equation (4-2) (Taylor et al. 2005).

However, lead concentrations were not reduced as much as with the inhibitor addition and

exceeded the action level in Phase II, as seen in Figure 4-7.

118

Overall, the orthophosphate inhibitor treated PDSs were able to maintain lead release

below the action level of 0.015 mg/L for the 90th percentile samples at the doses evaluated. The

water quality effects were dampened due to most observations below the detection limit of the

lead measurement for both the medium and high dose of orthophosphate inhibitor addition. The

pHs+0.3 treatment helped to reduce the lead release and effects of water quality to some extent, yet

still exceeded the action level during Phase II.

Surface Characterization

For each phase, lead/tin coupons of 50% lead and 50% tin composition were exposed to

the medium inhibitor dose and each of the two pH control PDSs. The coupons were then

analyzed by X-ray Photoelectron Spectroscopy (XPS) to identify compounds present on the

surface in an effort to gain insight to the mechanism of lead control by orthophosphate inhibitor

addition. The elements found on the surface of the orthophosphate lead/tin coupons and both pH

control coupons are shown in Table 4-12. There were four coupons (one for each phase)

incubated over the course of the study in each of the PDSs.

Table 4-12 Elements found in XPS analysis

Element OP pHs pHs+0.3 (4 total) (4 total) (4 total)

Carbon 3 3 3 Calcium 2 2 1

Lead 4 3 4 Oxygen 4 4 4

Phosphorus 3 0 0 Tin 4 3 4

119

For the orthophosphate treated coupons, carbon was found on the coupons for all phases

except Phase II in the carbonate form. Calcium was found as CaCO3 for Phase III and both

Ca3(PO4)2 and CaO for Phase IV. However, CaO is highly soluble and would not exist as a solid

in an aqueous environment. Formation of the compound likely occurred while preparing the

sample for analysis, where compounds initially in a soluble form remained on the coupon surface

upon drying.

Oxygen appeared on the scan of each phase of the orthophosphate coupons. It was

determined to be of the forms of O2-, CO32-, and OH- for every phase, while PO4

3- was found

only for Phase IV. However, the phosphorus element was determined to be present in the

surface scale on lead/tin coupons for Phases I, II, and IV. As would be expected from

application of OP in the system, the phosphorus was present as phosphate and is less sensitive to

the deconvolution of the oxygen element.

Elemental lead was detected during Phase IV. For the scale on lead/tin coupons exposed

to OP, PbO2 was found to contribute as a lead corrosion product. During Phases III and IV,

PbO2 accounted for 18% and 31% of lead. PbO, hydrocerussite (Pb3(OH)2(CO3)2), and cerussite

(PbCO3) were identified in the scale on lead/tin coupons exposed to OP for Phases I, II, and III.

Phase IV represented the only observation in which cerussite was not identified. This same

lead/tin coupon also represented the only sample during the study to detect elemental lead. The

presence of elemental lead coincided with the identification of calcium phosphate. This

observation may be relevant in demonstrating the ability of calcium phosphate to suppress the

corrosion of lead. Tin was identified as SnO and, for Phase IV, SnO2. Elemental tin remained

on the samples from Phase I, II, and IV.

120

The lead/tin coupons were exposed to blend waters from pHs and pHs+0.3 during four

phases of operation found carbon due to the presence of carbonate for all phases excluding Phase

II. This pattern was identical to OP inhibitor PDS. Calcium was detected as calcium carbonate

during Phases III and IV for pHs and Phase IV for pHs+0.3. Similar observations were identified

for OP. However, there was no evidence of a calcium form other than the carbonate within pHs

or pHs+0.3. The detection frequency between pHs and pHs+0.3 may seem contrary to the positive

LSI of pHs+0.3. Although calcium carbonate was observed during the same phases for both pHs

and pHs+0.3, a quantitative assessment of the samples suggests that more deposition occurred

within pHs+0.3.

Compounds associated with oxygen were determined as O2-, CO32-, and OH- for all

phases excluding Phase III for pHs and Phase IV for pHs+0.3. The coupon analyzed for pHs

during Phase III only detected CO32- within the oxygen scan. The failure to detect any form of

lead, tin, or zinc (elements commonly identified with O2-) suggests a predominance of calcium

carbonate on the scale. As shown in Table 4-12, phosphate based compounds were not detected

by XPS for pHs and pHs+0.3. The absence of phosphorus from lead/tin coupons taken from the

pH PDSs indicates phosphorus scale came from the orthophosphate inhibitor.

Lead corrosion products for pHs and pHs+0.3 were found to be similar regarding the

detection of PbO2, PbO, hydrocerussite, and cerussite. Differences between the inhibitors could

be attributed to the absence of PbO2 during Phases I and IV for pHs+0.3. However, the detection

of PbO2 is consistently low in all lead scales for all pHs and pHs+0.3 samples. The lead corrosion

products detected between lead/tin coupons exposed to inhibitor and the pH control lines were

very similar.

121

Corrosion products of tin were found in the form of oxides for both pHs and pHs+0.3.

Results were consistent with the inhibitor in that a majority of the corrosion products were found

as SnO. However, SnO2 was detected during Phase IV for pHs. Elemental tin was identified

during Phase I and Phase II for pHs, and during Phase I and Phase IV for pHs+0.3. Variations of

elemental tin composition in the scale for the pH only inhibitors were relatively similar between

samples and accounted for approximately 9% of the average elemental tin formed in the scale.

Consideration of the lead compounds PbO2, PbO, Pb3(OH)2(CO3)2, and PbCO3 were

sufficient to describe all lead/tin coupons. This would imply that introducing the inhibitors used

during the study into the system did not significantly change the composition of the lead

corrosion products. PbO was found in all scales on lead coupons regardless of treatment, which

suggests PbO is involved with lead release with or without the presence of inhibitors for these

water quality conditions. During the study, the percentage of PbO2 was the least of all lead

compounds identified.

From Figure 4-8, OP shows a variation in PbO2 ranging from 2.7% to 31.5% and

averaging 13.9% through all phases. The relative small fraction of PbO2 in the lead scale

indicates relatively significant amounts of lead are not released directly from PbO2 surfaces and

that PbO2 was not the direct controlling phase for lead in this work. PbO accounts for 24.5% of

lead scales on average. At least one observation for each treatment has a percentage of PbO

accounted for over 30% of the lead in the scale on the lead coupons, which demonstrates the

formation of PbO scale in lead coupons for all conditions for this work. The fraction of

hydrocerussite in the scale for all lead/tin coupons was greater than 20%, which shows

hydrocerussite as the major solid form in the scale of lead/tin coupons for this work. The

122

fraction of cerussite was less than the fraction of hydrocerussite in the scale on all lead/tin

coupons. However, there were observations in which cerussite (PbCO3) accounted for more than

45% of the lead contained within a coupon. The relative distribution of solid lead forms indicates

that hydrocerussite, cerussite, or PbO was the controlling solid phase for lead release in this work

regardless of corrosion control treatment or blend. However, the phosphates on the OP treated

coupons suggest less soluble phosphate products could have formed as well and been present in

the corrosion layer. These will be further evaluated.

0

10

20

30

40

50

60

70

80

90

100

OP pHInhibitor

% A

rea

of C

ompo

und

PbO2 PbO Pb3(OH)2(CO3)2 PbCO3

Figure 4-8 Distribution of lead compounds

123

Thermodynamic Modeling

Thermodynamic modeling was performed to validate the lead controlling solid phase.

The controlling solid phase found from XPS analysis was dominantly hydrocerussite

(Pb(OH)2(CO3)2). However, the phosphate found on coupons exposed to orthophosphate

inhibitor suggests other solids may have formed as well and been present in the corrosion layer.

Assuming the presence of Pb(II) species, the following dissolved lead ions and complexes were

incorporated into the thermodynamic modeling efforts. Equilibrium modeling considered the

same dissolved lead species as that of TBW I (Taylor et al. 2005). Incorporation of any other

complexes cited by the literature proved to be redundant. The resulting model is presented in

Equation (4-3).

[ ] [ ] ( )[ ] ( )[ ]( )[ ] [ ] ( )[ ] [ ]+−−

−++

++++

+++=

3223

03

24

30

22

T

PbHCOCOPbPbCOOHPb

OHPbOHPbPbOHPbPb (4-3)

A pe-pH diagram was developed assuming the PbO, Pb(OH)2, PbCO3, and

Pb3(CO3)2(OH)2 solid species and dissolved species discussed previously. In addition, this

model considered phosphate-based solids to be significant in the corrosion layer and complexes

to be a significant fraction of the dissolved lead. The phosphate-based complexes proved to be

insignificant with respect to PbT, consistently representing less than 1% of PbT for PO4T as high

as 2 mg/L as P. The basic-lead(II) phosphate, hydroxypyromorphite, is often assumed to be the

solid responsible for inhibiting lead release (Schock 1989). The reaction used during the

analysis is shown in Equation (4-4).

124

( ) ( ) OHPOPbHsOHPOPb 2

34

2345 35 ++=+ −++ (4-4)

The pe-pH diagram taking into account orthophosphate inhibition of lead, with PbT = 10-6

M and PO4T = 1 mg/L-P, is shown in Figure 4-9. The diagram suggests hydroxypyromorphite to

be the predominant form of the corrosion layer, and not hydrocerussite, for domains typical of

drinking water conditions, depicted within the shaded cube. The shaded cube represents the

region of the diagram that would be typical of drinking water having an alkalinity between 50

and 300 mg/L as CaCO3. The pH spans from 7.5 to 8.5, while the pe spans from 9.0 to 12.5 for a

system maintaining a monochloramine residual.

125

-20

-10

0

10

20

4 6 8 10 120

50100

150200

250300

pe

pH

Alkalinity

(mg/L as CaCO3 )

pe-pH Diagram for Leadat varying Alkalinities

(PbT = 10-6 M)

PbO2(s)(for pe above boundary of Pb(II) species)

Pb2+

PbHCO3+

Pb5(PO4)3OH(s),Hydroxypyromorphite

Pb3(CO3)2(OH)2,Hydrocerussite

Pb(OH)42-

Pb(OH)3-

Pb(s)(for pe below boundaryof Pb(II) species)

Pb2+

PbHCO3+

PbCO3o

Pb(CO3)22-

Pb(OH)2o

Typical DrinkingWater Conditions

Figure 4-9 Pourbaix diagram for lead species with 1 mg/L P inhibitor addition

126

Although increasing PbT would widen the predominance boundary of

hydroxypyromorphite and hydrocerussite, the coexistence boundary between the two solids

would remain the same, provided both solids are favorable. This implies that for a dose of 1

mg/L as P, hydroxypyromorphite formation would be more favorable over hydrocerussite within

household plumbing. With respect to the prediction of the solid present, an increase in

orthophosphate dose would extend the boundaries of hydroxypyromorphite, including the

coexistence boundary, thus further ensuring its favorability over hydrocerussite in household

plumbing.

While the pe-pH diagram implies that hydroxypyromorphite is less soluble than

hydrocerussite, the pe-pH approach lacks application for determining the extent of inhibition.

For this reason a pC-pH diagram was developed while varying CT to determine the response of

PbT. The resulting pC-pH diagram is shown Figure 4-10. Figure 4-10 assumes that the pe is

such that only Pb(II) species exist within the pH range of interest, a valid assumption as

demonstrated by the typical drinking water conditions identified by the shaded cube in Figure

4-9.

127

-5

-4

-3

-2

-1

0

1

7.6

7.8

8.0

8.2

8.4

0.51.0

1.52.0

2.5

log(

Pb2+

-ppb

)

pH

Total Phosphorus (mg/L-P)

Thermodynamic Model assuming Pb5(PO4)3OH

as controlling solid

OP-low doseOP-medium doseOP-high doseEquilibrated samples

Thermodynamic Model

Model adjusted fordiffusion limitation

Equilibrated sample forOP-low dose

Equilibrated sample forOP-med dose

Equilibrated sample forOP-high dose

Figure 4-10 pC-pH diagram assuming hydroxypyromorphite as the controlling solid

128

Figure 4-9 was developed assuming hydroxypyromorphite as the controlling solid. The

log of the Pb2+ ion concentration was plotted so the effects of carbonate complexes could be

accounted for, but the total phosphorus could remain a variable. The upper plane shows the

hydroxypyromorphite model prediction of Pb2+ with varying pH and phosphorus concentrations.

However, due to the small lead coupon used in the copper loops to represent solder, lead release

was diffusion limited in the sampling observations. Therefore, the diffusion equation shown in

Equation (4-5) was used to calculate a diffusion factor for the lead concentrations that would be

expected from such a small quantity of lead.

⎟⎟

⎠

⎞

⎜⎜

⎝

⎛−=

tDxerfCC

effsx 4

1 (4-5)

where: Cx = concentration at distance x and time t Cs = equilibrium concentration x = distance t = time Deff = diffusion coefficient erf = the error function

The lead coupon was located about 1 ft from the sample port in the copper loops. 1 L

samples were taken after a 6 hr stagnation time, drawing about 5 m of water from the loops.

Using this information and Equation (4-5) a diffusion adjustment factor of 0.002 was

determined. Therefore Cobs = 0.002 Cs. The thermodynamic model was adjusted for this

diffusion limitation and is shown as the lower plane in Figure 4-10. Plotted in Figure 4-10 are

the observations of lead release from each of the OP treated PDSs. These observations agree

well with the diffusion limited model.

129

One coupon from each PDS was also left to incubate for 14 months after removal from

the PDSs in a 100 mL container of the PDS water. These three coupons dissolved lead

concentrations were measured and are also plotted in Figure 4-10 as equilibrated samples. Since

they have had time to come to equilibrium, they agree well with the thermodynamic model

without adjustment for diffusion. As was seen with the observations during the study, the

dissolved lead concentration in the low dose OP PDS was greater than the dissolved lead

concentrations from the medium and high dose OP PDSs. These results suggest

hydroxypyromorphite model is well suited as the controlling solid phase and formation may be

the mechanism for lead release control with orthophosphate corrosion inhibitor addition.

Conclusions

• Empirical modeling of water quality to predict total lead release showed that total

phosphorus, pH, alkalinity, chloride and temperature were significant parameters.

The addition of phosphorus and an increase in pH were found to decrease the total

lead release while the alkalinity, chloride and temperature were found to contribute to

total lead release. The range of phosphate dosages used in this study was 0 (control)

to 2 mg/L as P.

• Observations found all PDSs treated with orthophosphate inhibitor to did not exceed

the action level of 0.015 mg/L. The low dose of OP had higher lead release than the

medium and high dose, both of which had many observations below the measurement

detection level for lead of 0.001 mg/L. This shows the ability of the inhibitor to

dampen the effects of water quality on lead release. Elevated pH treatment to pHs+0.3

130

also reduced lead release and slightly dampened water quality effects compared to the

pHs PDS, yet was found at times to exceed the action level.

• XPS analysis of lead coupons exposed to the medium OP dose and both the pH

control PDSs showed hydrocerussite, Pb3(OH)2(CO3)2, to be the dominant solid found

on the coupons regardless of corrosion control treatment and blend. This suggests

hydrocerussite is the controlling solid for lead release. However, phosphate forms

were found on the orthophosphate coupons that were not identified on the pH control

coupons and lead release was significantly reduced in the OP treated PDSs.

• Thermodynamic modeling was performed to find the thermodynamically favorable

controlling solid responsible for corrosion layer composition. Hydroxypyromorphite,

Pb5(PO4)3OH, was found to be less soluble in the region of water quality typical of

drinking water distribution systems than hydrocerussite. Observations of lead release

followed a model assuming hydroxypyromorphite as the controlling solid with

diffusion limitations taken into account. Observations from equilibrated samples

followed the trends of the equilibrium model. Therefore, the controlling solid is

determined to be hydroxypyromorphite. Hydroxypyromorphite forms a solid on the

lead surface in the presence of orthophosphate inhibitor to reduce lead release.

131

References

APHA, AWWA, and WEF (American Public Health Association, American Water Works Association, and Water Environmental Association). 1998. Standard Methods for the Examination of Water and Wastewater. 20th ed. Washington, D.C.: APHA.

AwwaRF. 2001. Role of phosphate inhibitors in mitigating lead and copper corrosion. AwwaRF, Denver.

AwwaRF. 1996 (2nd ed). Internal Corrosion of Water Distribution Systems.. AwwaRF, Denver.

Boffardi, B.P. 1995. Lead corrosion. Journal New England Water Works Association, 109(2):121-131.

Boffardi, B.P. 1990. Minimization of lead corrosion in drinking water. Materials Performance, 29(8):45-49.

Dodrill, D. M. and M. Edwards. 1995. Corrosion control on the basis of utility experience. Journal American Water Works Association, 87(7):87-99.

Edwards, M., and L.S. McNeill. 2002. Effect of phosphate inhibitor on lead release from pipes. Journal American Water Works Association, 94 (1):79-90.

Federal Register. 1991. Drinking Water Regulations; Lead and Copper Rule. 56 FR 26460

Gardels M.C. and M.R. Schock. 1981. Corrosion indices: invalid or invaluable? In Proceedings of the AWWA Water Quality Technology Conference. Dever, Colo.: American Water Works Association.

Giani, R., W. Keefer, and M. Donnelly. 2005. Studying the effectiveness and stability of orthophosphate on Washington D.C.’s lead service lines. In Proc. of AWWA WQTC Conference. Quebec City, Quebec: AWWA.

Gregory, R. and P.J. Jackson. 1983. Reducing lead in drinking water. Report 219-S, Water Research Centre, Stevenage Laboratory, Elder Way, Stevenage, Herts.

Hozalski, R.M., E. Esbri-Amador, and C.F. Chen. 2005. Comparison of stannous chloride and phosphate for lead corrosion control. Journal American Water Works Association, 97(3):89-102.

Hunt, D.T.E and J.D. Creasey. 1980. The calculation of equilibrium trace metal speciation and solubility in aqueous systems by a computer method with particular reference to lead. Technical Report 151. Medmenham, England: Water Research Centre.

132

Lee, R.G., W.C. Becker, and D.W. Collins. 1989. Lead at the tap: sources and control. Journal American Water Works Association, 81(7):52-62.

Nriagu, J.O. 1974. Lead orthophosphates-IV. Formation and stability in the environment. Gerchim. Cosmochim. Acta, 38:887

Pinto, J.A., A.S. McAnally, and J.R.V. Flora. 1997. Evaluation of lead and copper corrosion control techniques. Journal of Environmental Science and Health, A32(1):31-53.

Reiber, S.H. 1991. Galvanic stimulation of corrosion on lead-tin solder-sweated joints. Journal American Water Works Association, 83(7):83-91.

Schock, M.R. 1989. Understanding Corrosion Control Strategies for Lead. Journal American Water Works Association, 81(7)88.

Schock, M.R. 1980. Response of lead solubility to dissolved carbonate in drinking water. Journal American Water Works Association, 72(12):695.

Sheiham, I. and P.J. Jackson. 1981. The scientific basis for control of lead in drinking water by water treatment. Journal of the Institution of Water Engineers and Scientists, 35(6):491.

Singley, J.E. B.A. Beaudet, D.W. DeBerry, J.R. Kidwell, D.A. Malish, and P.H. Marker. 1985. Corrosion Prevention and Control in Water Treatment and Supply Systems, Pollution Technology Review No. 122, Noyes Publications, Park Ridge, NJ.

Taylor, J.S., J.D. Dietz, A.A. Randall, S.K. Hong, C.D. Norris, L.A. Mulford, J.M. Arevalo, S. Imran, M. LePuil, S. Liu, I. Mutoti, J. Tang, W. Xiao, C. Cullen, R. Heaviside, A. Mehta, M. Patel, F. Vasquez, and D. Webb. 2005. Effects of Blending on Distribution System Water Quality. Denver, Colo.: AwwaRF and Tampa Bay Water.

133

CHAPTER 5 COMPARISON OF PHOSPHATE INHIBITORS FOR IRON, COPPER

AND LEAD CORROSION CONTROL

Abstract

This study evaluated the effects of phosphate-based corrosion inhibitor addition on iron,

copper, and lead release from pilot distribution systems (PDS) exposed to different blends of

groundwater, surface water, and desalinated seawater. Three phosphate-based corrosion

inhibitors were employed; blended orthophosphate (BOP), orthophosphate (OP), and zinc

orthophosphate (ZOP). Non-linear empirical models were developed to predict iron, copper, and

lead release from each PDS treated with different doses of inhibitor ranging from zero (control)

to 2 mg/L as P. The predictive models were developed using water quality parameters as well as

the inhibitor dose. Using these empirical models, simulation of the water quality of different

blends with varying alkalinity and pH were used to compare the inhibitors performance for

remaining in compliance for iron, copper, and lead. Increasing pH was shown to reduce iron,

copper and lead metal release, while increasing alkalinity was shown to reduce iron release but

increase copper and lead release. The ZOP inhibitor was not predicted by the empirical models

to perform as well as BOP and OP at the low dose of 0.5 mg/L as P for iron control, and the OP

inhibitor was not predicted to perform as well as BOP and ZOP at the low dose of 0.5 mg/L as P

for lead control. The three inhibitors evaluated performed similarly for copper control.

Therefore, BOP inhibitor showed the lowest metal release at the low dose of 0.5 mg/L as P for

control of iron, copper and lead corrosion.

134

Introduction

With increasing water demands and more stringent drinking water regulations, many

utilities are turning to desalinated sources to supplement their surface and groundwater supplies.

Tampa Bay Water (TBW) and the University of Central Florida (UCF) studied the effects of

blending multiple alternative source waters on distribution system water quality (Taylor et al.

2005). In this study, three different phosphate inhibitors are evaluated for iron, copper and lead

corrosion control of different blends of groundwater, surface water and desalinated seawater.

Literature suggests for iron control, orthophosphate forms a protective scale while

blended orthophosphate is thought to sequester iron (Wagner 1992; Benjamin et al. 1990;

Boffardi 1988; Wagner and Kuch 1984; Huang 1980; Pryor and Cohen 1951). Zinc

orthophosphate has been reported to be a more effective inhibitor for iron corrosion control as

compared to blended orthophosphate and orthophosphate based inhibitors (Bancroft 1988;

Swayze 1983; Bailey 1980; Mullen and Ritter 1980; Mullen and Ritter 1974; Murray 1970;

Powers, Cahalan, and Zalfa 1966; Kleber 1965). Other studies have found no benefit of zinc

orthophosphate as compared to blended orthophosphate and orthophosphate for iron corrosion

control (McNeill and Edwards 2000; Volk et al. 2000; Malcolm Pirnie 1998; Williams 1990;

Wagner and Kuch 1984; Swayze 1983; Huang 1980). Lytle and Snoeyink (2002) reported that

polyphosphates were found to reduce turbidity and apparent color associated with iron release as

compared to the use of orthophosphate based inhibitors. In another study, McNeill and Edwards

(2000) found orthophosphate and zinc orthophosphate increased iron release under stagnant

conditions. Blended orthophosphate was also found to increase iron in four of five conditions

tested.

135

For copper release, Edwards, Hidmi, and Gladwell (2002) found both blended

orthophosphate and orthophosphate decreased copper release at a dose of 1 mg/L as P over a

three year period. At a pH of 7.2 and an alkalinity of 300 mg/L as CaCO3, the three inhibitor

types increased copper release. At high pHs and lower alkalinities blended orthophosphate was

not found to be as beneficial as orthophosphate, due to complexation with copper. Similarly,

another study (McNeill and Edwards 2004) found orthophosphate to reduce soluble copper

levels, with little effect on particulate copper release. Blended orthophosphate was found to

increase particulate copper levels and zinc orthophosphate did not offer more advantages than

orthophosphate and increased release of particulate species after an 8 hour stagnation period.

However, another study found blended orthophosphate to be a more effective inhibitor for

copper corrosion than orthophosphate by having more effective protective scale formation

(Souissi and Triki 2007). Lytle, Schock, and Sorg (1996) found copper levels were effectively

reduced by both zinc orthophosphate and orthophosphate. Schneider et al. (2007) also

determined orthophosphate was as effective as zinc orthophosphate at reducing copper corrosion.

For lead release, Lin et al. (1997) found increasing pH as well as adding inhibitor is

effective at reducing dissolved lead concentrations. Another study found orthophosphate and

blended orthophosphate decreased particulate lead species, while zinc orthophosphate increased

particulate lead species (McNeill and Edwards 2004). Zinc orthophosphate was also found to

perform better than blended orthophosphate at reducing lead levels in waters with various pHs

and hardness (Mass et al. 1991). Lee, Becker, and Collins (1989) found zinc orthophosphate

reduced lead levels at the water qualities tested and blended orthophosphate performed as well as

the zinc orthophosphate when pH was greater than 8.0. Edwards and McNeill (2002) and

136

Hozalski, Esbri-Amador, and Chen (2005) found blended orthophosphate to increase particulate

and soluble lead release as compared to orthophosphate after 8-hr and 72-hr stagnation periods

over a broad range of water qualities. Lytle, Schock, and Sorg (1996) and Schneider et al.

(2007) found lead levels were effectively reduced by both zinc orthophosphate and

orthophosphate. Dodrill and Edwards (1995) found orthophosphate, but not blended

orthophosphate, inhibitor to reduce lead release for alkalinity less than 30 mg/L as CaCO3 and

pH less than 7.4. Higher pH and alkalinity waters did not see an improvement with inhibitor

addition. At alkalinities between 30 and 74 mg/L as CaCO3 and pH greater than 7.4, blended

orthophosphates adversely affected lead release.

This study evaluated the effects of blended orthophosphate (BOP), orthophosphate (OP),

and zinc orthophosphate (ZOP) inhibitor addition to blended treated surface, ground, and

seawater sources of varying blend percentages. The effects of water quality were evaluated and

models predicting total iron, copper, and lead release using water quality and total phosphorus

doses were developed. These models were then used to evaluate each inhibitors’ performance

under simulated and variable water quality conditions.

Experimental Methods

Experimental Design

Experimentation was conducted with the use of pilot distribution systems (PDSs) built

from actual pipelines extracted from TBW member governments distribution systems

(Hillsborough County, Fla.; Pasco County, Fla.; Pinellas County, Fla.; City of New Port Richey,

Fla.; City of St. Petersburg, Fla.; and City of Tampa, Fla.). Details regarding the PDS and prior

137

study results are reported elsewhere (Taylor et al. 2005). Each PDS runs in parallel with

segments of PVC, lined cast iron, unlined cast iron, and galvanized steel pipes that were placed

sequentially to simulate actual distribution systems. The materials and their respective diameter

and length are shown in Table 5-1. Each PDS was fed blends of groundwater, surface water, and

desalinated seawater along with different types and doses of corrosion inhibitor.

Table 5-1 Pipe materials in PDSs

Pipe Material Length (ft)

Diameter (in)

PVC 20 6 Lined Cast Iron 20 6

Unlined Cast Iron 12 6 Galvanized Steel 40 2

Images of materials as well as the PDSs used to represent full scale distribution systems

are shown in Figure 5-1. The pipe materials used, shown in the left image, are galvanized steel,

PVC, lined cast iron, and unlined cast iron from left to right. The image on the right shows the

PDSs with the materials connected in series from influent (closest to viewer) to effluent (furthest

from viewer) of the system. The pipe materials were connected in sequence of PVC, lined cast

iron, unlined cast iron, then galvanized steel.

138

Figure 5-1 Pipe materials (left) and parallel PDSs (right)

The PDSs were fed blends of conventionally treated groundwater (GW), enhanced

coagulation-sedimentation-filtration (CSF) treated surface water (SW), and desalinated seawater

by reverse osmosis (RO). The description of the three finished source waters is presented in

Table 5-2.

Table 5-2 Finished source water descriptions

Water Source System Description

GW Groundwater Ground water source. Treatment by aeration, disinfection by free chlorine with a residual of 5 mg/L after a 5 minute contact time. 5.0 mg/L chloramine residual.

SW Surface water

TBW treatment plant: Treatment by ferric sulfate coagulation, flocculation, settling, filtration, disinfection by ozonation and chloramination. Project site: adjustment of chloramine residual to 5.0 mg/L chloramine residual.

RO Groundwater Treatment by membrane reverse osmosis, aeration, disinfection by free chlorine with a residual of 5 mg/L after a 5 minute contact time. 5.0 mg/L chloramine residual.

139

The GW unit used raw well water from the Cypress Creek well field owned by TBW.

The GW was treated with aeration, disinfection, and pH stabilization. Aeration was achieved in

the GW by pumping the raw water to the top of the finished water tank through a spray nozzle.

Sodium hypochlorite was used for primary disinfection and was dosed to provide a 5 mg/L

residual after a 5 minute contact time. Afterwards, ammonium chloride was added to produce a

5 mg/L monochloramine residual. Ammonia was added in the form of NH4Cl at a 5:1 ratio. The

Cl2: NH3 ratio was initially 4:1 to protect against DBP formation. This ratio was increased to 5:1

in after 6 months of operation to reduce free ammonia.

SW was treated at the TBW Regional Surface Water Treatment Facility by enhanced

coagulation, ozonation, biologically activated carbon (BAC) filtration, aeration, and

chloramination. The SW was hauled weekly to the field facility for use and temporarily stored in

two 7000 gallon storage tanks before being transferred to the SW finished water tank. In the SW

finished tank, the chloramine residual was adjusted to 5 mg/L as Cl2.

The RO pilot plant was housed in a trailer at the testing facility and utilized raw

groundwater for the feed stream. The RO treatment pilot system required the addition of TDS,

calcium, and alkalinity to the RO permeate to represent the finished water produced by the TBW

Regional Desalination Facility. RO pretreatment consisted of 2.7 mg/L antiscalant addition

(Hypersperse MDC700TM, GE Water, Minnetonka, Minn.) followed by 5-micron cartridge

filtration. The RO membrane unit was operated at 72-73% recovery, producing 9.3 gpm

permeate flow, which was aerated by a 10-inch diameter aeration tower filled with tripack plastic

packing. After aeration, 50 mg/L of sea salt was added to the aerated permeate stream to

simulate the TBW desalination process. Calcium chloride and sodium bicarbonate were also

140

added to meet the calcium and alkalinity specifications. The finished was stabilized with sodium

hydroxide to 0.1 to 0.3 pH units above pHs.

The effects of water quality were evaluated by varying the blend quarterly, while

seasonal effects were evaluated by maintaining the same blend in the summer and winter. The

quarterly phases and percentages of the blends are shown below in Table 5-3. The average water

quality of each of the source waters in each phase is shown in Table 5-4. The effects of season

are seen in the temperature as well as rainy and dry season effects on the surface water between

Phases I and III. The blends with a high percentage of groundwater in Phases I and III are

characterized by high alkalinity and pH. Phase II had the highest percentage of surface water

and is characterized by high sulfate concentrations. Phase IV has average water quality

parameters due to the equal percentage of GW and SW.

Table 5-3 Blend percentages for each phase

Phase Quarter % GW % SW % RO I Feb-May 2006 62 27 11 II May-Aug 2006 27 62 11 III Aug-Nov 2006 62 27 11 IV Nov 2006-Feb 2007 40 40 20

Table 5-4 Average water quality by phase

Phase pH Alkalinity (mg/L CaCO3)

Chloride (mg/L)

Sulfate (mg/L)

Temperature (°C)

I 8.0 161 45 62 21 II 7.9 104 67 103 26 III 8.0 150 68 66 26 IV 7.9 123 59 76 21

141

The feed rate of the blend into each PDS was maintained to achieve a two-day hydraulic

residence time (HRT). Pumps maintained the blend flow as well as the inhibitor addition into

each PDS. The PDSs each were fed different inhibitor types and doses. The inhibitors were

dosed to the PDSs at three different levels, categorized as low dose, medium dose, and high dose.

Blended ortho-phosphate (BOP), orthophosphate (OP) and zinc orthophosphate (ZOP) were each

maintained at a target dose of 0.5 mg/L as P for the low dose, 1.0 mg/L as P for the medium

dose, and 2.0 mg/L as P for the high dose. Control PDSs were not fed any chemical inhibitor;

one was maintained at pHs and a second was treated with elevated pH, maintained at pHs+0.3.

The PDS at pHs+0.3 was maintained at a positive LSI to assess the affect of elevated pH treatment

as a means of iron, copper and lead release control.

Phosphate Inhibitors

The physical properties of each inhibitor used during the project are shown in Table 5-5.

The selected blended orthophosphate (BOP) product was Sodium Polyphosphate SK-7641

(Stiles-Kem/Met-Pro Corporation, Waukegan, Ill.). Manufacturer suggested the BOP product

contained approximately 40% orthophosphate and 60% polyphosphate. However, monitoring of

BOP dose administered during operation indicated 60-80% orthophosphate. Periodic

determination of the actual ratio was conducted to verify the correct dose of the product. The

orthophosphate inhibitor was Inhibit-All WSF-36 (SPER Chemical Corporation, Clearwater,

Fla.). The zinc orthophosphate inhibitor was CP630 (Sweetwater Technologies, Temecula,

Calif.), and was prepared by dissolving zinc sulfate into phosphoric acid solution in a 1:5 Zinc to

PO4 ratio.

142

Table 5-5 Inhibitor product properties

Parameter BOP OP ZOPPercent Active Product 36% 36% 44%Bulk Density (lbs/gal) 11.5 11.25 10.8 Specific Gravity (at 72oF) 1.3 1.35 1.45 pH 1% solution (at 72ºF) 6.3-6.6 5.1-5.4 <1 Recommended Dose (mg/L) 1-2 as P 1-4 as P 2-3 as P Recommended pH Range 6.0-8.5 6.8-7.8 7-8 Solubility in water (g/g H2O) 60/100 N/A N/A Shelf Life (months) 6 None 6 Storage Limitation Indoors None None

Data Collection

Water quality parameters were collected and analyzed weekly in the influent and effluent

of the PDSs. The water quality parameters monitored and methods used for samples transported

to the UCF laboratory for analysis are shown in Table 5-6. Table 5-7 shows the water quality

parameters monitored and methods used for the parameters analyzed in the field laboratory at the

testing facility. The method detection limit (MDL) is also shown. Methods noted as SM are

from APHA, AWWA, and WEF (1998).

143

Table 5-6 Water quality parameters and methods performed at University Laboratory

Parameter Method Reference Method Description MDL

Aluminum SM 3120B ICP Method 0.001 mg/L Bicarbonate SM 2320B Titration Method 5 mg/L Calcium SM 3120B ICP Method 0.1 mg/L

Chloride SM 4110 Ion Chromatography with Chemical Suppression of Eluent Conductivity 0.1 mg/L

Color SM 2120A Or Hach 8025 Cobalt-Platinate Method (with spec) 1 CPU

Conductivity SM 2510B Laboratory Method 1 μmho/cm Copper SM 3120B ICP Method 0.001 mg/L Iron SM 3120B ICP Method 0.001 mg/L Lead SM 3120B ICP Method 0.001 mg/L Magnesium SM 3120B ICP Method 0.1 mg/L Nitrogen (NH3,TKN) SM 4500-Norg Macro-Kjeldahl Method 0.1 mg/L

NPDOC SM 5310C Persulfate-UV Oxidation Method 0.1 mg C/L pH SM 4500-H+ B Electrometric Method ± 0.01 pH units Phosphorus SM 3120B ICP Method 0.001 mg/L Silica SM 3120B ICP Method 0.001 mg/L Sodium SM 3120B ICP Method 0.1 mg/L Solids (TDS) SM 1030E Estimation of TDS by major ion sum 1 mg/L

Sulfate SM 4110 Ion Chromatography with Chemical Suppression of Eluent Conductivity 0.1 mg/L

Turbidity SM 2130B Nephelometric Method 0.01 NTU UV-254 SM 5910 UV Absorption at 254 nm 0.0001 cm-1 Zinc SM 3120B ICP Method 0.001 mg/L

144

Table 5-7 Water quality parameters and methods performed at Field Laboratory

Parameter Method Reference Method Description MDL Alkalinity SM 2320 B Titration 5 ppm Ammonia-N SM 4500-NH3 C Membrane Probe Method 0.1 ppm Chloride SM 4500-Cl- B Argentometric Titration 1 mg/L Chlorine, free SM 4500-Cl G or Hach 8021 DPD colorimetric 0.1 ppm Chlorine, total SM 4500-Cl-G or Hach 8167 DPD colorimetric 0.1 ppm Color, apparent SM 2120 B Visual Comparison

(by spectrometer) 1 CPU

Conductivity SM 2510 B Conductivity Bridge 1 μmho/cm Hardness (total, calcium) SM 2340 C EDTA Titration 5 mg/L

Nitrate Hach 8192 Cadmium reduction 0.1 mg/L Nitrite Hach 8507 Diazotization 0.1 mg/L Oxygen, Dissolved (DO) SM 4500-O G Membrane probe 0.1 mg/L

pH SM 4500-H+ B Electrometric ± 0.01 pH units

Phosphate-P (Reactive) SM 4500-P E. or Hach 8048 Ascorbic Acid Method 0.1 mg/L

Silica, SiO2 (reactive) SM 4500-SiO2 or Hach 8185 Molybdosilicate Method 0.1 mg/L

as SiO2 Temperature SM 2550 B Direct reading 0 deg C Turbidity SM 2130 B Nephelometric 0.01 NTU UV254 SM 5910 A UV spectrometry 0.0001 cm-1

Portions of the flow from each PDS were fed to a corrosion loop consisting of 30 feet of

5/8 inch copper tubing with one lead/tin coupon to represent solder. Each loop holds

approximately 1.8 L of water. The copper tubes were flushed every morning with 2 gallons of

PDS water. Weekly samples were collected after a six-hour stagnation period in order to

simulate tap monitoring as described in the Lead and Copper Rule. The corrosion shed and

housed corrosion loops are show in Figure 5-2.

145

Figure 5-2 Corrosion shed (left) and corrosion loops (right)

Results and Discussion

Iron

Empirical Modeling

The average water quality for each phase of operation is presented in Table 5-8. As

previously discussed, Phases I and III are enriched with groundwater, having high alkalinity and

silica, while Phase II is enriched with surface water having a high sulfate concentrations. Phases

I and IV were in winter, while Phases II and III were in summer.

Table 5-8 Average water quality of each phase

Blend Silica Alkalinity Chloride Sulfate Temperature (mg/L-SiO2) (mg/L-CaCO3) (mg/L) (mg/L) (°C)

Phase I 10.8 161 45.4 62.6 21.1 Phase II 5.0 104 68.1 104.3 26.2 Phase III 10.4 150 68.6 66.8 25.6 Phase IV 6.3 123 59.5 76.8 21.2

146

A non-linear empirical model was developed using this water quality data and inhibitor

dose for each of the phosphate-based inhibitor PDSs and the two controls, pHs and pHs+0.3. The

model was developed using non-linear least squares regression and is presented in Equation

(5-1). The model makes use of dummy variables, that are either one or zero, depending on

absence or presence and the type of inhibitor used. The pHs term is used to represent both

control PDSs at pHs and pHs+0.3, where no inhibitor is added.

25772.04

319.0597.0527.02

320.0

051.03095.03

0003.1

)0107.00112.01060.91091.7(

−−

−

−−−

×××××

+×+

××+××=

T

s

SOClAlkSiO

pHTPZOPTPOPTPBOPFeTotal

(5-1)

where: Total Fe = total iron, mg/L BOP = BOP inhibitor (0, 1) OP = OP inhibitor (0, 1) ZOP = ZOP inhibitor (0, 1) pHs = pH control (0, 1) TP = total phosphorus, mg/L SiO2 = silica, mg/L as SiO2 Alk = alkalinity, mg/L as CaCO3 Cl = chloride, mg/L SO4 = sulfate, mg/L T = temperature, °C

The water quality parameters of silica, alkalinity, chloride, sulfate, and temperature

remained significant in the model, with total phosphorus significant when inhibitor addition is

used. Total phosphorus was found to be beneficial for controlling iron release for both the BOP

and ZOP inhibitors, but not for the OP inhibitor as shown by the sign on the total phosphorus

exponents. However, these exponents are range from -0.320 to 0.051, suggesting the iron release

is not sensitive to inhibitor dose, especially for the BOP and OP inhibitors. Silica, chloride,

147

sulfate, and temperature are shown to contribute to iron release while, alkalinity is shown to

reduce iron release by the negative exponent on the alkalinity term.

The model fits the data with an R2 of 0.30. The fit is shown in Figure 5-3 by treatment.

The diagonal line represents an ideal perfect agreement between the observations and predictions

by the model.

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45Actual Total Fe (mg/L)

Pred

icte

d To

tal F

e (m

g/L)

BOP OP ZOP pHs pHs+0.3

Figure 5-3 Model predicted versus actual total iron concentration

Inhibitor Performance

Figure 5-4 shows the average iron release from each of the PDSs by inhibitor and dose.

The error bars represent the minimum and maximum observations and the iron secondary

148

standard of 0.3 mg/L is noted. The pH controls use “low dose” to show pHs and “med dose” to

show pHs+0.3. As shown in Figure 5-4, the treatments except for the medium and high dose BOP

inhibitor and the low dose OP inhibitor have observations exceeding the iron secondary standard

of 0.3 mg/L.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

BOP OP ZOP pH controls

Inhibitor

Tota

l Fe

(mg/

L)

low dose med dose high dose

Secondary Standard = 0.3 mg/L

Figure 5-4 Comparison of inhibitor treatments on total iron release

A one-way ANOVA test reveals iron release was reduced equivalently by BOP inhibitor

at the doses evaluated, OP inhibitor at the low and high dose, ZOP inhibitor at the high dose, or

pH elevation to pHs+0.3. This suggests phosphate inhibitor addition does not appear to reduce

iron release as compared to pH elevation. However, if pH elevation is not practical (due to

149

calcium carbonate scaling) inhibitor addition is just as beneficial. Use of ZOP inhibitor required

a higher dose to achieve the same levels of iron release as the BOP and OP inhibitor at lower

doses.

Water Quality Simulations

Water quality simulations were performed making use of the empirical model developed

in Equation (5-1). The average water quality for each blend used in the simulation was presented

in Table 5-8. This water quality along with the minimum total phosphorus dose of 0.5 mg/L as P

were input to the empirical models and plotted with varying pH and alkalinity, shown in Figure

5-5 to Figure 5-8 and Figure 5-9 to Figure 5-12, respectively. The figures show a different line

for each inhibitor type as well as one for the “control” which is modeled assuming no inhibitor

was added.

Varying pH in Figure 5-5 to Figure 5-8 was modeled by the change in alkalinity with

varied pH. As shown in the figures, increasing pH has an effect of reducing iron release. The

highest iron release is predicted in Phase II (Figure 5-6), due to the high alkalinity, chloride,

sulfate, and temperatures observed during this phase with an enriched groundwater blend. This

is a condition where the secondary standard of 0.3 mg/L is predicted to be exceeded at pHs

below 7.2 with a 0.5 mg/L as P dose of ZOP inhibitor. The ZOP inhibitor addition is shown to

have the highest iron release at the low dose, followed by the controls with no inhibitor added,

OP inhibitor and BOP inhibitor with the lowest iron release, as was seen with the data presented

previously.

150

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0pH

Tota

l Fe

(mg/

L)

BOP OP ZOP Control

Secondary Standard = 0.3 mg/L

Figure 5-5 Phase I water quality iron simulation varying pH

151

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0pH

Tota

l Fe

(mg/

L)

BOP OP ZOP Control

Secondary Standard = 0.3 mg/L

Figure 5-6 Phase II water quality iron simulation varying pH

152

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0pH

Tota

l Fe

(mg/

L)

BOP OP ZOP Control

Secondary Standard = 0.3 mg/L

Figure 5-7 Phase III water quality iron simulation varying pH

153

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0pH

Tota

l Fe

(mg/

L)

BOP OP ZOP Control

Secondary Standard = 0.3 mg/L

Figure 5-8 Phase IV water quality iron simulation varying pH

A similar simulation was done while varying alkalinity, and again, increasing alkalinity

helps to reduce iron release. The secondary standard of 0.3 mg/L is predicted to be exceeded in

both Phases II and III (Figure 5-10 and Figure 5-11. respectively) at alkalinities below 100 mg/L

as CaCO3 with the 0.5 mg/L as P dose of ZOP inhibitor. This is due to the high chlorides and

temperatures observed in these blends. Again, the ZOP inhibitor addition is predicted to have

the highest iron release, followed by the control, OP inhibitor and BOP inhibitor with the lowest

predicted iron release.

154

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

80 90 100 110 120 130 140 150 160 170 180Alkalinity (mg/L-CaCO3)

Tota

l Fe

(mg/

L)

BOP OP ZOP Control

Secondary Standard = 0.3 mg/L

Figure 5-9 Phase I water quality iron simulation varying alkalinity

155

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

80 90 100 110 120 130 140 150 160 170 180Alkalinity (mg/L-CaCO3)

Tota

l Fe

(mg/

L)

BOP OP ZOP Control

Secondary Standard = 0.3 mg/L

Figure 5-10 Phase II water quality iron simulation varying alkalinity

156

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

80 90 100 110 120 130 140 150 160 170 180Alkalinity (mg/L-CaCO3)

Tota

l Fe

(mg/

L)

BOP OP ZOP Control

Secondary Standard = 0.3 mg/L

Figure 5-11 Phase III water quality iron simulation varying alkalinity

157

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

80 90 100 110 120 130 140 150 160 170 180Alkalinity (mg/L-CaCO3)

Tota

l Fe

(mg/

L)

BOP OP ZOP Control

Secondary Standard = 0.3 mg/L

Figure 5-12 Phase IV water quality iron simulation varying alkalinity

Overall, iron release was simulated to be lowest with the BOP inhibitor addition at a dose

of 0.5 mg/L as P. The control is shown to reduce iron release further than the ZOP inhibitor,

which is modeled assuming no inhibitor was added. Increased pH, and therefore alkalinity, also

reduces iron release levels below the secondary standard.

158

Copper

Empirical Modeling

The average water quality for each phase of operation is presented in Table 5-9. As

previously discussed, Phases I and III are enriched with groundwater, having high pH, alkalinity

and silica, while Phase II is enriched with surface water having low alkalinity and silica. Phase

IV had average water quality of the other phases because it had equal fractions of groundwater

and surface water.

Table 5-9 Average water quality of each phase

Blend Silica pH Alkalinity Chloride (mg/L-SiO2) (std. units) (mg/L-CaCO3) (mg/L)

Phase I 10.8 7.94 161 45.4 Phase II 5.0 7.78 104 68.1 Phase III 10.4 7.91 150 68.6 Phase IV 6.3 7.77 123 59.5

Empirical modeling was performed on this water quality data and inhibitor dose for each

of the PDSs dosed with phosphate-based inhibitor and the two control PDSs at pHs and pHs+0.3 to

predict copper release. The non-linear least squares regression resulted in Equation (5-2). This

model is similar in form to the iron model presented previously. Dummy variables are used to

indicate the presence or absence and type of inhibitor added to the PDS as either a one or zero.

159

421.0263.1409.4445.02

304.0338.0204.0

)676.8

671.0936.3156.3(

ClAlkpHSiOpH

TPZOPTPOPTPBOPCuTotal

s ××××+

×+×+×=−−

−−−

(5-2)

where: Total Cu = total copper, mg/L BOP = BOP inhibitor (0, 1) OP = OP inhibitor (0, 1) ZOP = ZOP inhibitor (0, 1) pHs = pH control (0, 1) TP = total phosphorus, mg/L SiO2 = silica, mg/L as SiO2 pH = -log[H+] Alk = alkalinity, mg/L as CaCO3 Cl = chloride, mg/L

The model found silica, pH, alkalinity and chloride to be significant water quality

parameters for predicting copper release, as well as total phosphorus when inhibitor was added.

Total phosphorus is shown to reduce copper release with the three inhibitors, as shown by the

negative exponent on the total phosphorus terms. Silica and pH also reduce copper release while

alkalinity and chloride contribute to copper release. The fit of the model to the data is presented

in Figure 5-13 by inhibitor treatment. The data fits the model with an R2 of 0.79. The diagonal

line represents an ideal agreement between the actual and predicted observations by the model.

160

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0Actual Total Cu (mg/L)

Pred

icte

d To

tal C

u (m

g/L)

BOP OP ZOP pHs pHs+0.3

Figure 5-13 Model predicted versus actual total copper concentration

Inhibitor Performance

Figure 5-14 shows the average total copper release observed in this study by inhibitor

treatment and dose. The “low dose” pH control was at pHs and the “med dose” pH control was

at pHs+0.3. The error bars represent the 90th percentile of observations for the Lead and Copper

Rule. The copper action level of 1.3 mg/L at the 90th percentile of monitored household taps is

noted in Figure 5-14.

161

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

BOP OP ZOP pH controls

Inhibitor

Tota

l Cu

(mg/

L)

low dose med dose high dose

Action Level = 1.3 mg/L

Figure 5-14 Comparison of inhibitor treatments on total copper release

A one-way ANOVA suggests that the inhibitor treatments evaluated at the differing doses

were able to reduce copper concentrations below that obtained by the pHs treatment. The

elevated pH treatment did not reduce copper release as much as the tested inhibitors except for

the low dose of ZOP. The other inhibitors at evaluated doses reduced copper levels lower than

the elevated pH treatment, with the high doses having the greatest control of copper release.

Therefore, inhibitor treatment is a valid addition for reducing copper levels further than pH

elevation alone.

162

Water Quality Simulations

Water quality simulations were performed making use of the empirical model developed

for copper release presented in Equation (5-2). The average water quality for each blend was

presented in Table 5-9. This water quality along with the total phosphorus dose of 0.5 mg/L as P

were input to the empirical models and plotted with varying pH and alkalinity, shown in Figure

5-15 to Figure 5-18 and Figure 5-19 to Figure 5-22, respectively. The figures show a different

line for each inhibitor type as well as one for the “control” which is modeled assuming no

inhibitor was added.

Varying pH in Figure 5-15 to Figure 5-18 is shown to decrease copper release with

increasing pH. The three inhibitors performed nearly equivalently at a dose of 0.5 mg/L,

regardless of the water quality of the blend. The control without inhibitor addition is shown to

exceed the action level of 1.3 mg/L on average below a pH between 7.2 and 7.6, depending on

the blend. Phase III (Figure 5-17) required the highest pH, above 7.6, to maintain copper levels

below the action level without inhibitor addition, likely due to the high alkalinity and chloride

observed in this blend. Phase II (Figure 5-16) required the lowest pH of 7.2 to maintain copper

levels below the action level, due to the low levels of alkalinity observed in this blend.

163

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0

pH

Tota

l Cu

(mg/

L)

BOP OP ZOP Control

Action Level = 1.3 mg/L

Figure 5-15 Phase I water quality copper simulation varying pH

164

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0

pH

Tota

l Cu

(mg/

L)

BOP OP ZOP Control

Action Level = 1.3 mg/L

Figure 5-16 Phase II water quality copper simulation varying pH

165

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0

pH

Tota

l Cu

(mg/

L)

BOP OP ZOP Control

Action Level = 1.3 mg/L

Figure 5-17 Phase III water quality copper simulation varying pH

166

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0

pH

Tota

l Cu

(mg/

L)

BOP OP ZOP Control

Action Level = 1.3 mg/L

Figure 5-18 Phase IV water quality copper simulation varying pH

Similar results are presented in Figure 5-19 to Figure 5-22 varying alkalinity for each

blend. Contrary to the results for iron release, increasing alkalinity increases copper release.

However, the inhibitor addition at a dose of 0.5 mg/L as P resulted in copper levels below the

action level for the blends evaluated in this study. The control, without inhibitor addition, is

predicted to exceed the copper action level in three of the four blends at alkalinities from 125 to

170 mg/L as CaCO3. Phase II (Figure 5-20) saw the lowest alkalinity to maintain copper release

below the action level, below 125 mg/L as CaCO3, due to the low pH and silica and high

chloride levels observed in this blend. Phase I (Figure 5-19) experienced the highest alkalinity to

167

maintain copper release below the action level, with no observations in the range of this

simulation exceeding the action level, due to the high silica and pH and the lowest chloride

concentrations observed in the study.

0.0

0.5

1.0

1.5

2.0

2.5

80 90 100 110 120 130 140 150 160 170 180

Alkalinity (mg/L-CaCO3)

Tota

l Cu

(mg/

L)

BOP OP ZOP Control

Action Level = 1.3 mg/L

Figure 5-19 Phase I water quality copper simulation varying alkalinity

168

0.0

0.5

1.0

1.5

2.0

2.5

80 90 100 110 120 130 140 150 160 170 180

Alkalinity (mg/L-CaCO3)

Tota

l Cu

(mg/

L)

BOP OP ZOP Control

Action Level = 1.3 mg/L

Figure 5-20 Phase II water quality copper simulation varying alkalinity

169

0.0

0.5

1.0

1.5

2.0

2.5

80 90 100 110 120 130 140 150 160 170 180

Alkalinity (mg/L-CaCO3)

Tota

l Cu

(mg/

L)

BOP OP ZOP Control

Action Level = 1.3 mg/L

Figure 5-21 Phase III water quality copper simulation varying alkalinity

170

0.0

0.5

1.0

1.5

2.0

2.5

80 90 100 110 120 130 140 150 160 170 180

Alkalinity (mg/L-CaCO3)

Tota

l Cu

(mg/

L)

BOP OP ZOP Control

Action Level = 1.3 mg/L

Figure 5-22 Phase IV water quality copper simulation varying alkalinity

Overall, the phosphate-based inhibitors utilized in this study are effective at a dose of 0.5

mg/L as P at controlling copper release below the action level of 1.3 mg/L regardless of water

quality. Without inhibitor addition, blend water quality must be maintained at a high pH and low

alkalinity to not exceed the action level.

171

Lead

Empirical Modeling

The average water quality for each phase of operation is presented in Table 5-10 As

previously discussed, Phases I and III are enriched with groundwater, having high pH and

alkalinity, while Phase II is enriched with surface water having low alkalinity. Phases I and IV

were in winter, while Phases II and III were in summer.

Table 5-10 Average water quality of each phase

Blend pH Alkalinity Chloride Temperature (std. units) (mg/L-CaCO3) (mg/L) (°C)

Phase I 7.94 161 45.4 21.1 Phase II 7.78 104 68.1 26.2 Phase III 7.91 150 68.6 25.6 Phase IV 7.77 123 59.5 21.2

An empirical model for predicting lead release was also developed similar to the iron and

copper models developed previously. A non-linear least squares regression of the water quality

and inhibitor dose data from the phosphate-based inhibitor PDSs and the two control PDSs at

pHs and pHs+0.3 resulted in the model in Equation (5-3). The model makes use of dummy

variables where a one or zero is used to indicate the presence or absence and type of inhibitor

used.

172

911.2701.1864.1913.6

7680.613

410.18417.08

)10191.110722.310019.210836.1(

TClAlkpH

pHTPZOPTPOPTPBOPPbTotal

s

××××

×+××+

××+××=

−

−−−

−−−−

(5-3)

where: Total Pb = total lead, mg/L BOP = BOP inhibitor (0, 1) OP = OP inhibitor (0, 1) ZOP = ZOP inhibitor (0, 1) pHs = pH control (0, 1) TP = total phosphorus, mg/L pH = -log[H+] Alk = alkalinity, mg/L as CaCO3 Cl = chloride, mg/L T = temperature, °C

The model found pH, alkalinity, chloride, and temperature to be significant water quality

parameters for predicting lead release, as well as total phosphorus when inhibitor was added.

Phosphorus was found to help reduce lead release for the three inhibitors, as shown by the

negative exponent on the total phosphorus terms. pH was also shown to reduce lead release

while alkalinity, chloride, and temperature contribute to lead release.

Figure 5-23 shows the fit of the data to the model. The data fits the model with an R2 of

0.64. The diagonal line represents an ideal agreement between the actual and predicted

observations by the model. Many observations of lead release were below the detection limit of

0.001 mg/L. This is apparent by the many low points on this graph.

173

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.00 0.02 0.04 0.06 0.08 0.10 0.12Actual Total Pb (mg/L)

Pred

icte

d To

tal P

b(m

g/L)

BOP OP ZOP pHs pHs+0.3

Figure 5-23 Model predicted versus actual total lead concentration

Inhibitor Performance

Figure 5-24 shows the average lead release for each inhibitor treatment by dose. The

“low dose” pH control was at pHs and the “med dose” pH control was at pHs+0.3. The error bars

represent the 90th percentile of observations for the Lead and Copper Rule. The action level of

0.015 mg/L at the 90th percentile of observations is noted in Figure 5-24.

174

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

BOP OP ZOP pH controls

Inhibitor

Tota

l Pb

(mg/

L)

low dose med dose high dose

Action Level = 0.015 mg/L

Figure 5-24 Comparison of inhibitor treatments on total lead release

A one-way ANOVA shows inhibitor addition for the doses evaluated, except for the low

dose OP inhibitor, reduced lead levels below both the pHs and pHs+0.3 treatments.

Water Quality Simulations

Water quality simulations were performed making use of the empirical model developed

for lead release shown in Equation (5-3). The average water quality for each blend was

presented in Table 5-10. This water quality along with a total phosphorus dose of 0.5 mg/L as P

were input into the empirical model and plotted with varying pH and alkalinity, shown in Figure

175

5-25 to Figure 5-28 and Figure 5-29 to Figure 5-32, respectively. The figures show a different

line for each inhibitor type as well as one for the “control” which is modeled assuming no

inhibitor was added.

Similar observations to the iron and copper release simulations are seen with lead release.

Shown in Figure 5-25 to Figure 5-28, increasing pH is predicted to decrease lead release. Lead

release is predicted to exceed the action level for the control without inhibitor in Phase II below a

pH of 7.6 (Figure 5-26) and in Phase III below a pH of 8.5 (Figure 5-27). The OP inhibitor

addition at a dose of 0.5 mg/L as P is also predicted to exceed the action level below pH of 7.4 in

the Phase III blend (Figure 5-27). Phase III had high alkalinity, combined with high chloride and

temperature, which are shown to contribute to lead release. The BOP and ZOP reduce lead

release below the action level regardless of water quality at a dose of 0.5 mg/L as P. ZOP

maintained the lowest lead levels for the four blends evaluated.

176

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0

pH

Tota

l Pb

(mg/

L)

BOP OP ZOP Control

Action Level = 0.015 mg/L

Figure 5-25 Phase I water quality lead simulation varying pH

177

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0

pH

Tota

l Pb

(mg/

L)

BOP OP ZOP Control

Action Level = 0.015 mg/L

Figure 5-26 Phase II water quality lead simulation varying pH

178

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0

pH

Tota

l Pb

(mg/

L)

BOP OP ZOP Control

Action Level = 0.015 mg/L

Figure 5-27 Phase III water quality lead simulation varying pH

179

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0

pH

Tota

l Pb

(mg/

L)

BOP OP ZOP Control

Action Level = 0.015 mg/L

Figure 5-28 Phase IV water quality lead simulation varying pH

Similar simulations were performed by varying the alkalinity. As was seen with the

copper release, increasing alkalinity increases lead release. For the Phase II blend, the action

level was exceeded with the control above alkalinities of 110 mg/L as CaCO3 (Figure 5-30). The

OP inhibitor addition at a dose of 0.5 mg/L as P is also predicted to exceed the action level in

Phase II above alkalinities of 165 mg/L as CaCO3 (Figure 5-30). The Phase II blend was

characterized by high chloride, high temperatures and lower pH, contributing to lead release. In

the Phase I blend, no inhibitor evaluated or the control was predicted to exceed the action level

for the simulation range tested (Figure 5-29). Phase I was characterized by low chloride, low

180

temperatures and higher pH. Both the BOP and ZOP inhibitor additions at a dose of 0.5 mg/L as

P maintained lead release below the action level of 0.015 mg/L regardless of water quality, with

ZOP maintaining lead levels below detection.

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

80 90 100 110 120 130 140 150 160 170 180

Alkalinity (mg/L-CaCO3)

Tota

l Pb

(mg/

L)

BOP OP ZOP Control

Action Level = 0.015 mg/L

Figure 5-29 Phase I water quality lead simulation varying alkalinity

181

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

80 90 100 110 120 130 140 150 160 170 180

Alkalinity (mg/L-CaCO3)

Tota

l Pb

(mg/

L)

BOP OP ZOP Control

Action Level = 0.015 mg/L

Figure 5-30 Phase II water quality lead simulation varying alkalinity

182

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

80 90 100 110 120 130 140 150 160 170 180

Alkalinity (mg/L-CaCO3)

Tota

l Pb

(mg/

L)

BOP OP ZOP Control

Action Level = 0.015 mg/L

Figure 5-31 Phase III water quality lead simulation varying alkalinity

183

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

80 90 100 110 120 130 140 150 160 170 180

Alkalinity (mg/L-CaCO3)

Tota

l Pb

(mg/

L)

BOP OP ZOP Control

Action Level = 0.015 mg/L

Figure 5-32 Phase IV water quality lead simulation varying alkalinity

Overall, the ZOP inhibitor addition maintained lead release below detection for the

blends evaluated. The BOP inhibitor also maintained the lead levels below the action level of

0.015 mg/L at a dose of 0.5 mg/L as P. The OP inhibitor performed better than the control, but

at the low dose was not always able to maintain lead below the action level in periods of high

alkalinity and low pH. High pH and low alkalinity must be maintained in order to keep lead

release below the action level if no inhibitor was added.

184

Conclusions

• Empirical modeling suggests iron release is affected by total phosphorus, silica,

alkalinity, chloride, sulfate and temperature. Increases in total BOP and ZOP and

alkalinity are shown to help reduce iron release, while increases in total OP, silica,

chloride, sulfate and temperature are shown to contribute to iron release. The iron

release has minimal sensitivity to the effect of phosphorus dose from each of the

inhibitors. At a low dose of 0.5 mg/L as P, the BOP and OP inhibitors are predicted

to maintain iron release below the secondary standard of 0.3 mg/L regardless of water

quality, with BOP having the lowest predicted iron release at a 0.5 mg/L as P dose.

Adding no inhibitor resulted in lower iron release than adding 0.5 mg/L as P of ZOP

for iron control.

• Empirical modeling suggests copper release is affected by total phosphorus, silica,

pH, alkalinity, and chloride. Increases in total phosphorus, silica and pH are shown to

reduce copper release, while increases in alkalinity and chloride are shown to

contribute to copper release. At a dose of 0.5 mg/L as P, the phosphate-based

inhibitors are predicted to maintain copper release below the action level of 1.3 mg/L

regardless of water quality. The no inhibitor option often exceeded the action level in

periods of low pH and high alkalinity.

• Empirical modeling suggests lead release is affected by total phosphorus, pH,

alkalinity, chloride, and temperature. Increases in total phosphorus and pH are shown

to help reduce lead release, while increases in alkalinity, chloride, and temperature

185

are shown to contribute to lead release. At the low dose of 0.5 mg/L as P, the BOP

and ZOP inhibitors are predicted to maintain lead release below the action level of

0.015 mg/L regardless of water quality with the ZOP inhibitor maintaining lead

below the measurement detection limit. The no inhibitor control and 0.5 mg/L as P

dose of OP inhibitor addition often exceeded the action level in periods of low pH

and high alkalinity.

• Due to contradicting effects of alkalinity on iron compared to copper and lead,

inhibitor addition is recommended to control metal release for the water blends

evaluated in this study. The BOP inhibitor was effective at control of iron, copper

and lead, by maintaining metal release below regulated levels for the blends at a dose

of 0.5 mg/L as P. Higher doses may be required for OP to control lead release and for

ZOP to control iron release.

• Cost was not evaluated in this study as it was just a comparison evaluation. Since

higher doses may be required for OP to control lead release and for ZOP to control

iron release, cost analyses for each product would be prudent before selection.

Higher doses of some products may result in a lower operation cost to obtain the

same iron, copper and lead corrosion control.

186

References

APHA, AWWA, and WEF (American Public Health Association, American Water Works Association, and Water Environmental Association). 1998. Standard Methods for the Examination of Water and Wastewater. 20th ed. Washington, D.C.: APHA.

Bailey, T.L. 1980. Corrosion control experiences at Durham, North Carolina. In Proceedings Seminar on Corrosion Control in Drinking water Systems. Hollison, Mass.: New England Water Works Association.

Bancroft, D.A. 1988. Corrosion control program in Danvers, Massachusetts. Journal New England Water Works Association, 102:163.

Benjamin, M.M., S.H. Reiber, J.F. Ferguson, E.A. Vanderwerff, and M.W. Miller. 1990. Chemistry of Corrosion Inhibitors in Potable Water. Denver, Colo.: AWWA Research Foundation and American Water Works Association.

Boffardi, B.P. 1988. Potable water treatment and monitoring for corrosion and scale control. Journal New England Water Works Associations, 102(2):111.

Dodrill, D.M. and M. Edwards. 1995. Corrosion control on the basis of utility experience. Journal of American Water Works Association, 87(7):74-85.

Edwards, M., L. Hidmi, D. Gladwell. 2002. Phosphate inhibition of soluble copper corrosion by-product release. Corrosion Science, 44(5):1057-1071.

Edwards, M., and L.S. McNeill. 2002. Effect of phosphate inhibitors on lead release from pipes. Journal American Water Works Association, 94(1):79-90.

Hozalski, R.M., E. Esbri-Amador, and C.F. Chen. 2005. Comparison of stannous chloride and phosphate for lead corrosion control. Journal American Water Works Association, 97(3):89-103.

Huange, D.J.-S. 1980. Polyphosphate for corrosion control in water distribution system. PhD dissertation, University of Missouri, Columbia.

Kleber, J.P. 1965. Use of bimetallic glassy phosphates for corrosion control. Journal American Water Works Association, 57(6)783.

Lee, R.G., W.C. Becker, and D.W. Collins. 1989. Lead at the tap: sources and control. Journal American Water Works Association, 81(7): 52-62.

187

Lin, N.H., A. Torrents, A.P. Davis, M. Zeinali, and F.A. Taylor. 1997. Lead corrosion control from lead, copper-lead solder, and brass coupons in drinking water employing free and combined chlorine. Journal of Environmental Science and Health, A32(4):865-884.

Lytle, D.A. and V.L. Snoeyink. 2002. Effect of ortho- and polyphosphates on the properties of iron particles and suspensions. Journal American Water Works Association, 94(10):87-99.

Lytle, D.A., M.R. Schock, and T.J. Sorg. 1996. Controlling lead corrosion in the drinking water of a building by orthophosphate and silicate treatment. Journal of the New England Water Works Association, 110(3):202-217

Malcolm Pirnie. 1998. Central Avra Valley Storage and Recovery Project, Task 6: Water Quality Management Programs, Iron Release Testing Results. Phoenix.

Mass, R.P., S.C. Patch, D.J. Kucken, and B.T. Peek. 1991. A multi-state study of the effectiveness of various corrosion inhibitors in reducing residential lead levels. In Proceedings of Annual Conference. Dever, Colo.: American Water Works Association.

McNeill, L.S. and M. Edwards. 2004. Importance of Pb and Cu particulate species for corrosion control. Journal of Environmental Engineering, 130(2):136-144.

McNeill, L.S. and M. Edwards. 2000. Phosphate inhibitors and red water in stagnant pipes. Journals of Environmental Engineering, 126(12):1096.

Mullen, E.D. and J.A. Ritter. 1980. Monitoring and controlling corrosion by potable water. Journal American Water Works Association, 72(5):286.

Mullen, E.D. and J.A. Ritter. 1974. Potable water corrosion control. Journal American Water Works Association, 66(8):473.

Murray, W.B. 1970. A corrosion inhibitor process for domestic water. Journals American Water Works Association, 62(10):659.

Powers, J.T., E.M. Cahalan, and A.J. Zalfa. 1966. Eliminating red water with bimetallic glassy phosphate. Journal New England Water Works Association, 80(3):282.

Pryor, M.J. and M. Cohen. 1951. The mechanism of the inhibition of corrosion of iron by solutions of sodium orthophosphate. Journal Electrochemical Society, 98(7):263.

Schneider, O.D., M.W. Lechevallier, H.F. Reed, and M.J. Corson. 2007. A comparison of zinc and nonzinc orthophosphate-based corrosion control. Journal of American Water Works Association, 99(11):103-113.

188

Souissi, N. and E. Triki. 2007. A chemiometric approach for phosphate inhibition of copper corrosion in aqueous media. Journal of Material Science, 42:3259-3265.

Swayze, J. 1983. Corrosion study at Carbondale, Ill. Journal American Water Works Association, 75(2):101.

Taylor, J.S., J.D. Dietz, A.A. Randall, S.K. Hong, C.D. Norris, L.A. Mulford, J.M. Arevalo, S. Imran, M. LePuil, S. Liu, I. Mutoti, J. Tang, W. Xiao, C. Cullen, R. Heaviside, A. Mehta, M. Patel, F. Vasquez, and D. Webb. 2005. Effects of Blending on Distribution System Water Quality. Denver, Colo.: AwwaRF and Tampa Bay Water.

Volk, C., E. Dundore, J. Schiermann, and M. Lechevallier. 2000. Practical evaluation of iron corrosion control in a drinking water distribution system. Water Research, 34(6):1967-1974.

Wagner, I. 1992. Influence of operating conditions on materials and water quality in drinking water distribution systems. Corrosion and related aspects of materials for potable water supplies. In Proceedings Institute of Materials Conference, London.

Wagner, I and A. Kuch. 1984. The influence of water parameters of corrosion rate, scale deposition and iron(III) uptake in unprotected iron pipes. Water Supply, 2(3/4):SS11.

Williams, S.M. 1990. The use of sodium silicate and sodium phosphate to control water problems. Water Supply, 8:195.