EPA-670/2-75-036
May 1975
Environmental Protection Technology Series
STUDY OF CORROSION PRODUCTS IN THE
SEATTLE WATER DEPARTMENT TOLT
DISTRIBUTION SYSTEM
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
EPA-670/2-75-036
May 1975
STUDY OF CORROSION PRODUCTS IN THE
SEATTLE WATER DEPARTMENT TOLT DISTRIBUTION SYSTEM
By
Robert A. Dangel
Water Supply Research Laboratory
Program Element No. 1CB047
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
REVIEW NOTICE
This report has been reviewed by the National Environmental Research Center,
Cincinnati, and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for
use.
ii
-------
FOREWORD
Man and his environment must be protected from the adverse effects of pest-
icides, radiation, noise and other forms of pollution, and the unwise
management of solid waste. Efforts to protect the environment require a focus
that recognizes the interplay between the components of our physical environ-
ment—air, water, and land. The National Environmental Research Centers
provide this multidisciplinary focus through programs engaged in
• studies on the effects of environmental
contaminants on man and the biosphere, and
• a search for ways to prevent contamination
and to recycle valuable resources.
This study examined the changes in potable water quality in a large urban
distribution system. The data supports the conclusion that corrosion control
is needed to reduce the pick-up of metals by the water.
A. W. Breidenbach, Ph.D.
Director
National Environmental
Research Center, Cincinnati
iii
-------
ABSTRACT
Samples from the Seattle Water Department's Tolt distribution system were
analyzed for chemical and bacteriological parameters. Changes from the raw
water quality were observed, particularly in trace metal concentrations and
other parameters related to corrosion. Distribution mains were found to be
adequately protected from corrosion by cement and bituminous linings whereas
service lines and household plumbing were actively corroded.
Metals in the ug/1 concentration range were determined by a flameless atomic
absorption technique.
IV
-------
CONTENTS
FOREWORD iii
ABSTRACT iv
ACKNOWLEDGEMENT vii
CONCLUSIONS 1
INTRODUCTION 2
SAMPLING PROCEDURE 3
ANALYTICAL METHODS 4
RESULTS 5
DISCUSSION 12
COMPARISON OF SAMPLE VALUES WITH THE U.S. P.H.S.
DRINKING WATER STANDARDS 16
APPENDIX 17
REFERENCES 21
-------
TABLES
1 Recovery of Dissolved Metals added to Distilled Water
and a Composite Seattle Tap Water by Conventional and
Heated Graphite Atomizer Atomic Absoprtion 6
2 Wet Chemistry Data 7
3 Metals Data g
4 Additional Data and Comparison to U. S. P.H.S. Standards 9
5 Notes and Pipe Data 10
6 Means and Maximums of Running and Standing Samples
Compared to Raw Water Values 11
7 Comparison of Corrosion Products in Running Samples
Taken from Mains and Buildings !4
8 Water Quality Changes attributed to Asbestos-Cement Pipe 15
9 Number of Samples Exceeding P.H.S. Drinking Water Standards 16
10 Typical Furnace Parameters 20
vi
-------
ACKNOWLEDGEMENT
The Seattle Water Department was extremely helpful in selecting sample sites,
collecting samples and reviewing the report. Thanks especially to John
Courchene, Brian Hoyt, and James Chapman, Water Quality Division. Dr. Marvin
Gardels, EPA Water Supply Research Laboratory contributed to many parts
of the project; G. J. Vasconcelos, Northwest Water Supply Research Lab-
oratory, analyzed the bacteriological samples; William Mullen, EPA Region X,
and Dr. James Symons, EPA NERC-Cincinnati, contributed to the review of
the report; Dorothy Jacobs and Janet Gavin, EPA Region I, typed the manu-
script.
The author is currently Fluoridation Engineer, Water Supply Branch, EPA
Region I, JFK Federal Building, Boston, Massachusetts 02203.
vii
-------
CONCLUSIONS
When the iron, copper, zinc, lead, and cadmium means for the standing
samples are compared with the raw water values (Table 6), the corrosiveness
of Tolt water is confirmed. Comparison of the running and standing means
indicates that most of the metal pick-up is occurring in the service lines
and plumbing inside buildings. This occurs despite the fact the residence
time in the distribution system is about 1 week and only overnight in the
building plumbing. The data also indicate that the distribution mains are
adequately protected by their cement and bituminous linings.
A combination of an alkalinity increase of 1 mg/1 (as CaCO~) and a turbidity
increase of 0.1 FTU from running to standing samples is indicative of
corrosion. In the absence of a metal analysis, this could be used as a
qualitative index.
Laboratory analysis of corrosion products correlates well .with the mater-
ials in contact with the water. Buried pipe could be identified by comparing
the influent and effluent metal concentrations.
Tolt water passing through asbestos-cement pipe exhibited radical changes in
pH, alkalinity, calcium, and conductivity, which increased with longer
exposure to the pipe. Although the samples are within the limits of the
P.M.S. Standards, they are indicative of rapid pipe wear. As the cement bind-
er is dissolved, asbestos fibers may possibly be leaching from the pipe
walls. The water quality changes in asbestos-cement pipe are in marked
contrast to the inertness of cement-lined cast iron pipe in contact with
Tolt water.
Standing samples were collected without regard to the number of hours the
water was exposed to the building plumbing. Further study is needed to
determine whether a standard residence period is necessary to ensure
reproducible data.
Bacteriological samples were part of the standing group. They contained
corrosion products which might have inhibited bacterial growth. Further
work should include running samples to determine whether the organisms
detected in the standard plate count grow primarily in the distribution
system or in the building plumbing.
The data and conclusions presented in this report should help in further
study of both the Tolt and Cedar distribution systems. Also, they will
provide baseline data for future evaluation of corrosion control chemicals
in the Tolt distribution system.
-------
INTRODUCTION
Seattle obtains its water from two surface sources, the Cedar River and the
south fork of the Tolt River. The watersheds are in mountainous areas closed
to public access.
Before collection, the runoff has a short contact time with the soil. The
hardness, alkalinity, salinity, and trace metal content of these waters is
remarkably low and the dissolved oxygen content approaches saturation. The
resulting water is an excellent solvent and exhibits aggressive corrosion
tendencies.
Treatment consists of screening, gaseous chlorination, fluorldation
(25% H2SiFg), and rechlorination at distribution reservoir outlets. The
addition of these chemicals to the recommended levels causes an alkalinity
decrease of 2.5 to 3.5 mg/1 as CaCo^ , and lowers the pH 0.2 to 0.4 units
in Cedar water and 0.4 to 0.8 units in Tolt water. Besides transforming
bicarbonate alkalinity to carbonic acid, the hypochlorous acid and hypochlorite
ions may increase the water's activity on metal because of their oxidizing
power.
Seattle, recognizing the corrosive tendencies of its water, requested
technical assistance from the U. S. Environmental Protection Agency to
determine the severity and location of the corrosion problem. The avail-
ability of the facilities at the Northwest Water Supply Research Laboratory
made the study possible.
-------
SAMPLING PROCEDURE
The Tolt distribution system was chosen for this study because it accounts
for a larger per capita percentage of the red water complaints than does the
Cedar system. The Tolt's lower pH, alkalinity, and hardness also make it
the more corrosive of the two supplies. At the time of the study, the
Tolt supplied about one third of the water used in Seattle.
The water mains in the city were predominately cement and bituminous lined
and were assumed to be adequately protected from the corrosive waters.
Service connections and residential plumbing were not similarly protected
and were assumed to be the source of the< corrosion products. To test this
hypothesis, two types of samples were collected in the early morning hours
from the source, transmission mains, distribution system, and residences.
Standing samples collected from the first water to run out of the faucet
represented water in contact with the household piping for at least one
night. Running samples collected after a 30-second bleeding of the lines
represented water from the mains.
The standing sample consisted of a quart of water to be used for deter-
mining trace metals, another quart for wet chemistry, and then smaller
amounts for dissolved oxygen and bacteriological analyses. The water was
then allowed to run before additional quarts were taken for trace metals
and wet chemistry analyses. Concentrated nitric acid (1.5 ml) was added
to the trace metal samples as a preservative.
Bacteriological testing was not the major thrust of this study, and one set
of tests per location was believed to be adequate.
Samples were collected in the early morning hours by members of the Seattle
Water Department. Numbers 1 through 4 and 6 through 12 were collected
on October 12, 1972. Numbers 5 and 13 through 34 were collected on October
26, 1972. Analysis of both sets was completed on December 8, 1972.
-------
ANALYTICAL METHODS
The chemical constituents chosen for analysis were those directly related
to corrosion or corrosion products. In addition, bacteriological and
fluoride determinations were performed. In the preliminary work, chromium,
nickel, and cobalt were found to be below JQug/1 and, therefore, were not
included in the analysis sequence. Residual chlorine concentration was not
measured because it fluctuates too widely to be correlated with corrosion
data.
The procedures selected for the analytical methods were:
Temperature was measured at the time the sample was collected.
Dissolved oxygen was determined by the Azide modification of the Winkler
method in Standard Methods. '*•' Analyses were performed by the Seattle
Water Department.
Conductivity was determined with an A. R. Thomas Model 15B1 Serfass
conductivity bridge and a Beckman 0.1 factor conductivity cell.
Color rarely matched the color standards of the Hellige Aqua Tester.
Where possible, color densities were measured. Otherwise, samples appearing
different from the distilled water blank, when compared in matched 50-ml
Nessler tubes, were recorded as having color.
Fluoride was determined with a Corning fluoride ion electrode on a Corning
Model 101 meter. All samples were mixed 1:1 with TISAB buffer. Commercial
standards were used.
pH was determined with a glass electrode and silver/silver chloride ref-
erence on a Corning Model 101 meter. The low buffering capacity of the
water made the determination difficult. A sample at pH 5.4 would drift to
pH 6.0 in 10 to 30 minutes while being stirred in an open beaker. This drift
indicated a loss of carbon dioxide.
Total alkalinity was determined by the potentiometric method, titrating
to pH 4.5 and 4.2 with 0.0020N HCl. A reagent one-tenth the recommended
strength was needed because of the low aIkaUnities.
Turbidities were determined with a Hach 2100A unit. Calibration was ob-
tained from standards provided with the instrument at 0.61 and 10 Formazine
turbidity units. The values were not corrected for the background light
scattering of the blank which was about 0.04 FTU.
-------
Chloride was determined with a Corning chloride ion electrode and a Corning
Model 101 meter. Standards were prepared by diluting commercial products.
Bacteriological samples were analyzed by Standard Methods techniques. The
plate count is reported as organisms/ml, incubated at 35°C. Coliform
and pseudomonas are reported as organisms/100 ml.
Metals were determined by flame atomic absorption and graphite furnace
atomic absorption on a Perkin Elmer 303, with a strip chart recorder.
Standards were obtained by diluting commercial products. Values obtained in
percent absorption were converted to absorbance and then to concentration
from a parabolic calibration curve by computer. Cadmium, lead, manganese,
and the lower values for copper and iron were obtained by the furnace
technique, with the Perkin-Elmer HGA 2000 Heated Graphite Atomizer. Calcium,
magnesium, zinc, high copper, and high iron values were obtained by the
flame technique. The detection limits for zinc and cadmium were 0.015 mg/1
and 0.4 ug/1 respectively. A lanthanum chloride solution was added to the
calcium and magnesium samples and standards to eliminate chemical inter-
ferences.
Trace metal standards from the Methods Development and Quality Assurance
Research Laboratory, National Environmental Research Center, Cincinnati,
were analyzed in conjunction with this study to check the accuracy of the
determinations. Six concentrates were used to spike deionized water and
also a composite of Seattle tap water which was made by combining a number of
the standing samples. Both sets were analyzed by flame and furnace atomic
absorption. The deionized water contained no background metals, and the
values obtained were solely from the spike. The results for the spike in
the tap water samples were obtained by subtracting the metal concentration
of the tap water from the total of the tap water and the spike. For these
waters, the accuracy of the furnace technique is excellent, even at the
1 pg/1 level (as shown in Table 1). A separate report describing the tech-
nique and operating parameters for the graphite furnace can be found in the
appendix.
RESULTS
The data are presented in Tables 2 through 5 and in Table 6 are the means
and maximums. Values are reported as mg/1 or yg/1, depending on the range
of concentrations. "S" is for standing and "R" is for running samples.
Standing samples were not collected from sample stations numbers 2 through 4,
7 through 12, 16, 18, and 28. These locations were taps directly from
distribution mains and other places that did not have significant volumes
of water in long contact with the piping materials.
-------
TABLE 1
RECOVERY OF DISSOLVED METALS ADDED TO DISTILLED WATER AND A
COMPOSITE SEATTLE TAP WATER BY CONVENTIONAL AND HEATED
6RAPHITE ATOMIZER ATOMIC ABSORPTION
Metal Known Flamec Furnace" Known Flame Furnace
DWe TwlS DW TO
SAMPLE 1
Cd 71 100 100 72 75 1
Cr 370 370 370 408 408 7
Cu 302 270 140 301 339 7
Fe 840 800 860 880 8339 24
Pb 367 480 480 295 301 37
Zn 281 297 267 7
SAMPLE 2
Cd 14 N N 13 14 2
Cr 74 90 90 77 74 15
Cu 60 60 -30 56 41 12
Fe 350 220. 280 336 339 10
Pb 101 hdl^ bdl 90 86 25
Zn 70 57 44 11
DW TW DW TW
SAMPLE 5
.4 N*1 N 1-4 1.8
.4 bdl bdl 8 7
.5 40 30 77
bdl bdl 22 23
bdl bdl 36 38
.0 12 3
SAMPLE 6
.8 N N 2.5 2.9
bdl bdl 15 15
bdl bdl 11 12
bdl bdl 9 1
bdl bdl 25 24
14 7
SAMPLE 3 BACKGROUND LEVEL IN COMPOSITE TAP WATER
Cd 18 N N 17 18
Cr 93 90 90 100 95
Cu 75 80 -10 78 59
Fe 438 420 430 422 449
Pb 84 bdl bdl 75 74
Zn 70 77 59
SAMPLE 4
Cd 78 100 100 79 84
Cr 407 420 420 475 475
Cu 332 300 170 364 383
Fe 700 670 730 682 709^
Pb 334 310 310 275 270
Zn 310 328 300
a All values in pg/1
b Amount added to DU and TW samples
C Conventional atomic absorption
d Heated graphite atomizer atomic absorption (HGA
e Recovery of metal added to deionized water
]$ Recovery of metal added to composite Seattle tap
the background levels shown in the Table
g Total value above range of HGA 2000. Flame value
h Signal too noisy to read
Ji Below detection limit
bdl 0.8
bdl 0.7
470 550
250 272
bdl 4.9
606
2000)
water in the presence of
substitued in the subtraction.
-------
TABLE 2 - WET CHEMISTRY DATA
Sample
Number Sample Site
1
2
3
4
5'
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Hoyt, 10602 NE 137 PI, Kirkland
Tolt Regulating Basin, Before
Treatment
Tolt Treatment Bldg, After Cl2,F-
Tolt Pipeline Air Valve #9
Seeklander, 7015 14th NW
Duvall Shops
Tolt pipeline, Air valve #21
" " Air valve #24
" " LK Forest Pk Res.
" " NE 195 & 35 NE
Foy Pump Station
North Gate Pump Station
Jessup, 625 N. 180th
Duvall, Fire Station
Bulter, Duvall
N. City Pump Station
Richmond Beach Library
Sample Station
Taylor, 19527 Stone Ave. N
Courchene, 1622 N 51st
Lehman, 116 N 78th
Brehan, 8526 19th Ave. NW
Moore, 935 N 128th
Larson, 10041 14 Ave. NW
Buckingham, 12733 8th NW
Thompson, 8751 16th NW
Schwind, 8351 22nd NW
Bitter LK Res. Sample Station
Vining, 9523 Evanston Ave. N
Scholz, 11727 Corliss Ave. N
Fanson, 12336 3rd Ave. NE
Bringhurst, 843 NE 78 NE
Broswick, 537 NE 81st
Philbrick, 11702 22nd Ave. NE
High
Low
Mean
% Relative Std Deviation
Std Deviation
a Running Sample
b Standing Sample
Temperature,
"c
12
12
13
13
14
13
13
13
13
13
13
13
12
12
11
12
14
14
12
15
13
14
13.5
13
13
13
13.5
12
12.5
12.5
13
13
13
12
14
11
13
Su
18
15
12
16
10
14
13
16
18
17
17
10
17
13.5
19.5
13.5
12.5
22
13.5
14
14
19.5
10
15
D.O.,
mg/1
R
10.5
10.6
10.7
10.5
10.0
10.3
10.5
10.4
10.4
10.4
10.5
10.5
10.7
8.7
10.6
10.8
10.1
9.9
8.3
10.4
10.5
10.5
10.8
10.3
10.2
10.3
10.0
10.4
10.5
10.7
10.3
10.7
10.6
10.6
10.8
8.3
10.3
5.0
0.51
S
9.2
9.0
10.0
5.6
8.9
9.6
7.1
7.0
5.2
9.7
9.6
9.5
7.8
10.2
7.6
4,7
10.2
4.9
9.9
10.3
6.4
10.3
4.7
8.2
23.4
1.9
Conductivity,
urnho, 20 "C
R
22
18
20
20
21
23
21
21
20
20
21
20
22
42
25
18
23
23
24
21
22
23
21
22
22
22
21
23
22
21
22
21
21
20
42
18
22
19
4.2
S
24
21
23
29
42
28
28
27
21
24
23
24
26
33
23
27
23
22
26
22
22
22
42
21
25
19
4.8
Turbidity,
FTU
R
0.48
0.45
0.43
0.72
0.41
0.50
1.8
2.4
0.45
0.45
0.47
0.43
0.45
1.0
0.47
0.48
0.48
1.2
0.90
0.47
0.50
0.44
0.54
0.48
0.46
0.58
0.52
0.49
0.52
0.46
0.50
0.46
0.46
0.48
2.4
0.41
0.63
67
0.42
S
0.58
0.45
0.62
1.8
0.94
13
0.60
1.3
0.88
3.3
0.50
28
0.84
6.3
0.83
1.3
3.8
0.50
2.8
0.67
0.55
3.0
28
0.45
3.3
188
6.2
pH
R
5.3
5.4
5.5
5.3
6.0
5.2
5.4
5.3
5.8
5.1
5.3
5.3
5.9
8.9
5.9
6.0
5.9
5.9
6.3
5.7
5.7
5.8
5.6
5.7
5.8
5.9
5.8
5.8
5.8
5.7
5.7
5.6
5.7
6.0
8.9
5.1
5.8
S
5.6
5.8
5.4
5.8
9.1
5.8
6.2
6.7
5.9
6.0
5.8
6.0
6.0
6.1
5.8
6.1
5.9
5.8
5.7
5.7
5.6
5.7
9.1
5.4
6.1
Alkalinity,
mg/1 CaCOi
R
2.3
\tt
•x/2
0.9
1.1
2.7
^2
0.9
.2.4
0.9
1.3
1.4
2.6
i/14
4.1
2.8
3.5
2.5
4.9
1.1
2.2
2.4
1.1
1.6
1.9
2.5
1.7
2.2
2.5
2.0
2.0
1.7
1.6
2.3
14
0.9
2.5
89
2.2
S
4.2
2.0
3.3
7.2
VL4
6.6
6.8
6.3
2.2
4.3
3.6
2.1
5.4
^6
2.7
5.6
3.4
2.0
4.5
3.0
1.4
2.5
14
1.4
4.5
61
2.8
Chloride,
mg/1
R
1.6
1.0
1.7
1.6
2.6
1.7
1.6
1.7
1.6
1.7
1.6
1.6
2.3
2.6
2.4
2.3
2.5
2.2
2.5
2.4
2.5
2.6
2.2
2.7
2.6
2.4
2.3
2.9
2.1
2.2
2.4
2.1
2.1
2.2
2.9
1.0
2.1
21
0.44
S
1.6
2.3
1.7
2.4
2.4
2.3
2.2
2.2
2.4
2.6
2.6
2.2
2.5
4.6
2.6
2.4
2.4
2.3
2.6
2.1
2.2
2.3
4.6
1.6
2.4
23
0.55
-------
TABLE 3 - METALS DATA
00
Sample
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
High
Low
Mean
% Rel Std
Dev
Std Dev
R
0.10
0.002
0.24
0.005
0.069
1.67
0.006
0.008
0.003
0.097
0.049
0.010
0.035
0.007
0.030
0.029
0.12
0.019
0.042
0.082
0.043
0.15
0.072
0.061
0.018
0.021
0.022
0.053
0.17
0.022
0.36
0.23
0.028
0.12
1.67
0.002
0.12
240
0.29
Cu,
ms.fi
S
1.02
0.10
1.92
0.14
0.21
0.028
1.09
0.038
0.24
0.091
0.20
2.05
0.12
0.87
0.080
0.037
0.17
0.14
0.13
1.10
0.12
0.092
2.05
0.028
0.45
134
0.61
Mn
Ugi
r&
10
18
16
16
14
16
45
33
9
9
9
7
10
11
11
13
7
29
23
14
15
7
10
7
6
7
14
8
10
10
15
9
9
13
45
7
14
61
8
»
!l
8
10
19
19
19
6
26
5
47
22
43
12
79
11
20
15
19
20
16
47
19
19
42
79
6
24
71
17
1
i
R
0.16
0.17
0.15
0.27
0.21
0.17
1.1
0.75
0.16
0.14
0.17
0.15
0.21
0.26
0.18
0.19
0.19
0.35
1.2
0.26
0.30
0.19
0.24
0.17
0.15
0.27
0.26
0.16
0.24
0.20
0.36
0.18
0.20
0.22
1.2
0.15
0.28
87
0.24
fe,
Ha/1
S
0.32
0.48
0.53
1.6
0.25
1.5
0.19
0.70
1.1
2.0
1.2
>70
0.68
0.37
0.91
0.48
2.0
1.9
4.2
0.40
5.4
2.3
5.4^
0.19
1.4C
98
1.3
P
V
R
2
1
1
17
4
2
22
•21
1
1
1
2
4
1
2
3
5
2
3
3
12
3
2
11
2
2
2
1
3
2
17
6
13
6
22
1
5
114
6
b,
g/1
9
4
4
36
13
12
26
12
17
16
11
170
71
51
13
17
22
23
170
25
108
26
170
4
39
125
49
Zn
mg>
R
0.05
*
*
*
0.05
*
*
*
*
*
0.03
0.05
0.28
0.79
0.07
*
*
*
1.73
0.08
0.27
0.08
0.11
0.08
0.15
0.03
0.14
*
0.26
0.04
0.55
0.08
0.06
0.11
1.73
<0.015
0.15
)
a
s
0.16
0.31
0.39
3.65
0.09
2.07
0.06
3.42
0.81
2.24
0.79
32.6
2.12
5.46
1.05
2.78
2.33
1.14
4.48
0.62
1.09
1.45
5,46C
0.06
1.74
87
1.5
C
K
R
*
*
*
*
*
*
*
*
*
*
*
*
*
*
A
*
*
*
*
*
0.8
*
*
*
*
*
*
*
0.4
A
0.5
*
A
A
0.8
<0.4
<0.4
d,
fiZ!
s
*
0.7
0.4
1.4
A
0.4
*
0.6
0.8
4.9
0.6
14
0.8
0.5
25
2.1
4.0
0.8
2.4
1.0
4.2
1.8
25
<0.4
2.0
M
m
R
0.31
0.33
0.33
0.32
0.32
0.32
0.32
0.31
0.29
0.29
0.30
0.30
0.28
0.08
0.18
0.28
0.29
0.30
0.29
0.30
0.30
0.26
0.29
0.26
0.26
0.38
0.31
0.26
0.29
0.29
0.30
0.31
0.31
0.29
0.38
0.08
0.29
17
0.05
8t
K/l
S
0.31
0.33
0.37
0.29
0.08
0.21
0.26
0.29
0.32
0.31
0.26
0.27
0.27
0.38
0.24
0.41
0.27
0.29
0.35
0.31
0.30
0.29
0.41
0.08
0.29
23
0.07
C
m
R
2.71
2.07
2.06
2.13
2.66
2.06
2.13
2.13
2.22
2.10
2.11
2.21
2.87
7.27
3.94
2.22
3.03
3.01
2.68
2.58
2.43
3.03
2.53
2.95
2.88
3.22
2.48
2.86
2.76
2.70
2.42
2.38
2.43
2.34
7.27
2.06
2.69
34
0.91
a,
R/l
S
2.66
2.54
2.11
2.64
7.34
3.82
2.13
2.74
2.50
2.41
2.88
2.20
2.88
3.81
2.89
3.24
2.61
2.66
2.80
2.40
2.66
2.78
7.34
2.11
2.94
37
1.08
a Running Sample
b Standing Sample
C Sample 23 S not included in the computation.
A Below detection limit.
-------
TABLE 4 - ADDITIONAL DATA AND COMPARISON TO PHS STANDARDS
Sample
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
High
Low
Mean
% Rel
a
b
Fluoride
Color
IB
20
Yes
Yes
20
45
Yes
Yes
Yes
Yes
Yes
Std Dev
Running sample
Standing sample
ms
F~^
0.95
0.05
0.86
0.88
0.94
0.90
0.92
0.90
0.96
0.98
1.00
1.00
1.08
0.84
0.98
0.10
1.02
0.94
1.00
0.96
1.00
1.00
1.00
1.00
0.92
0.96
0.94
1.00
0.94
1.00
0.96
1.02
0.98
0.54
1.08
0.10
0.90
25
r/1
S
0.94
0.92
0.94
1.00
0.96
1.10
1.02
0.98
1.06
1.00
0.90
0.96
0.94
1.25
0.98
0.94
0.94
0.92
1.06
1.00
1.00
0.94
1.25
0.90
0.99
7.8
Std
Plate
Count Coliform
/ml /100 ml
3 <1
11 <1
5 *1
1 *1
7 *!
25 <1
7 <1
11 <1
3 <1
1 <1
4 <1
1 <1
160 <1
150 <1
700 <1
6 <1
11 <1
4
-------
TABLE 5 - NOTES AND PIPE DATA
Sample
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
H "
0 18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Note*
Ra- S copper pickup
Raw Tolt Water - very clean
Some lead
Mild R - S corrosion
Copper high R - S, galvanized alao
Partially standing sample? Lead, Turbid, Iron
ii ii it M it it
Clean sample
" " , some copper
Clean sample
Clean sample
Large iron, lead & zinc R-S Increase
Mild Increase copper, less zinc R-S; low tig & high Ca in R&S
Large iron, zinc Increase, plate count high
Clean sample
Large R-S copper, lead increase
High turbidity, some iron, partially standing?
High iron, zinc, manganese in both R&S
Large iron, zinc pickup
Large iron, zinc, cadmium increase
1 Increase, copper in both R&S
Worst standing sample; not in zinc -6 iron averages
Large iron, lead, zinc increase
Large copper, lead, zinc pick up. High plate count
Large iron, zinc, cadmium pickup
Large zinc Increase
Clean sample
Large iron, zinc increase, copper in R&S
High plate count, large iron, lead, zinc Increase, copper loss
Large copper pickup, iron, zinc increase
Large iron, lead, zinc increase
Large iron, zinc increase
Assumed
Piping Material
Cuc
Cu
Fe ,
Fe/Znd
Cu,Fe/Zn
Fe,Pbe
Fe.Pb
Cu
Fe/Zn, Pb
Cement, Cu
Fe/Zn
Cu
Fe
Fe/Zn
Fe/Zn, Cu
Fe/Zn
Fe/Zn, Cu
Fe/Zn, Cu,Pb
Fe/Zn, Pb
Fe/Zn, Cu
Fe/Zn
Fe/Zn
Fe/Zn, Cu
Fe/Zn
Fe/Zn, Pb
Cu, Fe/Zn
Fe/Zn
Fe/Zn
Actual Piping Materials
Service Line
Length, type, age, size
75' unkd, 8 yrs
40' plastic, 4 yr,3/4";18' iron 56 yr,3/4"
Copper
2" galv
42' galv, 20 yr,3/4";24' plastic, 7 yr,3/4"
Mixed galv & copper service & plumbing,
asbestos cement mains
20' plastic, 10 yr
10' copper, 7 yr,3/4"i490' unk 3/4"
18' copper, 7 yr,3/4";34' unk 3/4"
30' iron, 60 yr, 3/4"
45' galv, 18 yr,3/4"
30' copper, 9 yr,3/4"j23' unk 3/4"
42' copper, 12 yr,l";27 galv, 12 yr,3/4"
42' galv,20 yr.l"; 42' unk
34' galv,32 yr,3/4";30' galv, 3/4"
10-15' 3/4" copper
20' copper, 23 yr,3/4";44' unk 3/4"
40' steel, 32 yr,3/4"j25' unk 3/4"
44' steel, 22 yr,3/4"j64' unk
40' copper, 3/4";39' unk
40' steel, 48 yr,3/4";25' steel.V'
20' copper. 11 yr,3/4";24' galv, 11 yr,3/4"
House Plumbing
Type, age, size
copper, 8 yr,V'
galv, V
Copper & galv, 10 yr,3/4",>j"
galv, 20 yr,V
galv, V
galv, 27 yr.V
galv, 50 yr.V1
galv.V
galv, 18 yr.V
galv, 35 yr.V
galv, 12 yr.V
galv, 14 yr,3/4"
galv, 32 yr.V
galv, 15 yr,V',3/4"
galv, 35 yr,3/4"
galv, 15 yr,3/4"
galv, 40 yr.V
unk
galv, 11 yr,3/4"
0. Running sample
6 Standing sample
C Copper
d Galvanized iron
-------
TABLE 6 - MEANS AND MAXIMUMS OF RUNNING
AND STANDING SAMPLES COMPARED TO
RAW WATER VALUES
Parameter
Temperature, "c
PH
D . 0 . , ppm
Conductivity, umho
Turbidity, FTU
Alkalinity, mg/1 CaC03
Chloride, mg/1
Fluoride, mg/1
Plate count /ml
Calcium, mg/1
Magnesium, mg/1
Copper, mg/1
Iron, mg/1
Zinc, mg/1
Cadmium, ug/1
Lead, ug/1
Manganese, ug/1
Running
Mean Max.
11
5.8
10.3
22
0.63
2.5
2.1
0.90
Not
2.69
0.29
0.12
0.28
0.15C
<0.4
5
14
14
8.9
10.8
42
2.4
14
2.9
1.08
sampled
7.27
0.38
1.67
1.2
1.73
0.8
22
45
Standing
Mean Max.
10
6.1
8.2
25
3.3
4.5
2.4
0.99
l,030a
2.94
0.29
0.45
1.4'3
1.746
2.0
39
24
19.5
9.1
10.3
42
28
14
4.6
1.25
31,000
7.34
0.41
2.05
5.4°
5.46b
25
170
79
Raw
Water
12
5.4
10.6
18
0.45
%4
1.0
0.05
11
2.07
0.33
0.002
0.17
£ 0.015
<• 0.4
1
18
CL The plate count mean is distorted by one high value. Only 4 plate
counts were above 200/ml. The geometric mean is 14/ml.
b Sample 23-S is deleted from the mean and maximum.
C Values less than the detection limit are counted as zero.
11
-------
The acid used to preserve the samples for metal determination brought the
sample pH to about 2. Suspended particulate corrosion products dissolved
by the acid are part of the apparent concentration. Sample 23-S was
filtered before analysis because it contained a large amount of suspended
corrosion products. Its extreme values were not included in the Zn and Fe
means because of the possibility of the acid effect.
DISCUSSION
Average air temperature during October was close to the average running water
temperature. The standing samples were warmer because the water was
collected from piping systems in buildings.
Dissolved oxygen in the running samples was 95% to 100% of the saturation
value. The standing average was 20% below saturation, which indicated
consumption in corrosion reactions. No correlation of decreased oxygen and
an increase in the corrosion products was apparent in the standing group,
however.
Conductivity was slightly increased in the standing samples.
t
Mean turbidity increased fivefold from running to standing samples. It was
a good indicator of corrosion.
pH wes determined by the C02 equilibrium. Because the water was poorly
buffered, the determination was instrumentally difficult. No significant
difference existed between running and standing samples.
Mean alkalinity almost doubled from running to standing. Alkalinity increase
was a good indicator of corrosion. In general, a sample set that showed an
increase of 0.1 FTU and 1 mg/1 alkalinity as CaCog from running to standing,
also exhibited corrosion products. Sample 30 was an exception.
The chloride concentration showed little variation. The difference between
the October 12 (1 through 4, 6 through 12) and October 26 (5, 13 through 34)
groups may have been caused by the residual chlorine in the earlier samples
that had not yet reacted to form chloride. Instrumental error was also
possible. A later check on the October 12 group showed higher values:
2-R, 1.4 mg/1; 7-R, 2.3 mg/1; and 9-R, 2.3 mg/1.
The fluoride concentration showed little variation and was close to the
recommended 1.0 mg/1. The Seattle Water Department reported that the two
samples with low fluoride levels were the result of temporary shutdown of
the feed equipment at the treatment building.
The number of standing samples exhibiting color was indicative of the
corrosion occurring in unprotected pipes. This is supported by the fact that
none of the running samples, which represent water from the mains, showed
color.
12
-------
All samples were negative for coliform and pseudomonas. One high plate
count (31,000) raised the mean to 1030 organisms/ml. With the exception of
that sample, the mean was 94 organisms/ml, well below the proposed 1974
Drinking Water Standard of 500 organisms /ml. The geometric mean of all
samples was 14 organisms/ml.
Calcium and magnesium levels were remarkably constant through the system and
also between running and standing samples. The water did not deposit a
scale.
The metals all showed increases from the raw water to the running samples
with the exception of some manganese values. The increase from running to
standing samples was significantly larger. Copper and galvanized iron pipe^
the most common pipe materials, were vigorously attacked. Even in the
worst case (Sample 23), however, the running sample contained low concen-
trations of corrosion products.
High manganese levels correlated well with high iron concentrations.
Manganese is a constituent of iron pipes. The mean for iron increased
five fold from running to standing samples.. The manganese mean doubled.
Cadmium, an impurity in the zinc coating of galvanized pipes, appeared in
cases of galvanized corrosion. The standing mean for cadmium was sig-
nificantly above the running mean. For zinc, the mean increased over tenfold.
No direct correlation between the zinc and cadmium values was established.
The copper mean increased fourfold from running to standing samples, and the
lead mean increased eightfold. The source of lead was not identified, as no
lead pipes were reported. Sweat-solder in copper plumbing may have been the
source and could have been identified by testing sites with only solder-
joined copper pipes.
Table 7 compares the corrosion-related parameters for the two types of
running samples. The samples collected from reservoirs, transmission mains
and pumping stations (numbers 3, 4, 7 through 12, 16, 18 and 28) are in
the mains category. Those collected from buildings (numbers 1, 5, 6, 13
through 15, 17, and 19 through 27, 29 through 34) fill that category.
13
-------
TABLE 7 - COMPARISON OF CORROSION PRODUCTS IN RUNNING SAMPLES
TAKEN FROM MAINS AND BUILDINGS
Corrosion parameter Mains Buildings
Turbidity, FTU
Alkalinity, mg/1 as CAC03
Cu mg/1
Pb yg/1
Fe mg/1
Mn yg/1
Zn mg/1
Cd yg/1
0.88
2.2
0.047
7.1
0.34
19
< 0.015
< 0.4
0.52
2.1*
0,153
4.8
0.25
11
0.22
< 0.4
^Excluding samples 14 and 15 (*ee text).
Values from mains group are lower in copper and zinc than those from the
buildings, but higher in turbidity, lead, iron, and manganese. Alkalinity
and cadmium values are the same for both groups. The lack of uniform
differences between the two types of sampling points indicates that the short
sampling lines in the mains group were contributing corrosion products to
the water.
Further evidence supporting the theory that the mains were not contributing
corrosion products was provided by sample site 9. The sample came through
a 6-inch main off a distribution main and then through a 2-inch galvanized
line to a continuously running tap. The concentration of each metal in this
sample was identical to the concentration found in the raw water.
The notes (Table 5) provide a quick summary of the corrosion products in each
sample, including differences between the running and standing samples.
"Clean" means relatively free of metals, that is, similar to the running aver-
age.
An assumption was made of the type of piping material through which the
sample was drawn. The assumption is based on the corrosion products summar-
ized in the notes.
The identification of actual piping materials proved difficult and was not
completed. There are three sections of pipe between the main to the faucet.
(1) A service line from the main to the meter, which is
near the sidewalk.
14
-------
(2) A line from the meter to the building. This category was
rarely known. When the information was available, it
was included under the service line heading.
(3) Plumbing inside the building.
Where the actual pipe data were complete, correlation with assumed materials
was good. Analysis of corrosion products can provide a reliable method
of identifying buried pipes. Frequent mixtures of piping materials prevented
the calculation of a corrosion rate for any particular metal in this
distribution system.
Sample sites 14 and 15 were located in the Duvall Water District, which
buys Tolt water wholesale from Seattle. Water reaching the service line of
site 15 passed through 1 1/2 to 2 miles of 10-inch asbestos-cement main
at a high flow rate. To reach site 14, the water passed through the same
main, plus an additional 800 feet of 10-inch, 400 feet of 6-inch, and
1000 feet of 4-inch pipe, all made of asbestos-cement. Flow rates in the
smaller sections were not as high as in the 10-inch main. Both sets of
samples contained corrosion products from building plumbing. They also
exhibited differences from the mean of all stemples in parameters not directly
related to the corrosion of exposed metal. These data have been summarized
in Table 8.
TABLE 8 - WATER QUALITY CHANGES ATTRIBUTED TO ASBESTOS-CEMENT PIPE
Sample 14
Parameter
R S
Sample 15
R S
Mean of all Samples
R S
PH 8.9 9.1 5.9 5.8 5.8 6.1
Alkalinity mg/1
as CACC-3 * 14 o,14 4.1 6.6 2.5 4.5
Conductivity, ymho 42 42 25 28 22 25
Calcium, mg/1 7.27 7.34 3.94 3.82 2.69 2.94
Magnesium, mg/1 0.08 0.08 0.18 0.21 0.29 0.29
15
-------
Because parameters in this group remained constant from running to standing
samples, the changes from the mean were caused by the water reacting with
asbestos-cement pipe, not building plumbing. Sample site 14, with greater
exposure, showed large deviations in pH, alkalinity, conductivity, and
calcium. The depressed magnesium concentration may have been caused by ion
exchange on the pipe walls.
Increases in pH, alkalinity, and calcium have been observed by the Seattle
Water Department in water from freshly relined mains. The water quality
changes cease after a few weeks of flow, presumably after the uncombined
calcium oxide has been removed from the cement. This is in contrast to
asbestos-cement pipe, which is specifically manufactured to contain no
uncombined calcium oxide.
COMPARISON OF SAMPLE VALUES
WITH THE 1962 U.S. P.H.S.
DRINKING WATER STANDARDS
Eighteen of the twenty-two standing samples exceeded one or more of the
recommended limits of the 1962 P.H.S. Drinking Water Standards. Six of these
also exceeded the mandatory limits. Six of the thirty-four running samples
exceeded the recommended limits, but none were above the mandatory levels.
On Table 9, the breakdown is by constituent.
TABLE 9 - NUMBER OF SAMPLES EXCEEDING P.H.S. DRINKING WATER STANDARDS
Exceed mandatory Exceed mandatory
Parameter limits limits
Running Standing Running Standing
Lead 0 5
Cadmium 0 2
Iron 5 16
Copper 15
Zinc 0 2
Manganese 0 1
Turbidity 0 2
16
-------
APPENDIX
ATOMIC ABSORPTION WITH A HEATED GRAPHITE ATOMIZER
In the study of corrosion products in the Seattle drinking water distribution
system, metal concentrations far below the detection limit for conventional
flame atomic absorption were encountered. The normal techniques of
preconcentration by boiling or extraction were rejected in favor of the
graphite furnace modification of the atomic absorption method.
This technique was selected on the basis of reports in the literature (2-4)
and a demonstration of the Perkin-Elmer HGA 2000. A brief review of the
flameless atomic absorption technique was printed in American_L_ab_o_ratory
in August 1972 (5). Caldwell, Yee, and McFarren ^ reported in 1974
that concentrations of chloride, sulfate arid nitrate higher than the levels
in Seattle tap water caused suppression of the signal in lead analyses.
A Perkin-Elmer model 303 atomic absorption spectrophotometer was fitted
with an HGA 2000 Heated Graphite Analyzer. The absorbance signal from the
spectrophotometer was automatically recorded on a strip chart. The peak
heights were converted to concentration values from a curve fitted to a
least-squares parabolic fit of the known standards.
Glassware cleaned with acid was essential to prevent contamination of samples
or standards. Volumetric flasks, previously washed with detergent, were
prepared for microgram-per-liter-level standards by the following procedure:
(1) 24 hours filled with 8N HNO~, prepared with deionized water
(1:1 cone HN03),
(2) 4 rinses with deionized water,
(3) 24 hours filled with deionized water,
(4) refilled with deionized water for storage.
The clean flasks were segregated from other laboratory glassware and
were not exposed to tap water or detergents. As previously reported (7,8),
rubber stoppers contaminated the standards and they were not used. Before
flasks were refilled with a standard, they were rinsed with nitric acid, which
had been saved from the initial washing procedure, and deionized water.
Standards were prepared fresh daily and stabilized with 0.1 ml concentrated
nitric acid per 100 ml solution. Less than 5% concentration loss at the
10 ug/1 level was observed after 3-days of storage. Aliquots of standards
at concentrations below 10 mg/1 were transferred with plastic tapped pipets
(e.g., Eppendorf). After several operations, including rinses with
17
-------
deionized water, tips were discarded because they failed to empty completely.
The spectrophotometer was similarly adjusted to the settings for the
flame technique, with the following changes. When the hollow cathode
lamp was adjusted as described in the instruction manual, a 15% increase
in the photomultiplier gain from the "flame" setting was required to
compensate for the reduced light intensity. When properly aligned, the
furnace assembly physically blocked some of the incident light. This was
measured as less than 10% absorption, versus no obstacles in the beam.
The recorder was activated automatically by the temperature programming
unit.
Typical operating parameters for the furnace are shown in Table 1. Drying
time at 150° C was 40 seconds for 100 yl samples and 30 seconds for 25 ul
and 10 yl samples.
The length of char cycle depended on the amount of organic material in the
sample. If no organics were present, as in Seattle tap water, 15 seconds
was allowed for the recorder to establish a baseline.
The char temperature was increased in a series of experiments to determine
the highest temperature that did not result in volatilization and loss of
the metal. Transition metals were charred at higher temperatures than
shown in Table 10* without loss.
Absorption of the incident beam during the atomization cycle was strongly
dependent on the temperature of the graphite tube. The values shown in the
Table are a compromise of a number of factors that provided the best
precision and accuracy.
Peak height (absorption) increased with higher temperature, but the
width of the peak decreased. The temperatures chosen were low enough to per-
mit the recorder pen to follow the signal. To prevent carry over to the
next sample, the atomization temperature was high enough to volatilize the
metal completely. Above 2000°C, an intense white light, which flooded
the photodetector, was emitted by the graphite tube. This was avoided
by lowering the atomization temperature and duration. For a 5-second atom-
ization, the white light produced a shoulder on the side of a 50% absorption
peak.
The gas interrupt function automatically stops the flow of inert gas through
the graphite tube during atomization. This prolongs the time that the
metal vapor is exposed to the light beam and increases the sensitivity of the
method. The Table indicates where the gas interrupt was applied.
The graphite tube was heated to maximum temperature to vaporize residues
between runs when assaying a low-concentration (weight) sample after a
high concentration sample. It was required more often with the transition
metals or when precision was poor. Prolonged loss of precision indicated
the need for a new graphite tube.
18
-------
TABLE 10
Typical Furnace Parameters
Metal
Cd
Co
Cr
Cu
Fe
Mn
Ni
Pb
Zn
Char
Temp.
300°C
600
600
600
600
600
600
600
600
300
Atomization
Temp.
1800°C
2400
2500
2300
2400
2400
2400
2300
2000
1300
Atomization
Time
6 sec.
5
8
6
5
5
5
5
5
15
Gas
Interupt
No
Yes
Yes ,
No
Yes
No
Yes
No
Typical
Standard
10 jag/1
50
100
100
5
30
40
5
Typical
ul Standard
25 }il
25
25
25
100
25
25
25
Analysis
% Absorption
40%
45
41
44
70
33
36
34
-------
When the cooling water for the furnace was below 15°C, vapor from 100-yl
samples condensed on the external surfaces of the graphite cone. Some
of this water vaporized or spattered during atomization, which decreased
precision. The problem was eliminated by swabbing the cone with a lintless
batt and heating the tube to maximum temperature after four to six samples.
No condensation was observed with 10-pl or 25-ul samples.
Polyethylene tips for injecting the samples into the graphite tube were
discarded after 4 to 10 samples, when graphite embedded on the outside
of the tip caused wetting.
The deuterium background corrector,which eliminates interferences from
broad band absorption, was not needed for the drinking water samples tested
in this study.
The HGA. 2000 was operated over a one-thousand fold concentration range by
varying the sample volume. The time for a single analysis was longer than
by flame atomic absorption. Samples were screened by the flame method,
and those requiring scale explanation or noise suppression were retested
with the HGA.
20
-------
REFERENCES
1. Standard Methods for the Examination of Water and Wastewater
(13 Edition) APHA, New York.
2. Manning, D.C. and Fernandez, P., "Atomization for Atomic
Absorption Using a Heated Graphite Tube," Atomic Absorption
Newsletter, 9_, 65 (1970).
3. Fernandez, F. and Manning, B.C., "Atomic Absorption Analyses
of Metal Pollutants in Water Using a Heated Graphite Atomizer,"
Atomic Absorption Newsletter, 10. 65 (1971).
4. Davidson, and Secrest, W. L., "Determination of Chromium in
Biological Materials by Atomic Absorption Spectrometry Using a
Graphite Furnace Atomizer," Anal. Chem., 44, 1808 (1972).
5. Robinson, J.W. and Slevin, P.J., "Recent Advances in Instrument-
ation in-Atomic Absorption"; Amos, M.D., "Nonflame Atomization
in AAS - "A Current Review"; American Laboratory 4., (8) (August
1972).
6. Caldwell, J.S., Yee, L.M., and McFarren, E. F., "Evaluation
of Atomic Absorption'Graphite Furnace for Metals," American
Water Works Association Second Water Quality Technology
Conference, Dallas, Texas, December 1-4, 1974.
7. Robertson, D.E., "Role of Contamination in Trace Element Analysis
of Sea Water," Anal. Chem., 40, 1067 (1968).
8. Everson, R.J., "Zinc Contamination from Rubber Products,"
Atomic Absorption Newsletter, 11. 130 (1972).
21
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-670/2-75-036
3. RECIPIENT'S ACCESSIOWNO.
4. TITLE AND SUBTITLE
STUDY OF CORROSION PRODUCTS IN THE SEATTLE WATER
DEPARTMENT TOLT DISTRIBUTION SYSTEM
5. REPORT DATE
May 1975; Issuing Date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
Robert A. Dangel
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Northwest Water Supply Research Laboratory*
U.S. Environmental Protection Agency
Gig Harbor, Washington 98335
10. PROGRAM ELEMENT NO.
1CB047; ROAP 21AQF; Task 04
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
In-house
14. SPONSORING AGENCY CODE
16. SUPPLEMENTARY NOTES
*The Northwest Laboratory is now part of the Water Supply Research Laboratory in
Cincinnati.
16. ABSTRACT
Samples from the Seattle Water Department's Tolt distribution system were analyzed
for chemical and bacteriological parameters. Changes from the raw water quality
were observed, particularly in trace metal concentrations and other parameters
related to corrosion. Distribution mains were found to be adequately protected
from corrosion by cement and bituminous linings whereas service lines and household
plumbing were actively corroded. Metals in the yg/1 concentration range were
determined by a flameless atomic absorption technique.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
*Potable water
*Corrosion products
Distribution systems
*Seattle (Washington)
Flameless atomic absorp-
tion
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
30
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
22
^UAGOVBIIMEHTMlliailffiOFFKE: 1975-657-593/5382 Region No. 5-11
------- |