This article was downloaded by: [University of California, San Diego]

On: 13 June 2015, At: 05:46

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Journal of Environmental Science and Health, Part

A: Toxic/Hazardous Substances and Environmental
Engineering
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/lesa20

Corrosion control in water supply systems: Effect of

pH, alkalinity, and orthophosphate on lead and copper
leaching from brass plumbing
a b
Y. S. Tam & P. Elefsiniotis
a
Watercare Services Ltd. , Auckland, New Zealand
b
Department of Civil and Environmental Engineering , University of Auckland , Auckland,
New Zealand
Published online: 04 Sep 2009.

To cite this article: Y. S. Tam & P. Elefsiniotis (2009) Corrosion control in water supply systems: Effect of pH, alkalinity, and
orthophosphate on lead and copper leaching from brass plumbing, Journal of Environmental Science and Health, Part A:
Toxic/Hazardous Substances and Environmental Engineering, 44:12, 1251-1260, DOI: 10.1080/10934520903140009

To link to this article: http://dx.doi.org/10.1080/10934520903140009

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Journal of Environmental Science and Health Part A (2009) 44, 1251–1260
Copyright
C Taylor & Francis Group, LLC

ISSN: 1093-4529 (Print); 1532-4117 (Online)

DOI: 10.1080/10934520903140009

Corrosion control in water supply systems: Effect of pH,

alkalinity, and orthophosphate on lead and copper leaching
from brass plumbing

Y.S. TAM1 and P. ELEFSINIOTIS2

1
Watercare Services Ltd., Auckland, New Zealand
2
Department of Civil and Environmental Engineering, University of Auckland, Auckland, New Zealand
Downloaded by [University of California, San Diego] at 05:46 13 June 2015

This study explored the potential of lead and copper leaching from brass plumbing in the Auckland region of New Zealand. A five-
month field investigation, at six representative locations, indicated that Auckland’s water can be characterized as soft and potentially
corrosive, having low alkalinity and hardness levels and a moderately alkaline pH. More than 90% of the unflushed samples contained
lead above the maximum acceptable value (MAV) of 10 µg/L (New Zealand Standards). In contrast, the copper level of unflushed
samples remained consistently below the corresponding MAV of 2 mg/L. Flushing however reduced sharply metal concentrations,
with lead values well below the MAV limit. Generally, metal leaching patterns showed a limited degree of correlation with the
variations in temperature, dissolved oxygen and free chlorine residual at all sampling locations. Furthermore, a series of bench-scale
experiments was conducted to evaluate the effectiveness of pH and alkalinity adjustment, as well as orthophosphate addition as
corrosion control tools regarding lead and copper dissolution. Results demonstrated that lead and copper leaching was predominant
during the first 24 hr of stagnation, but reached an equilibrium state afterwards. Since the soluble fraction of both metals was
small (12% for lead, 29% for copper), it is apparent that the non-soluble compounds play a predominant role in the dissolution
process. The degree of leaching however was largely affected by the variations in pH and alkalinity. At pH around neutrality, an
increase in alkalinity promoted metal dissolution, while at pH 9.0 the effect of alkalinity on leaching was marginal. Lastly, addition
of orthophosphate as a corrosion inhibitor was more effective at pH 7.5 or higher, resulting in approximately 70% reduction in both
lead and copper concentrations.
Keywords: Water supply, corrosion control, lead, copper, metal leaching, orthophosphate, alkalinity.

Introduction by leaching from fittings.[3,4] In particular, lead has been

Corrosion (or plumbosolvency) is one of the major chal-
lenges water utilities face worldwide. From an asset man-
agement perspective, the principal concerns for a water
utility are twofold: (i) corrosion of the distribution sys-
tem pipes which may eventually lead to water leakage
and loss of hydraulic capacity due to the accumulation
of corrosion products,[1] and (ii) premature failure of water
pipes which impacts the structural integrity of the distri-
bution systems.[2] From a health viewpoint, drinking wa-
ter suppliers are concerned about corrosion of plumbing
fittings, which may result in the staining of fixtures and
expose consumers to heavy metal contamination caused
Address correspondence to P. Elefsiniotis, Department of
Civil and Environmental Engineering, University of Auckland,
Private Bag 92019, Auckland 1030, New Zealand. E-mail:
t.elefsiniotis@auckland.ac.nz
Received December 9, 2009.
recognized as a cumulative poison to the human body,
which can cause irreversible damage to the nervous system
and create neurological problems especially to infants and
children.[5,6]
To minimize such potential hazards to public health
and infrastructure, the metal content in water supplies has
been regulated by various authorities worldwide. For in-
stance, strict limits on lead concentrations in drinking water
have been promulgated by the United States Environmen-
tal Protection Agency (U.S. EPA) due to lead’s toxicity.[7]
In addition, the Guidelines for Canadian Drinking Water
Quality have a maximum acceptable concentration (MAC)
of 10 µg/L for lead.[8] The European Commission (EC)
also plans to tighten the drinking water lead standard
to 10 µg/L by 2013, in line with the World Health Or-
ganization (WHO) guideline value.[9] Presently, the max-
imum acceptable value (MAV) of lead in New Zealand
is 10 µg/L,[10] which is consistent with several interna-
tional drinking water standards.[8,9] Regarding cooper, the
 1252
corresponding MAV is 2 mg/L for health reasons with a
guideline value (GV) of 1 mg/L for aesthetic purposes.[10]
Corrosion is a complex phenomenon as its behavior is
dependent on water chemistry as well as a number of envi-
ronmental factors including pH, alkalinity, chlorine resid-
ual, temperature, dissolved oxygen, chloride, phosphate,
and organic matter.[11,12] Although certain factors, such as
pH and alkalinity, may be controlled during water treat-
ment to reduce the impact of corrosion,[13] other factors
such as temperature are difficult to control. To estimate the
corrosion potential of a certain water type, several indices
have been introduced, including the Langelier Saturation
Index (LSI) and the Aggressiveness Index (AI). The LSI is
commonly used to predict whether calcium carbonate will
precipitate or dissolve in a particular case, thereby reducing
Downloaded by [University of California, San Diego] at 05:46 13 June 2015
the water’s tendency for corrosion.[14]
At an LSI of zero, the water is in equilibrium with cal-
cium carbonate. A negative LSI suggests that the water
dissolves calcium carbonate, and a positive LSI value indi-
cates that the water will tend to deposit calcium carbonate.
Usually a LSI between – 0.5 and 0.0 is desirable for infras-
tructure as well as public health protection.[15] On the other
hand, the AI can predict the degree of deterioration in wa-
ter quality that arises from pipe dissolution by measuring
pH, alkalinity and calcium hardness.[16] If the AI ≤ 10.0,
the water is categorized as highly aggressive; if the AI lies
between 10.0 and 11.9, the water is considered moderately
aggressive; and if the AI ≥ 12.0, the water is classified as
completely non-aggressive.[17]
Currently, the most common methods for corrosion con-
trol involve either adjustment of pH and alkalinity or ad-
dition of corrosion inhibitors such as orthophosphate and
silicate. The pH and alkalinity adjustment affects the ther-
modynamic properties of the carbonate system to achieve
reduction in the corrosive nature of the water.[18,19] Alterna-
tively, addition of corrosion inhibitors aids the formation of
protective coating on a pipe, thereby preventing corrosive
water from attacking the metal surface.[20−22]
Both field and bench-scale investigations can be used to
determine water quality and quantify the corrosion poten-
tial, thus enabling water utilities to gain an insight into
corrosion problems and implement the necessary correc-
tive strategies. The objectives of this research are therefore
as follows: (i) conduct a field study to assess the degree of
corrosion in the drinking water of the Auckland region in
New Zealand and (ii) evaluate the effect of pH, alkalin-
ity and inhibitor (orthophosphate) addition on corrosion
control in bench-scale experiments.
Materials and methods
Field investigation
The primary focus of the field work was to investigate the
leaching potential of lead and copper in water conduits
Tam and Elefsiniotis
(asbestos-cement and concrete-lined steel) in Auckland’s
drinking water distribution system. The experimental pro-
cedure followed was established by the New Zealand Min-
istry of Health.[10] Specifically, each brand new yellow brass
sample tap (containing 58.0% copper, 37.9% zinc and 3.78%
lead) was completely filled with water, kept stagnated for
4 hr and then flushed to waste prior to installation to the
water mains. This procedure was repeated for a week to
minimize variability.
To evaluate the effect of overnight storage on metal re-
lease in brass fittings, all sample taps were flushed for 2
min, 24 hr prior to sampling to achieve a uniform resi-
dence time. On each sampling occasion, both unflushed
samples (i.e., first-draw water) and flushed samples (i.e.,
running water collected after 2 min) were taken. In total,
84 samples (14 from each of the 6 sampling locations) were
collected on a weekly basis during a 5-month period. The
sampling locations were selected based on the ease of acces-
sibility, provision of a representative coverage of different
water types, and pipe materials used for conveyance. For
brevity purposes, the six locations have been designated as
S1 through S6 (Table 1).
Bench-scale experiments
The bench-scale experiments examined the performance of
corrosion control techniques under stagnant conditions to
simulate the dead-end sections (i.e. worst-case scenario) of
a distribution system. Pipe sections made up of concrete-
lined steel were connected with identical brass fittings uti-
lized in the field study as illustrated in Fig. 1. The pipes had
been in service for approximately 3 years and no apparent
tubercles were observed. Each pipe section had an inner di-
ameter of 0.20 m, a length of 0.50 m and a carrying capacity
of approximately 16 L. All pipe sections were covered with
a polyurethane transparent lid to ensure minimal airspace
between the surface of the water sample and the covering
to prevent contact with the atmosphere.
Initially all pipes were rinsed three times with 0.1N
sodium hydroxide (NaOH) to remove organic deposits and
Table 1. Sampling locations, pipe material and corresponding
water source.
Sampling location Pipe material Water source
Ardmore & Concrete-lined steel (CLS) Dam & river
Waikato (S1)
Waikato (S2) Concrete-lined steel (CLS) River
Ardmore (S3) Concrete-lined steel (CLS) Dam
Ardmore (S4) Asbestos-cement (AC) Dam
Waitakere (S5) Concrete-lined steel (CLS) Dam
Ardmore, Concrete-lined steel (CLS) Dam
Waikato, Huia
& Waitakere
mix (S6)
 Corrosion control in water supply systems 1253

During the corrosion inhibition tests, disodium hydro-

gen orthophosphate (Na2 HPO4 ), containing 21.8% P, was
used as a corrosion inhibitor. To evaluate the effect of or-
thophosphate addition, a passivation dose of 1.6 mg/L P
was applied for 10 consecutive days to ensure that a pro-
tective layer was formed on the appropriate pipes prior to
orthophosphate treatment.[22,23] Following this, an amount
of 0.8 ± 0.05 mg/L P was added to all samples. The exper-
imental design was identical to the one followed previously
(Table 2), except that there was no need to conduct experi-
ments at pH 9.0 with orthophosphate addition (explained
later).

Analytical methods
Downloaded by [University of California, San Diego] at 05:46 13 June 2015

During both the field and the bench-scale parts of the study,
Fig. 1. Experimental set-up for bench-scale tests.
pH, temperature, dissolved oxygen (DO), and free chlorine
residual were measured on-site during sampling. All other
then rinsed five times with distilled water. Since it was ev-
parameters such as lead and copper (both total and solu-
ident that the naturally low alkalinity levels (i.e. 20 to 50
ble), total alkalinity, total hardness, total dissolved solids
mg/L as CaCO3 ) in Auckland’s water had induced concrete
(TDS), chloride, and sulfate were determined in the labo-
leaching in the distribution system,[23] a higher range of al-
ratory. Each water sample was stirred using a cleaned glass
kalinity was applied in this study. To that effect, sodium
rod prior to obtaining the aliquot for analysis to ensure
bicarbonate (NaHCO3 ) was added to achieve concentra-
uniform distribution.
tions of 100, 125, and 150 mg/L as CaCO3 , respectively,
With respect to the on-site measurements, pH and
while the control samples had an alkalinity of 20 mg/L as
temperature were recorded after 20 sec while stirring
CaCO3 . At each alkalinity level tested, the pH was subse-
the solution using a portable kit (pHTestr30, Eutech
quently adjusted to 7.0, 7.5, 8.0, and 9.0 by bubbling carbon
dioxide gas into the water and was measured at 1-hr, 4-hr, Instruments
R ). The DO was measured with a portable dis-
24-hr, 48-hr and 72-hr intervals. To minimize any potential solved oxygen meter (Microprocessor Oximeter Oxi 325
R )
pH changes, the pH was continuously monitored and ad- using an oxygen sensor (CellOx325). Free chlorine residual
justed during the initial 4 hr, and then corrected to within was measured at a wavelength of 530 nm using a HACH
±0.1 unit of the initial values at the time of sampling there- colorimeter by the colorimetric version of the N,N-diethyl-
after. The range of pH and alkalinity tested is shown in p-phenylenediamine (DPD) method.[24]
Table 2. All tests were conducted in duplicate to enhance For total metal (lead and copper) determination, sam-
the statistical reliability of the measurements. ples were preserved with concentrated nitric acid at pH <
2 and then digested in accordance with the nitric acid di-
Table 2. Range of pH and alkalinity tested during bench-scale gestion #3030E method prior to analysis.[24] Samples for
experiments. soluble metal analysis were collected into cleaned 100-ml
polyethylene bottles and then filtered through a 0.45 µm fil-
pH Alkalinity (mg/L as CaCO3 )
ter. Lead and copper were analyzed by inductively coupled
7.0 100 plasma mass spectrometry (ICP-MS). Alkalinity and hard-
125 ness were determined by digital titration methods (Hach
150 #8203 and #8213, respectively) and TDS by a gravimetric
Control method (Hach #8277).[24] Chloride and sulfate determina-
7.5 100 tion was carried out by a computer-controlled Dionex DX-
125 120 ion chromatograph equipped with an AS9-HC column.
150 Further details on all sampling and analytical procedures
Control
are available elsewhere.[23]
8.0 100
125
150
Control Statistical methods
9.0 100 Regression analysis was performed in Microsoft Excel
125 Analysis ToolPak
R . The statistical significance of a wa-
150
ter quality parameter was evaluated based on the p-value
Control
obtained from regression analysis. In general, a p-value
 1254 Tam and Elefsiniotis
Table 3. Average water quality parameters at each sampling quality indices (LSI and AI). In general, Auckland’s wa-
location. ter is soft in nature, with low alkalinity (19 to 55 mg/L as
CaCO3 ), low hardness (25 to 60 mg/L as CaCO3 ), and a
Parameters S1 S2 S3 S4 S5 S6
moderately alkaline pH (7.80 to 8.33). The majority of the
Temperature, ◦ C 21.4 22.5 21.4 21.3 20.7 20.5 water samples from the six sampling locations were com-
pH 8.04 7.99 7.97 7.95 8.33 7.80 parable, with the exception of location S2, where higher
DO, mg/L 7.71 6.89 7.51 7.64 7.90 7.80 alkalinity, hardness, sulfate and TDS concentrations were
Residual 0.64 0.56 0.79 0.81 0.63 0.54 consistently recorded. The higher alkalinity, hardness and
chlorine, mg/L TDS values observed in location S2 can be attributed, at
Total alkalinity, 25.4 54.8 19.2 19.7 25.4 21.6 least partially, to lime addition at the Waikato water treat-
mg/L as
ment plant. Furthermore, the LSI values indicated that
CaCO3
Total hardness, 31.2 60.2 24.8 24.6 42.4 34.7
all samples were corrosive in nature. This observation is in
mg/L as agreement with the interpretation of the AI, which revealed
CaCO3 that the water was moderately aggressive.
Figure 2 displays the distribution of total (i.e. soluble
Downloaded by [University of California, San Diego] at 05:46 13 June 2015

Total dissolved 91.2 154 81.4 84.1 107 103

solids, mg/L plus non-soluble) lead release following a 24-hr standing
Chloride, mg/L 16.2 19.3 15.3 15.3 23.8 23.0 time in unflushed samples. It was found that more than
Cl 90% of the values exceeded the MAV of 10 µg/L and were
Sulfate, mg/L 11.5 20.1 8.27 8.22 12.4 13.1 classified as “aggressive” according to the New Zealand
SO4 standards.[10] Samples from locations S1 and S2 exhibited
Langelier −1.10 −0.52 −1.34 −1.35 −0.67 −1.41 similar water characteristics and therefore had higher me-
Saturation
dian lead concentrations. The other samples (locations S3
Index (LSI)
Aggressiveness 10.8 11.4 10.5 10.5 11.2 10.5
to S6) had lower lead values and the relative concentrations
Index (AI) were reasonably consistent.
It is worth noting that the S2 samples, despite having a
relatively high alkalinity at 55 mg/L as CaCO3 , also ex-
between 0.1 and 0.03 is considered to indicate a moder- hibited a median lead concentration of 28.0 µg/L, which is
ate correlation, while a p value below 0.03 reveals a strong the second highest mean value (Fig. 2). It has been reported
correlation.[25] that alkalinity up to 75 mg/L as CaCO3 may reduce cor-
rosion in surface waters.[18,26,27] Furthermore, the LSI of
the S2 water was −0.52 which suggested the least corrosion
Results and discussion
potential; nevertheless, the median lead concentration was
relatively high. It should be noted that there has been some
Field investigation
concern raised about the reliability of LSI to predict the
Results summarized in Table 3 include the mean values of aggressiveness of water.[17,28] Results from this study tend
the various water quality parameters that were monitored to support the notion that LSI may not be a dependable
during the field investigation, plus the corresponding water indicator of water quality.

80

70

60
Lead Concentration (µg/L)

50

40

30

20

10

0
S1 S2 S3 S4 S5 S6
Location

Fig. 2. Median total lead concentrations (including the 25th and 75th percentiles) in unflushed samples from the field locations.
 Corrosion control in water supply systems 1255
40

35

Mean Lead Concentration (µg/L)

30

25

20

15

10

0
S1 S2 S3 S4 S5 S6
Location
Downloaded by [University of California, San Diego] at 05:46 13 June 2015

Unflushed sample Flushed sample

Fig. 3. Mean total lead concentrations in unflushed and flushed water samples from the field locations.

It has been reported that lead leaching can be also af-
fected by the presence of sulfate and chloride ions, as well as
TDS, since water with a high TDS content becomes a better
conductor.[29,30] Results revealed much higher average sul-
fate and TDS concentrations in location S2, compared to all
other samples (Table 3). The chloride concentration, how-
ever, appeared to be within the same range in all locations.
To examine the effect of flushing on metal leaching,
flushed samples also were taken (followed the procedure
described previously) and compared to the unflushed ones.
The mean lead concentrations detected in both types of
samples are shown in Fig. 3. It is apparent that flushing
greatly reduced lead dissolution by 90 to 99%, resulting
in concentrations below the MAV. The significant drop in
lead levels confirms that lead dissolution occurred in the
brass fittings.
An attempt was made to correlate the effect of the pa-
rameters measured to lead dissolution data. Regarding pH,
no clear correlation between pH and lead dissolution was
observed in the field study. The mean pH values were rea-
sonably consistent in the range of 7.80 ± 0.48 to 8.33 ±
0.38. Regression analysis indicated that pH did not affect
lead leaching from brass taps (P-value between 0.105 and
0.924). With respect to temperature, only location S6 re-
vealed a strong correlation with the amount of lead leached
from brass fittings (P-value = 0.019). In location S3, both
temperature and DO appeared to have an influence on lead
release to a varying degree (P-value = 0.081 and 0.040, re-
spectively). Furthermore, the S1 and S2 samples suggested
a moderate correlation between lead dissolution and the
amount of free chlorine residual (P-value = 0.074 and
0.078, respectively).
In addition to lead, copper was analyzed and the re-
sults have been summarized in Fig. 4. Location S1 had
the highest median total copper concentration (22.5 µg/L)
and S4 had the lowest one (11.5 µg/L). In all locations,

50

45

40
Copper Concentration (µg/L)

35

30

25

20

15

10

0
S1 S2 S3 S4 S5 S6

Location

Fig. 4. Median total copper concentrations (including the 25th and 75th percentiles) in unflushed samples from the field locations.
 1256
the maximum copper levels fluctuated between 31 µg/L
and 46 µg/L. It is therefore apparent that the unflushed
water samples collected had copper levels well below the
corresponding MAV limit of 2 mg/L, according to the
New Zealand Standards.[10] As observed in the case of
lead, flushing reduced the copper concentration in water
by more than 90%. This provides additional evidence that
copper leaching originated from the brass plumbing.
Although investigating the role of natural organic mat-
ter (NOM) on metal leaching was not part of this research
project, as background information, the mean total organic
carbon (TOC) content of the samples tested within the
Auckland region was 1.26 ± 0.56 mg/L. It should be noted
that the presence of NOM in surface waters, even at rela-
tively low concentrations (i.e., less than 2 mg/L as TOC) has
Downloaded by [University of California, San Diego] at 05:46 13 June 2015
been indicated to enhance lead and copper solubility.[31,32]
In contrast, organic matter has been also found to pro-
mote coating in water pipes, thus reducing the corrosion
potential.[8] It is therefore apparent that the significance of
NOM on corrosion has not been conclusively determined.
Bench-scale experiments
Effect of stagnation time on lead and copper release
The effect of stagnation time on lead and copper leached
from the brass taps at varying pH and alkalinity values was
investigated by comparing metal concentrations at differ-
ent stagnation times (i.e., 1-hr, 4-hr, 24-hr, 48-hr and 72-
hr). Since the field study revealed that flushing reduced the
metal concentrations drastically, only unflushed samples
were analyzed during the bench-scale investigation. Results
illustrated in Fig. 5, at an alkalinity level of 100 mg/L as
240
220
200
Lead concentrations ( g/L)
180
160
140
120
100
80
60
40
20
0
0 5
Fig. 5. Effect of pH and stagnation time on total lead dissolution at an alkalinity of 100 mg/L as CaCO3.
Tam and Elefsiniotis
CaCO3 , provide a representative example of the pattern
of total lead release as a function of the initial pH value.
Overall, lead levels increased rapidly during the first 4 hr
of stagnation; however, they appear to fluctuate around
an “equilibrium” state after 24 hr, at all pH levels tested,
except at the highest pH of 9.0. This denotes that a pH
higher than 8.0 may limit the leaching potential of lead.
Similar trends were observed in the tests with higher ini-
tial alkalinity content as well (i.e., of 125 and 150 mg/L as
CaCO3 ), suggesting that a moderate increase in alkalinity
has practically no effect on lead leaching. The pattern of
lead release observed in this study is typical of that reported
in the literature.[13,33,34]
The effect of stagnation time on copper leaching was
found to be similar to that described for lead (Fig. 5). That
is, a substantial increased in copper release was observed
within the first 24 hr, which was followed by an “equilib-
rium” state. The leaching of copper was more pronounced
at the lower pH values investigated (i.e. 7.0 and 7.5), be-
coming practically negligible at pH 9.
During the bench-scale experiments, DO, free chlorine
residual and temperature were also monitored. The initial
DO values varied between 6.7 to 7.2 mg/L, while the fi-
nal DO at the end of the 72 hr period was in the range
of 5.9 to 6.2 mg/L, while the temperature fluctuation was
minimal (17 ± 1◦ C) throughout the tests. The decrease
in DO concentration over the stagnation period followed
first-order kinetics (Fig. 6), with an average rate constant
k DO of 0.0036 ± 0.0004 hr −1 . Moreover, chlorine decay
was also found to follow a first-order reaction (kCl = 0.16
± 0.01 hr−1 ), with no free chlorine residual detected af-
ter 24 hr of stagnation. Both the DO and chlorine deple-
tion patterns observed are consistent with several other
studies.[35,36]
10 15 20 25 30 35 40 45 50 55 60 65 70 75
Stagnation Time (hr)
pH 7.0 pH 7.5 pH 8.0 pH 9.0
 Corrosion control in water supply systems
2 concentrations were significantly lower, accounting for an
1.95
1.9
1.85
ln(D O )
1.8
1.75
1.7
2
R = 0.9728
1.65
0 10 20 30 40 50 60 70
Downloaded by [University of California, San Diego] at 05:46 13 June 2015
Time (hr)
Fig. 6. Typical dissolved oxygen (DO) depletion profile .
Impact of pH and alkalinity on lead release
The total lead concentrations as a function of pH and al-
kalinity, at the end of the 72-hr stagnation period, have
been summarized in Fig. 7. Results demonstrated that at
the lower pH values tested (i.e., 7.0 and 7.5), an increase in
alkalinity resulted in a substantial increase in lead release.
For instance, at pH 7.0, samples with alkalinity in the range
of 100 to 150 mg/L as CaCO3 , exhibited a lead concentra-
tion 3 to 5 times higher than that of the control, which had
an alkalinity of 20 mg/L as CaCO3 . The relative increase in
lead leaching became less pronounced as the pH increased,
and it was practically eliminated at pH 9.0. This suggests
that at pH around neutrality, an increase in alkalinity pro-
motes lead dissolution and corrosion, while at pH > 8.0
the effect of alkalinity on lead leaching is marginal. With
respect to soluble lead behavior, a trend similar to the one
just described was observed. Although the corresponding
Fig. 7. Total lead concentration as a function of pH and alkalinity
at a stagnation time of 72 hr.
1257
average of 11.6 ± 7.1% of the total lead content, almost
all samples (expect the control ones) were above the MAV
limit of 10 µg/L.[10] It is therefore apparent that only a small
fraction of the total lead was in the dissolved state which
indicates that the presence of insoluble lead compounds
play an important role in water chemistry, as explained in
the next paragraph.
The amount of lead in drinking water depends upon
the solubility of several lead corrosion by-products such as
cerussite [PbCO3 ], hydrocerussite [Pb3 (OH)2 (CO3 )2 ] and
lead hydroxide [Pb(OH)2 ].[26,37,38] Although any of these
compounds can play an important role under a given pH
80 and alkalinity range, it has been reported that hydrocerus-
site is the dominant corrosion by-product formed in drink-
ing water supplies.[29,39] According to the Pourbaix dia-
gram, hydrocerussite is formed under typical conditions
existing in drinking water systems (i.e., pH 7.0 to 9.0 and
alkalinity 20 to 200 mg/L as CaCO3 ).[40] Moreover, con-
sidering the consistency of the experimental trends with
the theoretical solubility model, which states that increased
lead solubility occurs at higher alkalinity values, it is rea-
sonable to postulate that Pb3 (OH)2 (CO3 )2 may have played
a predominant part in controlling lead dissolution during
this study.
Impact of pH and alkalinity on copper release
Figure 8 illustrates the total copper concentrations as a
function of pH and alkalinity, at the end of the 72-hr stag-
nation period. As was observed in the case of lead, at an
“elevated” alkalinity ranging between 100 and 150 mg/L as
CaCO3 , copper leaching displayed its highest potential at
pH 7. At a given alkalinity level, the copper concentrations
measured declined progressively as the pH increased and
reached their lowest values at pH 9.0. It is apparent that a
Fig. 8. Total copper concentration as a function of pH and alka-
linity at a stagnation time of 72 hr.
 1258 Tam and Elefsiniotis

combination of pH around neutrality and alkalinity in the Table 4. Percent decrease in total lead concentration due to or-
100 to 150 mg/L as CaCO3 range can induce an adverse im- thophosphate addition as a function of pH and alkalinity.
pact on drinking water quality due to copper dissolution.
Decrease in total Pb concentration (%)
This phenomenon can be attributed to the formation of
soluble cupric bicarbonate [Cu(HCO3 )] and other carbon- Alkalinity (mg/L as CaCO3 )
ate complexes.[41−43] Regarding the soluble copper content,
it averaged 28.9 ± 12.1% of the total copper. Although pH 20 (Control) 100 125 150 Mean value
soluble copper appeared to be a comparatively “minor”
7.0 57.5 42.7 17.1 26.5 36.0 ± 15.4
component in the water; however, expressed as a percent- 7.5 88.2 73.9 67.9 64.5 73.6 ± 9.1
age, it was more than twice the corresponding soluble lead 8.0 81.4 67.6 60.5 68.8 70.0 ± 7.5
value of 11.6%. The concentration of copper detected in
oxidizing systems depends on the water chemistry and the
relative stabilities of copper (I) and copper (II) aqueous
species. According to the Pourbaix diagram, within the that the presence of orthophosphate reduced remarkably
typical pH range between 6.5 to 8.0 for water supplies, the lead dissolution potential.
Downloaded by [University of California, San Diego] at 05:46 13 June 2015

the principal stable copper species are cuprite [Cu2 O] and
malachite [Cu2 (OH)2 CO3 ].[44]
Effect of orthophosphate addition on lead and copper
release
The effect of orthophosphate addition (0.8 mg/L as P)
as a corrosion inhibitor was investigated performing a se-
ries of tests, at the same conditions as described previously
(Table 2). Since it was found that there was a limited metal
leaching potential at pH 9.0 (Figs. 7 and 8), no further ex-
periments at this pH value were conducted. The total lead
concentrations as a function of pH and alkalinity at the
end of the 72-hr stagnation period are depicted in Fig. 9.
Overall, the pattern was similar to that observed without or-
thophosphate addition (Fig. 7), with lead reaching a peak
at the lowest pH of 7.0 and the highest alkalinity of 150
mg/L as CaCO3 tested. Comparing the relative concentra-
tion values in Figs. 7 and 9, however, it becomes apparent
To further elucidate and quantify the performance of the
inhibitor, the percent decrease in total lead concentration
was computed at each condition studied, and the results
have been summarized in Table 4. Orthophosphate was
found to be much more effective at pH 7.5 or higher, result-
ing in approximately 70% corrosion reduction. In contrast,
at a neutral pH, its effectiveness was reduced to almost half
of the above value (i.e., a 36% decrease). The variation in
alkalinity had a rather moderate effect on corrosion, with
the control samples demonstrating the greatest improve-
ment at all pH values investigated. In addition, the soluble
lead concentrations were below the MAV limit of 10 µg/L
in all samples analyzed,[10] and accounted for only 2.4 ±
1.9% of the total lead content.
Figure 10 displays the total copper concentrations of the
tests with orthophosphate addition. As expected, the be-
havior was comparable to the one followed by lead (Fig. 9),
showing a sharp decrease in copper release as the pH in-
creased in the high alkalinity samples, and very low levels
of copper in the control samples. The percent decrease in

Fig. 10. Total copper concentration as a function of pH and

Fig. 9. Total lead concentration as a function of pH and alkalinity alkalinity at a stagnation time of 72 hr (with orthophosphate
at a stagnation time of 72 hr (with orthophosphate addition). addition).
 Corrosion control in water supply systems
Table 5. Percent decrease in total copper concentration due to
orthophosphate addition as a function of pH and alkalinity.
Decrease in total Cu concentration (%)
Alkalinity (mg/L as CaCO3 )
pH 20 (Control) 100 125 150 Mean value
7.0 11.5 48.1 9.7 40.9 27.6 ± 17.2
7.5 66.3 72.7 56.4 72.7 67.0 ± 6.7
8.0 77.3 68.8 52.7 70.9 67.4 ± 9.1
total copper as a result of orthophosphate addition has
been included in Table 5. It is clear that application of the
Downloaded by [University of California, San Diego] at 05:46 13 June 2015
inhibitor reduced significantly the corrosion potential, by
an average of 67% at pH 7.5 and 8.0, and by 28% at pH 7.0.
Furthermore, soluble copper averaged at 21.9 ± 6.6% of
the total copper detected, a slightly lower value than that
observed without orthophosphate addition.
According to solubility models for lead and copper, the
optimal pH for orthophosphate film formation ranges be-
tween 7.0 and 8.0 on lead surfaces, and between 6.5 and
7.5 on copper surfaces.[8,41,45] Although the findings from
this investigation are in agreement with respect to the opti-
mal pH range indicated for lead, in the case of copper, or-
thophosphate addition exhibited the highest performance
at pH 8.0, which falls outside the predicted range. This
study explored the corrosion pattern at a 72 hr stagnation
period; however, it is possible that metal dissolution behav-
ior could differ in the long term. As a final note, it has been
reported that although the use of orthophosphate may re-
duce copper levels in the short term, in the long term the
formation of more stable protective films such as those con-
taining malachite [Cu2 (OH)2 CO3 ] and tenorite (CuO) may
be prevented.[46−48]
Conclusions
The field investigation indicated that Auckland’s water is
aggressive in nature, with a low alkalinity and hardness
content. Over 90% of the unflushed samples contained lead
above the corresponding MAV of 10 µg/L; however, copper
was consistently well below the regulatory limits. Flushing
reduced metal concentrations by more than 90%, with lead
values below the MAV level. The bench-scale experiments
revealed that the lead and copper leaching behavior fol-
lowed similar trends during the stagnation period, showing
a sharp increase within the first 24 hr and then approaching
an equilibrium state. The relatively small soluble fraction
for each metal (12% for lead, 29% for copper) denotes that
the role of non-soluble compounds is prevalent during dis-
solution. Furthermore, the highest degree of leaching was
consistently observed at the lowest pH 7.0 tested. As pH
increased, a considerable decrease in lead and copper val-
1259
ues was apparent, with the lowest concentration obtained
at pH 9.0. It was also found that dissolution of lead and
copper was intensified with an increase in alkalinity, par-
ticularly at pH ≤ 7.5. The addition of orthophosphate as
corrosion inhibitor was more effective at pH ≥ 7.5 resulting
in approximately 70% reduction in both lead and copper.
Overall, regarding the effect of pH and alkalinity, the pat-
terns were similar to those observed without orthophos-
phate addition, albeit at lower metal concentrations.
Acknowledgments
The authors wish to thank the staff at the Environmen-
tal Engineering, and Environmental Science Laboratories,
University of Auckland as well as at Watercare Service
Ltd. for their assistance throughout this study. Funding
was provided by a Technology in Industry Fellowship and
Watercare Services Ltd.
References
[1] Brongers, M.P.H. Appendix K. Drinking water and sewer systems.
Corrosion control costs and preventive strategies in the United States,
FHWA-RD-01-156, U.S. Department of Transportation Federal
Highway Administration, Houston, TX, 2002; 1–33.
[2] Cantor, A.F.; Bushman, J.B.; Glodoski, M.S.; Kiefer, E.; Bersch,
R.; Wallenkamp, H. Copper pipe failure by microbiologically influ-
enced corrosion. Mater. Perform. 2006, 45(6), 38–41.
[3] Churchill D.M.; Mavinic, D.S.; Neden, D.G.; MacQuarrie, D.M.
The effect of zinc orthophosphate and pH-alkalinity adjustment on
metal levels leached into drinking water. Can. J. Civil Eng. 2000,
27(1), 33–43.
[4] Ahmedna, M.; Marshall, W.E.; Husseiny, A.A.; Rao, R.M.;
Goktepe, I. The use of nutshell carbons in drinking water filters
for removal of trace metals. Water Res. 2004, 38(4), 1062–1068.
[5] Clement, M.; Seux, R.; Rabarot, S. A practical model for estimat-
ing total lead intake from drinking water. Water Res. 2000, 34(5),
1533–1542.
[6] Elci, L.; Kartal, A.A.; Soylak, M. Solid phase extraction method for
the determination of iron, lead and chromium by atomic absorption
spectrometry using Amberite XAD-2000 column in various water
samples. J. Hazard. Mater. 2008, 153(1–2), 454–461.
[7] United States Environmental Protection Agency. Revised Guidance
Manual for Selecting Lead and Copper Control Strategies. Report
No. EPA-816-R-06-001; U.S. Environmental Protection Agency,
Washington, DC, 2003.
[8] Health Canada. Corrosion control in drinking water distribution
systems. Canada, 2007. http://www.hc-sc.gc.ca/ewh-semt/alt
formats/hecs-sesc/pdf/pubs/water-eau/consultation/corrosion/
corrosion-eng.pdf (accessed April 2009).
[9] World Health Organization. Guidelines for Drinking-water Quality
3rd Ed. Volume 1: Recommendations. WHO, Geneva, Switzerland,
2004; 145–196.
[10] New Zealand Ministry of Health. Drinking-water Standards for
New Zealand 2005, The Ministry of Health: Wellington, New
Zealand, 2005.
[11] Cardew, P.T. A method for assessing the effect of water quality
changes on plumbosolvency using random daytime sampling. Water
Res. 2003, 37(12), 2821–2832.
 1260
[12] Xiao, W.; Hong, S.; Tang, Z.; Taylor, J.S. Effects of blending on total
copper release in distribution systems. J. Am. Water Works Assoc.
2007, 99(1), 78–88.
[13] Boyd, G.R.; Shetty, P.; Sandvig, A.M.; Pierson, G.L. Lead in tap
water following simulated partial lead pipe replacements. J. Environ.
Eng. 2004, 130(10), 1188–1197.
[14] Melidis, P.; Sanozidou, M.; Mandusa, A.; Ouzounis, K. Corrosion
control by using indirect methods. Desalination, 2007, 213(1–3),
152–158.
[15] Prisyazhniuk, V. Prognosticating scale-forming properties of water.
Appl. Thermal Eng. 2007, 27(8–9), 1637–1641.
[16] Al-Rawajfeh, A.E.; Al-Shamaileh, E.M. Inhibition of corrosion
in steel water pipes by ammonium pyrrolidine dithiocarbamate
(APDTC). Desalination, 2007, 206(1–3), 169–178.
[17] Chan, M.W.M. A critical study on corrosion and corrosion control
methods for piping systems in Hong Kong. Ph.D. Thesis. The Hong
Kong Polytechnic University, Department of Building Services En-
Downloaded by [University of California, San Diego] at 05:46 13 June 2015
gineering, Hong Kong, 2003.
[18] Edwards, M.; McNeill, L.S. Effect of phosphate inhibitors on
lead release from pipes. J. Am. Water Works Assoc. 2002, 94(1),
79–90.
[19] Edwards, M. Controlling corrosion in drinking water distribution
systems: a grand challenge for the 21st century. Water Sci. Technol.
2004, 49(2), 1–8.
[20] McNeil, L.S.; Edwards, M. Phosphate inhibitors and red water stag-
nant pipes. J. Environ. Eng. 2000, 126(12), 1096–1102. 1
[21] Becker, A. The effect of corrosion inhibitors in drinking water in-
stallations of copper. Mater. Corros. 2002, 53(8), 560–567.
[22] Ebrahimi Mehr, M.; Shahrabi, T.; Hosseini, M.G. Determination of
suitable corrosion inhibitor formulation for a potable water supply.
Anti-Corros. Methods Mater. 2004, 51(6), 399–405.
[23] Tam, Y.S. Investigation into the options to control plumbosolvency
in drinking water. Master of Engineering Thesis. University of
Auckland, Department of Civil and Environmental Engineering,
Auckland, New Zealand, 2007.
[24] HACH Water Analysis Handbook-ISO9001 Certified. HACH
Company, Loveland, CO, USA, 1997.
[25] Berthouex, P.M.; Brown, L.C. Statistics for Environmental Engi-
neers. 2nd Edition, Lewis Publishers, Boca raton, FL, USA, 2002.
[26] Schock, M.R. Response of lead solubility to dissolved carbon-
ate in drinking water. J. Am. Water Works Assoc. 1980, 72(12),
695–704.
[27] Dodrill, D.M.; Edwards, M. Corrosion control on the basis
of utility experience. J. Am. Water Works Assoc. 1995, 87(7),
74–85.
[28] Benjamin, M.M.; Sontheimer, H.; Leroy, P. Corrosion of iron and
steel. In: Internal Corrosion of Water Distribution Systems, 2nd
Ed.; AWWA Research Foundation and DVGW Technologiezen-
trum Wasser: Denver, CO, 1996; 29–68.
[29] Tang, Z. Effect of source water blending on iron and lead release:
thermodynamic and statistical modeling. Ph.D. Thesis. University
of Central Florida, Civil and Environmental Department, Orlando,
FL, 2003.
[30] Volk, C.J. Corrosion control in drinking water systems. In: Water
Encyclopedia, John Wiley & Sons, Inc.: New York, 2005; 5–11.
Tam and Elefsiniotis
[31] Korshin, G.V.; Ferguson, J.F.; Lancaster, A.N. Influence of natural
organic matter on the corrosion of lead brass in potable water.
Corros. Sci. 2000, 42(1), 53–56.
[32] Korshin, G.V.; Ferguson, J.F.; Lancaster, A.N. Influence of natu-
ral organic matter on the morphology of corroding lead surfaces
and behavior of lead-containing particles. Water Res. 2005, 39(5),
811–818.
[33] Britten, A.; Richards, W.N. Factors influencing plumbosolvency in
Scotland. J. Instit. Water Eng. Sci. 1981, 35(4), 349–364.
[34] Hulsmann, A.D. Particulate lead in water supplies. Journal of the
Institution of Water and Environ. Manage. 1990, 4, 19–25.
[35] Boulay, N.; Edwards, M. Role of temperature, chlorine and organic
matter in copper corrosion by-product release in soft water. Water
Res. 2001, 35(3), 685–690.
[36] Huang, Y.H.; Zhang, T.C. Effects of dissolved oxygen on formation
of corrosion products and concomitant oxygen and nitrate reduc-
tion in zero-valent iron systems with or without aqueous Fe2+ .
Water Res. 2005, 39(9), 1751–1760.
[37] Leroy, P. Lead in drinking water—Origins; solubility; treatment. J.
Water Supply: Res. Technol.—Aqua, 1993, 42(4), 223–238.
[38] Miranda M.L.; Kim, D.; Hull, A.P.; Paul, C.J.; Overstreet Galeano,
M.A. Changes in blood lead levels associated with use of chlo-
ramines in water treatment systems. Environ. Health Perspect. 2007,
115(2), 221–225.
[39] Schock, M.R.; Wagner, I.; Oliphant, R.J. The corrosion and sol-
ubility of lead in drinking water.” In: Internal Corrosion of Water
Distribution Systems. 2nd Ed.; AWWA Research Foundation and
DVGW Technologiezentrum Wasser, Denver, CO, 1996, 213–308.
[40] Conio, O.; Ottaviani, M.; Formentera, V.; Lasagna, C.; Palumbo,
F. Evaluation of the lead content in water for human consumption.
Microchem. J. 1996, 54(4), 355–359.
[41] United States Environmental Protection Agency. Effect of pH, DIC,
orthophosphate and sulfate on drinking water cuprosolvency, Report
No. EPA/600/R-95/085; U.S. Environmental Protection Agency:
Cincinnati, OH, 1995.
[42] Edwards, M.; Jacobs, S.; Dodrill, D.M. Desktop guidance for miti-
gating Lead and Copper corrosion by-products. J. Am. Water Works
Assoc. 1999, 91(5), 66–77.
[43] Letelier, M.V.; Lagos, G.E.; Reyes, A. Chemical characterization
of blue stains in domestic fixtures in contact with drinking water.
Environ. Monitor. Assess. 2008, 139(1–3), 307–315.
[44] Merkel, T.H.; Grosz H.-J.; Werner W.; Dahlke T.; Reicherter
S.; Beuchle G.; Eberle S.H. Copper corrosion by-product re-
lease in long-term stagnation experiments. Water Res. 2002, 36(6),
1547–1555.
[45] Schock, M.R. Understanding corrosion control strategies for lead.
J. Am. Water Works Assoc. 1989, 81(7), 88–100.
[46] Schock, M.R.; Clement, J.A. Lead and copper control with non-
zinc orthophosphate. J. New England Water Works Assoc. 1998,
112(1), 20–33.
[47] Edwards, M.; Sprague, N. Organic matter and copper corrosion
by-product release: A mechanistic study. Corros. Sci. 2001, 43(1),
1–18.
[48] Cantor, A.F.; Park, J.K.; Vaiyavatjamai, P. Effect of chlorine on
corrosion. J. Am. Water Works Assoc. 2003, 95(5), 112–123.