WATER QUALITY MONITORING
                    IN
           DISTRIBUTION SYSTEMS
                    by
            Nina I. McClelland
      National Sanitation Foundation
        Ann Arbor, Michigan  48106
                    and
                K. H. Mancy
        The University of Michigan
        Ann Arbor, Michigan  48109
          Contract No. 68-03-0043
              Project Officer

              James M. Symons
      Water Supply Research Division
Municipal Environmental Research Laboratory
          Cincinnati, Ohio  45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
    OFFICE OF RESEARCH AND DEVELOPMENT
   U.S. ENVIRONMENTAL PROTECTION AGENCY
          CINCINNATI, OHIO  45268

-------
                       DISCLAIMER
     This report has been reviewed by the Municipal Environ-
mental Research Laboratory, U.S. Environmental Protection
Agency, and approved for publication.  Approval does not
signify that the contents necessarily reflect the views and
policies of the U.S. Environmental Protection Agency, nor
does mention of trade names or commercial products constitute
endorsement or recommendation for use.

-------
                        FOREWORD
     The Environmental Protection Agency was created because
of increasing public and government concern about the dangers
of pollution to the health and welfare of the American people.
Noxious air, foul water, and spoiled land are tragic testi-
mony to the deterioration of our natural environment.  The
complexity of that environment and the interplay between its
components require a concentrated and integrated attack on
the problem.

     Research and development is that necessary first step
in problem solution and it involves defining the problem,
measuring its impact, and searching for solutions.  The Muni-
cipal Environmental Research Laboratory develops new and im-
proved technology and systems for the prevention, treatment,
and management of wastewater and solid and hazardous waste
pollutant discharges from municipal and community sources,
for the preservation and treatment of public drinking water
supplies, and to minimize the adverse economic, social,
health, and aesthetic effects of pollution.  This publication
is one of the products of that research; a most vital communi-
cations link between the researcher and the user community.

     Concern for protecting the public health by assuring
the quality of public drinking water supplies is apparent with
passage of the Safe Drinking Water Act  (Public Law 93-523) ,
signed by the President on December 16, 1974.  This is the
first federal act dealing in depth with providing safe
drinking water for public use.  Through this project, the
feasibility of measuring levels of drinking water contaminants
at tkz con^ame^.'-4 tap has been demonstrated.  Subsequent use
of the mobile water quality monitoring laboratory will per-
mit EPA to assure compliance with many of the maximum levels
for inorganic chemicals proposed as national interim primary
drinking water standards (CFR Vol. 40, No. 51—Friday, March
14, 1975).
                            Francis T. Mayo, Director
                            Municipal Environmental Research
                            Laboratory

-------
                        ABSTRACT
     A mobile laboratory with 18 integrated, computer con-
trolled parametric systems for monitoring potable water
quality in distribution systems was developed and field
evaluated at ten locations in four United States cities:
Chicago, Illinois; Ann Arbor and Detroit, Michigan; and
Philadelphia, Pennsylvania.  Temperature, conductivity, pH,
chloride, dissolved oxygen, free and total residual chlorine,
turbidity, corrosion rate, free and total fluorides, alka-
linity, hardness, nitrate, copper, cadmium, lead, and cal-
cium carbonate deposition rate are measured using commercially
available and newly developed sensor systems.
                              IV

-------
                             CONTENTS
Foreword

Abstract

List of Figures

List of Tables

Acknowledgments


Sections

I            Conclusions                                         1
II           Recommendations                                     4

III          Introduction                                        5

               General                                           5
               Study Aims and Objectives                         5
               Study Plan                                        6

IV           Prototype Monitor                                   8

               General                                           8
               Theory                                            8
               Performance Characteristics                       8
               SRM-Housed Systems                               14
                 General                                        14
                 Electronics                                    15
                 Recorder                                       18
                 Sensors                                        18
                 Interpretation of Hardness and Nitrate Data    19
               Commercial Systems                               24
                 Free and Total Residual Chlorine               24
                   General                                      24
                   Theory                                       24
                   Experimental                                 27
                   Results and Discussion                       31
                 Turbidity                                      33
                   General                                      33
                   Experimental                                 33
                 Corrosion Rate                                 33
                   General                                      33
                   Theory                                       36
                   Experimental                                 39
                   Results and Discussion                       40
                                v

-------
                                                              Page

               Other Systems                                     41
                 Calcium Carbonate Deposition Test              41
                   General                                       41
                   Theory                                       41
                   Experimental                                 50
                   Results  and Discussion                       54
                   Field Experience                             59
                 Alkalinity                                     62
                   General                                       62
                   Theory                                       62
                   Experimental                                 70
                   Results  and Discussion                       72
                   Special  Case Study                           75
                 Free and Total Fluorides                       75
                   General                                       75
                   Theory                                       76
                   Experimental                                 77
                   Results  and Discussion                       78
                   Automated  System                             87
                   Continuous Fluoride Monitor                 106
                   Field Studies                               111
                 Trace Metals                                  121
                   General                                      121
                   Theory                                      123
                   Experimental                                126
                   Results  and Discussion                      130
                   Special  Study                               130
V            Mobile Laboratory                                 140
               General                                         140
               Physical Description                            140
               Instrumentation                                 141
               Organization                                    141
               Computer System                                 142
               Field Studies                                    149
VI           References                                        168
VII          Project Publications                              172
VIII         Appendices                                        174
                                VI

-------
                             FIGURES
No.                                                         Page
 1     NSF/EPA Mobile Water Quality Monitoring Laboratory     7
 2     Schematic Arrangement of Schneider Robot Monitor      14
 3     Schematic Diagram of SRM Flow Cell Cubicle, Top View  16
 4     Schematic Illustration of Sensor Positions and Over-  17
         all Arrangement of SRM-Housed Systems
 5     Typical Strip Chart Recorder Output of Hardness Data  21
 6     Conductivity Correction Nomogram                      23
 7     Distribution of Hypochlorous Acid and Hypochlorite    26
         Ion in Water at Different pH Values and Tempera-
         tures  (2)
 8     Schematic Illustration of Residual Chlorine Analyzer  29
 9     Schematic Illustration of Cl2 Analyzer                30
10     Free and Total Residual Chlorine Variations at Gulf   32
         Oil, Philadelphia
11     Schematic Illustration of Turbidity System            34
12     Schematic Illustration of Turbidimeter in the Mobile  35
         Laboratory
13     Schematic Illustration of Corrosion Rate Monitor in   40
         the Mobile Laboratory
14     Fluid Flow in the Rotating Electrode System           45
15     Concentration Profile at Electrode Surface under      47
         Steady State Conditions, in the Absence of CaCOs
         Film
16     Oxygen Current - Time Curves during CaC03 Deposition  49
17     Stability Index Monitor for Saturation Equilibrium    51
         Measurement
18     PIR Rotator - Electrode System                        53
19     CCDT System in the Mobile Laboratory                  53
20     Current versus Potential Sweep                        55
21     Calcium Carbonate Deposition Test                     56
22     Correlation of CCDT Slope with Ryzner Stability Index 57

                               vii

-------
No.
23
24
25
26
27
28
29
30

CCDT versus Ryzner Stability Index
Effect of Caustic Feeder Cycling at Chicago CWFP
on CCDT and pH Monitored at Site No. 1, Chicago/
Lehigh
[H ] versus C032 formality, given applying Equa-
tion 98
Experimental and Theoretical Relationships between
Alkalinity and pH
Schematic Illustration of Alkalinity Monitor in
the Mobile Laboratory
Good Quality Recording of Continuous Alkalinity
Analysis
Alkalinity Calibration Curve
Characterization of Fluorides and Preparation of
Page
61
63
67
68
71
73
74
79
         Buffers
31     Effect of pH on Fluoride Ion-Selective Electrode       80
         Measurement
32     Calibration Curves of Different Levels of pH           82
33     Effect of pH on Fluoride Ion Electrode Sensitivity     85
34     Complexometric Titrations of Fluoride with Aluminum    86
35     Aluminum Complexing Effect and Fluoride Recovery by    89
         Different Masking Agents
36     Rate of Fluoride Recovery with CDTA Masking Agent      90
37     Rate of Fluoride Recovery with Citrate Masking         91
         Agent
38     Schematic of Automated Sample Analysis for Free and    92
         Total Fluoride
39     Effect of Pump Surges on Reference Electrode           95
40     Effect of Flow Rate versus Response Time               97
41     Recorder Output for Calibration of Free and Total      98
         Fluoride (5 minutes)
42     Recorder Output for Calibration of Free and Total      99
         Fluoride (2 minutes)
43     Recorder Output for Calibration of Free and Total     100
         Fluoride (1 minute)
44     Calibration Curves for Automated Fluoride Electrode   102
         Measurement
45     Reproducibility of Fluoride Measurement in Absence    104
         of Complexing Cations

                              viii

-------
No.                                                        Page
46     Reproducibility of Fluoride Measurement (1.0 mg/1)  105
         in Presence of Aluminum
47     Schematic Illustration of Free and Total Fluorides  110
         Monitor in the Mobile Laboratory
48     Fluoride System Output including a Calibration      112
         Series
49     Typical Calibration Curves for Fluorides,  Prepared  113
         from Output in Figure 48
50     Special Fluoride Study at Chicago/Lehigh Station    118
51     Map Identifying Relative Locations of Monitoring    119
         Sites in Chicago
52     Output from Chicago Fluoride Monitor, recorded at   120
         Chicago/Greenleaf
53     Special Fluoride Study at Calumet Harbor            122
54     Plate-Strip Sequence for Differential Anodic        125
         Stripping Voltammetry
55     Schematic Diagram of DASV Flow Cell used in Mobile  128
         Laboratory System
56     Schematic Illustration of Trace Metals Monitor in   129
         the Mobile Laboratory
57     Typical Voltammograms of a Standard Solution and    131
         Tap Water Analyses
58     Sweep Rate Response                                 132
59     Household Trace Metals Survey, Cadmium              134
60     Household Trace Metals Survey, Lead                 135
61     Household Trace Metals Survey, Copper               136
62     Schematic Illustration of Right Side Interior of    143
         Mobile Laboratory
63     Schematic Illustration of Left Side Interior of     144
         Mobile Laboratory
64     Photograph Showing Overview of Mobile Laboratory    145
         Interior
65     Relationship of Computer System to Analytical       147
         Sensor Systems
66     Computer System Software Relationships              148
67     Flowchart of INIT                                   150
68     Flowchart of NSFC                                   151
69     Flowchart of PUN                                    153
70     Flowchart of BELL                                   154
                              IX

-------
No.                                                         Page
71     Flowchart of SMPL                                    155
72     Flowchart of FLOW                                    156
73     Flowchart of DAQ                                     157
74     Flowchart of Subroutine A (DAQ)                      158
75     Flowchart of DATA                                    159
76     Normal Data Flow                                     162
77     Data from Main Flushing at Chicago/Calumet Harbor    165
78     Relative Locations of Philadelphia Monitoring Sites  166
79     [H+] as a Function of C032~ and C12 Formality        185
80     Experimental and Theoretical Relationships between   186
         Alkalinity and Potentials
81     AE versus [Clal in Alkalinity Determination          187
82     Potential versus Alkalinity Calibration Series       191
                               x

-------
                             TABLES
No.                                                         Page
 1     Sensors in the Prototype Monitor                       9
 2     Electrochemical Sensors for Water Quality Monitoring  10
 3     Performance Characteristics of Mobile Laboratory      12
         Sensor Systems
 4     Summary of Data from Analyses of Test Solutions       58
 5     Results of Field Studies                              60
 6     [H ]  versus Alkalinity, applying Equation 98          69
 7     Experimental and Theoretical Relationships between    70
         Alkalinity and pH
 8     OH~ Selectivity Coefficients (K)                      81
 9     Shift in Intercept of Calibration Curves at Low pH    84
         Values
10     Aluminum Complexing Effect and Fluoride Recovery by   88
         Different Masking Agents
11     Iron Effect and Fluoride Recovery by Masking Agents   93
12     Reproducibility of Automated Measurements at Differ- 106
         ent Fluoride Levels in Presence and Absence of
         Aluminum
13     Fluoride and Iron Determinations in Water Samples    107
         (mg/1) from Jackson, Michigan, 2067
14     Characterization of Fluoride in Water Samples from   108
         Wyoming, Michigan, by the Manual and Automated
         Electrode Methods
15     Statistical Data Related to Manual and Automated     109
         Fluoride Electrode Measurements in Water Samples
         from Wyoming, Michigan
16     Characterization of Fluoride in Water Samples from   115
         Ann Arbor, Michigan (April 6, 1971)
17     Chronological Characterization of Fluoride in Water  116
         Distribution System of Ann Arbor, Michigan
18     Summary of Free and Total Fluoride Data (Chicago/    117
         Lehigh Station)
19     Electrode Deterioration                              133

                                xi

-------
No.                                                        Page
20     Electrode Response After Regeneration               133
21     Trace Metals Survey Sites                           137
22     Summary of Raw Data from Household Trace Metals     138
         Survey
23     [H+] as a Function of C032~ and C12 Formality       188
24     Experimental and Theoretical Relationships between  189
         Alkalinity and Potentials
25     Theoretical Relationship between Alkalinity and     189
         Potential for Solutions with and without Free
         Residual C12
26     Changes in Potential from Addition of 1.0 mg/1      190
         Free Residual C12
27     Effect of Thiosulfate                               192

-------
                          ACKNOWLEDGMENTS

Sincere appreciation is expressed to the following persons who
provided input to the project:

Gordon G. Robeck and James M. Symons, Ph.D., Director and Chief,
Water Supply Research Division, respectively, who served as proj-
ect officers.  Through their efforts, continuation support was
made available to equip the mobile laboratory and computerize
monitoring operations.  The merits of adding practical applica-
tion to analytical feasibility are readily apparent.  The overall
guidance and support provided by both Mr. Robeck and Dr. Symons
are greatly appreciated.

Further guidance and considerable constructive criticism were
provided by members of the Advisory Committee.  Their assistance,
both individually and collectively, is also very much appreciated.

Commercially available instrumentation with potential applica-
bility for the proposed monitoring system was evaluated early
in the project.  The cooperation of manufacturers who generously
consigned equipment for this purpose is gratefully acknowledged:

                 Aqua Test Corporation
                 Capital Controls Company, Inc.
                 Hach Chemical Company
                 Schneider Instrument Company

During the various field assignments, personnel at each respec-
tive local water utility provided liason with NSF/EPA staff oper-
ating the mobile laboratory.  Each of these persons responded
willingly to needs of the permanent crew, and demonstrated en-
thusiastic support for the monitoring effort.  Virtually every
request was granted.  Thanks is expressed on behalf of the entire
project staff for their untiring assistance:  Harvey Mieske, Ann
Arbor; Richard A. Pavia, N. J. Davoust, Charles Halter, Ben F.
Willey, and Green Whitney, Chicago; Carmen F. Guarino, Charles E.
Vickerman, Joseph V. Radziul, Alan Hess, Charles Pierce, and
David Gotshall, Philadelphia; and Gerald Remus and Albert Shannon,
Ph.D., Detroit.

To assure smooth transfer of operations from NSF to EPA, three
persons in EPA were assigned responsibility for operating the
mobile laboratory:  Marvin C. Gardels, Ph.D., Robert Thurnau,
and Daniel F. Watkins.  The EPA crew visited Ann Arbor to get


                               xiii

-------
acquainted with NSF project staff, become familiar with systems
in the mobile laboratory, and participate in the final meeting of
the Advisory Committee.  During the monitor's month-long visit
to Philadelphia, the EPA crew rotated for one-week periods and
accompanied the NSF operator to and from Philadelphia.  EPA parti-
cipation was scheduled to provide each operator with experience
in shutting down, moving, and starting up in a new location.  It
was reassuring to observe the enthusiasm of the EPA crew for the
task ahead.  After more than four years of effort, the mobile
laboratory and each of its systems were, themselves, like part of
the project staff.  A great deal of pride and effort went into
their development.  It is gratifying to know that future opera-
tions are in the hands of competent, concerned professionals.

Two weeks were scheduled to install - and develop software for -
the on-board mini computer.  The experts said it could not be
done, but three dedicated people proved them wrong:  William B.
Everett, NSF; Charles A. Khuen, Ph.D. and James Seydel, Ph.D.,
The University of Michigan.  These people met, assigned tasks,
and agreed that the task was virtually impossible in the time
available to them, then proceeded to ignore the clock and complete
their assignments on schedule.  For their initiative and persis-
tence, sincere thanks to each of them for a job very well done.

Finally, the staff for this project worked tirelessly to achieve
each of the expressed objectives.  Their loyalty and dedication
to the project through each phase of effort was indeed commend-
able.  Special thanks are expressed to John R. Adams, who devoted
many extra hours to operating the mobile laboratory on its
various field assignments; Robert R. Wood; Warren K. Schimpff,
Ph.D.; Thomas L. Schwenk, M.D.; Diane Daniel and Sue Buske, proj-
ect secretaries,  Nina I. McClelland, Ph.D. was project director,
and K. H. Mancy, Ph.D. and Dr. McClelland, coprincipal investiga-
tors.
                                xiv

-------
                             SECTION I

                            CONCLUSIONS

It is analytically feasible, using automated electrochemical
techniques, to detect and measure changes in quality which do
occur in potable water during transmission through distribution
systems.  A fully operational prototype drinking water quality
monitor, with 18 integrated, computer controlled, parametric
systems installed in a mobile laboratory was delivered by the
National Sanitation Foundation to the National Environmental
Re'search Center, Cincinnati, Ohio upon successful completion of
the project, "Water Quality Monitoring in Distribution Systems."
All systems were evaluated under actual field conditions by
operating the mobile laboratory for two and a half months at
remote sites on transmission lines in Ann Arbor and Detroit,
Michigan; Chicago, Illinois; and Philadelphia, Pennsylvania.

The reliability and applicability of on-board systems in the
mobile laboratory for detecting and measuring significant changes
in quality at remote monitoring sites along distribution networks
was clearly established.  Because of relatively infrequent
sampling schedules and time lapse between sampling and laboratory
analysis, automated, onsite measurements are potentially more
meaningful than data acquired from perimeter surveys.  Widespread
use of the NSF/EPA mobile laboratory and other similar systems
could be expected to contribute significantly to diagnostic
activities and preventive quality assurance programs consistent
with surveillance and enforcement requirements of the Safe Drinking
Water Act.

Correlation of data acquired by the mobile laboratory with measure-
ments recorded at the water purification plant provides for cal-
culation of residence times in the distribution system at periods
of varying demand, indicates the need for flushing at or near
deadends in the system, and assists the treatment plant operator
in evaluating the need for-and effect of-changes in treatment
plant operation.

Equipment design specifications developed early in the project were
oriented principally to laboratory use.  Computer control and
mobile laboratory application became objectives only after the
analytical feasibility of individual parametric systems was clearly
established.  Specific observations and analysis of data acquired
during field operations suggest the following refinements in
planning  second generation equipment intended for field use; e.g.,

-------
1.  Computer capabilities can- and should- be greatly
    expanded.  At the least,  programs should be
    developed for better utilization of the automatic
    sampler; e.g., sampler operates when predetermined
    levels of specified parameters are exceeded;  and
    on-board electronic data  processing capability
    including computer graphics should be developed.
    In addition,  either a cassette or magnetic tape
    system should be provided in lieu of the paper
    tape option purchased for this project.  Selection
    was made on the basis of  economy at the time  of
    purchase; however, paper  tape output is extremely
    difficult to handle in the field.  Cassettes  are
    attractive for ease in mailing from remote loca-
    tions for additional processing on a larger com-
    puter away from the mobile laboratory.

2.  The original DASV system  was designed for labora-
    tory operation.  Its complex arrangement with
    switching/timing capability is unnecessary with
    the mini computer available.  A commercial po-
    tentiostat is entirely adequate for DASV mea-
    surements , operational control should be a
    function of the computer.  In addition, pH ad-
    justment capability should be provided for the
    DASV monitor to extend its capability for mea-
    suring additional metals.

    A single Technicon pump could be used for all
    systems requiring reagent addition; i.e., DASV,
    alkalinity, and fluorides.  Recurrent bubble
    entrapment problems in the fluoride systems
    should be eliminated.  Further study of solution
    ground, peristaltic pump  noise, and the role  of
    organics with respect to  the alkalinity monitor
    is indicated.  Prepotentiometer electronics
    might be added to filter  the signal.

3.  Liquid junction ion-selective electrodes; i.e.,
    hardness and nitrate should not be used when
    advanced electrode technology provides more re-
    liable electrodes.  Although it is unlikely that
    solid state probes will be feasible in the near
    future, improved bodies with better electrical
    shielding are promised by the manufacturers.
    Use of an electrode other than the divalent
    cation electrode for measuring hardness, and
    placing hardness outside of an SRM system should
    be considered.  The overall usefulness of nitrate
    measurements in distribution system monitoring
    should be reviewed.  It might be desirable to
    retain a nitrate electrode "for use as needed"
    and substitute routine monitoring of sodium or
    calcium.
                       2

-------
          4.   Use of an electrode multiplexer should be con-
              sidered.   One multiplexer channel could be de-
              dicated to bench measurements and calibrations
              while other channels are used for Technicon
              systems measuring special interest parameters.
              Linking multiplexer, potentiometer, and computer
              would reduce the number of potentiometers re-
              quired in a mobile laboratory.

          5.   Better provision for external grounding, re-
              placing the plastic adapter for intake water
              supply, and acquiring a pump for use in low
              line pressure areas in the field are recommended
              changes in the NSF/EPA mobile laboratory.  Addi-
              tional airconditioning capability is also a
              desirable improvement.

In developing a "second generation" mobile water quality monitor-
ing laboratory, considerable savings in time and cost can be
anticipated as a result of experience with this study.

-------
                            SECTION II

                          RECOMMENDATIONS

1.   The NSF/EPA mobile laboratory should be placed in regular
    service and made available to communities where water quality
    deterioration in the distribution system is known to occur.
    For systems in which the quality of water at the tap could be
    improved by modifying treatment plant operation, the mobile
    laboratory is an excellent diagnostic tool.  It can be used
    to identify changes and, with continued monitoring, to assure
    the effectiveness of selected corrective measures.

2.   The NSF/EPA mobile water quality monitoring laboratory should
    be made available to states for support of their new respon-
    sibilities, defined by the federal Safe Drinking Water Act
    (Public Law 93-523); and used routinely by EPA for public
    drinking water supply surveillance activities.

3.   Additional mobile water quality monitoring laboratories should
    be constructed.  This would provide for operation of two or
    more systems on the same transmission line as well as wide-
    spread, simultaneous monitoring capability.  Federal, state,
    district, and large local water utility resources should be
    committed to this objective.

-------
                           SECTION III
                          INTRODUCTION
General
The primary objective of any drinking water utility is to pro-
vide an adequate supply of high quality water to its users.
The quality of water leaving the treatment plant is assured by
routine analytical procedures at the local utility and periodic
surveillance by state water supply agency personnel.

It is generally assumed, however, that quality of a treated
water supply changes during transportation and delivery to the
user's tap.  These changes may occur as a result of:   (1) physi-
cochemical or biochemical transformations in the municipal- or
household-pipeline; i.e., spontaneous changes in the water in-
dependent of pipeline characteristics; or (2) physicochemical
interactions of the water with the municipal- or household-pipe-
line.  Examples of these effects include "unstable" water which
deposits calcium carbonate (CaCOs) film as a function of its
residence time in the piping system  (a physicochemical trans-
formation in the water itself), and uptake by the water of avail-
able metals from the piping system (a physicochemical inter-
action of water with the pipeline).

Regardless of the type of change which occurs during distribution,
quality at tke, U.A&SI'& tap must be assured.  In 1965 the Research
Committee of the American Water Works Association (AWWA) recom-
mended that instrumentation for measuring corrosion and stability
be developed for distribution system monitoring applications.
The research study, "Water Quality Monitoring in Distribution
Systems," undertaken by the National Sanitation Foundation
under grant and contract support from the U.S. Environmental
Protection Agency, was a direct result of this recommendation.


Study Aims and Objectives

Principal objectives of the NSF study were to develop basic
design criteria and operational specifications for a continuous
monitoring system which could measure changes in water quality
characteristics in distribution systems and provide for analyti-
cal quality control of water purification processes.  Specific
aims included:  (1) establishing the analytical feasibility of
commercially available sensor systems for potable water quality

-------
monitoring applications, (2) developing new sensor systems as
required, (3) assembling a prototype monitor, (4) establishing
performance characteristics for each integrated parametric
system in the prototype monitor, (5) installing the monitor in
a mobile laboratory, and (6) operating the mobile laboratory
at selected sites under actual field conditions.

The NSF/EPA mobile water quality monitoring laboratory, shown
in Figure 1, was designed,  constructed, and successfully oper-
ated at the municipal water treatment plant in Ann Arbor,
Michigan; four sites in metropolitan Chicago; a remote distri-
bution system site in Detroit; and four sites in Philadelphia.
It was delivered to EPA/NERC, Cincinnati, Ohio in October 1973
for operation by EPA in future distribution system monitoring
and research assignments.


Study Plan

Eleven persons with expertise in water utility management served
as an advisory committee to establish priorities and provide
overall guidance to the project staff.   (The advisory committee
membership is listed as Appendix A.)  Parameters, selected by
the advisory committee for inclusion in the prototype monitoring
system, included temperature, conductivity, pH, chloride, dis-
solved oxygen, free and total residual chlorine, turbidity,
corrosion rate, free and total fluorides, alkalinity, hardness,
nitrate, copper, cadmium, lead, and calcium carbonate deposi-
tion rate.

Consistent with project aims and objectives, commercially avail-
able sensor systems were loaned to the project by their respec-
tive manufacturers.  The analytical feasibility of including
these systems in the prototype monitor was carefully evaluated
in the laboratory.  No effort was made to modify existing-or
develop new-sensors when commercially available systems were
shown to be applicable.

-------
  o
 -P
  03
  J-l
  o
 X!
 tn
 C
 -H
 M
 O
 -P
 •H
 £
 O
 g
 -P
 •H
 O1
 (U
 -P
 0)
rH
•H
X!
 O
 i=

<
ft
W
\
fci
01
 0)

-------
                           SECTION IV
                        PROTOTYPE MONITOR
General
Eighteen parametric sensors are included in the prototype moni-
toring system developed during this study.  Nine of these are
commercially available and were used without modification.
(See Table 1.)   Five of the nine - temperature, conductivity,
pH, chloride, and dissolved oxygen - are original equipment
with the Schneider Robot Monitor(RM-25), an integrated instru-
ment system designed for monitoring the quality of water in
lakes and streams.  These five sensors were shown to be highly
reliable with no requirements for special timing sequence or
preconditioning of the sample stream.

Other commercial sensors used without modification include free
and total residual chlorine analyzers from Capital Controls
(Models 871 and 872, respectively), a Hach CR low range light
scatter turbidimeter  (Model 1720), and an instantaneous corrosion
rate monitor by Petrolite Corporation.

Orion ion-selective electrodes for fluoride, hardness, and
nitrate are included in the prototype monitor.  The fluoride
system, which provides for both free and total fluoride measure-
ments, requires preconditioning of the sample stream.  Sensors
for hardness and nitrate, though housed in Schneider modules,
required development of support electronics and a flow interrupt
timing/switching device.


Theory

Basic theoretical relationships describing electrochemical
sensors for a number of parameters in the prototype monitor are
summarized in Table 2.  Theoretical considerations for other
systems are described in greater detail in respective sections
of the report.


Performance Characteristics

Performance characteristics of each sensor in the prototype
system are described in Table 3.  Terms are defined as follows:

-------
PS
o
EH
H
!S
O
a

w
EH
O
EH
O
A
CM

H
ffi
EH
CO
PS
o
CO
sz
w
co
cu
CO CU
CU rH 0
SH rd tn G
•H --H C CU
3 O -H 2
D1 CU g (J1
cu CU-H cu
PS CO EH CO
•d
cu
c
0
•H
CU 4-1
H -H
CU 1 ^d
g CU G
fd SH O
co CM u
rH
cu fd cu
•H -rl rH
•W O X!
•H SH fd g
T3 CU >H CU
O g-H 4->
g g fd co
G 0 > >i
!D u <; co
















cu
pi
!>i
EH
J_|
O
CO
G
CU
CO





SH
CU
4J
OJ
g
rd
SH
fd
CU
























^~>.





















J_l
o
4->
CO
•H
g

CU
X!
EH



CU
(-1
rj
•P
rd
rM
cu
Cu
g
(U
EH
























V












rH
rH
(U
O

J>^
4->
•H
•H
4->
O
rj
'c
o
o

u
1
<


K*1
4->
•H
>
•rH
4->
u
E3
'd
G
o
o
























%.

















cu
•g
M
4J
o
cu
H
w

CO
CO
fd
rH
o












K
cu
























^i.








cu
J>
•H
4->
U
CU
rH
CU


pj
O
H

(U
4-)
rd CU
•P T3
CO O
}_j
•d -P
•H 0
rH CU
O rH
CO W






cu

•H
SH
O
rH
XI
a
























^.











cu
•g
SH
4->
u
cu
rH
W
U
•H
SH
4J
1
g
fd
4->
rH
o






*d
cu
£>
•H G
O CU
co tn
CO >i
•H X
Q O















^v








s.


















rH
H
0
CJ

O
•H
C
fd
£>
rH
fd
a

rH
fd
rj
•d
•H
CO CU
CU G
PH 'H
^
CU O
CU 
rH
fd
o
rH
(d
rj
t3
•H
CO
CU CU

•H
rH >H
fd o
4-J H
O X!
H U
























s>.



















!H
cu
4J
Q)
g
o
rH
CU
X!
Cu
cu






!>i
4J
•H
•d

X!
r)

H
























^v










CU
o
G
fd
4J
4j
•H
g
'd
^
£
o
•H
4-1 (U
N O1
•H -H
LJ r~*
^^ M
rd X!
rH O
O CU
CM EH
CU
4->
fd
Pi

C
O
•H
CO
O

SH
0
U















^v

















CU
J>
-H
4->
O
CU
tH
CU
CO

G
0
H

CU
4J
rd cu
-p -d
CO O
^
'd 4-*
•H 0
rH CU
O rH
CO W

CU
T)
•H
J_l
0

rH
pL|

cu
CU

PM















^>.

















CU
J>
•H
4->
U
CU
rH
CU
CO

G
O
H

CU
4-)
fd cu
•P n3
CO O
r<
-d -P
•H o
rH Q)
O rH
CO W
CU

•H
^_|
0

rH
Pn

rH
rd
4->
0
EH















^>















fj
O
•H
4J
rrj
c
•rH
g
O
u

u
•rH
M
4-> CU
CU T3
g o
G JH
0 4J
•H 0
4J CU
G rH
CU H
4->
O W
CM Cu




^i
•P
•H
G
•H
rH
fd
X
rH
.










^






















G CU
O ^
H O
J^
G 4-1
0 0
•H Q)
4-> rH
O H
C
rj Q)
I'D >
•H
T3 4->
•H rj
3 CU

•H CU
rl CO






CO
CO
CU
CH
•d

fd
ffi




"^V

































G CU
o fd
H O
J_J
G 4->
O 0
•H CU
4-> rH
U W
G
3 CU
rD >
•H
13 4-1
•H O
3 CU
CT'rH
•H CU
i-H CO







CU

fd
rM

•H
&




^.










^













tn
G
-H
Cu
Cu
•H
J-l
4-j
CO
o
•H — *
•d >
O CO
G <
r< Q
H
fd >i
•H ^ [
4-> 4-1
G CU
CU g
(j C
^1 J-H
cu fd
MH 4-1
lii i
•H O
Q >







g
3
•H
g

fd
U




%.










•v



































J>
CO
fZ£
Q










•d
rd
0)





V










%.



































^>
CO
f^J
Q








j^
CU

Cu
0
u



































tn
G
•H
4J
fd cu
[ \ r^
o o
fyj SH
4J
0 D
•H CU
4J rH
fd W
4-)
CO U
O CO
•H -H
4-> Q
C
cu tn
4J G
O -H

CO
CU
EH

G
CU O
•P-H
g fd -P
3 G-H
•H O CO
O XI O
rH SH D.
rd rd CU
O O Q

-------
Table 2.  ELECTROCHEMICAL SENSORS FOR WATER QUALITY MONITORING
Sensor Type

Conduc tome trie
Po ten tiome trie
- Glass electrode
(e.g. , pH)
- Membrane electrodes
(e.g., ion-selective
electrodes)
- cationic
- anionic
Voltammetric Membrane
Electrodes
Equation

L = Kc I Ci Xi zi (1)
RT f 2i/Zil
E = K + £i In a. + KTa x D (2)
m zt L i J J J
pH = -log aR+ (3)
pM+ = -log aM+ (4)
pA = -log aA~ (5)
^ = [z FAPm E] a02 (6)
where:
          L
          Ke
          C.
          z. =
          E  =
           m
          F  =
          K. =
           D
          id =
          A  =
          b  =
specific conductance
cell constant
ionic concentration
ionic equivalent conductance
ionic valency
measured electrode potential
Faraday constant
selectivity coefficient
diffusion current
electrode surface area
membrane permeability coefficient
membrane thickness
                              10

-------
(1) V
-------
CO
E
LU
CO

CC
LU
CO
CC
o
oc
o
CO
LLJ
_J

OQ
o
u.
o

CO
o

CO

CC
LU

u
<
CC
<
X
o

LU
u
CC
o
LI-
CC
01
Q.


CO
 CO
Long Term Stability,
Recalibration and
Maintainance Requiremsnts



JC
en +-
y c u
c w H-
o 
h- Q




6
'o
O)
ol
u.
c
O)
*U (A
£: * t:
** a»
> ~

tj ~~
u
IT.

•ST
"aT
O)
c
re
CC
C
o
OJ ~
C) £
a j

^
ai
t;
£
These liquid junction ion se-
lective electrodes have been
shown to have stable poten-
tials which permit continuous
use for periods up to seven
days without recalibration;
c ~ ^S
•o-aji.! £


!2.H IB' c 3

o °^_ o •£  C

111 5 IE
2 3 ? ? c 3
c S jc E .2 ^

0 = o£ 3 E

o>
Eo
sy
"-T °
Ol 1
E o >
" i S"
O S  in — . to q
ro to ro ^J i—
u II ii M n
CM CM CM p4 tM
C a> 3 •- ra
N U- U Z U
cT 'SO
T3 CO
ro c ro
S oo
0} . | '^
c °^ O CT
O « *— W

c * O
o 5v>
2 r o


1
i
however, due to the shape of
the electrode tip, twice daily
attention is required to insure
that bubbles have not formed
on the tips. Complete rejuven-
ation of the probes must be
performed every 30 days.






Z



CM
CM
•o
(U

o jj
w C
I 8.
fe E
£ O
1- U





q
z






CM O
0 °- ™


O CM 10°
0 O i
X O O
VI
Q, _
^ ^^
u P
eu *"
0 §
O . —

"ot
E
*•


0)
C
Z
7 days; No known main-
tainance






Z



CM
CM
T3
0)
L_ +J
2?
E a
fc E
j; o
H 0




^
I"
CM
+ 1







2

O
,11
I
O
Ol
E
§
I
a*
E
q
to


I
U
7 days; No known main-
tainance I




to

Z




Z




^

Z




u.
0
o
+ 1










c
o
Z

u.
0
o
CM
A

.
<;
z
£
3
Q.
E
u
7 days; Internal flow chan-
nel and electrode surfaces
must be cleaned monthly
u
"£ ._
E .E


Q. 3 ~3* CJI

llfil



N
in
•D
i- *j
S 5
I Q.
£ E
f O
H O
§
o
X
S
o
u
I
*T










0)
c
o
Z
5r
o
o o5
o o£
CM «t o E
o o £

.
<;
z
>.
*>
C
•o
c
o
o
7 days; No known main-
tainance






"i
Z



CM
CM
•o
Q>

||
IS.
uE
JC O
I-O


w
'c
3
CM
q
o
+1










ra
Z

CM
CM
a

.
 E
-C O
HO



Ol
E
o
+ 1










c
0
Z


CM«t_
AAl
Ol
E

d


T)
>£
as
.- X
QO
Daily checking of calibra-
tions by comparison with
„
1"s1
Ol *^

"" ^_> <
o "i "t;
if!
a°]

— ^ tj
O. \t\ ^
ro -5*1
in « c
•o
k. -4-t
OS
•A C
II
a> E
£ O
1-0



"Si
E
q
+f










c
o
Z


o o <
fr
00 (
'is
E

5
d
n
3
S
at **
v^
ito
required. Reagent reservoirs
must be refilled every 7 days
>
i
i^

: « c
Sg-o
i "^
: 13
) >T3
j J3 n
H
3 C
>•£ o

:|1




























CI
3 E
Ii
E
2
d
3
3
oce
5.2
KO
7 days; Periodic cleaning of
photocell windows maintains
sensitivity. Light fouling of
the windows does not effect
accuracy within the 7 day
calibration period






Z




Z




^J

Z



D
^

3
">
q
o


i
3
                                                      12

-------
c
o
u


Long Term Stability,
Recalibration and
Maintainance Requirements


x;
"en*-
u c o

01 
HQ



.5
'M
'o
0)
£

M
Selectivity Coefficieni
or
Interferences


S
at
O»
c
IX
c
0

t5 4-*
v •-
Q 3

a)
•s
E
1
j

T3 Ol ^."t?
Daily introduction of stan-
dards maintains accuracy an
demonstrates integrity of th
flow system. Electrodes are
known to be stable over ver;
long periods of time. Reager
reservoirs must be refilled
every ten days
.c
en _
£^

a)
3
" ?

"• .2 E, cT 0

*;  « .5 c s-
u •*•
:o S g
-sic?
i£ 2 i- ~ t
c- £
* O ** w
"J CO _ 2 JC
+ .5


"Si
N
0
+ 1


: c o-^








ir.8 ® ° !n 2
O i^ O • O '
1, 0 „ 0 n 0
i i " i "
o o o

1
o.
I
o


"en
E
o
o
2
o
IL
£






"Si
E
o
1
o
1

•8
Reagent reserviors must be
refilled every four days. Re-
calibration must be perform
for each new set of reagents
f I
£ ° »•£
£-a-£ QC
0 *J ^ CO .2

1 .1 = ^
O "O O -^"O
_ Q u .Q ra

Z



c
o
c
c
3
to
o
u
n
n =
+ | S
E
i-n • S
Residual chlorine inte
feres, but is eliminate.
by addition of thiosul
fate to the sample stn
to
c ^.O
S. °*^
^ 0^
en op >
O "TO CT
U> U CM S


^
Dl
o

>.
'§
15

f




Standards are analyzed auto
matically every six hours.
DASV electrodes must be
replated daily
V
at
3

£2S-

trt j
0 "

*-o
-~ (o y A ra
IS
??
11;
ro -o -*-
in S a
Z




q
z
















(A
4>
C.
Trace organics can in-
terfere by giving false
peaks. Accurate iden-
tification of peak po-
tential indicates the
oresence of interferen

=^
I"
f
o
5i
oi
E
O
o.
o

1
I
•o
fl>
u

=5.
F
I
_
01
E
i
o


•o
s

d
E
I
,_

E
1
o

w
a
5

'S w
Each analysis is discrete.
Electrode must be rejuvenat
between each run by cleanir
with dilute nitric acid





Z
.EC ,
C j2 ,O i^. o
"co 0) '£ ^ 13 c
™ *" > w C O



C
o
c
JC
c
D




<;
z


Unknown



Z



z


§
01 C
SS .
No calibration or maintaina
requirements. Internal elect
ic standard demonstrates in>
strument integrity.





Z
ill*.
Hill

w
3 w n O
13 c u '5
» ° .E °

| ._ ^ 0 a
H > -« .E 2



'o
^ 0)
in ?
+t 2


0>
C
o
o
•a >• ^"
n o Q- 2
•S **" ° a
oflll



z

1
IA
O
o
u
                                                                            o  •«

                                                                            I  I

                                                                            2  f
                                                                            o  ».
                                                                            -Q  o
                                                                            o  13
                                                                            E  ~
                                                                           s  I
                                                                            i  s
                                                                            i  ?
                                                                            I  c
                                                                            E  5
                                                                            S  £
                                                                                    1
                                                                                    I
                                                                               S
                                                                               a
                                                                               5
 f  §
 £  «
 o  .a
 s  c
 •i  •-
 >-  c
                                                                            11    I
                                                                                     t
                                                                                    1
                                                                                    a
                                                                                    3
                                                                                    a
                                                                                    a.
 *  ~
 c  c
 E  3

 I  sl

 i  il  i
q  "i
z  z
—  CM
                                              13

-------
SRM-HOUSED SYSTEMS
General

The Schneider Robot Monitor  (SRM) is a modular  instrument  system
with three stacked sliding drawer cubicles  (Figure  2).   Sensors,
flow cells, flow controlled valves  (except  the  flow interrupt
assembly for hardness and nitrate), inlet reservoirs,  an outlet
header, all interconnecting tubing  and fittings,  a  splash  pro-
tected terminal board for sensor connections, and a high pres-
sure header and tubing designed for automatic sensor cleaning
are housed in the bottom cubicle.   (To facilitate visual inspec-
tion of the entire flow system, all tubing  is transparent  plastic,
and flow cells, reservoirs and sample headers,  machined  from
solid blocks of Plexiglas, are polished  to  be transparent  and
clear.  There are no metal plumbing components  in contact  with
the sample stream.)
   SCHNEIDER
   ROBOT
   MONITOR
   FROM
   COMPUTE*
MUITI POINT

RECORDER

MODULE
                               ELECTRONICS

                               MODULE
                               SENSOR

                               FLOW CELL

                               MODULE
                                                      TO COMPUTER
                                                    DRAIN
  Figure  2.   Schematic  arrangement of Schneider Robot Monitor.
                                14

-------
Flow through the cubicle is illustrated in Figure 3.  The sample
stream enters the cubicle through the inlet reservoir and is
distributed evenly to all flow cells.  To ensure adequate mixing
and to facilitate cleaning, flow cells are funnel-shaped at the
inside bottom.  They are capped at the top by neoprene stoppers
through which the sensor assemblies are inserted.  Each flow
cell is provided with an air bleeder hole to prevent air locking.
The effluent from all flow cells is collected by the outlet
header and discharged to a drain through a common effluent line.

Each individual SRM-housed sensor except temperature is installed
in a separate flow cell.  For convenience in calibrating the
dissolved oxygen electrode, the temperature sensor is installed
as part of the DO electrode assembly.  In the prototype system,
six flow cells are engaged and two positions open, as shown
schematically in Figure 4.


Electronics

All sensor systems are wired to their respective electronic
circuitry (analyzers) through the electrode terminal board in
the flow cell cubicle.  This board consists of 37 barrier strip
terminals and four steatite insulated terminals for coax cable
tie points.  Individual sensor assemblies are provided with
appropriate leads or terminals attached for lead connection.

The eight plug-in type modular electronic analyzers are housed
in the middle cubicle.  The panel-rack assembly is slide mounted
as two separate drawers for easy access.  The panel through
which power is supplied to the analyzers and a test signal
supply for the recorder are located in the analyzer cubicle.

A solid state operational amplifier is the active element in
each analyzer system.  This type of circuitry is particularly
useful for stable amplification of small DC voltages and per-
mits the use of ancillary items for temperature compensation,
linearization of sensor characteristics, etc.  A built-in cali-
bration check and moisture sealed panel meter are provided with
each analyzer module.  Panel meters have individually calibrated
hand drafted scales with an accuracy of ;+0.5 percent.

Outputs from the electrometers are used as inputs to a second
stage operational amplifier.  Offset is introduced in the
second stage to give a proper zero scale, and provision is made
for handling either polarity or measuring electrode input.
Temperature compensation with a wide range of adjustment is
accomplished with a thermistor in the input or feedback circuits
of a third stage operational amplifier.
                               15

-------
                        Out
Sensor Cleaning
    Network
            n
Outlet
Header
       Flow  Control Valve
                                                   Inlet Header
                                                   Flow Cells
                  Front
           To  flow cell

           From  flow cell

           Downward flow
Note:  4 to 8 flow
       cells may be
       used in the
       cubicle.
    Figure 3.   Schematic diagram of the SRM flow cell cubicle,
               top view.
                               16

-------
           s S
liljlii  I	
j I i i 11 [_ J^yM -±.4=f $




Of
o
h-
Z
0
H-
o
co
O
Of
Of
111
Q
111
z
I
u
«/>

r





L





i
L
r. r

















$
^
u.

( l_- .. - -v|r- - ^£ -J<«
1 	 ^--^.-.^
•^ -Tp r^
!_~~ . _x"__ i 	 r

1
r"^~ 1
i"
1
1

1 ' i
j
i 1 1
i 4 1
L._, L. L_,
1 I
_l 	 1 	
1 I '
in ' ^
C I
0) '_
g1
1

^H CN
c
(U 1 'r,
Oj I — °
0 *
\


L
Ul _j Ul
2 $


._Zj 1 i fi
1 r1
ji fO
[ J j 1 tn •
T T f * ^ F
! j i 	 |jj 1 -H dJ
I U """1 ' § 1 -H M
r1 1 | 1 *i S^1
iii i Si **
III 1 7 U (1)
o l o tn
t; i co
	 1-r. "J , 0) ^
i j i in I tn i
-J J ' J «w §
1^1 ow
1 ' ' (-•
III on
j i f 4 --4
'—, ' ! ! ^a
1 -P e
._ _J i tn QJ
1 ! si
I.I rH (0
1 | .J SJ
r~ • oo V-i
i s a"
Q 1" 2 ^ . .
" •- O — (0 rH
I u •- S (0
fl) SJ
1 -• X! Q)
S r^ n
ro *" "^ *
I Q U, ^*
a. "~~ _
< « ^
I O ^ 
-------
Recorder

Sensor output from the SRM-housed systems is displayed (in
addition to the analyzer panel meters) on an Esterline Angus
24-point servo recorder (Model E1124E), located in the upper
cubicle.  The recorder has a 50 millivolt range and is used
with 60 seconds per point printing speed and one inch per hour
chart drive.
Sensors

The five original SRM sensors - temperature, conductivity, pH,
dissolved oxygen, and chloride - are operated and maintained in
conformance with the manufacturer's instructions contained in
manuals supplied with the mobile laboratory.  They sense the
species of interest, generally without interference from any
other species which may be present in the test water.  No rea-
gents are added for their normal operation, but standard solu-
tions are required weekly for calibration.  In field operation,
no unusual problems were observed with sensors for temperature,
pH, and chloride.  Very occasional erratic readout or discontin-
uities in strip chart data from the conductivity system were
caused by bubble entrapment on the sensor surface.  The dis-
solved oxygen sensor is flow sensitive.  Short-term noncumulative
drift in output was observed occasionally when flows dropped to
critically low levels.  In general, one gpm flow at 20 psi is
maintained past the electrode.  Long-term cumulative drifts in
calibration indicate degeneration of electrode response and the
need for rejuvenation.

Unique problems were addressed in developing operational systems
for hardness and nitrate.  Like the DO membrane electrode, the
hardness and nitrate probes are flow sensitive; however, they
produce stable output only under quiescent conditions.  They must
be c.cmp£e.£e£t/ isolated from hydrodynamic effects of the sample
stream.  A flow interrupt system, including chassis, power sup-
ply, and relay for switching AC voltage to a three-way solenoid
valve was installed in the mobile laboratory.  Test water passing
through the valve goes either to waste (during interruption of
flow) or to exchange the sample in the hardness and nitrate flow
cells.  The relay power assembly is activated by digital output
from the computer through the program "FLOW," which sets a digital
output bit high, waits 15 minutes, sets the bit off, waits 45
minutes, then begins the cycle again; i.e., test water flows
through the sample cell for 15 minutes, then the flow cell is
quiescent and the sample completely isolated from test water
flowing to other sample cells for 45 minutes.  During this in-
terval, electrode response equilibrates and valid measurements
of hardness and nitrate are recorded just prior to the time when
samples are exchanged by new flow into the cells.
                                18

-------
In addition to being flow sensitive, hardness and nitrate
electrodes are, as a result of their design and construction,
extremely subject to bubble entrapment and accumulation of
static electricity.  Bubbles are easily trapped in the concave
bottom configuration of Orion liquid junction electrodes.  To
overcome this effect, box-shaped bubble traps constructed from
heavy gauge, coarse-mesh stainless steel screen are attached
to the hardness and nitrate electrodes by a plastic ring which
fits tightly over the outside of the electrode body.  The bottom
of the trap is covered with a flat sheet of plastic to keep
bubbles from reaching the electrode from the bottom, and the
entire trap assembly is filled with glass beads into which the
electrode is seated.  The glass beads effectively prevent any
movement of bubbles to the electrode while allowing completely
free movement of test solution to the sensing surface.

An alternate electrode design (marketed by Corning) was also
evaluated.  Its flat bottom surface was an attractive feature,
but exceptionally high impedance made it totally impractical for
use in the prototype system.  Although the Orion 92 series
liquid junction electrodes have a tendency to accumulate static
electrical charge, a new design with lower impedance and im-
proved shielding has been announced.  It is strongly recommended
that, when they are available, the feasibility of substituting
new liquid junction series electrodes in the mobile laboratory
be evaluated by EPA.

The nitrate liquid junction ion-selective electrode is expected
to give valid analytical_results at nitrate activity somewhat
less than ten mg/1 of NO~; however, chloride and bicarbonate ions
interfere when they exceed the nitrate concentration by a factor
of ten.  The nitrate probe is calibrated over two decades of
nitrate activity; i.e., 1.0 mg/1 at the bottom of the scale,
ten mg/1 at midscale, and 100 mg/1 of NO! activity at full scale.
Most United States tap waters contain less than 1.0 mg/1, and
normal response of the electrode is near the bottom of the scale.
As nitrate concentration in the test water increases, electrode
response is proportional.  If the nitrate concentration should
increase to a dangerous level (e.g., 45 mg/1), the electrode
can be expected to track the change with precision.  Because of
the logarithmic response of the probe, good precision is attain-
able at very low as well as high levels.

Preparation of reagents used with SRM-housed systems is described
in Appendix B.


Interpretation of Hardness and Nitrate Data

In interpreting hardness and nitrate data, activity coefficients
and complexation are important considerations.  Physical charac-
teristics of the liquid junction ion-selective electrode make it


                               19

-------
difficult to design a sample flow cell without imposing pres-
sure problems on the electrode; thus, it is impossible to add
ionic strength buffers and masking agents vis-a-vis the fluoride
monitoring system.  (Addition of an ionic strength buffer would
provide for measuring standards and sample at constant ionic
strength giving direct measurement of ionic concentration;
masking agents would free bound Ca2  and Mg2  from sulfate and
bicarbonate for direct sensing by the electrode.)  Electrode
measurements are adjusted by applying constant factors, deter-
mined experimentally.

Two corrections must be applied:  one for the activity coeffi-
cient and the second,  for that percentage of the total amount
of Ca2  and Mg2  which may be complexed.  If the two factors are
constant over the monitoring period, the equation for the cor-
rection of free uncomplexed species activity to total concentra-
tion is:

            ,+        (M' + )F
          [M2+]T  =  	L_                               (7)
                     fM2+ «M2 +

where:
          (M2 )   =  the monitored value
          fM2+    =  the partition coefficient or the
                     fraction of the divalent metal ion
                     concentration which is unbound
          aM2+    =  the average activity coefficient
                     for divalent metal ions in the
                     monitor stream

Another factor to be considered in operating the hardness moni-
toring system is the effect of the flow interrupt device, which
alternately causes flow/no flow in the system.  Figure 5 is  an
example of typical strip chart recorder output for hardness  data.
Measurement points on half of the output are connected to better
illustrate the nature of electrode response to the flow inter-
rupt system.  During the flow portion of the cycle, response of
the electrode is erratic or noisy; when flow is stopped, the
signal becomes quiet and quite stable.  Hardness is measured
after the signal has stabilized.

On the strip chart recording, shown in Figure 5, the divalent
cation electrode is calibrated to read from 3.16 x lO^M CaCl2
to 3.16 x 10"3M CaCl2 over the 12 inch scale.  This places
1.0 x 10~3M CaCl2, or 100 mg/1 of equivalent CaC03/ at exactly
midscale  (60.0 recorder chart units).  To properly analyze the
data, electrode and EDTA titration data are compared to determine
the fraction of hardness which is free to be sensed by the
electrode.  The complete procedure for converting strip chart
recorder output for free divalent cation activity to total
hardness concentration is as follows:
                               20

-------
Figure 5.'  Typical strip chart recorder output of' hardness*data.
                                21

-------
1.  Determine the free divalent cation
    activity by comparing chart recorder
    output to a calibration curve.
2.  Use the nomogram provided by the
    Orion Company (Figure 6)  to convert
    activity to concentration measurements
    a.   Determine free activity
    b.   Determine conductivity
    c.   Place a straightedge along
        a line through the free hardness
        activity and the measured speci-
        fic conductance and read the per-
        cent correction for total hardness
        concentration where the straight-
        edge crosses the third line.
    d.   Correct the free hardness activity
        to free hardness concentration:

        [Me2*] .  .    , = (Me2 + ),   + C. (Me2 + )c       (8)
              electrode       free   f      free

3.  Determine the fraction of the total hard-
    ness concentration which is complexed and
    unavailable to be sensed by the divalent
    cation electrode.

        [Me  ]electrode   f                         .  .
        [Me2+]EDTA
4.  Determine the total hardness concentration:
                        ,       ,e   ,        ,,nx
                        free    fv     f ree \    (10)
          total ~|           f              J

Example:
1.  Free divalent cation activity = 100 mg/1
    equiv CaCOs .
2.  a.  Free activity = 100 mg/1 equiv CaCOs.
    b.  Conductivity = 320 yu/cm.
    c.  Cf = 5.5 percent or 0.055.

    d.  [Me2+]      = (100 mg/l)+0.55(100 mg/1)
                   =  100 mg/1 +5.5 mg/1
                   =  106 mg/1

    f
    r =
                  electrode   106 mg/1
                 - = - •" — =
            [Me  ]EDTA
                      22

-------
30
1000
900
800
700
600
500
400
500
ZOO
100
90
80
70
60
50
40
30
20
10
8
6
4
2
1
total hardness
— activity
ppm CaCO3
—
-
-
-
-
-
—
-^
— *-.









                           specific
                           conductivity
                           micromhos (25°C)

                             10000--
                              900O--
                              8000--
                              7000--

                              60OO--

                              50OO--

                              4OOO--


                              3000--


                              2000--
                              1000-
                               900-
                               800-
                               700-
                             ^600-
400-
300-
200-

 100-
 50-


 10-
  I-
% correction
for total hardness
concentration
          50
          48
          46
          44
          42
          4O
          38
          36
          34
          32
          3O
          28
          26
          24
          22
          20
          18
          16
          14
          12
          ' 10
           8
           6
           4
           2
           0
       Figure 6.   Conductivity correction  nomogram.
                                 23

-------
              4    rMe2+l      = ! (100 mg/l)+.055(100 mg/1)
                        total   (          0.82

                             = 129 mg/1  of  equiv  CaCOa
Nitrate measurements are not expected to require the degree of
correction described for the hardness system.  Nitrate is not
found in a complexed state and its concentration usually need
not be known with great accuracy, unless it is present near
the maximum permissible level specified in the Drinking Water
Standards (1).   Unfortuantely, NSF cannot present any actual
field data obtained from the nitrate sensing system; i.e., the
mobile laboratory was not operated in an area where nitrate
activity exceeded the detection limit of the sensor.
COMMERCIAL SYSTEMS

FREE AND TOTAL RESIDUAL CHLORINE

General

Residual chlorine is the single most important parameter in
drinking water quality.  Its presence in a finished water is
generally considered to indicate that disinfection is complete,
and the water is safe to drink.

At the water treatment plant, chlorine application is highly
mechanized and carefully controlled.  It must be added to the
treated water at a level sufficient to accomplish disinfection
but not sufficient to incite taste and odor complaints from the
consumer.  A total residual of 1.0 mg/1 in water leaving the
treatment plant is generally assumed to meet these objectives.

In traveling through the distribution system, it is entirely
possible that the level and type of chlorine residual will
change.  These changes may occur as a function of other parameters
which characterize the quality of the water itself, or as a re-
sult of conditions within the transmission pipeline.  It is
therefore desirable, if not necessary, that residual chlorine
levels be monitored in the distribution system.
Theory

Regardless of the form in which chlorine is added, it reacts
with the water to form hypochlorous acid or hypochlorite ion:

          C12 + H20 *==* HOC1 + H+ + Cl~                       (11)
         gaseous         hypo-
         chlorine        chlorous
                        acid
                                24

-------
          Ca(OCl)2 + H20 S5=5 Ca2 + + 2 OC1~ + H20              (12)

          calcium                     hypo-
          hypo-                       chlorite
          chlorite                    ion


          NaOCl + HZ0 ^=5= Na+ + OC1~ + H20                    (13)

          sodium
          hypo-
          chlorite


The equilibrium condition of HOC1 and OC1  is described by
Equation 14:

          HOC1 ^==B OC1~ + H+                                  (14)
                          K. = 2.5 x 10~8
                           i

Thus, the ratio of HOC1 and OC1  present following disinfection
is determined by pH of the treated water:


         log  [OC1"] = PH + log K                              U5)
              [HOC1]

The effect of pH on [HOC1]: [OC1~] species distribution is illus-
trated graphically in Figure 7.  From this figure, it is apparent
that a mixture of the two species exists in most United States
drinking water supplies.  In either form, it is referred to as
"free residual chlorine."

The term "total residual chlorine" includes both free and
"combined" forms.  The most common combined chlorines are mono-
chloramine and dichloramine, found in the presence of ammonia:

          NH3 + HOCl -> NH2 Cl + H20                           (16)
                       mono-
                       chlor-
                       amine

          NH2C1 + HOCl -> NHC12 + H20                          (17)
                       dichlor-
                       amine

Trichloramine occurs in tap water only infrequently; e.g., with
extreme pH or breakpoint chlorination:

          NH2C1 + HOCl -> NC13 + H20                           (18)
                         tri-
                         chlor-
                         amine
                                25

-------
                                           90
                                       10  11
                                           100
   Figure 7.  Distribution of hypochlorous acid and hypochlorite
              ion in water at different pH values and tempera-
              tures (2) .
The rate at which the various forms of combined chlorine achieve
disinfection is significantly less than that of free chlorine.

Nine analytical methods for measuring residual chlorine are
described in Standard Methods (3).  One of these, the amperometric
technique, is used in the mobile laboratory in both the free
and total residual chlorine analyzers, and in the manual titrator
used to standardize the analyzers.

In the amperometric procedure, a galvanic cell is developed;
i.e., a current flows between immersed electrodes when a free
oxyhalogen acid; e.g., HOCl (hypochlorous acid), is present in
the test water.  Null point amperometric titration describes the
principle of this method.  Residual chlorine is quantitatively
removed from solution by reaction with a titrated reducing agent;
e.g., phenylarsine oxide.  As reducing agent is added, current
resulting from the presence of HOCl in the test water is reduced.
Titrant is added until there is no further decrease in cell
current.  The titrant required to reach this end point (i.e.,
no further current decrease) is directly proportional to the
level of residual chlorine present in the water.
                                26

-------
Equation 19 describes the principle of electrolytic reduction of
an oxyhalogen acid (HOX) at an inert cathode in the cell:

          Cu°|Cu2+||x~JHOX, Au (or Pt)                        (19)

          (or Cu°|Cu2+[|C1~|HOC1, Au  (or Pt))                  (20)

Half reactions for Equation 20 include:

          Anodic:  Cu° = Cu2+ + 2e                            (21)
                              E° = -0.337 volts
                               a
          Cathodic:  HOC1 + H+ + 2e = Cl~ + 2H2O              (22)

                              E° = 1.28 volts
                               c
Combining Equations 21 and 22, the overall reaction is:

          Cu° + HOC1 + H+ = Cl~ + Cu2+ + H20                  (23)

and the standard cell emf will be,

          E°    = E° + E° = 0.94 volts                        (24)
           CG-L X    3.    C
Consequently, the cell potential at equilibrium, at 25°C, will
be'                                   2 +
          E     = 0.94 + 0.03 log  -^	] [C1 3                (25)
           Celi                    [HOC1][H+]

Thus, cell emf is a function of [Cu2], [HOC1], and  [H+].  The
potential of the large surface area (nonpolarizable) copper anode
is a constant (as  [Cu2 ] in the bulk of the test solution approaches
zero) and is determined by the rate of dissolution of copper  from
its surface.  In practice, [H ]  is kept constant by the addition
of pH 4.5 sodium acetate/acetic acid buffer.  Cell emf, then, is
determined by - and varies with - the level of HOC1 in the testwater,

Experimental

Presently there are at least four amperometric-type residual
chlorine analyzers which are commercially available.  In two  of
the four systems, potential is measured across an extremely high
resistance  (rather than current flowing in the galvanic cell, mea-
sured across a low, fixed resistance).  Three of the four systems
use gold/copper electrode couples; the fourth, a platinum/copper
couple.  The gold or platinum electrode serves simply as an inert
surface at which the HOX is reduced;  thus, the theory of operation
of each of the four instruments is essentially the same.  Cell
                                 27

-------
geometry for each of the systems includes a. large surface area
copper cylinder  (the anode), surrounding a wire or rod of the
noble metal  (cathode).

Cleanliness of the anode is vitally important.  Fouling causes
surface area variations and large drifts in potential over short
time intervals.  As a result, electrode cleaning devices are
included with each instrument.   One of the systems has the cell
filled with plastic balls.   A motorized arm drives the balls in
the cell.  Two of the units are filled with plastic pellets
which abrade a rotating electrode.  In the fourth unit, grit
in the sample cleans the electrodes, agitated by velocity of
the flowing sample stream.

It is also important to maintain constant ionic strength in
the sample stream.  Significant changes in total ionic strength
(measured as total dissolved solids or as conductivity) result
in HOX activity coefficient variations without changes in analy-
tical concentration.  The high strength sodium acetate/acetic
acid buffer adjusts ionic strength in addition to controlling pH.

Signal variations resulting from temperature effects on cell emf
require compensation.  A typical galvanic residual chlorine
analyzer is illustrated schematically in Figure 8.  In this
figure, R2 is a variable resistance in parallel with the sensing
electrode.  It can be manually adjustable  (assuming that tempera-
ture of the sample stream remains constant between calibrations),
or a temperature dependent transducer  (thermistor) can be used
in which resistance varies automatically with changes in temper-
ature.

Total residual chlorine is measured by adding potassium iodide
to free combined residuals:

          NH2C1 + I~-*—r NH2I + Cl"                            (26)
          mono-         mono-
          chlor-        iod-
          amine         amine

          NHC12 + 2I~^=?NHI2 + 2C1~                          (27)
          dichlor-
          amine

Because iodamines are highly unstable  in aqueous solutions, the
reactions continue to produce hypoiodous acid, an oxyhalogen
acid which is  sensed by the galvanic cell:

          NH2I + H2O^=^NH3 + HOI                             (28)
                              hypo-
                              iodous
                              acid

          NHI2 + H2Oi=;NH2I + HOI                            (29)

                                 28

-------
                                  MEASUREMENT CIRCUITRY
                                         POTENTIOMETRY
                    HIGH R, R4
       r
                 I6  LOWR, R3
               	A/WV-
                 H-)
                 CU
                                         GALVANOMETRY
                             AU
                                      TEMPERATURE COMPENSATION
                                        -AUTOMATIC (THERMISTOR)
                                        -MANUAL
                                      CALIBRATION RESISTANCE
Figure  8.   Schematic illustration of  residual chlorine
             analyzer.
                              29

-------
Two Capital Controls residual chlorine  analyzers,  Models 871
for free residual and 872 for total  residual,  and  a Wallace and
Tiernan  (W&T) amperometric  titrator,  Series  A-790, are used in
the mobile laboratory.  The Capital  Controls analyzers are wall
mounted galvanic units, operated  continuously.   The electrodes
are cleaned by 200 polyvinylchloride (PVC) balls rotating in
the test stream.  A schematic illustration of these units is
shown in Figure 9.  The manual W&T titrator  is  used for daily
calibration of the continuous analyzers.

Each continuous analyzer has a gold/copper couple.  Chlorine is
the active electrolyte and  produces  a current measured with a
microammeter.  Reagents are introduced  to the analytical cell
by a cam feeder which operates every ten  seconds.   Reagent
reservoirs are two liter bottles  with sufficient capacity for
approximately seven days.   In both units, a  buffer addition
adjusts pH of the test water to  4.5, and  adjusts ionic strength.
The free chlorine analyzer  responds  directly to free chlorine
in the sample stream.  A second  reagent,  potassium iodide, is
added in the total chlorine analyzer as a masking  agent to con-
vert combined to free forms which can be  sensed by the electrode,
Preparation of reagents used for  monitoring  residual chlorine
is described in Appendix B.
           ROTATING
            STRIKER
        CYLINDRICAL
          COPPER
         ELECTRODE
        DRAIN
                                             SAMPLE
                                              INLET
                                                     SCREEN
     ELECTRIC
     SIGNAL
       TO
     MONITOR
                                                OVERFLOW
                                                TO DRAIN
                                    GOLD
                                  ELECTRODE
       Figure 9.  Schematic illustration  of  C12  analyzer,
                                30

-------
Results and Discussion

During more than three and a half months of continuous field
operation and up to three years of intermittent operation in the
laboratory, the residual chlorine analyzers operated with com-
plete satisfaction.  Maintenance was limited to routine weekly
preparation of buffer and masking agent reagents and daily cali-
bration.

In Chicago, the first two monitoring sites were on heavily used
distribution mains where residence time was normally on the
order of 15 to 20 hours, and total chlorine residuals, consis-
tently in the range of 0.7 to 0.9 mg/1.  Each of these sites was
approximately ten miles from the filtration plant.  At the
third and fourth monitoring sites, Des Plaines and Calumet
Harbor, the unit was located far out, near the ends of the distri-
bution systems.  At Des Plaines, the water had been transported
nearly 20 miles through main transmission lines and the residence
time was unknown.  Free chlorine averaged 0.14 mg/1 and total
chlorine, 0.28 mg/1.  At Calumet Harbor, the monitoring site was
only eight miles from the water treatment plant, but near the
end of a large main used mostly for industrial fire flows.  As
a result, residence time in the main was approximately 35 hours.
Free chlorine at this site averaged 0.24 mg/1 and total, 0.40
mg/1 over a nine-day period of observation.

Changes in residual chlorine levels during a main flush at the
Calumet Harbor site are described on pages 164 and 165.


At Philadelphia International Airport, there was opportunity to
evaluate analyzer response to ammoniated samples.  Free chlorine
ranged from "not detectable" to 0.1 mg/1, and was generally
<0.05 mg/1.  With no apparent difficulty, the free chlorine
analyzer differentiated between free chlorine and loosely com-
bined chloramines.  (At the operating pH; i.e., 4.5, the pos-
sibility of converting some combined forms to free chlorine
was questioned (4).)  This observation was confirmed by ampero-
metric titration at pH seven.

At the fourth location in Philadelphia, the Gulf Oil Refinery,
sudden variations in chlorine residual were observed.  A portion
of the recorder output from this site is shown in Figure 10.
The capacity of these analyzers to detect sudden changes in
both increasing and decreasing concentrations is clear from the
output, confirmed by titration at the time of occurrence.  The
variations were attributed to the mixing of water from two dif-
ferent storage locations.
                               31

-------
    iTxtrations
     support these
     data
                     Titrations
                     support these -
                     data         __
                          ~T>   ~-j" "
                              L_4_J__u
                                                  —1_
                                             T
                              1  i   ,  :  .  , ,
                              f~~~ 1 Total
                               5-1 Residual —
                                i Chlorine   i
                             11—i—f—- -
                              I y- I
                              1"D*1—
Figure 10.
Free and total residual  chlorine variations
at Gulf Oil, Philadelphia.
                           32

-------
TURBIDITY

General

Turbidity is defined as, "an expression of the optical property
of a sample which causes light to be scattered and absorbed
rather than transmitted in straight lines through the sample"
(5).  It is caused by the presence of suspended matter in the
sample stream.  The American Water Works Association has developed
a goal of 0.1 Jackson Turbidity Units (JTU) for drinking water
(6); the EPA standard is five JTU or less  (1).


Experimental

The Hach CR low range turbidimeter, Model 1720, continuous flow
nephelometer is used to measure turbidity in the mobile labora-
tory.  The measurement is accomplished by passing a strong beam
of light through the sample.  Turbidity, present as fine parti-
cles, scatters a portion of the light beam which is measured by
two photocells submerged in the sample.   If the sample is free
of turbidity, no light is scattered and no light reaches the
measuring cells; conversely, the presence of turbidity in the
sample results in light being scattered.  Some of this light
falls on the photocells, and is registered on the readout meter.

The meter is calibrated with a reflectance rod of 5.0 JTU.
Schematic illustrations of the turbidity system and the turbidi-
meter in the mobile laboratory are shown in Figures 11 and 12.

This system was essentially trouble free throughout its use in
the NSF project.  Data acquired routinely during field assign-
ments are reported in the respective reports of these activities
(7,8,9) .
CORROSION RATE

General

Corrosion is a serious-or potentially serious-problem in most
potable water distribution systems.  Taste, odor, and appearance
problems resulting from corrosion of transmission lines may make
water unacceptable to the consumer, and corrosion induced pipe-
line perforations may permit back siphonage of contaminated
groundwater into the system.

A major investment of the water supply industry is the distri-
bution network.  Continuous effort to produce a water with no
tendency toward pipeline corrosion is a principal treatment
objective.
                               33

-------
          REMOVABLE HEAD
         (CONTAINING ENTIRE
       ELECTRICAL ASSEMBLY)
           TO DRAIN
       SAMPLE DRAIN
 TO MASTER INDICATOR
                                           113V-COCY IN
                                         HOTOCELLS
PARTICLES OF SUSPENDED
MATTER REFLECT LIGHT
WHICH IS MEASURED BY
THE PHOTOCELLS.

STANDARD
REFLECTANCE
ROD (SHOWN IN PLACE)
         DRAIN PLUG

                 SAMPLE INLET


Figure  11.   Schematic illustration of  turbidity system.
                                 34

-------
               *
 o
 -U
 m
 n
 o
 X!
 m
                         '"I
                                     X
                                     Q
                                                           i
                                                           (U
                                                          X!
                                                          -p
                                                           G
                                                          •H
 (1)
 -P
 (U
 g
 •H
 T3
 -H
 XI
 M
                                                          o
                                                          -H
                                                          -p
                                                          (0
              o
                     —I
r-t
rH
•H

 O
•H
-P
 (0
 e
 0)
X!
 O
CO
O

GO
3
o.

O


O
                                                          CN
                                                          H

                                                          (U
Cn
                          35

-------
Theory

Electrode potentials, expressed as intensity measurements of
reversible equilibrium conditions, and the Butler equation for
electrochemical kinetics, are basic considerations in describing
the current-potential relationships associated with corrosion.

It is assumed that equilibrium conditions are maintained in the
thermodynamic development of electrode potentials; i.e., there
are both charge balance and mass balance in electrode reactions
for a metal  (electrode) in a solution of its ions:

          M fe^ M2+ + 2e~                                      (30)

Thus, at equilibrium, the same velocity occurs regardless of
the direction of the reaction.  The electrode potential is
determined by the metal ion activity, expressed by the Nernstian
relationship:

                   RT     [0x]
          E = E° + — In —2	                                (31)
                   2F     [Red]

where:

          E  = measured electrode potential
          E° = standard electrode potential

In corrosion, however, the anodic and cathodic reactions are
usually different.  Although there is charge balance, there is
also mass flux.  The anodic mass flux is, in fact, corrosion.

A typical corrosion cell is shown in the following series of
equations:

         Anodic:    Me -> Me2+ + 2e~                           (32)

         Cathodic:  2H+ + 2e~ + H2                            (33)
or
                    1/2O2 + H20 + 2e •+ 40H                    (34)

Thus, there  is a net mass flux, hydrogen gas evolves from the
system, and  corrosion products are formed from metal dissolution.
The metal is not at equilibrium potential  (as it  is in Equation
30).  It exists in transient or steady state with its potential
a function of the rate determining step in the reaction.  If  the
rate determining step is  the anodic process, the  potential of
the metal shifts in a positive direction; if it is cathodic,
excess electrons will accumulate on the metal, causing a negative
shift in potential.  Because of the kinetic effect and the
possibility  that another  cathodic reactant may be available,  a
corroding metal surface in neutral solutions rarely acquires
equilibrium  potential values.


                                36

-------
In addition to the energetics of corrosion; i.e., the tendency
of a reaction to proceed as measured by its thermodynamic pro-
perties, reaction kinetics are vitally important considerations.
The basic equation for electrochemical kinetics is expressed as:
                                      RT     •                 (35)
where :
          i  = polarization current density
          ip = exchange current density
          a° = charge transfer coefficient
          n  = polarization overvoltage

The overvoltage is defined by:

          n = E. - E                                          (36)
               i    eq

where :
          E. = electrode potential with an impressed
           1   current, i, flowing through the cell
          E  = electrode equilibrium potential  (with
            ^  no external current flowing through the
               cell)

The exchange current, i , represents the spontaneous rates at
which the anodic and cathodic reactions are proceeding.  Because
these are not identical for corrosion conditions, i  indicates
the rate of metal dissolution for the anodic reaction and the
rate of the cathodic reaction, whatever it might be.

Thus, i  is an exact measure of the rate of reaction if the true
current°density relationship is known.  It is difficult to mea-
sure because true surface area is not easy to determine.  It
is usually known within a factor of two, so use of o.ppo.A.en;t
surface area in calculating i  does not invalidate the results.

The effect of major current excursions on naturally corroding
metal surfaces is important in measuring corrosion rate by
polarization techniques.  Once an electrode has been used for a
complete polarization curve, its surface has changed irreversibly
"from its original freely corroding condition.  If, however, only
small increments of current or potential are applied to the
electrode surface , the perturbation to the naturally occurring
corrosion rate is small and might not affect the long term
corrosion rate of the sample being measured.

Thus, the overvoltage of a reversible electrode reaction is a
linear function of the applied current density for values of
overvoltage only slightly different from the reversible potential.

                                37

-------
This relationship can be developed from Equation 31, using, for
small values of x, the approximation:
                   x
                                                              (37)
For small values of overvoltage, n/ Equation 35 can be rewritten
as:
                        2*nl - fn- (1-a)%
                        RT  J   L      RT
                                                              (38)
                .   F_

                ""0 RT
          T
          I di
          L
              l
RT
i
 ~
i
                                                              (39)
                                        (40)
                n+o         -0

From the empirical Tafel relationship:

          n = a + b log i
                                                              (41)
where:
          a and b = constants for a particular system
               i  = i  + i
                p    a    c
Equation 40 becomes:


           'dn 1
                       b  b
                        a  c
                                2.3  i
                                                              (42)
where:
          subscript "a" = anodic Tafel slope for b
          subscript "c" = cathodic Tafel slope for b
The differential dri/di  is the polarization resistance, expressed
as ohms.  In corrosion measurements, the term i   (the exchange
current for reversible systems) is replaced by icorr-  The cor-
rosion current, icorr/ per unit area represents the rate of the
corrosion reaction and is directly proportional to weight loss
(the classical method for measuring corrosion).
Simplifying Equation 42:

          AE _ 	

          Ai   2.3(b.
                  babc
                        V
                                                              (43)
                                38

-------
or,
                          b  b
          log AE = lQg - a_^ -- lQg .                    (44)
              AI       2.3(b  + b )         corr
                            a    c

Plotting AE/AI versus log i     produces a slope equal to -1.
When the cathodic reaction is controlled by diffusion, b  is
approximated by infinity, and Equation 44 expressed as:
                                                              (45)
          AI   2.3i
                   corr
or
                =     .                                        (46)
           C0rr   2.3   AE
Experimental

A corrosion rate instrument, capable of monitoring the instan-
taneous corrosion rate of metals in their use-environment, is
installed in the mobile laboratory (Figure 13).   This instrument,
manufactured by the Petrolite Division of Petreco Corporation, uses
a "polarization of adjusted known surface area admittance" techni-
que with three identical electrodes machined from the material of
interest.  In the measurement cycle, a current is imposed between
the working electrode and an auxiliary counter electrode sufficient
to polarize the test electrode ten millivolts with respect to
the reference (third) electrode.  The magnitude of the current
required to effect the ten mv polarization is directly propor-
tional to the corrosion rate, and is interpreted as mils per year
of surface corrosion.  In use, this instrument is set up with
three active stations.  Station one is an internal calibration
device used to monitor correct operation of the instrument.
Stations two and three are used for actual corrosion rate mea-
surements.  For station two, a set of freshly polished electrodes
is installed in a flow cell, and corrosion rate measurements are
initiated immediately.  Data obtained with these electrodes over
several days or weeks of monitoring are used to provide informa-
tion about the rate of attainment of corrosion equilibrium in
the test medium.  For comparison, a second set of electrodes is
sent ahead to a new monitoring site for advance exposure  (passi-
vated) to the test water.  When the laboratory arrives, these
electrodes are installed in the test cell to provide comparative
equilibrium corrosion rate measurements.
                                 39

-------
 CORROSION  RATE  MONITOR
                                                      DRAIN
 TO COMPUTER


MEASURING
ELECTRONICS
STRIP- CHART
RECORDER
SENSOR
SWITCHING
RELAYS

SENSOR

FLOW

CELLS
 FROM COMPUTER
                                                       TAP

                                                       WATER

                                                       SAMPLE
Figure 13.  Schematic illustration of corrosion rate monitor in
            the mobile laboratory.
Results and Discussion

Mechanically, the corrosion rate monitor proved to be a trouble-
free system in Chicago; however, significance of  the Des Plaines
data is questionable, because the electrodes were not passivated
in the test waters.

The time required for passivation is a function of corrosion  and
deposition characteristics of the test water.  In Grand Rapids,
Michigan, for example, where water is heavily scale forming,
passivation is accomplished in a matter of hours.  In Ann Arbor,
where water is corrosive, months are required.

New (repolished) electrodes were used at three of the four
Chicago area test sites.  Moving the van on a weekly basis  (three
of the four locations) precluded passivating and  acquiring  long-
term data with the same electrode systems.  In preparing for  the
Philadelphia trip, new electrodes were sent ahead for passiva-
tion in the water in which they were to be used.  Because the
corrosion rate monitor is a four channel system,  passivated and
new electrodes can be used simultaneously to develop both kinetic
and long-term corrosion rate data.

Unfortunately, the Philadelphia trip produced no  meaningful cor-
rosion rate data.  The Chicago data were meaningful; i.e.,  daily
mean corrosion rates at the first location ranged from  2.2  to
3.0 mils per year, and at the second location, 2.0 to 2.2 mils
                                40

-------
per year.  These sites are served by water from the Central Water
Filtration Plant (CWFP),  where caustic is added for corrosion
control.  In Des Plaines, the daily mean ranged from 3.8 to 6.8
mils per year, reflecting different mix ratios of Chicago CWFP
and Des Plaines treated supplies.  At Calumet Harbor, the daily
mean was 5.0 to 7.6 mils per year for water treated at Chicago's
South Water Filtration Plant  (SWFP) where no corrosion inhibitors
are added in the treatment process.

These observations suggest that the Petrolite instrument has
good potential for application in the water supply industry;
however, improvement in its ruggedness and mechanical integrity
are recommended.
OTHER SYSTEMS

CALCIUM CARBONATE DEPOSITION TEST

General

Calcium carbonate  (CaC03) deposition is an important consideration
in water utility operation.  Excessive deposition in the trans-
mission pipeline reduces carrying capacity; conversely, corrosion
may occur when no protective film is formed.  Either event re-
sults in serious economic losses to the community; thus, a prac-
tical objective of the water treatment operator is to produce
a finished water which will deposit a light protective film in
the distribution network.


Theory
Water which is either over or undersaturated with respect to
CaC03 tends to undergo chemical change and approach a state of
equilibrium.  Thermodynamically , this tendency can be expressed
in terms of change of free energy  (AG) , as shown in the following
reactions :


          CaCO3 ,v + Ca2+ + CO32~

                     AG = RT ln            -
                                     K
                                      s
          CaC03/ x  + H+ + Ca2+ + 2HCQ3
                     AG = RT m        [HCOa"]                 (48)
                                   Ks[H+]
                                41

-------
          CaC03, v  + H2C03* + Ca2+ + 2HC03

                     AG = RT in  [Ca2+][HCQ3-]2K2              (4g)

                                  KjK  [H2C03*]
                                     s
                                        /
where :
          KI = first dissociation constant for H2C03*
          K2 = second dissociation constant for H2C03*
          K  = solubility product of CaC03
    [H2C03*f =  [H2C03]+[C02]

For oversaturated waters, AG is negative; for undersaturated
waters, AG is positive.

It is common practice to characterize the degree of departure
from calcium carbonate equilibrium conditions by calculating
the hydrogen ion activity at which, without change in total
alkalinity and calcium content, a water would be in equilibrium
with solid calcium carbonate (3); i.e.,


            +      [Ca2+]act[HC°3~]act
          [H+]   = - Set - act                        (5Q)

               q
The H  activity of the test water is correlated with a hypothe-
tical value for H .   When  [H ]     >  [H ]   , the water is aggres-
sive with respect to CaC03; when  [H ]   eq<  [H ]   , the water
will deposit CaC03.                    meas       eq

Langelier (10) proposed using the difference between measured pH
and calculated equilibrium pH as a measure of "stability;" i.e.,
the tendency of a water to be scale forming or corrosive.  He
defined the term "saturation index" (SI) as follows:

          SI  =  pH     - pH                                  (51)
                 ^ meas   ^ eq
where pH   is defined from Equation 50 as:
                                          - log [HCO,']        (52)
The concentration of HC03  can be calculated from analytically
measurable parameters; e.g.:

                     [Alk]  - Kw/rH+i +  [H+]
           [HC03~] = - - -   I   J -                   (53)
                              2K
                                   [H+])
                               42

-------
and

          [HC03~] = XiCT                                      (54)


where:
          C  = the analytical concentration for total
               carbonates

and
          *r    I HCO 3  I                                        / r- r- \
          Xi = 	                                        (5~>)

                 CT

Within the pH range of many natural waters  (7.5 to 9.0), Xi  is
close to unity and [HC03~] - CT.  Below pH 8.3,

          [HC03~] = [Alk]T                                    (56)


Larson and Buswell (11) added an ionic strength correction  (u):

  pH  = log(K /K2) -  log  [Ca2+]  .  - log[Alk]_ + 9.30
    6Q       S                 ciC"C           J.
                         2.5                                  (5?)
                     5.5y + 5.3/^7+1

Accordingly, from Equation 51, a positive saturation  index  indi-
cates a scale forming water; a negative index indicates a scale
dissolving water.

Another parameter, the "stability index"  (S) , proposed by
Ryzner  (12) , refers also to equilibrium pH:

          S  =  2PHeg - PH                                    (58)
According to Ryzner, whose work was principally empirical  and
based on weight changes of coupons in a test water, waters with
a stability index of <1 . 5 are scale forming; waters with indices
>7.5 are increasingly scale dissolving.
Calculation of the calcium carbonate stability index, using
either Equation 51 or 58, has been widely applied in the water
treatment industry; however, there are a number of  limitations
to .this procedure.  The calculated index provides information
about the extent of departure from equilibrium only; i.e. , the
directional tendency of a water to deposit or dissolve  a calcium
carbonate film.  Kinetic aspects of water stabilization or rates
of attainment of equilibrium are not considered; thus,  the rate
                                43

-------
and amount of calcium carbonate deposited or dissolved cannot
be predicted.  It has been demonstrated experimentally that,
in the absence of crystallization nuclei, oversaturated solu-
tions of calcium carbonate may remain oversaturated for years
(13,14,15).  In addition, the presence of Mg2 , polyphosphates,
or dissolved organic matter (present in most water supplies) may
significantly influence the rate of calcium carbonate precipi-
tation or dissolution.  This effect is not necessarily pH
dependent.
                7 +
The effect of Mg   on the rate of calcium carbonate nucleation
was studied by Pytkowicz  (13).  These studies indicate that
magnesium-free artificial seawater yields much shorter times of
nucleation (in order of minutes) than natural seawaters (in
the order of several hours).  Enrichment of water samples with
magnesium was found to significantly decrease the rate of cal-
cium carbonate nucleation.  This was explained as a result of
formation of magnesium carbonate complexes, which are highly
stable with higher alkalinities.  These complexes are very
effective in inhibiting calcium carbonate crystal formation.

Polyphosphates and trace organics interfere with the rate of
calcium carbonate deposition because of their tendency to form
complexes.  This causes a sequestering effect which keeps the
calcium ions in solution.  This effect can be demonstrated by
comparing results of Ca2  determinations using a calcium ion-
selective electrode and the EDTA titration technique.  The former
measures free Ca2  and the latter, total Ca2  concentration.
If EDTA titrations are greater than ion-selective electrode
measurements, all of the calcium present in the water supply is
not in the free ionic form.  It is distributed between the free
form and a fraction which exists as solubilized complexes,
generally organo- or phospho- in nature.

McClanahan (16) studied the role of dissolved oxygen and the
mechanism of CaCOs film formation with respect to cast iron
corrosion using a three electrode system, the rotating ring
disc electrode.

In the iron corrosion process, the metal undergoes anodic dis-
solution at iron starved regions  (Equation 59), and oxygen is
cathodically reduced to hydroxide at oxygen rich regions
(Equation 60).  The occurrence of hydroxide raises the pH in
the vicinity of the corroding metal surface and may cause the
ion product of CaCOs to surpass the solubility limits, depositing
CaCOs on the corroding metal surface:

          Fe,  N -*- Fe2+,   , + 2e~                              (59)
             (s)        (aq)

          1/2 O2 + H20 +  2e~ -> 2OH~                           (60)

          2Ca2+ + 2HC03~  + 20H~ -> 2CaC03 /  v +  2H2O            (61)
                                         \ s)

                               44

-------
 Thus,  it  is  clear  that  a water which  is  under saturated  with
 CaCOs  may deposit  CaCOs regardless  of its  Langlier  Index;  i.e.,
 in  theory, any pipe material  capable  of  supplying electrons may
 cause  CaC03  to deposit  from waters  which contain both calcium
 and carbonate ions .
 Rotating  R-cng

 The  rotating ring  disc  electrode  is  a  voltammetric  system with
 rigorously  defined hydrodynamic transport.   Consider  a flat disc
 rotating  with  uniform angular  velocity about an  axis  perpendicu-
 lar  to  the  plane of the electrode.   Because  of the  centrifugal
 force acting upon  it, the  layer of adjacent  solution  adheres to
 the  disc  and acquires a radial flow.   The  radial flow displace-
 ment is accompanied by  an  axial flow from  the bulk  of the test
 solution  to the surface of the electrode  ( 17) .   Figure 14 is a
 schematic illustration  of  fluid flow in the  rotating  electrode
 system.
                                       electrode
                           v
                            X
     Figure  14.   Fluid  flow  in  the  rotating electrode system.
Under constant rates of rotation, the thickness of the hydro-
dynamic boundary layer is a function of the angular velocity
of the rotating disc (w) and the kinematic viscosity of the test
solution (v) (18); thus, mass transport to the electrode surface
depends solely on the rate of rotation of the disc electrode.
                                45

-------
It is assumed,  based on the Nernst diffusion layer theory,  that
a major part of the concentration gradient occurs across  a
stagnant liquid layer,  the Nernst diffusion layer (S )(19).
The hydrodynamic boundary layer (SH)  can be equal to or greater
than S .
      n

The rate of mass transport; e.g., molecular oxygen in CCDT,  in
the x direction, normal to the electrode surface, can be  ex-
pressed as follows:


          i = D!^-vxi                                 (62)

where:
          C = dissolved oxygen concentration
          t = time
          D = oxygen diffusivity coefficient
         vx = axial liquid velocity

Under steady state conditions:

          i£ = 0                                             (63)
          at

and

          vx9£=Di!£                                       (64)
             9x     ax2

If the  rotating disc electrode is maintained at a potential at
which the following reaction will proceed:

          02 +  2H2O +  4e~ + 40H~                              (65)

the  concentration  profile  shown in Figure 15 will prevail in
the  vicinity of the electrode.

Because electrolytic reduction of molecular oxygen at the elec-
trode surface  (Equation  65) is more rapid than oxygen mass trans-
port, the concentration  of oxygen at the surface of the electrode
approaches  zero  (C =  C  ^  ).  In addition, a linear concentration
gradient is assumed to exist within the Nernst diffusion layer
 (Figure 15).

The  current flowing in the electrode is proportional to the con-
centration gradient at the electrode surface ,.

                                              (£)
                                                   X+o.
                                46

-------
         'B
       (X)
                                       CB  ~
                                Assumed linear
                                  concentration gradient
              Distance Normal to the Electrode Surface (x)


  Figure 15.  Concentration profile at electrode surface under
              steady state conditions, in the absence of CaC03
              film.
and can be expressed as:
where:
                                                              (66)
                      X->o
          i = diffusion current
          n = number of electrons transferred per mole
              of reactant
          F = the Faraday constant
          D = diffusion coefficient
          C = concentration of electroactive species
          X = distance to the electrode surface
Under steady state conditions:

         /3CN
                   C -C
                    B  o
                                                              (67)
                     N
                               47

-------
where :
          CR = bulk concentration
          C  = concentration at the electrode surface
          6  = diffusion layer thickness

If C  =0, the limiting current is:

                    CR
          i0 = nFAD -*£                                        (68)
           *        °N

The thickness of the diffusion layer can be estimated  from  (64) :

          6N = kD1/3oT1/2v1/6                                 (69)


Combining Equations 68 and 69:

          i = K nFAD2//3co1//2v~1//6C0                            (70)
                                 B

Thus, according to Equation 70, the current is  limited by dif-
fusion in an inter facial Nernst layer, the thickness of which
is solely dependent on the angular velocity of  the  rotating disc.
The theory of rotating ring disc electrodes is  discussed in much
greater detail in the literature  (18,20,21,22).


Ca£.Cxcam Carbonate, f-iim fotimat-ion

The steady state current for the electrolytic reduction of  dis-
solved oxygen at the surface of the rotating disc electrode re-
mains constant with respect to time, as shown in Curve 1, Figure
16.  This is true if P  , temperature, and the  speed of rota-
tion are constant, and 2the test solution contains  no  Ca2  .   The
oxygen limiting current decreases with time in  the  presence of
Ca2 , and the rate of decrease is directly proportional to  the
concentration of Ca2 , as shown in Curves 2 through 5,  Figure 16.
This decrease in current is attributed to the formation of  CaC03
film on the surface of the electrode, according to  the following
reactions :

          02 + 2H20 + 4e~ -*• 40H~                              (71)

          OH~ + HC03~*=^ C032~                                (72)
          Ca2+ + C032~^=5=CaC03 +                              (73)
Formation of CaC03 film on the electrode  starts  at the  center of
the disc and grows radially.  This  observation is  compatible with
the hydrodynamic regime prevailing  at  the surface  of  the elec-
trode.  The rate of film growth  is  proportional  to the  change of
oxygen diffusion current with time.  Curves  2 through 5 (Figure 16)


                                48

-------
                                               o
                                                •

                                               ro
 fi
 O
•H
4->
•H
 W
 O
                                                            O
                                                            U
                                                             (0
                                                            U
                                                   en
                                                   M
                                                   3
                                                   O
                                                   g
                                                  •H
                                                             U)
                                                             0)
 o

 (IJ
 g
•rH
  I

 -p

 cu
 S-l


 u



 tn

 X
 O
                                                             OJ
49

-------
exhibit an initial transient change in current followed by a
steady state or linear decrease, and finally the current reaches
a minimum limiting value.  At minimum current values, the elec-
trode surface is assumed to be completely covered by CaC03 film.
Acidification or addition of a sequestering agent to the test
solution solubilizes the CaC03 film, returning the current to
its original value.  The equation for diffusion current in the
presence of CaCOa film is:

          i = nFA(Pm/6F)CB                                    (74)


where :
          P^ = coefficient of CaC03 film permeability to
               molecular oxygen
          6-r, = thickness of CaCOs film
           r

It is interesting to note that after the CaCOs film is formed,
both P  and 6  are independent of system hydrodynamics.  Permea-
bility of the film to molecular oxygen is the limiting step in
transfer of oxygen to the electrode surface, and P  depends only
on CaCO3 crystallinity .                            m


Experimental

Co-iumn
In 1939, Enslow proposed a "continuous stability indicator"
( 23, 24) .  A stream of water was slowly but continuously passed
through a column of powdered chalk.  The pH and titratable
alkalinity were measured before and after contact in the column.
Changes in these parameters were related to the corrosivity or
scale forming capacity of the test water.  A contact time of
two hours was considered sufficient for the water to be in com-
plete equilibrium with calcium carbonate.

Early in the NSF project, an apparatus similar to Enslow "s column
was constructed from a 24-inch length of three-inch ID plastic
pipe.    (See Figure 17.)  Test water was recirculated through the
column using a peristaltic pump.  The column was packed either
with marble (calcium carbonate) chips or powdered calcium car-
bonate.  Ion-selective electrodes were used to measure influent
and effluent hardness.  Changes in pH , even with low flows and
high recirculation rates, were very small, never approaching
the hypothetical pH  .  Changes in hardness were negligible.  It
soon became apparenf^that very long contact times were required
for the water to come into equilibrium with calcium carbonate.

A second study was designed to demonstrate the rate of attaining
calcium carbonate equilibrium in various waters in contact with
different forms of calcium carbonate.  In a batch system, waters


                               50

-------
                       I
 CD
 G
•H
 S

 CD
-P
 CD
Q
     W
     CD
 dJ
 G
 (U
 CD
 CD
CQ
 CU
 O
 G
 CD
 H
 CD
•rH   -P
Q   G
     CD
          X
          CD
         T3
          G
         H
         •H

         •H
         -P
         CO
 CU
 H
 O
-P
•rH
 G
     UH
 U)
 a
•rH
.£
U

 CD
•H
£3
O
O
rd
                                                          \
                                                         •H
                                                         MH
                                                            O

                                                            CD
                                                            M
 3  O
 CO 4-)

 CD  rH
 >  CD
 CD -P
•H  fO

 U
 (0  rH  g
    O  3
 o m -H
-P     rH
    CD ,Q
 G  g -H
 O -rH rH
•rH -p -rH
-P     3
 (0 -P  D1
rH  U  CD
 3  fd
 CJ 4J  G
 rH  G  O
•H  O -H
 O  U -P
 CD     (0
 in -P  SH
    G  3
 M  CD -P
 O -H  fO
fe  U  CO
                                                d)
                                                                         -P
                                                                          G
                                                                          CD
                                                                          e
                                                                          CD
                                                                          n
                                                                          3
                                                                          03
                                                                          m
                                                                          0)
                                                                          g
                                                                         -H
                                                                          SH
                                                                         X)
                                                                         •H
                                                                         r-H
                                                                         •H
                                                                          3
                                                                          G1
                                                                          CD
 o
-H
-p
 m
 n
 3
-p
 <0
 to
 !H
 O
 VH
 O
-p
•H
 G
 O

 X
 0)
                                                                         -P
                                                                         -H
                                                                         -P
                                                                         cn
                                                                          cu
                                                                          tn
                                                                         •H
                                       51

-------
of defined chemical characteristics were mixed at constant rates
with marble chips or pure powdered calcium carbonate.  Tempera-
ture of the test samples was controlled in a water bath at 25
+0.2°C.  Hardness and pH were measured with respect to time using
a glass pH electrode and an Orion divalent cation selective
electrode.  Attainment of equilibrium was shown to be asymptotic,
and pH changes did not closely coincide with changes in [Ca2*] .
The time required to reach reasonably steady values of pH and
[Ca2 ]  was in the order of 30 to 60 hours.
Catc-ium Carbonate. VzpoA
-------
                                    Disc Contact
                 Motor  Pulley
                                             Electrode
                                             Pulley
                    Motor
                                 Ring
                                 Contact
                                            Electrode
                             Base
           Figure  18.   PIR rotator -  electrode system.
        CALCIUM  CARBONATE
        DEPOSITION TEST
         STRIP-

         CHART

         RECORDER
POTENTIOSTAT /

CURRENT MEASURING

BRIDGE
TO COMPUTER
                                                                    DRAIN
      Figure 19.  CCDT  system in the mobile  laboratory.
                                   53

-------
is controlled by mixing C02 with air - from which C02 has been
scrubbed by bubbling through a solution of concentrated NaOH -
in a Matheson 665 gas proportioner,  and monitoring pH with a
glass electrode system inserted in the sample cell.

When the test is initiated, 100 ml of the test water is immed-
iately withdrawn from the sample cell for analysis of hardness,
calcium, alkalinity, and pH.

Rotational speed of the electrode is 1600 rpm.  This speed was
selected as a compromise between the needs for maintaining con-
stant hydrodynamic conditions at the surface of the electrode
and mixing in the bulk of the solution, and the turbulence asso-
ciated with high rotational velocities.  At 1600 rpm, there is
little turbulence in the test solution.  Speeds greater than
1600 rpm provide no additional increase in diffusion current
attributable to changes in mass transport of oxygen to the
electrode surface.

The potential at which the cathode is controlled, -0.9 V versus
S.C.E., is the potential at which the smallest change in current
is observed over the largest region of change in potential.  A
recording of current as a function of potential sweep, used to
determine the operating potential, is shown in Figure 20.  The
test is continued for a period sufficient to establish the
linear portion of the current decay curve.  CCDT is reported as
the slope of the linear portion of the curve, in microampc per
minute.

During field operation of the mobile laboratory, CCDT was mea-
sured in a covered continuous flow cell in which the test water
entered at the bottom and overflowed through a port at the top.
pH control is not required in the continuous flow system.


Results and Discussion

Plots of diffusion current versus time for a series of formulated
water samples are shown in Figure 21.  Chemical characteristics
of each of these samples are shown in Table 4.  Curves 1 and 5
in Figure 21 were obtained from waters with Ryzner Stability
Indices  (S)=6.92 and 6.95 and CCDT values of 10.08 and 11.32
yamps/min, respectively.  These waters are expected to be scale
forming.  Curve 2 is from a water with S=8.76 and CCDT=0.8 yamps/
min, a water expected to be corrosive to CaCOs.  Curves 6, 7, and
8 are from waters with moderate scale forming characteristics, as
shown by their significant values of CCDT.

CCDT versus S for each of the formulated water samples is plotted
in Figure 22.  The utility of CCDT as a rapid technique for mea-
suring scale forming tendency is clearly demonstrated by the
high correlation of these data.


                               54

-------
800
700
Reduction of oxygen at
rotating gold electrode
coated with a thin CaCO.
film.
                               Same as A,  but in
                               purged water
                                         .'- \  —: -  t—,
                                            -   -
                                            _ i T - r
                                         -J-l "T'T
              L  	t „!	
                  1 _J _L i _4_ ;_
                            "
                         -h-M—I
                 - T-j-- r- -,—-(•--  -—,—-
                •  :iT4: .rriripirt::
                        :rrrz:±r±r
     .::•--::_.
            ^1.0      -0.8      -0.6      -0.4      -0.2
                    Potential (volts vs. S.C.E.)

          Figure 20.  Current versus potential  sweep.
                                0.0
                                55

-------
     600
 w
 
-------
C!
-H
E

en
a,
u
-H
s
QJ
Qi
O
rH
CO


EH
Q
U
U
   100
    10
   0.1
              Increasing scale
              formation tendency
       6.5
7.0       7.5         8.0        8.5

       Ryzner Stability Index
9.0
  Figure  22.   Correlation of CCDT slope with Ryzner  stability
               index.
                              57

-------
CO
z
g
i-
01
CO
111
CO
o
cc
>•
QC
<
D
CO
o>
3
C
IS ol
(/> 
.ti co C":
co c o
0^-^,0
|-§"
E _
D —
•Jj O)
S ra CO
^^8
i^<^
c
O *~t
•^ CM
 o a>


O £ o>
0) LL



CO
q
o
CM
OJ
CD
in
CO
5


CM
OJ



CM
CM
CO




o
OJ
in

in
00
co

CM
CM




^






CO
d
CO
CO
s
9


CM
in
CO



CO
r~-




in

o
CM
OJ

8
CM




CM






S S S 5
r- CN ^ ^
in o o oo
O) Is"1- CN t—
CD r-- oo co
§ £ § !2
? ? 9 9


OJ CM OJ O
oj CM ^j- ^r
r^ co co co



«- CO CO CM
CO O O OJ
oo co co r^




q •* «ct CM
in ^-' d in
CO CO <* CO

in o CM CM
OJ r-~ in oj
CD P~ CO •*

Tt CM CO CO
1^ OJ CO CM




in co r~ co



                             58

-------
The effect of pH and levels of calcium and bicarbonate ions on
CCDT was evaluated with solutions of Ca(HCO3)2 at pH=6.5, 7.5,
and 8.3.  In each of these solutions, bicarbonate concentration
was four times greater than the levels of calcium.  The solutions
were prepared by mixing CaClz and NaHCOa, and bubbling with a
mixture of oxygen, nitrogen, and carbon dioxide.  Bubbling was
continued throughout the test to maintain a constant pH.  Re-
sults of these analyses, plotted in Figures 21 and 22, demon-
strate that relationships do exist between CCDT and pH, Ca2 ,
and HC03~ levels in a test water.

Tap water samples, brought to the laboratory from a number of
Michigan cities as well as those included in the mobile labora-
tory field assignments, showed a wide variety of scaling char-
acteristics, measured by CCDT.  These data with pertinent
chemical characteristics of each test water are shown in Table 5.
CCDT values for each of the 18 samples in Table 5 are plotted
versus S as semilogarithmic coordinates in Figure 23.  The num-
bers circled in Figure 23 coincide with sample numbers assigned
in Table 5.  Dotted lines in Figure 23 separate the grid into
regions representing varying degrees of scale formation.  With
CCDT values >3, waters are known to form films which reduce
carrying capacities of transmission lines.  From CCDT >0.75 but
<3.0, films may be deposited which will, in time, affect carrying
capacities; however, small amounts of inhibitors added to the
water will significantly reduce the problem.  Waters with CCDT
values between 0.2 and 0.75 can be expected to produce thin hard
films of CaCOa which will effectively protect against corrosion.

The most apparent departure between observed and predicted
scaling using the Ryzner Index occurred at pH >9; i.e., lime
softened waters designated 4, 6, 8, and 9, and a Zeolite softened
water, sample 1.  The Zeolite treated water has high pH and high
alkalinity; thus, S was low although calcium hardness was re-
duced to 15 mg/1.  This water was not scale forming and no CCDT
slope was observed, but S ~ 6.6 predicted excessive scale forma-
tion.  The occurrence of S values overpredicting scale formation;
i.e., S indicates excessive scaling when none is observed, is
not uncommon for finished waters with high pH.


Field Experience

During field assignment of the mobile laboratory in the metro-
politan Chicago area, unique opportunity was provided to evalu-
ate the responsiveness of CCDT to water quality changes in the
distribution system.  Caustic is added routinely at the Central
Water Filtration Plant  (CWFP) for corrosion control.  For the
24-hour period beginning at 1200 hours on May 30, 1973, caustic
feeders were taken out of service.  The mobile laboratory was
operating at a four year old fire station, 10.5 miles from the
CWFP.  Two continuous flow CCDT measurements were made each day.


                                59

-------
CO
UJ

O
D

CO

Q
_l
UJ
O

CO
CO
UJ
oc
in

.2

 to
PO4
Addition
C/J
cr

to
t-
Q
O
0
v>
T
a
I
a
>
Total
Alkalini
E
'«
CD
c
0)
CO
£
D
•o
c3
(A
-5 £
o •§
i ^-
1— CO
I

z
00
LO
co
*—
o
T—
+
CD
c
o
z
O)
LO
Is-
§
00

CN
CM
CO

O

m
CO
00
CO

I
"5
o
CO
•
00
o
00
8
?
CD
0
z
5
00
•*
r-.
CD

CO
rj-

CO

o
CO
o
00
CN

I
I
<
1
z
LO
CO
CO
o
9
CO
LO
LO
LO
CO
LO
•3"
00

§

r~

CO
o
LO
O

I
+j"
'o
S
D
>
t~~
LO
r~.
c^
T
CD
O
Z
5)
00
LO
CM
O

CN
CO

LO

CO
CO
o
CN
O

Ypsilanti, Ml
z
§
CO
s
9
CN
CM
^
1^-
CO
8
CO

CO
en

CN

T
CO
o
CO
CO

i
w
~o
'OL
CO
IT
TJ
§
a
z
r—
en
r~
CD
o
T
r-
00
o>
CO
LO
o
o

CO
CO

LO

CO
CN
o
LO
CN

i
c
CO
1
r--
r—
en
r~-
LO
r-
?
CO
C
o
CN
r~-
CO
r-
•
^c
'vt
C
2
z
CN
CM
r^
'S-
^-
?
o
LO
r~
CO
CO
r-.
o
LO
r~~

CN
CO

CO
CO

en
en
o
CO
en

I
a.
1-
1
<
1
Z
00
en
r~.
3
9
g
^
>*
co
Is-
o
CO
r-

CN
CO

§

o
o
8

Petersburg, Ml
z
§
00
8
9
s
CD
CO
00
en
0
00

CO
o

o

CO
CO
o
8

*
5T
u_
o
_i
o"
I
6
z
en
00
00
co
<*
9
CM
CO
T—
•«t
CO
CO
O)
r-.

o

00

•<*
CO
o
Oi

*
sr
LL
1
_l
o"
O)
8
6
z
r-.
00
CN
O
?
CN
CN
CO
r^
CO
LO
r-~
00

en
•^•

r»-

CO
CO
o
CN

*
<
Q_
(0*
*
0}
-o
CD
£
z
CO
^•
r-
00
o
9
LO
•*
00
CO
r~
8
r-

CN
CN
CN

00
r~

CO
CO
0
5

00
o'
Z
C/5
>
Z
LO
LO
t~-
CN
O
?
8
r-
LD
r~
en
LO
r~.

CO
en

CO

CO
CN
o
en

a
E
8
co
£
£
E
to ^ •*
C CO §
S "a
-^ <+-

CO
CO
CO
CO
CN
?
s
CN
r*.
00
co
rv

co
8

R

5
o
o
LO

i
CD"
_c
8
                                                                60

-------
  10.0
-p
w
C
O
w
O
a
QJ
P

0)
-P
O
•H
U
U
1.0
   0.1
         Excessive
   u  Moderate
          Slight
         No  observable CCDT slope

            O

        _J	
                                               I
         6.5
               7.0       7.5        8.0        8.5

                     Ryzner Stability  Index
9.0
        Figure  23.   CCDT versus Ryzner stability index,
                               61

-------
Arrival of the distributed caustic-free water at the mobile lab
was marked by changes in observed levels of pH and CCDT.  Mean
CCDT for the previous day (May 29) was 0.52.  The lowest daily
mean CCDT, 0.26, occurred on May 31.  On June 3, a mean CCDT of
0.47 indicated that water with caustic had returned to the area.
At CCDT =: 0.50, light hard protective film formation is assured.
The low mean CCDT (0.26) indicated that more corrosive water
was present in the lines.  These data are plotted in Figure 24.

In extensive laboratory and field applications of CCDT, this
test was shown repeatedly to be capable of accurately predicting,
without respect to the metallic corrosion reaction, whether or
not films of CaCOa would form on the surface of distribution
system pipelines.


ALKALINITY

General

An automated potentiometric technique for measuring total alka-
linity in tap water was developed in this study.  The method
was fully evaluated in the laboratory and used successfully in
the mobile laboratory during field assignments.  As a result
of apparent faulty operation on Cincinnati tap water which has
high levels of residual chlorine, detailed theoretical relation-
ships were developed and the procedure slightly modified.


Theory

A linear relationship exists between the resultant pH of a
suitably selected phthalate buffer solution  (KHP) and the total
carbonate formality of a test water.  Theoretical aspects of
this relationship are described using an example test water.
Assume the alkalinity of the test water is in the range 30 to
60 mg/1 as CaC03  (3(10)'" to 6(10)~" F C032  ).  It has been
shown experimentally that five parts of the  test water and one
part stock phthalate buffer produce the linear relationship be-
tween resultant pH and sample alkalinity.  Initial formalities
of the buffer are:

          KHP  =  0.050
          HC1  =  0.019

and of the test water:

          Na2C03  = 6/5 C

and final formalities for the mixture  (buffer:sample =  1.5) are:

          KHP  =  8.33(10)~3


                               62

-------
                                                                                                                  13
ro

  •

CO
CN

  •


CO
                    00
o
  •

00
                            r
                           X
                  X
                  I
                  X
                    \
                               X —
          r
          X

           X
             v
                             "" X

                                \
                              — X
                                                                    EH
                                                                    Q
                                                                    U
                                                                    U
                                                                       CN
                                                                    cu
                                                                    a
                                                           fd
                                                          a
                                                          I
                                                             — x
                                                                   \
(Ti


O
00


O
         m
           •

         o
01
  •

o
CN

  •


O
                                                                         -P  _
                                                                         -P

                                                                         3
                                                                         O
                                               fd


                                               M S
                                               0) O =
                                               -P SH  co
                                               •H     o
                                                  a  3
                                                  a)  o
                                                  g  o
                                               f
                                                I
                                                                           oo
                                                                           o


                                                                           0)
                                                                           fi
                                                                                              (N

                                                                                              O


                                                                                               (D

                                                                                               fi
                                                                                                                  O
                                                                                                                  4J
                                                                                                                  •H
                                                                                                                  a
                                                                                                                  o
                                                                                                                  g
                                                        Ou
                                                        EH
                                                        P
                                                        O
                                                        O


                                                        ti
                                                        O


                                                        eu
                                                        o

                                                        o
                                                        en
                                                        m
                                                        o
                                                        •H
                                                        ^1
                                                        u
QJ

1?



•H
ro

K*1
fd
jg




o
oo

^
fd
S



CTl
CN

£•1
fd
-P
rd
tn
0) -H •
S rH 45
•H U tn
EH >i-H
O 45
CD
M ^
0) \
t5 O
0) Cn
(U fd
4-1 O
-H
0 45
•H O
-P
CD -
J3 rH
nj
O •
O
4H 3
O
0)
-P -P
0 -H
(U CO
                                                                                                                 M-l  -P
                                       U U Q EH
                                                                                                                 CN
                                                                                                                 •H
                                                                                                                 Cn
                                                     63

-------
             HC1  =   3.17(10)~3
          Na2C03  =   C

[H ]  can be calculated  as  a function of C, applying the  following
series of known relationships:

          Kl =  [H ] [HP  ] = 1.3(10)~3                          (75)
                  [H2P]
           2 —
   =  [H ] [P - L          ~6
                      2
          K2 =        -  =  3.9(10)~6                          (76)
                  [HP  ]
          K3 =  [H+] [HC°3"]  = 4.6(10)-                        (77)
                 [H2C03]


          K, =  [H  ] [C°32  ]  = 4.4(10)~11                       (78)
                  [HC03~]

          KW =  [H+] [OH~]  =  10~14                              (79)

           [Cl~] =  3.17 (10)~3                                  (80)

           [K+]  = 8.33(10)~3                                   (81)

           [Na+] =  2  C                                          (82)

           [H2C03]  +  [HC03~] + [C032~] = C                     (83)

           [H2P] +  [HP~]  +  [P2~]  = 8.33(10)~3                  (84)

                                    ~]+ 2[P2~] + [HC03"]+  2[C032~]  (85)
By assuming  [CO32~]  =  0,  [P2 ]  = 0, and  [OH ] =  0,  derivations can
be stated as follows:

from Equation  77:
                       +
           [H2C03]  =  -tS-L [HC03~]                              (86)
                     K3

from Equation  75:
                     +
           [H2P]  =  JJLJ. [Hp~]                                  (87)
from Equations  83  and 86:
[HC03~]
                                                               (88)
                     K3

                                 64

-------
and
           [HC0
                       4.6(10)  3  C
                     [IT]+  4.6(10)  3
from Equations 84 and  87:
                                                             (89)
           [HP  ]
                    +  1
                  K
                            =  8.33(10)
(90)
and
from
           [HP  ] =
                     1.083(10)  5
                                                             (91)
   Equations 79, 81,  82,  85,  89,  and 91:

         [H+] + 2 C +  8.33(10)~3=  3.17(10)~3
                                                   1.083(10)"5
                                                  [H+]+ 1.3(10)  3
                                      4.6(10)"3C
                                    [H+]+ 4.6(10)~3
                                                             (92)
then:
          L.C.D. =
                     [H+] 2  + 5.9(10)~3
                                           + 4.6(10)"3

                                           5.98(10)~6
                                                               (93)
where:
        L.C.D. E lowest  common denominator
    + 2C[H+]2 + 5.16(10)~3[H+]2 + 5 . 9 (10) ~ 3 [H+] 2 + 1. 18 (10) ~2C [
    + 3.044(10)~5 [H+] +  5.98(10)~6 [H+]  + 1.196(10)~5C
    + 3.086(10)~8- 1.083(10)~5 [H+]  -  4.982(10)~8 - 4.6(10)~3 C [
                                                             (94)
      - 5.98(10)~6C =  0
+] 3 + J1.106(10)~2+  2C\  [H+]2  + J2.
    - 1.896(10)~8 +  5.98(10)~6C = 0
                                      559(10)"5 + 7.2(10)~3C
                       5.98(10)~6C = 0

neglect the  [H ]3 term and  solve with quadratic equation:
                                                               (95)
                                  65

-------
       2.559(10)~5-7.2(10)  3C + / J2 . 559 (10) ~5+7 . 2 (10) ~3 c}2

                                   -4  J1.106(10)~2+ 2C(

                                   j-1.896(10)~8 + 5.98 (10)~6C|
                    2.212(10)
                                 -2
 4C
                                                               (96)
        -2.559(10)"5 -7.2 (10)"3 C  +   / 6 . 548 (10) ~ a °+5 .184 (10)~5  C2

                                     +3.685(10)"7 C +8.388 (10)"10

                                     -2.646UO)"7 C +1.517UO)"7 C
                                     -4.784 (10)  5 C
                        2.212 (10) 2  + 4C
                                                            (97)
        -2.559 (10)"5 - 7.2 (10)~3 C  +  /4 (10)"6 C 2 +2.556 (10) 7 C

                                      +1.489 (10)"9

                        2.212(10)" 2+ 4C
                                                           •(98)
The relationship between hydrogen  ion concentration and carbonate
formality given in Equation  98  is  shown to be linear by the data
in Table 6 and Figure 25.  This linearity extends to the rela-
tionship between resultant pH and  carbonate formality, as shown
in Figure 26.  In addition,  results  of an actual experiment are
compared with theoretical predictions in Figure 26.  These data
are shown in Table 7.

Thus, when the initial pH and total  carbonate formality of a test
water are known, concentrations of the species of interest can
be calculated as follows:
          H] [HC03-] = 4.6(10)-3.

          [H2C03]
•[H2C03J =
                                             K-
                                                 [HC03~]
K, =
              [C032"]
            [HC03~]
                                                      [HC03~]  (100)
                                66

-------
                                                                                         o
                                                                                         00
                                                                                         o
                                                                                         o
                                                                                         m
in
                   in
                                      o
                                        •
                                      m
                                               CO
                                                                            
•H
-P
•H
H
 (0
 e
 M
 O
O
U

en
                                                                                                     ffi
                                                                                                     m
                                                                                                     CM


                                                                                                      (U
                                        67

-------
                                 O
                                 00
                                             -P
                                             •H
                                              a
                                             •H
                                             iH
                                              05
                      O
                      
                                              cn
                                              C
                                              O
                                             •rH
                                             -P
                                              ai
  
-------
          [H2C03] +  [HC03~] +  [C032~] = C

Substitution of Equations 99 and 100 into  101  gives:

                                 K2  )
                                                      (101)
  [HC03~]  =
                            C +
              Kl
                                                      (102)
                       [HC03~]  =
                                        Ki [H+]C
                                          [H+]2
                                                  KiK2
Table 6.   [H+] VERSUS ALKALINITY, APPLYING EQUATION 98
Alkalinity
(mg/1 as CaCO3)

24
36
48
60
72
84
[H+]
(molarity)

5.33(10)'^
5.07 (lO)'1*
4.82(10)"^
4.57(10)"^
4.34(10)"'*
4.11(10)~'t
    *(C03  formality) x 10  = Alkalinity in
                              mg/1  as  CaCO3


    Linear regression:

          [H+] = {57.9133 - 0.237515*ALK)(10)~5
                        69

-------
   Table  7.  EXPERIMENTAL AND THEORETICAL RELATIONSHIPS

                  BETWEEN ALKALINITY AND  pH
Alkalinity
(mg/1 as CaC03)

• 24
36
48
60
72
84
pH
Experimental

3.303
3.347
3.391
3.434
3.476
3.519
Theoretical

3.273
3.295
3.317
3.340
3.362
3.386
Experimental

The sensing element in the alkalinity monitor is a combination
glass pH/reference electrode with a ground glass collar.  The
electrode fits directly into a glass flow cell.  Instrumentation
includes a Corning Model 101 digital pH/mv meter; Honeywell
Electronik 194 multi-speed, multi-range servo recorder for
signal conditioning; and an analog output to the computer.  A
Technicon II peristaltic pump is used for reagent/sample propor-
tioning.  A Valcor series SV-72 three-way miniature dri solenoid
valve switches between baseline and sample solutions.  AC line
voltage is switched to the valve by computer controlled relay.
A schematic illustration of the system is shown in Figure 27.

The Corning Model 101 is especially suited for use in the alka-
linity system because of its exceptionally quiet operation and
its provision for a solution ground.  The alkalinity measurement
is read from approximately 20 millivolts variation in a signal
with a total magnitude of only 200 millivolts; thus, the need
for quiet operation is apparent.  The peristaltic action of the
pump controlling flow through the system produces an electrical
signal.  This peristaltic noise is reduced by a solution ground
placed in the sample line.  Many ground configurations were
studied.  The most successful provides for contact with the
solution at two points - in the influent sample line, just before
the pump; and after the debubbler, just before the sensor cell.
These two grounds are connected together and terminate  at the
solution ground terminal in the pH meter.
                               70

-------
u. Z
tu -_
z 2
, fc-
ui 3
vi -"
< O
a -
u. 3
3 O
a in

















tu
i-
(n
<
s
o
*-





































0
z
ui x
£ u
s ±
< i
VI VI



















O
Z
•V ^
™ X
2! i
a
3

















N



-i „

8?



DC
O
t/i
Z
UI
it




i — — ~~ — 	 1 ^
M
1 °
^ ,„ ,,.] 	 _ 1 1
" i m
r I >• M
2*2 s 5
^ ^ *•• P*X

i — 1
Q)
„ , 	 i
QC r^
< 5
P>
o g
Z *=i
•» r^
? -H
h-
° a. 0
0 * -P
« 3 -rH
a. o. f->
S 0
g
4-1
•H
.— t
•H
i — 1
fO

rH

O
C!
O
•rH
-P
^ TO
UJ <_,
u -P
^-i
                              z
                           o  o
                           h-  U
x
u
r   o
«   U
                                                       O
                                                       •H
                                                       -P
                                                       (S
                                                       e
                                                       0)
                                                       X!
                                                       O
                                                       CN

                                                       0)
                                                       Cn
                                                      -rH
                                                       En
       71

-------
Because of the acid nature of solutions used in the system, it
is not uncommon for the ground wire to become coated with cor-
rosion products and lose contact with the solution.  Whenever
signal quality degenerates, the grounds are simply replaced.

The system operates on a 26-minute cycle controlled by the com-
puter.  At initiation of the cycle, the relay is switched to
the "on" position and the three-way valve is connected to the
baseline solution.  For the next 13 minutes, baseline solution
is drawn into the system.  Air bubbles are introduced beyond the
pump to separate the sample stream into discrete samples.  The
buffer is added and a three-inch glass mixing coil provides
mixing and reaction time before the sample reaches the electrode
system.  The sample stream is debubbled just before it enters
the sensor flow cell.  The bubble stream goes to waste under
siphon while the remainder of the sample is drawn past the
sensors, then to waste.  At the end of the 13-minute period, the
computer sets the relay to "off" and the three-way valve is open
to the test water stream.  A complete cycle of 26 minutes is
required to read both baseline and test water samples.

Preparation of reagents used with the alkalinity monitor is in-
cluded in Appendix B.


Results and Discussion

A good quality recording of "continuous" alkalinity analyses is
shown in Figure 28.  The analytical region of the curve is from
the tip of "drop" to the center of the equilibrated signal from
the sample.  Greatest precision is achieved by measuring peak
height immediately after initial leveling off.  The computer be-
gins its measurement after switching from baseline to sample and
continues to seek equilibrated signal for a period of six minutes,
In Figure 28, the height of the measured peak is 27.0 chart units
(C.U.).  Measurement of peak height to the nearest 0.5 C.U. and
reading alkalinity from a calibration curve produces data with
an accuracy of +2 mg/1, essentially the same as that which is
stated in Standard Methods  (3) for potentiometric titration of
alkalinity.

A calibration series of five or six standards is run for each
combination of newly prepared baseline or buffer solution.  Peak
heights in C.U. are plotted on the abscissa and concentration
on the ordinate.  A typical calibration curve is shown in Figure
29.  To ensure accuracy of the data, the standards are checked
by potentiometric titration and the continuous output compared
with a potentiometric titration of a tap water sample.
                                72

-------
                                                                                          -H
                                                                                           in
                                                                                           (0
                                                                                          -p
                                                                                          •H
ooeo
 (0

 tn
 3
 O
 3
 C
•H
•4-1
 fi
 O
 O

IH
 O

 tn
 0
•H
T)
 H
 O
 O
 0)
                                                                                           4J
                                                                                           •H
                                                                                           CP
                                                                                           O
                                                                                           O
                                                                                           O
                                                                                          oo
                                                                                          CN

                                                                                           0)
                                                                                           tn
                                                                                          •H
                                             73

-------
   24
   20
   16
w
4J
-H
-P
M
^u


£  12
Cn
a;
(13
(1)
Oi

          10         20        30         40        50         60

              Alkalinity (mg/1 equivalent  calcium carbonate)

           Figure  29.  Alkalinity calibration  curve.
                                  74

-------
Special Case Study

Following delivery of the mobile laboratory to the EPA/NERC
facility in Cincinnati, hookup was made to monitor Cincinnati
tap water.  For some time, the alkalinity system did not produce
data which was equivalent to that reported by NSF during most
periods of field assignment; i.e., the basic relationship be-
tween final pH and alkalinity could not be reproduced.  Titra-
tions with quiescent solutions allowed to equilibrate, stirred
solutions at equilibrium, and stirred solutions read at constant
time intervals produced apparently randomly scattered data points.
High levels of free residual chlorine were thought to be contri-
buting to this anomalous behavior.  The phenomenon had been ob-
served on two previous occasions - once when the laboratory was
located immediately adjacent to the water treatment plant in
Ann Arbor, Michigan, and once in the Calumet Harbor area of
metropolitan Chicago.

To quantify the role of free residual chlorine in potentiometric
measurement of alkalinity, an experiment was proposed and re-
sults of the experiment calculated theoretically.  The protocol
and related calculations are presented in detail in Appendix C.
Experimental data closely approximated the theoretical results
and established the anomaly could not be attributed to the pre-
sence of chlorine alone.  The apparent interference was elimin-
ated by adding thiosulfate  (five mg/1) to the buffer solution,
suggesting that chlorine may have contributed to the problem,
perhaps synergistically with an unidentified organic present in
the test water.  The problem has not recurred since EPA substi-
tuted a combination electrode with a calomel reference for the
silver chloride reference-combination electrode in the original
equipment.


FREE AND TOTAL FLUORIDES

General

Fluorides are added to treated water supplies as sodium fluosili-
cate  (NazSiFe), commonly referred to as sodium silica fluoride;
fluosilicic acid  (H2SiF6); or sodium fluoride  (NaF), each of which
dissociates to release free fluoride ions  (F~).  Prior to this
study, it was generally assumed that fluorides reached the con-
sumer in this form; i.e., as free ions, and that those which were
added during treatment were not different from those which may be
present in natural waters.  This assumption is invalid.  In most
chemically treated finished waters, fluorides are complexed with
polyvalent cations, such as aluminum or iron, which are commonly
added as coagulants or may be present as natural constituents of
the raw water.

In a survey of public water supplies in the 100 Largest Cities
in the United States  (25), residual aluminum and iron in distri-

                                75

-------
bution systems were reported to be as high as 1.5 and 1.7 mg/1,
respectively.

Fluoride is an important parameter of drinking water quality.
In high concentrations it may have a toxic effect; at low levels
it is beneficial; i.e., an effective preventer of dental caries
(26,27,28,29).  Chemical reactions between F~ and tooth enamel
are postulated as follows:
or
Cai o (POO e (OH) 2 + 2F -> Cai0(P04)6F2
hydroxylapatite        fluorapatite
                                                20H
Caio(POn)e(OH)2 + 20F
hydroxylapatite
8H+ +
                                 10CaF2 +
                                 calcium
                                 fluoride
                                          2H20
                                                   (103)
                                                    (104)
Reaction 103 is generally favored at low levels of fluoride; i.e.,
in fluoridated waters; reaction 104 occurs in the presence of
high concentrations of F~; e.g., dentifrices  (30).

The recommended level for fluoride in drinking water is less
than 1.7 mg/1  (1); thus, analytical detection methods must be
accurate at this level.  In the NSF project, the potentiometric
ion-selective electrode technique was studied in detail, and
automated for measuring both free and total fluorides in the
mobile laboratory.
Theory

The potentiometric membrane electrode selective for F~ can be
described schematically as follows:
  Internal
  Reference
  Electrode
  Ag/AgCl
   Internal
   Reference [XF]
   Solution
   AgCl(s)
                    Membrane
                      LaF'
                                      External
                                      Solution  [YF]
                                      Test
                                      Solution
                             External
                             Reference
                             Electrode

                             Saturated
                             Calomel
                             Electrode
                               (S.C.E.)
where:
          vertical lines represent boundaries  or  liquid
          junctions between components  of  the  cell.
The basis for ion-selective electrode measurements  is  the Nernst
equation:
                                        aj
                                                              (105)
E  = E  _ M in
 m    °   ZF
                  ap- +  I *a±
                         i
                                76

-------
where:
          E  = measured potential
          E  = a constant equal to the sum of the potentials
           °   at the Ag/AgCl electrode, the saturated calo-
               mel electrode, the liquid junction potential
               between the test solution and the reference
               electrode, and the potential across the mem-
               brane when F~ activity in the test solution
               is = 1.0
          R  = gas content
          T  = absolute temperature
          z  = number of electrons involved in the electrode
               reaction
         ap- = activity of F~
          a. = activity of ith interfering ion
          K1 = selectivity coefficient of interfering ion
               with respect to F~.

Activity (a) is related to concentration (c) as shown in the
following empirical expression:

          a = ac                                              (106)

where:

          a = activity coefficient

At 25°C, the slope of the line described by Equation 105 is
approximately 60 mv; however, in practice measurements are
usually read from calibration curves.  In calibrating ion-selec-
tive electrodes, changes in total ionic strength of the test
solution must be considered.  With the fluoride electrode, OH"
is the only species which interferes with F~ measurements.  The
effect is apparent when the level of OH~ is greater than one-
tenth the level of F~ present in the test water (31,32)._ Other_
anions commonly found in drinking water; i.e., Cl , SOi,2 , HCOs ,
N03 , and P0i*3~ do not interfere even when they are present at
levels 1,000 times the level of F~ (31,33,34,35,36).

It is important to differentiate between OH  interference and+
the complexing effect of certain cations; e.g., H , Al3 , Fe3
(31,34) .  Fluoride combined with these cations cannot be sensed
by the electrode, but this effect is different from the inter-
ference of OH~.
Experimental

Equipment used in studying and optimizing the fluoride ion-selec-
tive electrode measurement technique included an Orion digital
pH/mv meter, Model 801; Orion solid state fluoride ion-selective
electrodes, Model 94-09; saturated calomel reference electrodes;


                               77

-------
and polyethylene containers to minimize loss of F  by adsorption
onto glassware.  A Corning digital electrometer, Model 101, was
substituted for the Orion 801 early in the study.  The Corning
meter proved to be a much more reliable instrument, and with its
activity mode, appreciably simpler to use.  Calibration was set
so the upper limit read "100" for 5 x 10~5M F~ and the lower
limit, "20" for 10"5M F~; thus, percent recovery of F~ in the
test solution could be read directly.

In studying the effect of pH, a stock solution of sodium fluoride
was prepared from which suitable dilutions were made.  pH was
varied by addition of strong acids and bases, and ionic strength
was maintained at 0.5M by addition of appropriate volumes of
concentrated sodium nitrate  (NaN03) solution.  Potentials were
measured after equilibrium was established.  At low fluoride and
high pH levels, more than 15 minutes were required to reach
equilibrium.

Complexemetrie titration techniques were used to characterize
aluminum-fluoride complexes.  pH was controlled by an acetate
buffer.  Solubility of the complexes was suppressed with ethanol.

In measuring total fluorides, a masking agent is added to complex
Al3  more strongly than F~, liberating bound fluoride for sensing
by the electrode.  The relative efficiencies of four masking
agents was studied:  EDTA (disodium salt of ethylenediaminetetra-
acetic acid), CDTA (cyclohexanediaminetetraacetic acid), oxalate,
and citrate.

Equipment in the free and total fluoride monitoring system in-
stalled in the mobile laboratory included the Orion digital
pH/mv meter, Model 801 (for total fluorides) and a Leeds and
Northrup expanded scale pH/mv meter, Model 7415-EO  (for free F~);
two Orion solid state fluoride ion-selective electrodes,
Model 94-09; Ag/AgCl  (total) and calomel  (free) reference elec-
trodes; an Esterline Angus dual pen strip chart recorder; a
Technicon Pump II for reagent/sample proportioning; and Valcor
three-way miniature dri solenoid valves, series SV-72, for
switching between baseline and standard solutions.  AC line
voltage was switched to the valves by computer controlled relays.

A schematic procedure for measuring free and total fluorides with
ion-selective electrodes is  shown in Figure 30.  Preparation of
reagents is described in Appendix B.


Results and Discussion

pH E&6e.ct

The effect of pH on fluoride electrode measurements is shown in
Figure  31.  The OH~ interference effect is apparent when the


                               78

-------
                         SAMPLE
   Free Fluoride, A

 - Add free fluoride buffer
   (FFB)* to sample, 1:1

 - Potentiometric measurement
            Total Fluoride, B

          - Add total fluoride buffer
            (TFB)+ to sample, 1:1

          - Potentiometric measurement
                 Complexed Fluoride, C = B-A


 Preparation of Buffers:

   To about 700 ml double distilled water add:
                     *Free fluoride buffer,
                              FFB
                       Total fluoride buffer,
                                TFB
    Acetic acid

    Sodium nitrate

    Sodium citrate
 5.7 ml (0.1 M)

84.99 g (1.0 M)

     None
11.4 ml (0.2 M)

59.5  g (0.7 M)

58.8  g (0.2 M)
    Adjust the pH to 5.2 with concentrated solution of sodium
    hydroxide, and the volume to 1.0 liter.
Figure 30.  Characterization of fluorides and preparation of buffers
                                 79

-------
    250
                                                11
13
Figure 31.  Effect of pH on fluoride ion-selective electrode
            measurement.
                            80

-------
curves in Figure 31 show a nonlinear relationship and shift to-
ward more negative potentials.  Because F~ and OH~ have similar
changes and ionic radii (31), the electrode senses both (but to
varying degrees); thus, a false increase in apparent F~ activity
may occur at high levels of OH~.

The effect of H-F complexation is also apparent in Figure 31.
Hydrogen ions combine with fluoride ions to form hydrogen fluoride
(HF) and bifluoride ions (HF2~), neither of which are sensed by
the electrode (37,38).

According to the literature  (31,32), the fluoride electrode can
be used for measurements in drinking water at pH five to eight;
however, data from this study indicate a much more restrictive
pH range (Figure 31).  At 10~6M F~, the optimum range is 5.0 to
5.4; thus,  pH 5.2 is used in the NSF fluoride monitor and is
recommended for all ion-selective electrode measurements of
fluorides in water supplies.

The selectivity coefficient  (K) of OH~ with respect to F~ can be
calculated as follows:
                            _        _
          E = E  -    In [(F )  + K(OH )]
               O   r
                                                             (107)
Using the data from Figure 31, estimated values for K are shown
in Table 8.  Note that K increases as the level of F~ decreases;
i.e., the interference effect of OH~ is more predominant at lower
concentrations of F~.   At levels of F~ commonly found in drinking
waters, values of K are on the order of 0.70.  By calculation,
using Equation 107, it can be shown that at 10~SM F~, the OH~ in-
terference effect is significant only at pH > 5.4.

A further illustration of OH~ interference is shown in Figure 32
in which calibration curves at various levels of pH are plotted
as electrode potential versus concentration of F~.  Nonlinear
electrode response is greater as fluoride concentration decreases
and pH increases.  At low pH, electrode response follows the
classical Nernstian relationship.  Linearity was extended to
         at pH values of two and three.  The parallel shift in
10~6M F
          Table 8.  OH  SELECTIVITY COEFFICIENTS  (K)
F~ Concentration

10~6M - lO^M
10~"M - 10~3M
10~3M - 10~2M
K

0.70
0.07
0.05
                               81

-------
  -50
          0.019
   SO
g
"c
  100
0>
1
o
   150
  200
  250
- pH 12
        pH 2
           10
                 0.19
                       1.9
19.0
r6        |0-5        |0-4        !0-3
        Ruoride Concentration, Cj
mg/l-
   190
                                                    IO'2M
 Figure  32.  Calibration curves of different levels of pH,
                              82

-------
the intercept of the calibration curves below pH  five  can be
    .ained in terms of H  complexation  (37):
                     Ki(HF) v    0 Q                          (108)
                           pl\ i — ^ . .7
explained in terms of H  complexat

          (F~) (H+) -»- KI (HF) R  _ 2

          (HF) (F") -*• K2(HF2~) K  = o  77                       (109)

where:

          KI = first dissociation constant
          K2 = second dissociation constant
and:
where:
          CT = 2(HF2~) +  (HF) +  (F-)                          (110)
          C  = total concentration of  fluoride
For solutions of fluoride more dilute  than  0.05M,  (HF2  )  can be
neglected and Equation 107 rewritten as  (39):

          CT =  (HF) +  (F~)                                    (111)


Combining Equations 108 and 111 and solving for  F~:


          (F-) = CT 	^	—                                 (112)


Then, substituting Equation 112 into the  Nernst  equation  and
rearranging:

                   •prn       V"        T3T1
          E = E  - — In 	£•*	— In  C                 (113)
               0   F     K! +  (H+)   F      X

where the term - RT/F In KI/KI +  (H+), the  "shift term,"  is con-
stant for given pH and temperature and independent  of fluoride
concentration.  At pH > 5,  (H  ) is relatively  small,  the  ratio
KI/KI +  (H ) approaches 1.0, and  the shift  term  = 0;  i.e.,  no
shift in intercept occurs.  At pH < 5,  (H )  is large, and the
shift term is positive and increases with decreasing  pH.   Experi-
mentally determined values for shift in intercept at  25°C cor-
relate closely with calculated values, as shown  in  Table  9.
                               83

-------
  Table 9.   SHIFT IN INTERCEPT OF CALIBRATION CURVES AT LOW pH
            VALUES
pH Range

2-3
3-5
Calculated

36 . mv
14. mv
Experimental

40. mv
15. mv
It follows, then, that at constant pH, H  complex formation is
constant at all levels of fluoride and has no effect on the
slope of the calibration curve; i.e., at constant pH, changes
in the slope of a calibration curve developed within sensitivity
limits of the electrode are the result of OH  interference only.
The effect of pH on electrode sensitivity is shown in Figure 33.
Vsiva.le.nt Cat-ion
                2
                                     2
Both calcium (Ca2 )  and magnesium  (Mg2 ) ions form insoluble or
sparingly soluble salts of fluoride.  High levels of these
cations (145 mg/1 Ca2  and 120 mg/1 Mg2  as CaC03), the principle
hardness ions, are commonly found in distributed finished waters
(23).  Even in much higher concentrations than those reported
for treated waters,  neither Ca2  or Mg2  showed any effect on
fluoride electrode measurements.  Recovery of 5_ x 10~5M F~ was
complete in the presence of 10~2M Ca2  and Mg2 ; i.e., >400 mg/1
Ca2+ and >240 mg/1 Mg2+.  At these levels, any CaF2 or MgF2
which is formed is well below their respective solubility pro-
ducts; i.e., K  = 10.4 and 8.15, respectively.
              s
          Cat-ion

Trivalent cations; e.g., A13+ and Fe3  , rapidly complex with
fluoride.  The extent of complexation  is a function of pH and
relative levels of F~ and complexing species in the test solution.

Water which is treated by alum coagulation may contain from 3.3
to 1,500 mg/1 aluminum  (25).  Over the period March 1970 to
February 1971, aluminum levels in Ann  Arbor, Michigan tap water
varied from <20 to 500 mg/1, measured  by emission  spectroscopy .
The Al-F complex was characterized by  potentiometric  (ion-selec-
tive electrode) titration.  Results are shown in Figure  34.
Addition of acetate buffer  (Curve II)  to maintain  pH  5 . 2 signifi-
cantly improved end point detection.   With no pH control  (Curve  I) ,
addition of A13+ titrant lowered the pH of the test solution re-
sulting in further H  + F~  complexation.  Addition of ethanol
 (Curve III) also significantly improved the titration.  Both ethanol
and buffer added to the test solution  (Curve IV) produced the
                                84

-------
   20
 840

 i
 <3
    60
                                7

                                pH
                                                          13
Figure 33.  Effect of pH on fluoride ion electrode sensitivity.
                               85

-------
   -200
   -160  -
                               Titration of 10.0 ml of O.IM NaF,
                               brought up to 70.0 ml total volume
                               by the addition  of:
                               I.   60.0 ml of redistilled water
I.
IE.
                                    10.0 mi of I.OM acetate buffer
                                    + 60O ml distilled water

                                    50.0 ml absolute ethanol
                                    + 10.0 ml  distilled water
                                    10.0 ml acetate buffer
                                    iSO.O mi ethanoi
                             234
                             ml of O.IM AICI3 Titrant
Figure 34.   Complexometric titrations of  fluoride with aluminum.
                                    86

-------
optimum condition for titration.  Stoichiometric calculation of
the complex coordination number using the data from Figure 34
identified the complex as (AlF6)3~.

A number of masking agents have been proposed for potentiometric
measurement of total fluorides  (40,41,42).  Frant and Ross (40)
added citrate to total ionic strength adjustment buffer (TISAB);
Crosby, et al (41)  changed the composition of TISAB and added
EDTA; and Harwood (42) used CDTA.  Four separate masking agents
were studied early in the NSF project:  oxalate, citrate,  EDTA,
and CDTA.  Solutions of each were made from corresponding
equivalent weights.   Increments of aluminum were added to standard
solutions of 5 x 10~5M or 1.0 mg/1 F~.  When complexation was
complete, a masking agent was added and percent recovery of F~
calculated.  Electrode measurements were made in the presence of
l.OM acetate buffer (pH = 5.2) with constant ionic strength
(l.OM NaN03).  Results are shown in Table 10.

EDTA, CDTA, oxalic and citric acids have respective stability
constants with aluminum as follows:  16.1(43), 17.6(43), 14.6  (44),
and 20.0 (45).  EDTA and oxalate did not yield high recovery
rates.  Despite its relatively high stability constant, Al-EDTA
complex formation is too slow for EDTA to be an effective masking
agent in the measurement of fluoride.  Heating or standing over-
night showed no improvement.  The solubility of oxalate in TISAB
buffer solution was limited.  At 6.0 meq/1, CDTA yielded the
highest recovery, but higher levels of CDTA had no further effect
on recovery rates (Figure '35).  At higher concentrations of any
masking agent, citrate yielded highest recoveries.  As shown in
Table 10 and Figure 36, 60 meq/1 citrate provided good recoveries.
At 600 meq/1 citrate  (0.2M), 100 percent recovery at 0.5 mg/1
A13+ occurred after 6.5 minutes; at 0.1 mg/1 Al3^ recovery was
complete after three minutes.   (See Figure 37.)  Averaged addi-
tive effects with no synergism or potentiation were observed
when two or more of the masking agents were mixed.

Iron frequently occurs in drinking water in higher concentrations
than aluminum; i.e., levels from two to 1,700 mg/1 are reported
in the 100 Largest Cities in the United States  (25)•  The effect
of Fe3+ - F~ complex formation was also studied.  Results, shown
in Table 11, indicate that iron does not form stable complexes
with fluoride.  Nearly 100 percent recovery occurred with acetate
buffer alone.  At Fe3"^ >2.0 mg/1, competitive complexing effects
were observed.  Complete F~ recovery at very high concentrations
of Fe3+ occurred with citrate and, to a lesser extent, with CDTA.


Automated System

A schematic illustration of the automated free and total fluoride
measurement system is shown in Figure 38.  This system was oper-
ated in the laboratory to establish performance characteristics;


                                87

-------
Table 10.     ALUMINUM COMPLEXING EFFECT AND FLUORIDE RECOVERY
                                BY DIFFERENT MASKING AGENTS

Aluminum
Concentration
Acetate
Buffer

EDTA


Oxalate

CDTA


Citrate


meq/l
mg/l
0.1 Wl
1.0M
meq/l
6.
30.
60.
6.
30.
6.
30.
60.
6.
30.
60.
% Recovery of Fluoride
0.0
0.0
100
100

101
100
100
100
102
102
101
100
101
101
102
0.006
0.054
94
95

98
97
97
99
100
101
100
100
100
100
101
0.012
0.108
90
92

94
92
90
96
97
99
97
97
100
100
101
0.024
0.216
80
83

88
86
85
94
96
97
95
94
99
100
100
0.06
0.54
56
62

81
80
79
89
93
95
91
89
95
100
101
0.12
1.08
28
48

69
69
68
85
91
92
87
85
92
99
100
0.6
5.4
5
15

24
25
27
54
82
86
85
82
67
89
95
1.2
10.8
3
10

15
16
15
32
67
73
73
73
53
83
91
Buffered masking agents added to standard solutions containing 5 x

(0.95 mg/l), and variable aluminum concentration
                                                      fluoride,

-------
                            Al3* mg/l
                      3.6       5.4       72
                      0.4
                  0.6
               Al3* meq/l
0.8
1.0
                                            10.8
1.2
  I  Acetate  buffer, l.OM
 II  EDTA,  60 meq/l
III  Oxalate, 30 meq/l
                   IV  CDTA, 60 meq/l
                    V  Citrate, 60 meq/l
                   VI  Citrate, 200 meq/l
Figure 35.
Aluminum complexing  effect and fluoride recovery
by different masking agents.
                              89

-------
   100
   90
 o
 01
 •s
   80
 5
 o
 o
 o
 a:
   70
   60
•  6 meq/l

•  30 meq/i
   60 meq/l
                       8        12       16
                           Time in Minutes
           20
24
 Standard solutions  contain 1.0 mg/1 fluoride and
 0.5 mg/1 aluminum.
Figure 36.  Rate  of  fluoride recovery with CDTA masking agent.
                               90

-------
  100
£90

*k_
o


tZ

M-
o

>,80


S
o
o
  70
  60
     0
8        12


   Time in Minutes
16
20
24
Standard solutions  contain 1.0 mg/1 fluoride and


0.5 mg/1 aluminum.
Figure 37.  Rate of fluoride  recovery with citrate masking

            agent.
                              91

-------
«

Jj


2
        T3

         |E

        jZ "5
         <»>co
         a>
        it
                              t>

                              in
                                       a>
                                      •o

                                       oE
j»iqqr»q»o—•**
 £
 In
Is
o  Q}   fy   ao
<^  o   o   o
          S   -
                                                      E
                                                    S
                                                   a.
                      _a<
                      o.


                      I
                                                            CU
                                                            T5
                                                            •H
                                                            M
                                                            O
                                                            -P
                                                            o
                                                            -P
                                                            i
                                                            rH
                                                            (0
                                                            c
                                                            m

                                                            cu
                                                            H

                                                            &
                                                            (C
                                                            OJ
                                                            -P
                                                            (0
                                                            g
                                                            o
                                                            -P
                                                       o

                                                       u
                                                       -H
                                                       -P
                                                       (0
                                                       g
                                                       CL)
                                                       ^
                                                       0
                                                       00
                                                       ro

                                                       0)
                                                           -H

                                                           CM
                 92

-------
CO

z
LLJ
a
I

>-
03
cc
LU


8
LU
CC

LJJ
a
E
o
O

<
O
LJJ
01

Z
O
cc







•8

o
C
"S
L_
§
8
0)
OC
*












^
CN
CM
r~

(O
O
CM
O

CO
o
0
s
o
d

CN
o
0
CO
o
o
o
o
o
—
g
E



2
8
5
a
CM
CM
£
-
CO
CM
CM

CM
«-
LD
O

CM
CM
0


«~
0
o
o
O)
C
o
L.
§
<3
s

un
O)


oo
O)

O)
CD


8

o
o


0


o


o
o
1 1


<2 .
to j;
si
< OQ
§

O)
O)


§

o
o
^

o
o

8


o
o
T~

8
T—

o
o
8


<
8
r~ O> O
a> o> o

888
» *^ T~

O O O
o o o

o o o
o o o
t T™" '

o o o
o o o
^
o o o
o o o


888


888
T— r— r—

o o o
o o o
R S


o>
!;:
•(-»
O
                                         93

-------
i.e., buffer to sample ratio, flow rates, mixing times, etc.
required for design of the "continuous" fluoride monitor.  Flow
rates for each reagent/sample stream are expressed numerically
in the schematic.  Standard solutions and samples are placed in
individual sampler cups.  Automatic sampling is initiated after
stable baselines are established for reagents flowing through
the two flow cells.  The sample- or standard- is withdrawn,
segmented with air, and mixed with buffer in a mixing coil.- The
sample stream is debubbled; i.e., the segmented air and a portion
of the test solution are pumped to waste.  Residual test solu-
tion is passed through the flow cell for sensing by the electrode.

Sensed potentials for standards or samples from each electrode
system appear as output peaks from the recorder, one each for
free and total fluoride.  Calibration curves are used to read
the respective peak heights as fluoride concentrations.  Complexed
fluoride is reported as the difference between total and free.

The Nernstian semilogarithmic relationship has an important effect
on automated measurements; i.e., it precludes the use of zero
concentration fluoride as baseline solution in the measurement
system.  A fluoride standard solution must be used as the wash
between samples and standards.  A concentration of 0.1 mg/1 F~
is used for the wash/baseline solution.  This level of F~ is
well above the lower limit of detection for the electrode, is
in the linear range of electrode response, and is a level at
which the electrode response rate is relatively fast.  Fluoride
levels higher than the baseline concentration; i.e., F~ >0.1 mg/1
appear as positive peaks  (above the baseline), and F~ <0.1 mg/1
appear as negative peaks  (below the baseline).


E^e.c.t oft Pump on Re^e^ence Etic.tn.ode,

Early in the study, electrode potentials fluctuated in regular
rhythmic sequence with surges of the proportioning pump.  This
effect is shown in Figure '39.  These fluctuations were assumed
to be related to variation in junction potential of the reference
electrode, resulting from movement of the solution boundary at
the porous electrode junction.  The effect was overcome by
shifting the location of the pump to the debubbling line ahead
of the flow cell, and placing the reference electrode slightly
downstream after the fluoride sensing electrode.  With these two
changes, the surging effect of the pump was suppressed and a
smooth curve obtained as recorder output  (Figure 39,) .


ftu.£ 6e.fi:  Sample,  Rat
-------
jl
o
 fc
 %

_*J
u>
         8-S
V
                                                       ii
                                                      ^)i  \
                                                      ttuj   L
                                                        •o
                                                        o
                                                        o
                                                        
-------
water is 1.0 mg/1 and the buffer to sample ratio is 1:1, 1:2, or
1:5, the final concentration of fluoride in the sample stream is
0.5, 0.67, or 0.83 mg/1, respectively.  In addition, simple
dilution might lower the level of fluoride beyond minimum detec-
tion of the electrode or the linear range of electrode response.
A buffer to sample ratio of 1:5 was used throughout this study.


F£ow Rate.*

Early in the study, a net flow rate of 0.95 ml/minute through
the flow cell was used.  Electrode response was significantly
improved when the flow rate was increased to 2.04 ml/minute.
This effect is shown in Figure 40 .  Higher flow rates resulted
in no further improvement .
Coils are used to mix reagents with test waters in the fluoride
system, and to provide sufficient lag time to allow reaction
kinetics to proceed to completion.  A two-minute coil is adequate
for the free fluoride system, but a longer coil is required for
the masking agent to react with Al-F complexes in the total
fluoride system.  At A13+ = 1.5 mg/1, recovery of fluoride at
one mg/1 is 92.5 percent complete with a six-minute coil, and
100 percent complete with a nine-minute coil.  A 12-minute coil
is used for higher levels of complexes in the test water.
Conc.
-------
                          time-*
Sample and Wash Time:  A
                       B

           Flow Rate:  I
                      II
3 minutes
2 minutes

2.04 ml/minute
0.95 ml/minute
Figure 40.  Effect of flow rate versus response time,
                           97

-------
1
1
/
/
.X
^.rf»
\
\
___y
ILL
- • — -77
.— _i/

J
t
" 	 	 i^
v^
^J)
\s:
-Ji
c c~
* o> T
•o -§ y
J 1 L
Li. u.
"5 
§^
 C" ^
li. — o
o> o -^
E I

& ^c.
0 
c
QJ
0)
n •
m x5
u
m ro
o a)
O 0)
•H -P
-p 3
to a
i-l -H
rQ g
•H
rH LO
rt
0 -
0)
n e
O -H
m -P
-P X!
3 W
QJ rO
d
O 'd
M fO
OJ
T3 OJ
M rH
O DJ
u e
0) (0
Pi CO
•
H
•^r
Q)
M
3
iji
•H
98

-------
 o>  0
."2  -o

II


"o  o*
.o  .»-.
                                                   O
                                                   d
                                                   -P
                                                   O
                                                   -P
            r—
t

 E
                                                    cd

                                                    0)
                                                     3
                                                    (d G
                                                    M -H
                                                   ^3 g
                                                   -rH
                                                   •H CN
                                                    (d
                                                    o  -
                                                      
-------
            A. Calibration
                Total Fluoride
                Free Fluoride
            B.  Reproducibility  (1.0 mg/!  F)
                                 	 Total Fluoride
                                     Free  Fluoride
                   no Al
              0.5 rr-g/i Al
Figure 43.
Recorder output for calibration of free and  total
fluoride.   Sample  and wash time, 1 minute each.
                                   100

-------
show calibration curves made at five, two, and one minute
sample/wash times, respectively.  At one minute, reproducibility
was poor, with or without aluminum present (Figure 43) .   A
sample/wash time of two minutes each; i.e., 15 samples/hour, was
highly reproducible and adequate for automation.
Recorder outputs for the calibration procedure with sample/wash
times of two minutes each are shown in Figure 42.  Fluoride
standards with increments from 0.1 to 2.0 mg/1 were used.  Each
standard is represented by two peaks , one for free and one for
total fluoride.  Calibration curves drawn from corresponding
peak heights versus fluoride concentrations are shown in Figure
44.  A linear relationship was shown to exist for samples to
200 mg/1 or >10~2M F~ , but this level is much greater than any
level expected to occur in drinking water measurements.
Precision of the automated system was evaluated by calculating
the deviation (S) , in mv, from peak heights and determining the
corresponding S in concentration from the following relationships :
and:
where:
             = EQ - 2.3 p- log Ai                             (114)
          E2 = EQ - 2.3 p  log A2                             (115)
          EI = potential of fluoride concentration Ci
          A! = activity of fluoride concentration Ci
          E2 = potential of fluoride concentration C2
          A2 = activity of fluoride concentration C2


          Ei - E2 = 2.3 P. (log C2 - log d)                  (116)
                        r


           AE   = log C2 - log Ci                             (117)
          slope

Thus, with the potential difference resulting from fluoride mea-
surements, the slope, and original concentration  (Ci) known, C2
and AC can be calculated; and the difference in concentration
(AC ) at any other fluoride level (C ) resulting from a potential
variation (AE ) can be expressed as:
                               101

-------
                       I    I    I   I  I    I  I
                        Free Fluoride
                        Total  Fluoride
                           i    i   i  i   i i  i
               0.2            0.5         1.0
                    Ruoride Concentrations  Img/t)
2.0
Figure  44.  Calibration curves for  automated fluoride
             electrode measurement.
                             102

-------
          Acx • 57 cx AEx

i.e., at 25°C, F~ = one mg/1, and the slope is - 60 mv, the
effect of a one mv potential variation is:


         -^ = log C2 - 0                                      (119)


or:

          C2 = 1.04 mg/1                                      (120)

and:

          AC = 1.04 - 1.0 = 0.04 mg/1                         (121)

Reproducibility of fluoride measurements at various levels of
fluoride and no complexing agents is shown in Figure 45.  The
reproducibility with fluoride of 1.0 mg/1 and various levels of
aluminum is shown in Figure 46.  Standard deviations and coeffi-
cients of variation for peak heights in Figures 45 and 46 are
shown in Table 12.  The high precision of the method is clear
from the low standard deviations and coefficients of variation.
The concentration effect ranged from coefficients of 0.17 to
0.63 percent with a mean of 0.37 percent in the absence of
aluminum, and from 0.35 to 0.99 percent (1.0 mg/1 F + 1.5 mg/1
Al) with a mean of 0.67 percent with aluminum present.  The peak
height coefficient of variation is generally smaller than the
concentration coefficient, but increases as peak height decreases;
thus, concentration coefficients are better measurements of
precision.


ACC.UAO.CI/

Accuracy is measured by mean error (E), which is_defined as the
difference between the actual mean of all data (X) and the true
mean (T.V.); i.e.,

          E = X - T.V.                                        (122)

and:

          R.E. = XT~VT'V- x 100                               (123)


where:

          R.E. = relative error
                                103

-------
                                                        Total Fluoride
                                                        Free Fluoride
                  '   «
                  /I  //I  ,'/1  /I  ''A
                    \l
       fi  i\
      VJ
1.5 mg/l  Fluoride
       B.
                       	  Total Fluoride
                       	  Free  Fluoride
                             1.0 mg/l Fluoride
       C.
                       	  Total Fluoride
                       	  Free Fluoride
                     VI  VI
                             0.6 mg/l  Fluoride
Figure  45.   Reproducibility  of fluoride  measurement in  absence
              of complexing cations.
                                    104

-------
      A.
                             ~—  Total Fluoride
                             	  Free Fluoride
           VJ
                                                    ' I  /
                                                     i  /
                                                     i  i
                                                     i  i
                                                     i  i
                                                     i  i
                                                     \ i  \
                                                      \ i  *
                           O.I mg/l Aluminum
      B.
                             	  Total Fluoride
                             	  Free Fluoride
                              /i  A   *
             .V
/I
(
i


/-




i
i

/i
/
'


r






\
f
i



r






\\
i



r







i



f





\
I
/'
'



/-


i'
i
i
i
i
\!
v
                            V  V«   v   T-'
\J  NJ  VJ
                     V!
                           0.5  mg/l Aluminum
      C.
                             	  Total Fluoride
                             	  Free Fluoride
                /•i  /
               i   ;
\!
'*
        \
        '
                                N!
                                        !
/ '
/
1
i
i
i
i
I
j
j
i








\
\
\
/



















\
\
/



















\
\
\
i i
i
i
i
i
i
i
t



                           1.5 mg/l Aluminum
Figure 46.   Reproducibility of  fluoride measurement  (1.0 mg/l)
              in  presence of  aluminum.
                                  105

-------
Table 12.    REPRODUCIBILITY  OF AUTOMATED MEASUREMENTS AT DIFFERENT
           FLUORIDE LEVELS IN PRESENCE AND ABSENCE OF ALUMINUM


Original
Conc.(mg/l)
F + Al

1.5 + 0.0

1.0 + 0.0

0.6 + 0.0

1.0 + 0.1

1.0 + 0.5

1.0 + 1.5


Fluoride
Form
Free
Total
Free
Total
Free
Total
Free
Total
Free
Total
Free
Total
Fluoride Measured
Mean Value
Peak
Height
(Chart
division)
47.97
54.59
38.28
43.30
25.99
28.18
36.52
43.61
24.58
42.9
-9.26
43.32

F.Conc.
(mg/l)
1.50
1.50
1.00
1.00
.60
.60
.92
1.02
.58
.98
.12
1.00
Standard
Deviation

Peak
Height
.065
.158
.101
.074
.030
.118
.188
.185
.104
.119
.247
.178

F.Conc.
.004
.009
.004
.003
.001
.003
.008
.007
.002
.005
.001
.005
Coefficient
Variation

Peak
Height
(%)
.14
.29
.26
.17
.12
.42
.51
.42
.42
.28
2.66
.411

F.Conc.
(%)
.27
.63
.40
.30
.17
.47
.75
.73
.35
.50
.99
.71
In establishing the accuracy of the  fluoride  automated  system,
data were correlated with SPADNS data  for  the same  samples,  ex-
pressed as T.V.  Actual data obtained  for  test waters from two
Michigan cities using both SPADNS and  the  automated technique
are shown in Tables 13, 14, and 15  (pages  107, 108,  and 109,
respectively).


Continuous Fluoride Monitor

Figure 47 is a schematic illustration  of the  complete fluoride
monitoring system installed in the mobile  laboratory.   The
system operates on a 26-minute cycle,  controlled by  the computer,
At initiation of the cycle, the relay  is switched to the "on"
position and the three-way valve is  connected through to the
baseline solution.  For the next 13  minutes,  baseline solution
is drawn into the system.  Ahead of  the pump,  a Y-connector
separates the sample and sends it to both  the free  and  total
                                106

-------
r-
•
Z



o
o
DC
 D)



CO
UJ
QC
01
(A
Z
g
oc
UJ
H
LU
Q

Z
O
cc.
LU
Q
cc
o

_i
U.
_£

J2
CO
£ §1

o ~
s
CD
Q. — —— — —
CO
1
C c c-
.2 '| ^>

1 !~~
UJ
. 
03 -T3
0) ^ 0 ^

i § g E
LL [i]




Location

1 1
m
"'





CO
q
o



in


00
CN
d






CN
CN
d


CN
CN
d





L_
£
re
re
"3
5
-
in
,_'





CO
o
d



^f ^3 r*^ ^f
CN CN CN C53
T-' ^ d d

O o> r- T-
co q co in
d i-^ d d






in »- o oo
CN T-_ CO ^;
d i-' d d


in o en r»
CN T- CN Tfr
d i-' d d

a-  <->
3 o c
Q. Q. re
* I I .
i s| 5 a
« « § ^ •
^ ^ a o g5
^ ^ in S m
S £&i J= «
re re o ~
EL ZZ ^ S
CN co Tt in













CO 00 O CO
q o> q CN


co *— o o
co q r-; «-;
d i-' d ^






8 8 R ^
d «- d *-


O3 CD O «-
CM O> P- «-
d d d --1

•
Q.
co • o
O '~
<% Z - cu .2
^> .ti .i c
UJ O u. D
CO P- 00 O)













CN * CN
CO X CO
d-d

«— o> in
oo r^ r^
odd






OS 00 CO
p** p*» p**
odd


O) O) CM
r^ i*1""" r**«
odd





CD
(/>
3
O
-C c
o
•s >. s
c ^ oo
5 ? =
S re j?
OC _l CO
O «- CN

                                            107

-------
01
X
h-

CQ
z
<
g
i
o
o
5
o
DC
CO
LU
r? to

S8
DC I
LU H
I- LU
  LU
sg
ttui
gsi
-2
S5
N Q

S?
LU <
X <
o S
_QJ
JD
CO
c ^
o 01
± £
1 I
E £
D —
*
"P  01
£ S
OJ O
c3 <
D
U_
OJ
03
LL 2
CO
.22 o>
y- ^
to 3
00 z

c
o
1



1 § ' ^ "I 88 ' '
T P P P
I CN |0 || Og ||

|| CN CN ^r «- ooco coin
CDOl CDCM TfOO in^t
CNCN »-CO «-CM r-CN



mCN coco inm »— in inin
,--<- oo oo »-o oo



«- «- r^co -stco coco ^rco
«- «- oo oo oo oo


CN»- incN cooo oo^- r--co
^— r~ Is** r^ oo to oo r*** oo r**1






T-T- 0)10 t^CN CMCD I^O
T- ^- r^p*". oor^- or-* oo oo


^ OQ | | | | | | [|
|l co^- LOCO i>*oo O)O

s <-•

5 S ^ |
5 t '5 2
ro o> C £
•- o S) c
c -3 5 E
CU (O 3 O 03
g Q. I C3 CD
II II II

II II II

O ^" O) 00 CN 00
co r**" r^ co CD ^f
i- CN T- CM i- CN



Sco in co CD in
o o o o o



co co co CD in in
o o o o o o


in i- co in ' in CM






o> r^ r^ *— oo o
oo r^ oo co oo P^

< CD  2

y b bt"
1 LO
q
1 0

O3 CO
r~ co
«- CO



oo r>.
o o



CO CO
o o


CO CM






r«« o
oo r^


1 1
r^. oo



£
O)
o
Z
                                                            O5
                                                            CM
                                                            CM
                                                             I
                                                            in
*
8
o.

I

CO
                                                            O3
                                                            CN
                                                            CN
                                                            _
                                                            a
                                                            E
                                                             to
                                                             U.
                                                             OI
                                                                c
                                                                01
                                                                E
                                                                0}
                                                                (U
                                                                T3
                                                                O
                                                                D
                                                                C
                                                                (O
                                                                E

                                                                CD
                                                                .C
                                                                E
                                                                o
                                                                •g
                               108

-------
 Table  15.  SOME STATISTICAL  DATA RELATED  TO MANUAL  AND AUTOMATED
           FLUORIDE ELECTRODE MEASUREMENTS IN  WATER SAMPLES  FROM
           WYOMING, MICHIGAN
Parameter
Mean, x *
Std. Deviation, S
Correlation coefficient, r
Degrees of freedom, d.f.
t— value
Statistical difference
Free— F Measurement
Manual
0.737 mg/l
0.2367
Automated
0.704 mg/l
0.2252
0.9931
34
0.4436
N.S.t
Total -F Measurement
Manual
0.952 mg/l
0.3065
Automated
0.955 mg/l
0.3034
0.9989
34
0.0240
N.S.
 *AII the 18 samples were included.

 'Not significant - no difference could be established.

fluoride systems.  Beyond  the pump,  an air bubble is introduced
to separate the sample stream into many discrete samples and the
appropriate buffer reagent is added.   Next, a mixing coil pro-
vides mixing and reaction  time before sample reaches the sensors,
Mixing coil length is  especially important in areas where free
fluoride is significantly  different  than total.   The sample
stream is debubbled just before it enters the sensor flow cells.
The bubble stream goes to  waste under siphon while the remainder
of the sample  is drawn by  a fourth pump line past the sensors,
then to waste.  At the end of the 13-minute period, the computer
sets the relay to  "off" and the three-way valve  is open to the
tap water sample stream*   A complete  cycle of 26 minutes is re-
quired to read both baseline and tap  water samples.  Prepared
reagents are sufficient for ten days  to two weeks, depending on
system downtime.

Fluoride sensor flow cells are constructed from  glass, not
plastics.  As new cells are required,  they are aged (placed in
service for 24 hours to reach equilibrium with fluorides in the
test stream) to prevent loss of fluoride from adsorption.  Other
glass components; e.g., mixing coils,  are similarly aged.

Normal daily maintenance of the fluoride system  includes careful
examination of the recorder output to assure that the signal is
of high quality.  Bubbles  which become trapped in the flow cells
and on the surface of  the  electrode sensing element are the most
frequent problem.  They are removed by briefly pinching off the
                                109

-------
„. 2
z o
5 "
•• ?f —..,
• VI


VI
Q
— 0
" ""
Ui O U.
S 2 S
u. u. *

Ul
E .

< 0 «•
i-aS:
O -i 3
£ U. «


Ul
»-
VI
<
O
t-


j *i .? •< •* «.
Ul
Q
ee
O
3
_i
u.
_i
<
h-
0
»-
Q
Z
<
Ul
Ul
DC
U.



























(9
Z
Z
h-
i!
tt 3
x a.





















•t
ui
«i
«•
«
a
«
Ul
0
























>
X
v.
X
&

1 STRIP-CHART
•ECOIDER




•
4





M
^
0
u
O
3
>
i




•
111
»-
Ul
X






<
i



X

g
c,
c

~
111
u
ct
o
VI
Z
lil
VI
1
1
1

1




\
)




t*




























	





m
<



N,

i
i

1
(
I
1
1



- -/



~|

)






M
•<
^ M
? 2
•
3
•
IU
a

^
^
u
J
K
9
A
Z
u
A


> X
Z u,
^ H-
Z ui
«- S
[
STRIP-CHART
RECORDiR
S « .
^ " >
5 * <
M V> >




1
1
1
>-
<
^
IU
•1

X
Ul
1-
3
u.
X
0
u
O
U.




•c
IU
_ _ ^_ ^
~ - ^— -J
a.
S
o
	 	 u
O
»-

                                                      a
                                                     -H
                                                     -o
                                                      -P
                                                      -H
                                                      c
                                                      o
                                                      e

                                                      in
                                                      a)
                                                      T3
                                                      -H
                                                      0)
                                                      (D
                                                      M
                                                      o
                                                      O  >i
                                                     •H  H
                                                     -P  O
                                                      (0 -P
                                                      M  ai
                                                     •P  M
                                                      0)  O
                                                     •H
                                                         (D
                                                     O rH
                                                     -H -H
                                                     -P X5
                                                     nJ  O
                                                     e  e
                                                     0)
                                                     x;  Q)
                                                     O ,£
                                                     en -P
                                                     0)
                                                     tn
                                                     -H
110

-------
flow cell outlet and releasing it quickly, causing a flow surge
in the cell.  Persistent bubbles are dislodged by removing the
electrode from the flow cell and replacing it immediately.  The
millivolt meter is in the standby mode during this operation.

Standard solutions used for calibration of the strip chart and
computer output are 0.2, 0.5, and 1.0 mg/1 fluoride in distilled
water.  Peak heights are measured in chart units (or mv) and
plotted on semilogarithmic graph paper, with peak height on
the linear scale and concentration on the log scale.  Figure 48
is an example of fluoride system output and includes a calibra-
tion series.  Figure 49 is the calibration curve prepared from
this recording.

[F ] (mg/1)

0.1
0.2
0.5
1.0
Peak Height (C.U.)
Free

0.0
25.0
53.5
78.0
Total

0.0
17.0
36.5
52.5
Fluoride concentrations in tap water are determined by measuring
peak heights and comparing them with the calibration charts.  It
is important to note that the three or more points describing
the calibration curve may not be linear, particularly as elec-
trodes age.  It is not uncommon to observe a decreasing slope
with decreasing concentrations of fluoride.


Field Studies

The fluoride system was used to acquire data from several special
field studies.  Samples from Jackson, Wyoming, and Ann Arbor,
Michigan were analyzed in the laboratory prior to installing the
fluoride monitor in the mobile laboratory.


Jackson, M'i.efo't.gan

Raw water for the City of Jackson, Michigan is drawn from deep
wells (380 to 400 feet) in rock.  Treatment includes chlorination
and fluoridation only.  Fluoride is added as sodium silica
fluoride.  Polyethylene bottles were used to collect 12 grab
samples from the well field, the water treatment plant, and a
number of sites along a principal distribution system pipeline.

Fluoride in these samples was measured in the NSF laboratory by
ion-selective electrode and SPADNS; and iron, by phenanthroline.
Results are shown in Table 13.  Iron levels ranged from 0.27 to
                               111

-------
                                                          Cfl
                                                          0)
                                                          •H
                                                          S-l
                                                          0)
                                                          c
                                                          o
                                                          •H
                                                          -P
                                                          X!
                                                          •H
                                                          O

                                                          (0

                                                          tn
                                                          C
                                                         •H
                                                          u
                                                          c
                                                         •H

                                                         -P
                                                          3
                                                          O4
                                                         •4->
                                                          3
                                                          O
                                                          o;
                                                         -P
                                                          tn

                                                          cu
                                                         T5
                                                         •H
                                                          M
                                                          O
                                                         00
                                                         OJ
                                                         Cn
                                                         -H
112

-------
  60
   50
   40
-P
•H
c
3
u
-P
x:
tn
0)
   20
   10
         Free Fluoride
80
                             60
40
         Total Fluoride
    0  .1     .2   .3  .4  .5      1.0      .2         .5     1.0


                 Fluoride  Concentration (mg/1)
 Figure  49.   Typical  calibration curves for fluorides,  prepared

              from output  in Figure  48.
                              113

-------
2.0 mg/1, with a mean of 1.12 mg/1.  (The laundry, sample no. 11,
apparently used iron removal procedures to protect against stains.
This sample was excluded from the mean.)  Comparison of the data
for free versus total measured fluorides indicated no signifi-
cant complex formation at these levels of iron.  This is consis-
tent with previous laboratory observations.  The wide variations
in fluoride levels were attributed to inefficient mixing as the
water was fluoridated.  Sodium silica fluoride is mixed as a.
paste and fed to the water in pulses, creating a nonhomogeneous
mixture of fluoride and treated water.

The fluoride data acquired by ion-selective electrode and that
from SPADNS were highly correlated; i.e., correlation coefficient
equal to 0.994 and student's t equal to 0.1975, indicating there
was no statistically significant difference between the two
methods.


Wyoming, M-ick-igan

The population of Wyoming is 78,000.  Raw water is drawn from
380 to 400 feet deep wells in rock.  Treatment includes alum
flocculation, filtration, phosphate addition for corrosion con-
trol, chlorination, and fluoridation.  Both grab and weekly
composite samples were analyzed in the laboratory.  Results are
shown in Table 14.

Levels of complexed fluoride were clearly significant in the
Wyoming treated water supply.  In the distribution system, they
ranged from 14.8 to 33.3 percent of the total fluoride.  Key
samples were also analyzed for aluminum and iron by emission
spectroscopy  (Table 14).  There was close correlation between
aluminum levels and fluoride complexation  (correlation coeffi-
cient = 0.96).  The data from manual and automated measurement
of fluorides in Wyoming samples are compared statistically in
Table 15.  Again, the two methods were not significantly different,


Ann

Ann Arbor has a population of 84,000 and takes 80 percent of its
raw water supply from the Huron River, and 20 percent from wells
in drift, 35 to 56 feet deep.  Treatment includes alum floccula-
tion, lime softening, filtration, phosphate addition for corro-
sion control, chlorination, and-f luoridation.  Data from 15 loca-
tions sampled on April 6, 1971 are shown in Table 16.  Most of
the Ann Arbor area is represented in three samples.  Aluminum in
these samples was <0.02 mg/1; and iron, <0.1 mg/1.  No complexing
effects occur at these levels; i.e., free and total fluoride
levels are essentially the same.  It is further suggested by the
data in Table 16, that fluoride levels are similar throughout the
distribution system; i.e., they are not affected by length of
pipeline or main residence times.

                                114

-------
  Table 16.     CHARACTERIZATION OF FLUORIDE IN WATER SAMPLES FROM
                      ANN ARBOR, MICHIGAN (APRIL 6,1971)
Sample
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Location
River raw water
Well raw water
Plant finished effluent
Shell station, 1251 N. Maple
Fire station, 2130 W. Huron
Standard station, 1336 S. Main
Lakewood School, 344 Gray Lake
Residency, 2250 S. Seventh
Water Department Distribution
Center, S. Industrial Highway
Main fire station, 219 E. Huron
Northside School, 91 2 Barton
Bolgos Farms Store,
3601 Plymouth Road
Concordia College, 4090 Geddes
Mary D. Mitchell School,
Pittsview Drive
Newport Elementary School,
2776 Newport Road
Free Fluoride
(mg/l)
.14
.30
.96
.94
.95
.96
.95
.98
.98
.95
.93
.96
.96
.96
.98
Total Fluoride
(mg/l)
.14
.30
.97
.95
.96
.96
.94
.98
.98
.96
.93
.94
.97
.97
.99
Average values of fluoride analyses  from different locations in
Ann Arbor at various times during  the  period 1970 to 1972 are
shown in Table 17.  During 1970, fluoridation was inadequate.
Fluoride concentrations were considerably below the optimum level.
                                115

-------
Table  17.  CHRONOLOGICAL CHARACTERIZATION  OF  FLUORIDE  IN WATER
           DISTRIBUTION SYSTEM  OF ANN  ARBOR,  MICHIGAN
Date

July 28, 1972
April 3, 1972
January 20, 1971
September 16, 1971
April 6, 1971
March 22, 1971
March 9, 1971
November 6, 1970
October 4, 1970
October 15, 1970
October 5, 1970
Free Fluoride*
(mg/1)

0.98
0.98
1.08
1.11
0.96
0.97
0.94
0.43
0.59
0.29
0.49
Total Fluoride*
(mg/1)

.99
1.00
1.10
1.11
0.97
0.98
0.94
0.50
0.64
0.33
0.53
Complexed
Fluoride
Percent

NS
NS
NS
NS
NS
NS
NS
14.0
7.8
9.1
7.5
*Values presented are mean values of >10 samples


Aluminum analyses showed variations from <0.02 to 0.5 mg/1, and
iron varied from <0.015 to 0.9 mg/1 during that period.  Fluori-
dation has improved since early 1971.  Levels now approach
1.0 mg/1, the optimum level.  Aluminum is generally below
0.05 mg/1; and iron, below 0.3 mg/1.


Ck^icago, ILL-ino-ib

Raw water for the City of Chicago is entirely of surface origin.
It is drawn from Lake Michigan and treated in one of two plants,
the Central- or the South- Water Filtration Plants.  The combined
rated pumping capacity from these two plants is nearly three
billion gallons per day.  Distribution is made through more than
4,000 miles of tunnels and mains to all of Chicago and 72 subur-
ban communities.  Treatment at both plants includes alum-lime
flocculation, high rate filtration, chlorination, and fluorida-
tion.  Caustic for corrosion control is added at the Central
Plant.

During its first month's assignment in the Chicago area, the
mobile laboratory was located at a fire station, 10.5 miles from
the Central Plant.  Average daily mean, high, and low free and
total fluoride levels recorded on coincident days of normal opera-
                               116

-------
tion are summarized in Table 18.  Although aluminum is not mea-
sured in the mobile laboratory, the water treatment plant labora-
tory reported a level of 0.16 mg/1 in routine quarterly distri-
bution system samples from the north district.  It is apparent
that complexed fluorides are important in the Chicago water
supply.


     Table 18.  SUMMARY OF FREE AND TOTAL FLUORIDE DATA
                (CHICAGO/LEHIGH STATION)
Total Fluoride, (mg/1)
High

1.10
Low

.66
Average
Daily
Mean

.90
n

15
Free Fluoride, (mg/1)
High

.90
Low

.47
Average
Daily
Mean

.71
n

15
Beginning at 1200 hours on May 25, 1973, fluoride feeders at the
Central Plant were taken out of service.  The purpose of this
operational change was to observe the capability of the fluoride
monitor in the NSF/EPA mobile laboratory to detect the change,
and to determine residence times in the distribution system.
Results are reflected by a reduction in total fluoride levels
recorded from 1200 to 2000 hours.

This study was repeated on May 29th and 30th with fluoride feeders
cycled in- and out-of-service two times at four—hour intervals.
The first out-of-service period was from midnight May 28th to
0400 hours May 29th.  Results are plotted separately in Figure
50.  The data show that transmission time to the Lehigh location
during periods of low demand is 17 +_ 1 hours.  This is apparent
from the first recorded drop in total fluoride.  When the feeders
were placed back in service, recovery was incomplete because of
the low demand period and mixing in the distribution system.
Also as a result of these factors, the second decrease was signi-
ficantly lower than the first, and lasted until demand increased
during the morning hours.

The City of Chicago operates a high purity aluminum electrode
fluoride monitor (Fisher Porter) at another fire station 8.5
miles from the Central Filtration Plant.  The relative locations
of these stations are shown in Figure 51.  Data recorded by the
Chicago monitor during the May 29th-May 30th period are shown in
Figure 52.  The'actual levels of fluoride are not expressed on
the recorder output, but it is interesting to note the relative
                               117

-------
S/I

s/o

S/I
o

H
                                                        w
                                                        t-»


                                                        i
                              O
                                          O
                                                                     -l OJ
                                                                     0) U
                                                                     in -H
                                                               009T  ° »

                                                                     0 C  c
                                                                     0-H  o

                                                               QQfrT  w w  '^
                                                                     H H  (^
                                                                     0) 0)  4J
                                                                     TJ TJ  U)
                                                                     0) 0)
                                                                     (U 0)  ,£


                                                                     H II  -?
                                                                     CO CO  r"
                                                               OOOT  O H
                                                               0080
                                                               0090
                                                               QOfrO




                                                               0020




                                                               OOfrZ




                                                               OOZZ




                                                               0002




                                                               008T
                                                               0080
                                                                               o
                                                                               Cn

                                                                               O
                                                                              •H

                                                                              U

                                                                              -P
                                                                          CO
                                                                          o* o  en
                                                                          CD n
                                                                          2 .  cu
                                                                               u
                                                                               
-------
?-)  II  ^^ 1  ''•-'l.n.—,    ux J-
Figure 51.
             Map identifying relative locations of monitoring
             sites in Chicago.

                              119

-------
                                                cr>
                                                 Ul
                                                        (0
                                                        C
                                                        OJ
                                                        0)
                                                        H
                                                       O

                                                        O
                                                        tn
                                                        (0
                                                        O
                                                       -H
                                                       J3
                                                       U
 0)
tJ
 M
 O

 <1J
                                                       C
                                                       O


                                                       0)
                                                       T3
                                                       •H
                                                       M
                                                       O
                                                       O
                                                      •H
                                                      A
                                                      U
                                                       -P
                                                       3
                                                       CU
                                                       4J
                                                       3
                                                       O
                                                      CN
                                                      ID


                                                       (U
                                                       01
                                                      •H
120

-------
times at which peaks were recorded by the Chicago monitor versus
those by the NSF monitor:


                       1st minimum   Crest   2nd minimum
     Chicago Monitor      1650       1915   2250 (May 29)

     NSF Monitor          2000       2200   0600 (May 30)

(Reference to Figure 51 indicates more direct transmission to
the Chicago monitor than to the NSF mobile laboratory.)

A subsequent fluoride detection study occurred at the mobile
laboratory's fourth location in the Chicago area, Calumet, which
receives its water from the South Water Filtration Plant.  Be-
ginning at 1000 hours on June 25, 1973, fluoride feeders were
interrupted twice, with two consecutive on-off cycles.  These
data are plotted in Figure 53..  Note that times of recorded
changes at the South Plant outlet are shown in this figure.
The first change at the treatment plant outfall was detected
5-1/4 hours after the feeders were taken out of service.  The
first minimum level was recorded at the plant outlet 9-3/4 hours
after the feeders were out-of-service.  This level was recorded
by the NSF monitor after 31 hours.
TRACE METALS

General

Trace metals in public water supplies are of special concern be-
cause of their potential long-term chronic effects on the con-
sumer.  Surface and ground supplies vary in types and quantities
of metals constituents but, in general, both are significantly
lower than EPA Interim Primary Drinking Water Standards (1) .

Metals are not removed by conventional water treatment processes.
In addition to their occurrence in raw waters, they may be intro-
duced to treated waters as impurities in chemicals used in the
treatment process, or as products of corrosion in municipal dis-
tribution systems or individual household plumbing.

Simple analytical methods for detecting and measuring trace
metals in water supplies are not generally available.  Except
for neutron activation analysis (NAA) and anodic stripping voltam-
metry (ASV), direct measurement of trace metals cannot be made
without concentration of the test solution; i.e., separation, di-
gestion with acid, and concentration, which frequently modifies
the physicochemical characteristics of the species of interest.

Although activation analysis offers the high sensitivity required
for trace analysis, its use is limited by availability of reactor
facilities.   Further, it is an elemental analysis procedure that

                               121

-------



















5/1-
s/o-
s/i-
w
•
*
.
0

•
•
• ©
\
•
• ©

•
• 0
•
•*
D
• °0«—
l^«—
4> O) ^ —
S / o-1 — v 	 1 	 1 	 > 	 1 	 1 	 1 	 1 	 1 	 1 	
00tr2
"0002

"008T
"0091

"OOfT
"002T
"0001
"0080
"OOfrO
"0020
"OOt-2
"0022

"0002
^008T
"OOOT
"0020

"008T
OOOT
o
•H
Vj to
0) O
O Q)
3 C
O-M
(fl 01
M M
0) t
m
S
t
.1
CN
*5
                                                 ftf
                                                 S3

                                                 -P
                                                 0)
                                                 u
                                                 -p
                                                 (tf
                                                 -P
                                                 tn

                                                 0)
                                                 •O
                                                 -H
                                                 M
                                                 O
                                                 O
                                                 (U
                                                 ro
                                                 Q)
                                                 -H
                                                 Cn
122

-------
provides no information about oxidation state or degree and type
of complexation.

The analytical feasibility of anodic stripping voltammetry
for trace metals analysis in natural and wastewaters has been
established (46).   In addition to its high sensitivity, simpli-
city, and portability, ASV has the unique capability of charac-
terizing free versus complexed trace metals.


Theory

ASV consists of:  (1) a preconcentration (preelectrolysis, catho-
dic deposition, or plating) step, in which the metal ion (or ions)
of interest is reduced by controlled potential electrolysis and
deposited as an amalgam on the surface of a composite mercury
graphite electrode (CMGE), and (2) anodic dissolution  (reverse
electrolysis, or stripping), in which the metal is oxidized and
returned to solution.  ASV is, therefore, a nondestructive
method of analysis.

The plating step is carried out under reproducible conditions so
that concentration onto the electrode is quantitative  (total
plating), or a reproducible fraction of the desired component is
deposited from solution  (partial plating).   By controlling the
potential during this process, the more easily electrolyzed con-
stituents can be separated; thus, the plating step is achieved
by applying to the electrodes a potential which is cathodic of
the polarographic half-wave potential by approximately 0.3 to 0.4
volts (where the current has its limiting value), and maintaining
this potential for a given period of time under reproducible con-
ditions of stirring, type, and area of the microelectrode, and
composition of the medium.  After a short rest period  (which
allows the solution to become quiescent, the stripping process
is initiated.

The stripping process is performed by a voltammetric scanning
procedure which produces a response proportional to the amount
of material deposited.  The resulting "stripping voltammogram"
produces peaks, the heights of which are proportional to the con-
centrations of corresponding electroactive metal ions, and the
potentials, a qualitative indication of the nature of the species
present in the solution.  Thus, the important characteristics of
the voltammogram are:  the heights of the peaks  (peak current,
i , in microamperes), peak width at 1/2 i   (W 1/2 in volts or
mSllivolts), and peak potential  (E , in vBlts).  These charac-
teristics are affected by the type"of microelectrode used, and
by the rate of voltage change  (sweep rate, v, millivolts per
second)  employed in the stripping process.

Three electrodes are used in conventional ASV:  reference
(Ag/AgCl), counter  (a platinum wire), and test or working  (mer-

                               123

-------
cury-graphite) electrodes.  Test electrodes are made by plating
a thin film of mercury onto a polished, wax impregnated graphite
rod.  Because oxygen is an interference in trace level analyses,
the flow cell has provision for nitrogen deaeration.  The analy-
tical curve (voltammogram) includes peaks for background current
as well as for metals being measured.

Differential ASV  (DASV) is a four electrode system.  A second
working electrode is used to subtract extraneous background
signals from the voltammogram, significantly improving the sensi-
tivity of the technique.  The fourth electrode is identical to
the working electrode except that it is excluded from the plating
step.  During the stripping procedure, the background current
signal from the fourth electrode is subtracted from the analytical
current signal producing only current peaks for the metals of
interest.  The special DASV plating-stripping sequence is dia-
grammed in Figure 54.

In the voltammogram, the peak value of the current is proportional
to the concentration of the reduced species, the area and thick-
ness of the electrode, and the potential scan rate, as shown in
Equation 124:
          i  = (ZFA1   v)CR                                   (124)
where :
          i  = peak current

           •L = number of electrons transferred
           F = the Faraday
           A = electrode surface area
           1 = thickness of mercury film
           d> = 2!
           *   RT
           v = potential scan rate
           e = Naperian logarithm
          C^ = concentration of reduced species
           K.

Film thickness is determined by controlled electrolysis of the
mercury onto a graphite electrode.  The value of i  varies nearly
proportionally to the scan rate for very thin mercery films;
however, it varies significantly with film thickness at higher
sweep rates.  For lower values of one, i  varies only slightly
with film thickness.                    P

Equation 125 expresses peak position:

                                                              (125)
                               124

-------
                                                                                      -P
                                                                                      CD
             T
              *4
              CD
     o
 w  -—
 «  '5
 w   C
-8  s
 03

SI
 s
• M
H
                                                                                      -P
                                                                                      iH
                                                                                       O
                                                                         •H
                                                                         PU
                                                                         a
                                                                         •H
                                                                         S-l
                                                                         -P
                                                                         en

                                                                         o
                                                                         •H
                                                                         t3
                                                                         o
                                                                         £
                                                                         ftf
                                                                         •H
                                                                         -P
                                                                         C
                                                                         CD
                                                                         M
                                                                         0)
                                                                         m
                                                                         m
                                                                         -rH
                                                                                       5-1
                                                                                       o
                                                                         a)
                                                                         u
                                                                         C
                                                                         QJ
                                                                          CD
                                                                          tn
                                                                                       •H
                                                                                       !-l
                                                                                       -P
                                                                                       W
                                                                                       I
                                                                                       CD
                                                                                       -P
                          
       -H
                                           125

-------
where:

          E  = peak potential

           6 = Nernst diffusion layer
           D = diffusion constant
          E  = standard potentia-l

E , then, is a function of scan rate, film thickness, and the
Nfernst diffusion layer.  This theory does not apply for fast
scan rates with very thick films.  Peak width, which is important
in resolution, varies only slightly with scan rate when the
mercury thickness is small, and varies slightly with thickness
with a slow scan rate.
Experimental

Equipment items in the trace metals monitor in the mobile labora-
tory include a custom built, four electrode potentiostat/timer
(P/T) unit with automatic and manual modes; Technicon 16-channel
proportioning pump and long jacketed mixing coil; Varian recorder,
Model G2500; two Valcor miniature three-way valves; a customized
Plexiglas flow cell; and four electrodes, Ag/AgCl reference,
platinum  (Pt) counter, and two graphite working electrodes.

In the automatic mode, all functions are controlled by the timer.
There are six variable timed functions, five of which are necessary
for one complete cycle:

          1.  Warm up  (200 to 2,000 sec.)
              Allows time for system to warm up and sample
              to travel between its point of origin and the
              flow cell.  (This function is used only when
              system is turned on.)
          2.  Plating primary electrode  (100 to 1,000 sec.)
              Switches primary electrode to plating poten-
              tial and initiates nitrogen mixing.
          3.  Plating differential electrode (10 to 100 sec.)
              Switches differential electrode to plating
              potential and stops nitrogen mixing.
          4.  Preelectrolysis procedures  (1 to 10 sec.)
              Starts recorder and lowers pen.
          5.  Analytical step  (2 to 20 sec.)
              Starts potential sweep on both electrodes
              simultaneously.  (Sweep stops automatically.)
          6.  Wash period (30 to 300 sec.)
              Starts nitrogen mixing, allowing cell to be
              flushed with new sample prior to analysis.
               (System  switches back to function 2.)

In the mobile laboratory, the time for one complete DASV cycle
is 13 minutes.  Plating potentials for each function are set at

                               126

-------
a digital panel meter on the front panel of the control unit.  A
digital counter counts the number of analytical cycles which have
occurred and, at some preset number of cycles, switches a valve
at the beginning of function six, pumping standard instead of
sample through tha cell.

Sample is delivered to the flow cell, acid added, and standard
pumped with the Technicon proportioning pump.  Nitric acid is
mixed with the test water in a ratio sufficient to produce a
resultant pH of 2.5.  (Metals exist in water in three forms:
free ions, labile complexes, and nonlabile complexes.  ASV does
not measure the nonlabile complexes unless the sample is acid-
ified to either break down the nonlabile metal complexes or
exchange the H  for a metal ion.  Total metals are measured only
with acidification.)

A ten-inch Varian recorder with remote start/stop, remote pen
lift, event marker, and a Z-fold paper tray is used with the DASV
system.   (An event marker is available and could be used to re-
cord the beginning and end of a standardization sequence.)

A water jacketed mixing coil maintains constant sample tempera-
ture and aids in the control of outgassing.  Bubbles formed by
outgassing are removed through a debubbler before the sample is
acidified.

The flow of nitrogen is initiated by one of the Valcor valves in
the system; the second valve switches from sample to standard
in the standardization sequence.

A schematic illustration of the Plexiglas DASV flow cell in the
mobile laboratory is shown in Figure 55.  The internal volume of
the cell is ten ml.  Test electrodes are inserted through the
sides of the cell, directly opposite each other, and mounted
with epoxy to 1/4-inch hose conversion fittings for easy removal
for servicing.  Reference and counter electrodes, also directly
opposite, are inserted through the front and back of the cell.
The test solution is pumped in through the bottom and overflows
through a port in the stopper at the top.  Nitrogen enters
through a small diameter polyethylene tube inserted in the bottom.

Working electrodes are made from 1-1/2 inch lengths of 5/16 inch
diameter wax impregnated graphite rods.  The end of each is
polished for working electrode surface.  Complete instructions
for electrode preparation are included in the "Operation and
Maintenance Manual," provided with the mobile laboratory  (47).

A schematic illustration of the total DASV system is shown in
Figure 56.  Before plating is initiated, both working electrodes
are at rest potential and the solution is mixed.  To initiate
plating, the analytical electrode is adjusted to plating poten-
tial; the differential electrode remains at rest potential.  When


                               127

-------
                                          To Waste
        Rubber
        Stopper
Ag/AgCl Reference
Electrode
 Graphite
 Electrode
     Pt Counter
     Electrode
                                       Threaded
                                       Electrode
                                       Holder

                                       Graphite
                                       Electrode
                                                    Sample
                                                    Inlet
         Plexiglas Flow
         Cell Body
                      Nitrogen Mixing
                      Gas Inlet
  Figure 55.
Schematic diagram of DASV flow cell used in
the mobile laboratory.
                              128

-------
Ill
OK _j
III ft.
«• 5 *
2* 2



o
Q
Z
<
w»


























Q
^ «
Ul < Q
i!|
*n *n >
<
1



\
i
! {
	 , o E
? ^ 5
..1
u
<

O
z

*-
o ,
o s
« 3
tL. O.
X^




/ « - «.
'| i 5 3 > «
- 1 1 2t < 5
« i «. u. • > _ 5
i , ... , *
i
1
i I
£ i ' *
o j 1 J os..
z o 1 ; ^ - J 2
S =: S I | z" £ 5
i i i ;
i i
ii i
1 ; ;
• i
1 1 1
1 ! ^ a

    O
    z

UJ  0
it  o  s
t  2  O
Q  <  >


Z
O   IK
«   -1
>-   3
U   O

"   i
m   <
                                                                             ±  O
                                                                                                             O
                                                                                                            4-)
                                                                                                             fO
                                                                                                             M
                                                                                                             O
                                                                                                            (U
                                                                                                            H
                                                                                                            •rH
                                                                                                            ft

                                                                                                            i

                                                                                                            Q)
                                                                                                            XI
                                                                                                            4-)
                                                                                                            M
                                                                                                            O
                                                                                                            -P
                                                                                                            •rH
                                                                                                            C

                                                                                                            i

                                                                                                            in
                                                                                                            -P
                                                                                                            0)
                                                                                                            S


                                                                                                            QJ
                                                                                                            O
                                                                                                            (0
                                                                                                            M
                                                                                                            -P
                                                                                                            c!
                                                                                                            o
                                                                                                           -H
                                                                                                           -P
                                                                                                            (0

                                                                               L-.
                                                          U
                                                         -H
                                                         4->
                                                          rtj
                                                          a
                                                          0)
                                                         XJ
                                                          O
                                                         CO
                                                                                                           in

                                                                                                            OJ
                                                         tr>
                                                         •H
                                                    129

-------
plating is completed, mixing is stopped and the differential
electrode adjusted to plating potential.  After a short waiting
period (to attain solution quiescence and electrical stability),
the stripping step is initiated by simultaneous application of
the voltage sweep to both electrodes.  The sweep is stopped at
a predetermined potential and rest potential is again applied to
both electrodes.  Typical DASV voltammograms for standards and
a tap water sample are shown in Figure 57.  Preparation of rea-
gents for DASV analyses is detailed in Appendix B.


Results and Discussion

Performance characteristics of the DASV system were determined
experimentally.  Calibration curves from standard solutions of
5, 10, and 15  ppb  of each of the metals of interest (Cd, Pb,
and Cu),  run in triplicate, established'a linear response for
each of the metals.  The plating time response curve for discrete
samples is S-shaped with a limiting value representing total
plating of a metal.  In the continuous flow system, the plating
time is linear as a result of the continuous flow of fresh
sample; i.e., total plating cannot occur.

Continued use of working electrodes results in reduced signals
for each of the measured metals.  Periodic standardization is
required to observe changes in electrode sensitivity.  Data in
Table 19 relate to a tap water sample spiked with five mg/1 each
of Cd, Pb, and Cu.  The sample was cycled through the flow cell
continuously for a period of 18 hours.  Although electrode re-
sponse decreases with time (Table 19), the rate of decrease is
relatively constant; thus, results can be read from a calibra-
tion curve corrected for elapsed time.  Periodic rejuvenation
of the electrodes can be accomplished by a short period of
plating additional mercury, as shown in Table 20.  These data
are from a pair of electrodes used continuously for alternate
standard/sample cycling over a 48-hour period.  The electrodes
were rejuvenated by short mercury plating periods after each
24 hours of use.

Electrode response to variations in sweep rate is shown in
Figure 58.  Rapid sweep rates produce higher current response;
thus, 162.5 mv/sec. was used for much of the DASV acquired data.


Special Study

A special study --of the effect of residence times in household
plumbing on metals uptake in drinking water was undertaken during
this project.  NSF employees were asked to bring two samples to
the laboratory each day for five consecutive days.  One sample
was drawn from the tap immediately upon arising each morning;
i.e., before any other water was drawn  from the tap.  The second


                                130

-------
                                                   Tap Water
                                                   Analyses
Standards
0.005 mg/1
Cd. Pb, and Cu
Figure 57.  Typical voltairanograms of a" standard" solution and tap
            water analyses.
                                131

-------
  24  -
10 ppb lead

 5 min plate

10 ml  sample
  20 -
I 16
•rH

-------
               Table 19.  ELECTRODE DETERIORATION
Run

1

10

20

30

40
Cd ACd

33.0
6.0
27.0
4.0
23.0
5.0
18.0
4.0
14.0
Pb APb

44.0
1.0
43.0
1.0
42.0
0.5
41.5
0.5
41.0
Cu ACu

94.0
6.0
88.0
8.5
79.5
8.5
71.0
8.0
63.0
        Table 20.  ELECTRODE RESPONSE AFTER REGENERATION
                   (10 minute plating times; 10 ppb each
                   metal)

Start
24 hrs.
48 hrs.
Cd

23
24.5
27.0
Pb

22
25.5
27.0
Cu

50.5
54.0
55.0
sample was drawn from the same tap after the water had run contin-
uously for three minutes.  Samples were collected in acid-washed
polyethylene bottles provided by the laboratory.  For the analysis,
ten ml of sample was placed in the flow cell and acidified to
pH 2.5.  Plating time was five minutes.  Household systems in-
cluded in the study are described in Table 21.  Results are sum-
marized in Table 22 and plotted in Figures 59., 60, and 61.

Regardless of the sample or characteristics of the system from
which it was obtained, levels of Cd were insignificant (Figure 59).
None of the metals was found at a significant level in samples
from galvanized plumbing.  Relatively high levels of both Pb
and Cu were shown to accumulate with overnight residence in cop-
per plumbing systems, but these levels were quickly reduced with
flushing.  (Figures 60 and 61.)

This study should be repeated in cities with more aggressive
drinking water supplies.  Ann Arbor water has an unusually high
                               133

-------
      CD

      12
     •S
      3
      S
      O.
      O,
                                                                                        «o
            e

           •H



            O



            0)



            CO
TJ
 O
O
      1
      B



     I
 4)
 «4
• i^
en
                                                                                             M
                                                                                             3

SD  3
'S  .S
S  S
>  A
o

                                                                                                     -P
                                                                                                     0)
            Q)
            U
                                                                                                     O
                                                                                                     £1
                                                                                                     0)
                                                                                                     cn
           CTl
           in

           a)
                                                                                                     tn
                                                                                                     •H
                                                             qdd
                                                  134

-------
                                        I      I      I
                                                              immnniiiiiin
~o
 0
 0)
tfi
'£*
     o
     03

     13
     o
      in
      CM
                                                                                4)

                                                                                en
 a

I £
                                                                                          to
                                                                                          QJ
                                                                              CM
                                                                                   "•^
                                                                                          0)
                                                                                     rt
                                                                                     -P
                                                                                     (U
                                                                                     e

                                                                                     0)
                                                                                     o
                                                                                          O
                                                                                          A
                                                                                          0)
                                                                                          0
                                                                                          X
                                                                                          0)
                                                                                          •H
                                             IT)
                                                               CM
                                                    qdd
                                            135

-------
 
S-l
3
tn

w
i-H
(0
-p

i

OJ
0
(0
     1
     I
        in
        rt
o     in
CO     fM
                                •fe-\v
                                                                           
                                                                 I
                                                 00    (O
a)
en
^
O
S3
                                                                         vo


                                                                         Q)
                                                                         S-l
-H
fci
                                            qdd
                                            136

-------
                           Table 21. TRACE METALS SURVEY SITES
Site No.
Description
Approximate
Building Age
    (yrs)
Plumbing
   1         Single family home

   2         Single family home

   3         Single family home

   4         2nd story apartment in 2 unit dwelling

   5         2 story town house in new complex

   6         Single family home

   7         Commercial Building

   8         Single family home

   9         Single family home

   10        Single family home

   11         Single family home

   12        2nd floor apartment in new apartment complex

   13        Sub-basement in old University of Ml laboratory

   14        Basement in new University of Ml laboratory

   15        6th story in new University of Ml laboratory
                                                 10

                                                 13

                                                 17

                                                 35

                                                  2

                                                 49

                                                 10

                                                 20

                                                 12

                                                 20

                                                  1.5

                                                  1

                                                 30

                                                  2

                                                  2
               Galvanized iron

               Galvanized iron

               Galvanized iron

               Galvanized iron

               Copper

               Copper installed 2 years ago

               Copper

               Galvanized iron

               Copper

               Galvanized iron

               Copper

               Copper

               Galvanized iron-copper mixed

               Copper

               Copper
                                                137

-------
Table 22.  SUMMARY OF RAW DATA FROM HOUSEHOLD TRACE METALS SURVEY
GALVANIZED IRON PLUMBING
Site No.

1
2
3
4
8
10
Average
Cd ppb
O.R.*
0.16
0.11
0.12
0.06
0.5
0.29
0.2
3 +
0.08
0.08
0.08
0.05
0.11
0.04
0.07
Pb ppb
O.R.
2.68
2.84
1.24
1.4
1.35
1.84
1.9
3
1.75
1.46
1.15
1.2
0.5
0.48
1.1
Cu ppb
O.R.
0.64
2.78
0.35
0.22
1.65
1.58
1.2
3
0.42
0.39
0.45
0.14
0.54
0.26
0.4
COPPER PLUMBING
Site No.

5
6
7
9
11
12
13
14
15
Average
Cd p
O.R. i

0.02
—
0.25
—
0.01
—
0.26
—
0.06
pb
3

0.02
—
0.04
—
0.01
—
0.03
—
0.01
Pb ppb
O.R.
3.22
1.35
4.28
11.5
4.18
3.5
18.9
5.25
23.4
8.4
3
1.86
1.08
0.92
1.0
0.47
0.83
0.58
0.78
2.3
1.1
Cu ppb
O.R.
33.9
28.0
27.5
9.2
32.7
10.2
6.64
25.6
21.0
21.6
3
5.9
7.0
1.97
0.62
3.4
1.2
0.38
3.6
1.8
2.5
*O.R. = Overnight Residual
 3+ = 3   minutes later
                                   138

-------
pH (>_10.0 during the week of the study) .  A comparable  study with
tap water pH <_ seven is recommended.  Plastics piping systems
should be included in the study.
                                139

-------
                            SECTION V
                        MOBILE LABORATORY
General
In the final phase of the NSF water quality monitoring project,
the prototype monitor; i.e., all instrumentation and ancillary
equipment, was installed in a mobile laboratory and tested in
actual field operation.  Important design criteria included
ruggedness of instrument mountings and overall laboratory porta-
bility.  Ideally, the laboratory had to be easily and quickly
movable and capable of being installed virtually anywhere on a
potable water distribution network.


Physical Description

A heavy duty, intercity CMC delivery van with 4.9 meter long by
two meter high load space was selected to house the NSF/EPA
mobile water quality monitoring laboratory.  The van is designed
to carry heavy loads over long distances.  It rides on six wheels
(dual rear wheels) and includes heavy duty suspension to protect
the delicate instruments against extreme road hazards.  Its load
space is sufficiently large to provide adequate benchtop work
space, and sufficiently high to permit persons of average height
to work in comfort.  All cabinetry, electrical, and plumbing
installations were customized to NSF specifications by a local
general contractor.  Cabinetry was constructed from plastic
laminated heavy composition board, typical of that which is
found in household kitchens and benchtops, covered with "matte
white" formica.

The sink unit contains a closed cabinet and an electric heater
with thermostatically controlled blower.  An additional base-
board heater is provided on the opposite side, at the opposite
end of the van.  For cooling, a 12,000 BTU ARA mobile home air-
conditioner is mounted in the ceiling.

External power  (100 amps, 220v, single phase) is brought to the
van over a four conductor, 100 feet long by one inch diameter
cable which weighs slightly over 100 pounds.  When it is con-
nected to the van, the cable splits at the circuit breaker box
to two, llOv circuits.  The two main circuits are further divided
into nine smaller circuits for lights, airconditioner, left wall
outlets, right wall front and right wall rear outlets, heaters,

                               140

-------
hot water heater, computer, and  Schneider Robot Monitor systems
No overloads were observed during  six months of mobile labora-
tory operation.

NOTE:  A ptitnctpat Au.st.Qe.-type.  powe.fi tie.qu.tie.me.nt tA e.x.e,fite.d by
tke. kot wa.te.si kaa.te.ti.  Tkt& i.te.m tt> not c.onA-Lde,sie.d e.A
tn any ^utasie. psioje,c.t& oft tkt* k-ind.   One, ku.ndfie.d amp
womtd not be, sie.qu.tsLe.d 
-------
Figures 62 and 63 are schematic plan views of the equipped van
interior.  In Figure 62, "reagent reservoirs" contain, from left
to right, alkalinity baseline solution, alkalinity buffer, free
fluoride buffer, total fluoride buffer with masking agent, and
fluoride baseline solution.  Sensor flow cells and mixing coils
for free and total fluoride and alkalinity are mounted on the
platform to the left of the pump.

The flow cell, mixing coil, and run/by-pass switching values for
DASV are mounted on the H-type rack near the center in Figure 63.
The rotator, electronics module, and recorder for CCDT, and the
corrosion rate monitor are not fixed in place.  For travel,
these instruments are stored on thick pads on the floor.  A
photograph showing an overview of the interior arrangement of
the mobile laboratory is shown in Figure 64.


Computer System

All sensor systems in the mobile laboratory are controlled by an
on-board mini computer system which includes a Texas Instruments
digital computer, Model 960-A; Texas Instruments Silent 700/30
teleprinter; combination high speed paper tape reader perforator;
and a Computer Products wide range analog to digital  (A/D) con-
verter, Model RTP7480.  The computer contains 16K  (16,384 words)
of semiconductor memory, expandable internally to 32K and, with
the addition of an external chassis, to 64K.  An internal timer;
16 input/output digital switching board wired for interrupt; and
an internal communications register unit  (CRU) expansion chassis,
required for five to 20 peripheral connections, are options used
with the computer.  Seven CRU connections are used in the mobile
laboratory.

The gain of the wide range A/D converter is software programmable
and output resolution is 13 bits  (gain setting/4095) for full
scale input voltage ranges of +2, 5, 10 mv  ... +10.24v.  Input
from parametric systems to the converter, through Belden 8451
shielded twisted pairs plus ground cable, is made by connection
to multiplexer boards mounted in the converter.  Up to 16 eight-
channel multiplexer cards can be installed  internally - and 48
externally - to the converter chassis, for  a total of 512 analog
inputs.  Two multiplexer cards  (16 channels) are installed in
the mobile laboratory.

The high speed paper tape reader/perforator reads 300 and punches
75 characters per minute.  The Silent 700/30 teleprinter writes
on heat sensitive paper, operating at printing speeds up to 30
characters per second.  Its quiet operation makes it  ideally
suited for the mobile laboratory application.

The physical relationship of the Texas Instruments  (TI) 960-A
computer system to sensor equipment on the  mobile water quality


                                142

-------



















«
UI
»-
UI
X
>
2
*^
X
a









TOTAL
UI
UI
m











c

• c









RESIDUAL
CHLORIN
RESIDUAL
CHLORINE
w>
•c
o
at
UI
VI
UI
•c

UI
O
<
UI
K



^
X
a

>•
H-
ALKALIN



SCHNEIDER




ANALYZE!
ANALYZER











"2C
^
?c

C
•c
1 RECORDI
A






/

t 0
0 ^
• z
i i




a,
5
3
a.
..





I =
I <

VI
ui m
FLUORIC
RECORDI



ROBOT







f
I

<




(

f

(
(
..


MONITOR






























WHEEL







at
UI
^
<
ui
Z


IW
U
<
v>




p
<
<





«l
-^
UI
$

















                                       O
                                      -P
                                       a)
                                      rH
                                      •H
                                      X!
                                       o


                                      4-1
                                       0

                                       M
                                       o
                                      •H
                                       i-l
                                       0)
                                      -p
                                       c
                                      •H
                                       w

                                      .p
                                      X!
                                       tJ>
                                      •rH
                                       o
                                       o
                                      -H
                                      -p
                                       Hi
                                       n
                                      -p
                                       w
                                      •H


                                       O

                                      -P


                                       6
                                       0)

                                       u
                                      CO
                                      0)
                                      i-l
143

-------
ED

PE

PUN
HIGH S

PAPER

READER
COM
ANG
A/D

CONV
                       32
   u o
   u £
si
O w^
                          .
      o
      111
      0
                                3
§  i  i
                                       S
                                        .
                                                                 n
                                                                 o
                                                                 4J
                                                                 (0
                                                                 (-1
                                                                 O
                                                                 X!
                                                                 QJ
                                                                 H
                                                                 •H
                                                                 X)

                                                                 §
                          o
                          •H
                          n
                          0)
                          •4J
                          c
                          -H

                          
-------
                                           o
                                           •H
                                           M
                                           0)
                                           -P
                                           G
                                           •H
                                           O
                                           -P
                                           m
                                           H

                                           QJ
                                           X!
                                           O
                                           e

                                           4-1
                                           o
                                           •H

                                           M
                                           <1J

                                           O
                                           •H
                                           ^
                                           O
                                           Cn
                                           O
                                           -P
                                           O
                                           A
                                           CU
                                            Q)
                                           •H
                                           PM
145

-------
monitoring laboratory van is shown in overall view in Figure 65.
Digital input from the anodic stripping voltammetry equipment
is used to trigger the initiation of a data sampling cycle.

Within the data sampling cycle, appropriate digital outputs are
sent to the corrosion, alkalinity, and free and total fluoride
measuring equipment.  Independent programs operating from the
internal interval timer in the computer send digital output
signals to the sampler and control the flow through the nitrate
and hardness cells, located in the Schneider Robot Monitor.

In the data sampling cycle, the internal interval timer controls
the timing for reading the analog signals from each of the
different sensor instrumentation packages.  The data acquisition
program in the computer determines the gain to be used in the
analog-to-digital converter (ADC) in reading each analog input
channel.

Within the computer, all program operations are under the control
of the Texas Instruments supervisor program PAM  (Process Auto-
mation. Monitor).

Eight worker tasks have been installed to operate in the computer
under the control of PAM.  Interrelations between the worker
tasks, described as follows, are shown in Figure 66.

          1.  NSFC is a program containing utility routines,
              such as a decimal core dump routine.  It is
              built to facilitate adding additional options.
              After NSFC is called into execution (either
              by Control-X or NSFC in response to OP?), a
              two-character code is used to select the op-
              tion desired for execution.  NSFC must be
              assigned Task ID 30.
          2.  BELL is a simple program which rings the bell
              on the teleprinter repeatedly until the opera-
              tor halts the program.  Its purpose is to
              attract the operator's attention.
          3.  INIT initializes output storage area, checks
              PUN, starts flow control, periodic sampling,
              and data acquisition.  It is used for start-
              up when the system is initially loaded.  (By
              changing the operation of the ST option in
              NSFC to bid task 32-INIT- instead of task
              34-DAQ, the need to start system operation
              through job control can be eliminated.)
          4.  DAQ is the basic data acquisition program
              which reads the various analog input channels.
              It is initiated by an interrupt signal which
              indicates that DASV is ready to input data.
              During the next 13-minute interval, DAQ takes
              readings from each sensor system except alka-


                                146

-------
n
w
g
Q)
4-1
W
>1
W
                                                                                                      O
                                                                                                      in
                                                                                                     o
                                                                                                     •H
                                                                                                     -P
                                                                                                     a
                                                                                                     fl
                                                                                                     m

                                                                                                     o
                                                                                                     -P

                                                                                                     e
                                                                                                     0)
                                                                                                     4J
                                                                                                     W
                                                                                                     >!
                                                                                                     U)

                                                                                                     M
                                                                                                     0)
                                                                                                     cu
                                                                                                     g
                                                                                                     o
                                                                                                     u

                                                                                                     4-1
                                                                                                     o

                                                                                                     a,
                                                                                                     -H
                                                                                                     O
                                                                                                    -H
                                                                                                    -P
                                                                                                     fC
                                                                                                    rH
                                                                                                     Q)
                                                                                                    m
                                                                                                    \i>

                                                                                                     a)
                                                                                                     K
                                                                                                     3
                                                                                                     Cn
                                                                                                    -H
                                                  147

-------
 o
 <
O
z
                             z   ^
                        u   O   >-
                        u.   U   3
     Z
     O
«   -
O   5
                                           I  >
                                           5  «
                                           •*  IM
                                           00.
                                           O  =>

                                                                        z   *
                                                                                                         O4
                                                                                                         •H

                                                                                                         X!
                                                                                                         cn
•I
o
                                                                                                         O
                                                                                                         •H
                                                                                                         4J
                                                                                                         QJ
                                                                                                         (U
                                                                                                         S-l
                                                                                                         td
                                                                                                         O
                                                                                                         w

                                                                                                         e
                                                                                                         OJ
                                                                                                         tn

                                                                                                         s-i
                                                                                                         OJ
                                                                                                         -p
                                                                                                         =1
                                                                                                         cu
                                                                                                         e
                                                                                                         o
                                                                                                         u
                                                                                                          QJ
                                         148

-------
              Unity and corrosion rate, and switches
              the fluoride systems from baseline to
              sample.  During the second 13-minute
              period, it takes readings from all sensors,
              including alkalinity and corrosion rate.
              At the end of the full 26-minute cycle, it
              waits for another interrupt signal from
              DASV to begin the cycle again.
          5.  DATA obtains data from the buffer area of
              DAQ, processes the data as needed, and
              stores the results in the output storage
              area.
          6.  FLOW controls the flow of water to the
              nitrate and hardness cells in the Schneider
              Robot Monitor.
          7.  SMPL controls the sampler to collect hourly
              samples of the water flowing through the
              mobile laboratory.
          8.  PUN outputs data from the output storage
              area and clears that area for fresh data.
              Sufficient space is available in the allo-
              cated storage area to retain up to two
              hours of data.  DATA unsuspends PUN whenever
              more data is put into the output storage area.
              Currently, PUN outputs data and clears the
              area each time it is unsuspended; however,
              PUN can be modified to count the number of
              times it has been unsuspended since the
              last output of data, outputting data at
              greater intervals.

Flow charts of each of the worker tasks are shown in Figures 67
through 75.  The flow of information through the system is des-
cribed in Figure 76.

A detailed description of the operating system, DAQ, an operator
manual for the computer system on-board the NSF mobile laboratory,
and a paper entitled, "Computer Acquisition of Data from Paper
Tape Output of the TI960-A," are included with the "Operation
and Maintenance Manual" delivered to EPA with the mobile labora-
tory ( 47) .


Field Studies

To evaluate the performance of all on-board systems under actual
field conditions, the mobile laboratory was operated for one
week in Ann Arbor  (away from NSF), two months in the metropolitan
Chicago area, and one week in Detroit, Michigan during April,
May, June, and early July, 1973.  It returned to NSF in July
for installation and programming of the computer, then visited
four sites in Philadelphia, Pennsylvania during the month of


                               149

-------
f    START     A
CLEAI OUTPUT
DATA STORAGE
AREA
 BtO TASK 40
     (PUN )
WAIT A
BIT
 UNSUSPENO
 TASK 40
 • 10 TASK 39
     ISMPl)
 PRINT TIME
 MESSAGE
 BID TASK 38
     (FLOW)
 BID TASK 34
     IDAQ)
  T)ID TASK 40\   NO
  .EXIST ?
PRINT ERROR
MESSAGE
                                     Figure  67.   Flowchart  of  INIT,
                                   150

-------
(     START     )
It
w
1
WRITE
'MODE '
+
READ TWO
CHARACTERS
             KD
  WRITE

  'BAD ENTRY'
                                       WRITE TWO *

                                       SPACESIEAD FOUR

                                       CHARACTERS
                                      STARTING WITH
                                      FIRST NUMBER
                                      SEND DECIMAL
                                      ASCII OF TEN
                                      CORE LOCATIONS
                                      PER LINE UNTIL
                                      LAST LOCATION
                                      IS% SECOND
                                      NUMBER
                                      SEND NUMBER

                                      OF POINTS

                                      SENT
       Figure  68..  Flowchart of NSFC.
                          151

-------
V
•ID TASK 34
(OAQ )

1
C END )












SIT STOP FLAG
(WOtO 'M ' >
• IT 19
1
WilTI
'TIRMINATf

I
( »• )
V
•ID TASK 1
IPfBUO)

1
( •» )



Figure 68 (continued)
          152

-------
[    START     A

 III
     WRITE
     HEADER
                                         UNSUSPENDED
 SET PARAMETER
 COUNTER
GET LENGTH



4
^
NEXT
PARAMETER
A
 GET TIME AND
 VALUE, WRITE
 RECORD
 DECREMENT
 LENGTH
                             NO
                                                              I
                                                        SUSPEND
                                                        SELF
               Figure  69.   Flowchart  of  PUN
                             153

-------
   [     START    J
    SEND  "BELL"



    TO PRINTER
                  YES
CLEAR



STPFLG
    WAIT A BIT
Figure 70.   Flowchart  of BELL
                154

-------
         (    START     )
          SET DIGITAL
          OUTPUT BIT 3 : 1
           WASTE SOME
           TIME
           SET DIGITAL
           OUTPUT BIT 3 = 0
           WAIT  5 MINS.
           12  TIMES
Figure  71.  Flowchart  of SMPL.
                155

-------
                      START
                  SET  DIGITAL
                  OUTPUT BIT 4 = 1
                  WAIT 5 MINS.
                  3 TIMES
                  SET  DIGITAL
                  OUTPUT BIT 4:0
                  WAIT 5 MINS.
                  9 TIMES
Figure  72.   Flowchart of  FLOW.
                   156

-------
SET DO-3 OFF
SAMPLE, AVERAGE, AND STORE
AD15
SET DO-3 ON
SAMPLE, AVERAGE, AND STORE
AD 16
                                                     SET DO -? OFF
                            WAIT ONE MINUTE,CALL A
                                                     WAIT TWO MINUTES
SET DO -1 ON
                                                     WAIT FOUR MINUTES
                           WAIT 30 SECONDS
                           CALL SUBROUTINE A
                           GO TO START
              Figure  73.   Flowchart  of  DAQ.
                                   157

-------
                 SAMPLE, AVERAGE,AND STORE
                 CHANNEL ADI THRU AD 11
                 •ID DATA PROCESSING TASK
                           RETURN
Figure 74.   Flowchart  of Subroutine A  (DAQJt
                        158

-------
                 GET FLAGWORD, DATE/TIME , AND
                 ELEVEN READINGS FROM DAQ
SET POINTER
FOR NEXT
PARAMETER
                GET DATE /TIME AND FlUORIDE
                READINGS FROM DAO
                                                STORE DATE/TIME
                                                AND READING IN
                                                OUTPUT STORAGE
 SET POINTER
 FOR NEXT
 PARAMETER
            Figure  75.   Flowchart  of  DATA,
                              159

-------
GET DATE/TIME AND FIRST 1*
READINGS FROM DAQ
•UMPl 'GET DATA POINTER ' TO
• EGINNING OF CADMIUM WINDOW
LOOK
FOR MINIMUM
              DONE
           WITH 16
          READINGS IN
             UFFER ?
              DONE
           £ CADMIUM
            WINDOW
      GET 16 READINGS
      FROM DAO
              DONE
           WITH LEAD
            WINDOW
                7
                                                   OfT 16 READINGS
                                                   FROM DAQ
                                                            DON!
                                                        WITH COPPER
                                                          WINDOW
                                                   STORE DATE /TIME,
                                                   CADMIUM, LEAD,AND
                                                   COPPER IN OUTPUT STORAGE
                                                   GET DATE/TIME. CVCl I NO.
                                                   AND READING FROM
                                                   DAQ
                                                   STORE  DATE/TIME  AND
                                                   READING IN OUTPUT STORAGE
                                   Figure  75   (Continued)
                                      160

-------
GET DATE /TIME AND 16
READINGS FROM DAQ
                                                        FINI
                                                       O
                                                UNSUSPEND PUN
                          GET NEXT 16 READINGS
                          FROM DAQ
   AVERAGE LAST  32
   READINGS
READING :
"GROUND DROP" -
AVERAGE OF LAST
32
   STORE DATE/TIME AND
   READING IN OUTPUT STORAGE
    Figure  75  (Continued)
                             161

-------
  Z   a.  o.
  3   <  <
  0.0,1-
o
<   ^
O   vn
                                0   O
  O
  iu  O
  >-  U
                                    o
                                    z
                                    u
                                    O
                                                            fd
                                                           -P
  "»  O  uu

  <  u  «
                                    O
                                <
                                Q
r-


 0)
                                                            Cn
                                                           -H
                                  5
                            162

-------
September.  Acquisition of the computer  was  fully justified
during the Chicago assignments.   One man-day for every data-day
is a conservative estimate of the time required to transpose
parametric data from analog  strip chart  output to numbers which
could be analyzed.

The first stop was the Ann Arbor  municipal water treatment plant
where the principal objective was to gain proficiency with
getting the parametric systems on-line and calibrated, and
disassembling for the next move.   The van was parked outside,
adjacent to the filter building.   Although the visit was brief
and operators inexperienced, some meaningful data were collected,
and the troubleshooting experience was valuable.

The first field assignment away from Ann Arbor was in metropoli-
tan Chicago.  N. J. Davoust, Engineer, Chicago Division of Water
Purification, and T. E. Larson, Head, Chemistry Section, Illinois
State Water Survey, enthusiastically supported the project from
its inception and served as  members of the advisory committee.
The entire Chicago assignment was coordinated through N. J.
Davoust.

The trip from Ann Arbor to Chicago was made  without incident.
The van went directly to the first monitoring site, where it
was parked inside a four year old fire station in the northwest
part of the city.   (Engine Company 79 at 6424 N. Lehigh) , 10.5
miles from the Central Water Filtration  Plant.  All service lines
to the van were copper.  To  evaluate long-term performance char-
acteristics, especially sensor stability, the mobile laboratory
remained at this site for one month.

NOTE:  Ve.ta4.l.e.d ne.pon.tt> o £ data c.ot£e.c.tid at a.LL JLoc.at4.onA
pu.bt4.&he.d Ae.pa/iate.-ty and pfLOvtdtd to the. age.ncy th.tiou.Qh
the. ^4.e.td tn.4.pt> w
-------
The third stop was in the City of Des Plaines, adjacent to
northwest Chicago and O'Hare International Airport, within five
miles of the Lehigh  (first) location.  The van was located at
the Public Works Building  (1111 Campground Road), where distri-
buted water was normally a mixture of 75 percent purchased
Chicago water  (from the Central Filtration Plant) and 25 percent
Des Plaines treated well water.  The mix was altered to 1:1 to
demonstrate the ability of on-board systems in the mobile labora-
tory to detect subtle changes in the distribution system.  Their
responsiveness was reliably demonstrated.

The fourth and final monitoring site on the Chicago trip was a
nine year old fire station (Engine Company 80 at Doty Avenue and
127th Street) in the Calumet Harbor area, 41 miles from the
previous location in Des Plaines, and 6.1 miles from the South
Water Filtration Plant.  Water arrived at this location through
a one-mile section of six inch transmission line.  The fire
station was situated on the extreme dead end of the transmission
line.  The cast iron service line to the building was connected
to the mobile laboratory through a few feet of galvanized pipe.
Beginning at 1000 hours on June 29th, the main was flushed by
opening all connections in the fire station.  At 1115 hours,
changes in measured levels of turbidity, free and total residual
chlorine, temperature, and pH were appreciable, as shown in
Figure 77.  Relatively high levels of residual chlorine persisted
throughout the remaining period of observation.

Each of the last three sites in the Chicago area was visited for
just over a week.  Moving between them was accomplished with a
minimum of downtime.  Water and wastewater connections at a new
station were made by the laboratory operator; electric service
was provided by a city electrician.  With normal operation, all
systems were on-line and calibrated in little more than two hours
after arriving at a new location.

J. V. Radziul, Chief of Research and Development, Philadelphia
Water Department and member of the project advisory committee,
coordinated activities and planning for the Philadelphia trip.
During the month of September, the mobile laboratory visited
four sites in the City of Philadelphia, carefully selected to pro-
file water quality from the various sources and treatment facil-
ities providing water to Philadelphia.  The locations, identified
in Figure 78, include the Oak Lane Reservoir in the northern sec-
tion of the city  (site no. 1), Philadelphia International Airport
in the southern section of the city  (site no. 2), centrally lo-
cated City Hall  (site no.  3), and Gulf Oil Refinery, less than
one mile from  site no. 2,  but with a different  source of treated
water  (site no. 4).

Raw water for Philadelphia is entirely of surface origin but
drawn from two sources, the Schuylkill River and the Delaware
River.  Water  from the Delaware is treated at the Torresdale


                               164

-------
      2.2 -
      2.0 -
      1.8 -
      1.6 -
      1.4 -
  JTU 1.2-
      1.0 -
      0.80-
      0.60-
      0.40-
      0.20-
                                                    TURBIDITY
           • •  • •
                                           TOTAL RESIDUAL CHLORINE
0.55-
0.45- • •
mg/1 0.40-
0. 35-
0.30- _
0.25-

63.0- 4
62.5- •
61.5- • ^ • • *
61.0- • • •
60.5-

7.85-
7.70- • • • *
o' 1 ol i3 I o' ' o I ol 1 o 1 ol 1 o 1 o
O tl^O (DOOOOOOO
OOCOJJCM •<* VO 00 O CM Tl-
O -H r-l (8 i-( r-\ rH r-t CM CSJ CM
It. 5 June 29— »
H C

9 • •
• « 9



TEMPERATURE

•


pH


1 o ol I ol ol
o o o o
04 ^s* ^^ oo
o o o o
June 30 — >
Figure  77.   Data from main flushing at Chicago,  Calumet Harbor
                                165

-------
                                               TORRES DALC WflTt

                                               OUrCIM LOME WAT IT R.
                                               MIXED WITH roaaesoflLE
                                               w
                                        WATER SUPPLY
                                  MA(W TRANSMISSION LINES
                                    CITY OF PHILADELPHIA
Figure  78.   Relative locations of Philadelphia monitoring
             sites.
                             166

-------
Filtration Plant, third largest in the world (282 mgd), and de-
livered to most of the eastern half of the city.  Schuylkill
water is treated at Queen Lane (120 mgd) or Belmont  (78 mgd)
Filtration Plants, and distributed principally to western sec-
tions of the city.  Each of the treatment facilities is highly
automated.  Distribution is controlled by a "load control"
facility, a central point to which all data related to demand
are transmitted by microwave.  All water is treated by prechlor-
ination, presedimentation, alum-lime coagulation, rapid and slow
mixing, sedimentation, rapid sand filtration, chlorination, and
fluoridation.  In addition, Belmont water is ammoniated to im-
prove its taste and odor characteristics.  Finished water is
stored in three large reservoirs, Roxborough, Oak Lane (site
no. 1), and East Park, and elevated tanks at various locations
in the city.

The water at site no. 1 was treated at Torresdale and stored in
the reservoir.  Test water at the monitor was taken from the
reservoir or from the main filling the reservoir.  At the air-
port, site no. 2, water was from the Belmont filtration plant
exclusively.  Torresdale, or Queen Lane, or a mixture of the
two waters was monitored at City Hall.  At the Gulf Oil Refinery,
water was delivered principally from Queen Lane, with some mixing
from Torresdale possible.

During the first week in Philadelphia, little reliable data was
collected.  A record heat wave created virtually inoperable
conditions for the operator and many of the sensor systems.
Additional problems resulted from inadequate external grounding
of the laboratory.

A malfunction in the computer power supply caused repeated but
temporary "failures" throughout the Philadelphia assignment.
Numerous calls by TI service personnel resulted in little satis-
faction to the NSF staff.  In fact, a general comment regarding
failure of TI representatives to adequately diagnose and correct
frequent problems with the on-board computer system is in order.
A typical notation from the operator's log for site no. 3 cites
a malfunction in one of the A/D multiplexer cards which was not
replaced for three days after the TI field engineer was notified
of the problem.  In virtually every period of computer induced
downtime, the NSF operator or an NSF computer specialist - not
the TI field engineer - was responsible for diagnosing the
failure.  Although TI hardware offers many advantages for mobile
laboratory application, it is unlikely that this computer -system
would be selected for any future NSF project applications.
                               167

-------
                             SECTION  VI

                             REFERENCES

 1.   Environmental  Protection Agency.  Interim  Primary  Drinking
     Water Standards.   Federal Register.   £0(51), March 14,  1975.

 2.   Fair, G.  M., J.  C.  Geyer, and  D.  A.  Okun.  Water and  Waste-
     water Engineering.   John Wiley &  Sons,  Inc., New York,  1968.
     p.  17-31.

 3.   Standard  Methods for the Examination of Water  & Wastewater.
     13th Edition.   American Public Health Association, American
     Water Works  Association, and Water  Pollution Control  Feder-
     ation, New York,  1971.

 4.   McClelland,  N.  I.   Water Quality Monitoring  in Distribution
     Systems.   National Sanitation  Foundation,  Ann  Arbor,  Michigan,
     May 1971.   p.  169.

 5.   Turbidimeters.   Second  Revised Edition.  Hach  Chemical  Com-
     pany, Ames,  1973.

 6.   Quality Goals  for  Potable Water.  Journal  American Water
     Works Association,  September 1973.   p.  62.

 7.   Water Quality  Monitoring in Metropolitan Chicago:   Project
     Report.  National  Sanitation Foundation, Ann Arbor, Michigan,
     1973.

 8.   Water Quality  Monitoring in Detroit:  Data Presentation.
     National  Sanitation Foundation,  Ann Arbor, Michigan,  1973.

 9.   Water Quality  Monitoring in Metropolitan Philadelphia:   Proj-
     ect Report.  National Sanitation Foundation, Ann Arbor,
     Michigan, 1974.

10.   Langelier, M.  A.   The Analytical Control of  Anti-Corrosion
     Water Treatment.   Journal American  Water Works Association,
     28:1500,  1936.

11.   Larson, T. E.  and A. M. Buswell. Calcium  Carbonate Saturation
     Index and Alkalinity Interpretations.  Journal American Water
     Works Association, 34:1664, 1942.
                                 168

-------
12.   Ryzner,  J.  W.   A New Index for Determining the Amount of
     Calcium Carbonate Scale Formed by a Water.  Journal American
     Water Works Association, 36:472,  1944.

13.   Pytkowicz,  R.  M.  Rates of Inorganic Calcium Carbonate Nu-
     cleation.   Journal Geology,  73:196, 1965.

14.   Weyl, P. K.  Solution Kinetics of Calcite.  Journal Geology,
     66:163,  1958.

15.   Weyl, P. K.  The Carbonate Saturometer.   Journal Geology,
     69:32, 1961.

16.   McClanahan, M. A.  Mechanism of Cast Iron Corrosion Inhibi-
     tion by Calcium Carbonate Deposition in Water Distribution
     Systems.  Doctoral dissertation,  The University of Michigan,
     Ann Arbor,  1968.

17.   Schlicting, H.  Boundary Layer Theory.   McGraw-Hill Book Co.,
     1960.  p.  83.

18.   Levich,  V.  G.   Physicochemical Hydrodynamics.  Prentice Hall,
     Inc., 1962.

19.   Nernst,  W.  Z.   Physical Chemistry, 47:52,  1904.

20.   Riddiford,  A.  C.  Advances in Electrochemistry and Electro-
     chemical Engineering.  4:47, 1966.

21.   Napp, D. T.  Doctoral dissertation, University of Minne-
     sota, Minneapolis, 1967.

22.   Johnson, D. C.  Doctoral dissertation,  University of Minne-
     sota, Minneapolis, 1967.

23.   Enslow,  L.  H.   The Continuous Stability Indicator.  Water and
     Sewerage Works, 107, March 1939.

24.   Enslow,  L.  H.   The Continuous Stability Indicator and the
     Langelier Index.  Water and Sewerage Works, 283, July 1939.

25.   Durfor,  C.  N.  and E. Becker.  Public Water Supplies of the
     100 Largest Cities in the United States.  1962.  U.S. Geo-
     logical Survey Water Supply Paper 1812.   U.S. Government
     Printing Office, Washington, D.C., 1964.

26.   Zipkin,  I.  and F. J. McClure.  Fluoride Drinking Waters.
     U.S. Department of Education and Welfare Public Health
     Service Publication, 825:483, 1962.

27.   Hein, J. W., D. E. Gardner,  and G. B. Haydon.  Preliminary
     Investigations of the Effect of Sodium Monofluorophosphate


                               169

-------
     on Salivary Acid Production and Hydroxyapatite Solubility.
     Journal Dentist Research,  30:466,  1951.

28.   Cooley, W.  E.   Reaction of Tin (II)  and  Fluoride Ions with
     Etched Enamel.   Journal Dentist Research,  40:1199,  1961.

29.   Brudevold,  F.,  et al.   Uptake of Tin and Fluoride by Intact
     Enamel.  Journal American  Dentist Association, 53:159, 1956.

30.   Cooley, W.  E.   Applied Research in the Development of Anti-
     caries Dentifrices.   Journal Chemical Education, 47:177,  1970.

31.   Frant, M. S. and J.  W. Ross.  Electrode  for Sensing Fluoride
     Ion Activity in Solutions.  Science, 154:1553, 1966.

32.   Orion Instruction Manual,  Fluoride Ion Activity Electrode,
     Model 94-09.  Second Edition.  Orion Research Inc., Cambridge,
     Massachusetts,  1967.

33.   Bock, R. and F. Z. Strecker.  Direkte Electrometrische
     Bestimmung des  Fluorid-ions.  Analytical Chemistry (Germany),
     234:322, 1968.

34.   Harwood, J. E.   The  Use of an Ion Selective Electrode for
     Routine Fluoride Analysis  on Water Samples.  Journal Water
     Research, 3:273, 1969.

35.   Mesmer, R.  E.   Lanthanum Fluoride Electrode Response in
     Aqueous Chloride Media. Analytical Chemistry, 40:443, 1968.

36.   Rechnitz, G. A.  Ion Selective Electrodes.   Chemistry and
     Engineering News, 45 (25):146, 1967.

37.   Scrinivasan, K. and  G. A.  Rechnitz.   Activity Measurement
     with a Fluoride-Selective  Membrane Electrode.  Analytical
     Chemistry,  40:509, 1968.

38.   Ciavatta, L.  Hydrolysis of Fluoride Ion,  F~, in 3 M Na
     (CIO.*)" and 3 M K+(C1~, F~) Media.  Arkiv Kemi  (Sweden),
     21:129, 1963.

39.   Butler, J.  N.  Ionic  Equilibrium, A Mathematical Approach.
     Addison-Wesley Publishing  Co. Inc.,  1964.   Chapter 5-1,
     p. 115.

40.   Frant, M. S. and J.  W. Ross.  Use of a Total Ionic Strength
     Adjustment Buffer for Electrode Determination of Fluoride in
     Water Supplies.  Analytical Chemistry, 40:1169, 1968.

41.   Crosby, N.  T.,  A. L. Dennis, and J.  G. Stevens.  An Evaluation
     of Some Methods for  the Determination of Fluoride in Potable
     Waters and Other Aqueous Solution.  Analyst, 93:643, 1968.


                                170

-------
42.   Kelada,  N.  P.   Electrochemical Characterization of Free &
     Complexed Fluorides in Drinking Water and Effects of Aluminum
     & Iron on Fluoride Incorporation into Tooth Enamel.  Doctoral
     dissertation,  University of Michigan, Ann Arbor, 1972.

43.   Schwarzenback, G., R.  Gut,  and G.  Anderegg, 117.  Komplexone
     XXV.   Die Polarographische  untersuchung von austausch-
     gleichgewichten.   Neue daten der piltungskonstanten von metall
     komplexen der  Athylendiamin-tetraessigsaure und der 1,2-
     Diamin-cyclohexan-tetraessigsaure.  Helv. Chim. Acta (Germany),
     37:937,  1954.

44.   Lacroix, S.  Etude de  quelques Complexes et Composes peu
     Solubles des Ions.  Ann. Chim. (France), 4:5,  1949.

45.   Bertin-Batsch, C.   Etude par des Methodes variees de quelques
     Complexes Organiques de 1'Ion Ferrique.  Ann.  Chim (France),
     7:481, 1952.

46.   Peterson, T. L.,  D. 0. Brant, and K. H. Mancy.  Characteriza-
     tion of  Trace  Metals in Municipal Water Supplies by Thin
     Layer Anodic Stripping Voltammetry.   University of Michigan,
     Ann Arbor.   (Presented at 160th National Meeting of the
     American Chemical Society.   Chicago, September 1970.)

47.   Operation and  Maintenance Manual for Mobile Water Quality
     Monitoring Laboratory, National Sanitation Foundation,  Ann
     Arbor, Michigan,  1973.
                               171

-------
                            SECTION VII

                       PROJECT PUBLICATIONS

 1*   Committee  Report, AWWA  Corrosion  and  Stability Committee.
     "Water  Quality  Determination  in Distribution  Systems."   (Pre-
     sented  at  the 90th Annual Conference, American Water Works
     Association, Washington, D.C., 1970.  Journal American Water
     Works Association, 795, 1972.

 2.   Kelada,  N.  P.,  N. I. McClelland,  and  K.  H.  Mancy.   Character-
     ization of Fluorides in Water Supplies by  Selective Ion
     Electrodes.   (Presented at 160th  National  Meeting,  American
     Chemical Society, Chicago, Illinois,  1970.)

 3.   McClelland, N.  I. and K. H. Mancy.  Electrochemical Character-
     ization of Drinking Water.   (Presented at  the 141st Meeting,
     Electrochemical Society, Houston,  Texas, 1972.)

 4.   McClelland, N.  I. and K. H. Mancy.  Water  Quality Monitoring
     in Distribution Systems.   (Presented  at  the 92nd Annual  Con-
     ference, American Water Works Association,  Chicago, Illinois,
     1972.)

 5.   Schimpff,  W. K., N. I.  McClelland,  K. H. Mancy, and H. E.
     Allen.   An Automated Differential Anodic Stripping  Technique
     for Monitoring  Trace Metals in Distribution Systems.   (Pre-
     sented  at  the 92nd Annual Conference, American Water Works
     Association, Chicago, Illinois, 1972.)
 6.   McClelland, N.  I.  Monitoring of  Municipal Drinking Waters:
     Process to Consumer.   (Presented  at the  15th Eastern Analyti-
     cal Symposium,  New York, New  York,  1973.)
 7.   McClelland, N.  I.  Monitoring of  Drinking  Water Quality  in
     Distribution Systems.   (Presented at  the 2nd Joint  Conference
     on Sensing of Environmental Pollutants,  Washington, D.C.,
     1973.)

 8.   McClelland, N.  I., J. R. Adams, R.  R. Wood, and K.  H.  Mancy.
     Application of  Monitoring Technology  (for  Assuring) Drinking
     Water Quality.   (Presented at the 167th  National Meeting,
     American Chemical Society, Los Angeles,  California, 1974.)

 9.   Adams,  J.  R., R. R. Wood, and N.  I. McClelland.  Drinking
     Water Quality:   Application of Monitoring  Technology.   (Pre-
     sented  at  the  6th Central Regional Meeting, American Chemical
     Society, Detroit, Michigan,  1974.)

10.   McClelland, N.  I. and K. H. Mancy.  Water  Quality  Surveillance
     in Distribution Systems.   (Presented  at  94th Annual Conference,

                               172

-------
     American Water Works Association,  Boston, Massachusetts, 1974.)

11.   Symons,  J.  M., M.  C. Gardels,  K.  H.  Mancy, and N.  I.  McClelland,
     Continuous  Distribution System Monitoring to Study the Effects
     of Hardness and Other Water Quality  Parameters.  (Presented
     at 94th  Annual Conference,  American  Water Works Association,
     Boston,  Massachusetts, 1974.)

12.   McClelland, N. I.   Experiences in Monitoring Residual Chlorine.
     (Presented  at  94th Annual Conference,  American Water  Works
     Association, Boston, Massachusetts,  1974.)
                                173

-------
                          SECTION VIII
                           APPENDICES

                                                           Page
A.  Technical Advisory Committee                           175

B.  Preparation of Reagents                                176

C.  Special Alkalinity Case Study                          181
                              174

-------
                           APPENDIX A

                  TECHNICAL ADVISORY COMMITTEE
Harry A. Faber,  Chairman (1969-1972)
88 Main Street
Whitesboro, NY 13492

N. J. Davoust, Chairman (1973)
Engineer of Water Purification
Central Water Filtration Plant
1000 E. Ohio Street
Chicago, IL 60611

T. E. Larson
Head, Chemistry Section
Illinois State Water Survey
Box 232
Urbana, IL 61801

Blucher A. Poole
Consultant
5839 Brockton Drive
Indianapolis, IN 46220

Joseph V.  Radziul
Chief, Research & Development
Water Department
1270 Municipal Services Bldg.
Philadelphia, PA 19107

Verdun Randolph
Associate Director
Consumer Health Protection
Illinois Department of Public Health
535 West Jefferson Street
Springfield, IL 62761
Gordon G. Robeck
Director
Water Supply Research Division
U.S. Environmental Protection
  Agency
Municipal Environmental
  Research Laboratory
Cincinnati, OH 45268

Elroy F. Spitzer
Director
American Water Works Assoc.
Research Foundation
6666 W. Quincy Avenue
Denver, CO 80235

James M. Symons
Chief, Physical and Chemical
  Removal Branch
Water Supply Research Division
U.S. Environmental Protection
  Agency
Municipal Environmental
  Research Laboratory
Cincinnati, OH 45268

T. L. VanderVelde
Deputy Chief
Bureau of Environmental and
  Occupational Health
Michigan Department of Public
  Health
Lansing, MI 48914

Richard Woodward
Vice President
Camp, Dresser, & McKee
One Center Plaza
Boston, MA 02108
                                175

-------
                           APPENDIX B

                     PREPARATION OF REAGENTS
SRM-HOUSED SYSTEMS
Standard solutions are required for weekly calibration of SRM-
housed sensor systems.  No other reagents are used for routine
operation of these systems.  Standard solutions are prepared as
follows:
HARDNESS
b
         Calcium chloride stock solution

         Dissolve 11.099 g CaCl2 (anhydrous) in approximately
         one liter of distilled water.  This solution should
         be standardized against EDTA and adjusted to contain
         lO^'M CaCl2 concentration.

         Standard solutions

         To prepare standard solution, dilute 13.0 ml stock
         solution to 1,000 ml with distilled water, using a
         volummetric flask.  This solution contains 10 (~ ^M
         Ca2+ activity.  Dilute 1.075 ml stock solution to
         1,000 ml.  This solution contains 10 (-^M Ca2+ activity.
         Separate each of the standards into two 500 ml wide -
         mouth jars.  Label one jar of each standard "wash,"
         and the other, "standard."
NITRATE
     a.  Potassium nitrate stock solution
         Dissolve 1.6308 g KN03 in exactly 1.0 liter of dis-
         tilled water.

     b.  Standard solutions

         To prepare standard solutions, dilute 104.2 ml stock
         solution to 1,000 ml with distilled water, using a
         volummetric flask.  This solution contains 100 mg/1
         NO3~ activity.  Dilute 10.2 ml stock solution to
         1,000 ml.  This solution contains 10 mg/1 N0s~ activity.
                               176

-------
         Dilute 1.0 ml stock solution to 1,000 ml.  This
         solution contains 1.0 mg/1 NO3~ activity.  Separate
         each into "wash" and "standard" solutions.
CHLORIDE

     a.  Sodium chloride stock solution

         Dissolve 1.6482 g NaCl (dried at 140°C) in chloride
         free water and dilute to 1.0 liter.  This solution
         contains 1.0 mg Cl~/ml.

     b.  Standard solutions

         To prepare standard solutions, dilute 100 ml stock
         solution to 1,000 ml with chloride free water, using
         a volummetric flask.  This solution contains 100 mg/1
         Cl .   Dilute 10 ml stock solution to 1,000 ml.  This
         solution contains 10 mg/1 Cl~.  Separate each into
         "wash" and "standard" solutions.


CONDUCTIVITY

     a.  Standard reference potassium chloride solution

         Dissolve 745.6 mg anhydrous KCl in freshly boiled
         distilled-deionized water and dilute to 1,000 ml
         at 25°C.  This solution has a specific conductance
         of 1,413 umhos/cm at 25°C.  This solution is used
         to calibrate the laboratory conductivity meter, used
         to prepare standards of 1,000 and 100 ymhos/cm.

     b.  Standard solutions

         Add KCl solution to distilled water at 25°C to pre-
         pare standard solutions of 1,000 and 100 ymhos/cm,
         respectively.  Separate each into "wash" and "standard"
         solutions.
pH
     a.  pH 6.86 standard buffer
         (0.025 M potassium dihydrogen phosphate +0.025 M
         disodium hydrogen phosphate)
         Dissolve 1.179 g KHaPO^ and 4.30 g NazHPO^ in dis-
         tilled water and dilute to 1,000 ml.  This solution
         is pH 6.86 at 25°C.

     b.  pH 9.18 standard buffer
         (0.01 M borax)
                              177

-------
         Dissolve 3.80 g Na2B^O?'10 H20 in distilled water
         and dilute to 1,000 ml.  This solution is pH 9.18
         at 25°C.
RESIDUAL CHLORINE

     a.  Buffer solution
         (Sodium acetate-acetic acid buffer)

         Fill the buffer dispensing bottle approximately one-
         fourth full with distilled water.  Add 855 ml glacial
         HN03.  Add 460 g CH3COONa«3 H2O with vigorous mixing
         and stir till completely dissolved.  Fill bottle to
         top with distilled water.  This solution is used with
         free- and total- residual chlorine analyzers.

     b.  Potassium iodide solution

         Dissolve 20 g KI in distilled water and dilute to
         approximately two liters in the KI dispensing bottle
         (total chlorine analyzer only).


ALKALINITY

     a.  Buffer solution
         (Potassium hydrogen phthalate buffer)

         Dissolve 91.80 g potassium hydrogen phthalate
         (HOCOCelUCOOK) in approximately eight liters of dis-
         tilled water.  Dilute 13.4 ml concentrated hydro-
         chloric acid  (HC1) to 800 ml with distilled water.
         Mix the two solutions, add 0.0635 g sodium thiosulfata
         (Na2S203), and bring up to nine liters.  Adjust the
         pH to 3.1 using concentrated HC1.

     b.  Baseline solution
         (1)  Sodium carbonate stock solution
              Dissolve 0.477 g Na2C03 in C02  free distilled
              water and dilute to 1.0 liter.
         (2)  To prepare working baseline solution, dilute
              200 ml stock solution to 9.0 liters with C02
              free distilled water.  This solution is equi-
              valent to 10 mg/1 CaC03.

     c.  Standard solutions

         (1)  Sodium carbonate stock solution
              Dissolve 1.0590 g Na2C03 in C02  free distilled
              water and dilute to 1.0 liter.  Store in tightly
              stoppered glass bottle.  Make up fresh stock
              solution monthly.


                              178

-------
         (2)   To prepare standard solutions, dilute 1.0, 2.0,
              3.0, ... 9.0 ml stock solution to 100 ml each.
              These solutions contain 10, 20, 30, ... 90 mg/1
              equivalent CaCO3, respectively.  Fresh standard
              solutions must be prepared for each calibration
              series.
FLUORIDES

     a.  Buffer solutions

         (1)   Free fluoride
              Dissolve 850 g sodium nitrate (NaN03) in approxi-
              mately nine liters distilled water.  Add 57 ml
              glacial acetic acid (CH3COOH) and mix.  Bring
              volume up to 10.0 liters.  Adjust to pH 5.2 with
              concentrated sodium hydroxide (NaOH) solution.
         (2)   Total fluoride
              Dissolve 595 g NaN03 and 588 g sodium citrate
              (Na3C6H507*2 H20) in approximately nine liters
              distilled water.  Add 115 ml CH3COOH and mix.
              Bring volume up to 10.0 liters.   Adjust to
              pH 5.2 with NaOH.

     b.  Baseline solution

         Dilute 20.0 ml 100 mg/1 F~ Harleco standard to
         20.0 liters.

     c.  Standard solutions

         Dilute 2.0, 5.0, and 10.0 ml 100 mg/1 F~ Harleco
         standard to 1,000 ml each.   These solutions contain
         0.2, 0.5, and 1.0 mg/1 F~~,  respectively.  Store in
         plastic bottles.


TRACE METALS ANALYZER (DASV for Cd,  Pb, and Cu)

     a.  pH adjustment solution

         The strength of the pH adjustment solution will vary
         with the pH of the test water.  Add concentrated
         nitric acid (HNOs) d>iopu)-Lt> e. to distilled water to the
         predetermined level at which, when added to test
         water in the appropriate ratio, will result in a
         final pH of 2.5.

     b.  Standard solution

         (1)   Place 1.0 ml 1,000 mg/1 Cd Harleco standard,
              1.0 ml 1,000 mg/1 Pb Harleco standard, and
              1.0 ml 1,000 mg/1 Cu Harleco standard in a
              100 ml volummetric flask.  Bring to volume

                              179

-------
         with distilled water and mix.   This solution
         contains 10 mg/1 each of Cd, Pb, and Cu.
         Transfer to plastic bottle for storage.
    (2)   Dilute 1.0 ml of stock solution to 2.0 liters
         in the quartz reservoir, using test water as
         the diluent.  (The known addition technique
         is used for standardization in trace analysis.)
         This solution contains 0.05 mg each of Cd, Pb,
         and Cu.

c.  Mercury plating solution
    Dissolve one drop (approximately 0.05 ml) of reagent
    grade mercury (Hg) in concentrated HNOs  (as little
    as possible).  When solution is complete, dilute to
    approximately 100 ml with distilled water.   Store
    in a glass stoppered bottle.  Inject 2.0 ml of this
    solution into the flow cell filled with distilled
    water to plate  (rejuvenate) the electrodes.
                          180

-------
                           APPENDIX C

                  SPECIAL ALKALINITY CASE STUDY
Experimental Procedure

          1.  Prepare 0.050 F KHP and 0.019 F HCl
              standard phthalate buffer.
          2.  Prepare alkalinity standards from
              Na2CO3 in distilled water.
          3.  Prepare additional alkalinity standards
              using distilled with known concentra-
              tions of chlorine, prepared from gaseous
              chlorine.
          4.  Mix five parts of each alkalinity
              standard with one part buffer and
              determine resultant pH.
          5.  Plot H  versus alkalinity, developing
              a series of curves from the standards
              with varying concentrations of chlorine.


Theoretical Results

In the presence of free residual chlorine, initial formalities
for the buffer are:

          KHP  =  0.050
          HCl  =  0.019

and for the test water:

          Na2C03  =  6/5 C
             C12  =  6/5 E

Final formalities at buffer:sample = 1:5 are:

          HHP  =  8.33(10)~3
          HCl  =  3.17(10)~3
       Na2C03  =  C
          C12  =  E

Known relationships are the same as those expressed in Equations
75 through 79 and 81 through 84.  Equation 80 is rewritten as:

          [Cl~]  = 3.17(10)~3 +  [HOC1] + [OC1~]                 (126)


                                181

-------
and Equation  85:

           [H+]  4-  [Na+]  4- [K+] =  [OH~]  4-  [Cl"]  4

                               4-  [HC03~]  4- 2[C032~] 4-  [OC1~]   (127)
                  [Na+]  + [K+] =  [OH~]  +  [Cl ]  4- [HP ] + 2[P2~]
Other known  reactions include:


          K5  =  [HQC1] [H+] [el"] = 4.2(1(J)-*                     (128)

                    [C12]
          K6  =     ]        = 3.4(10)8                          (129)
                  [HOC1]
              =  [H ] [OC1            ~8
and

           2[C12]  4- [HOC1] 4-  [OCl"] 4-  [Cl"]= 2E 4- 3.17(10)~3   (130)

Assuming  [C032~]  = [P2~] =  [OH~] =  [C12]  = [OCl"] = 0, and  that
only the  equilibria expressed in Equations 131 and 132 occur  for
free residual  chlorine:

           C12  4-  H20 -> HOC1 4- H+ 4- Cl"                          (131)

           HOC1 4- H20 -* H30+ 4- OCl"                             (132)

then derivations shown in Equations  86 through 98 and  those in
Equations 133  and 134 can be expressed:

           [Cl"]  = 3.17(10)~3 4- E                               (133)

           [HOC1]  = E                                           (134)

Also, from Equations 80 and 86:

                   FH+1 (                    \
           [C12]  = -kg-J- ^3.17(10)  3 +  [HOC1] > [HOC1]            (135)


From Equations 79, 81, 82, 85, 95, and 98:


           [H+] 4- 2C 4- 8.33(10)~3 = 3.17(10)~3 4- E 4- -
                                                      [H'] 4-1. 3 (10)

                                       4.6(10)~3C
                                      •«• l" •» . A  ^-/^*S\  3
                                                                (
                                     [HT]+4.6(10)

           L.C.D.  = [H+]2 + 5.9(10)~3[H+]  + 5.98(10)~6         (137)
                                182

-------
                           1.083 (10) ~5        4.6 (10) 3C
                    —  ij ~ 	.	 —  	.	
IH+]3  + 2CIH+]2 +  5.16(10)^ [H+]2  - E[H+]2  + 5.9(10)~3  [H+]2


      +1.18(10)~2C IH+]  + 3.044 (10)~5 IH+]  - 5.9 (10)"3 E[H+]


      +5.98(10)"6 IH+]  +  1.196(10)~5C +  3.086(10)~8  - 5.98(10)~6E


      -1.083(10)~5 IH+] - 4.982(10)~8  -  4 . 6 (10) ~3 C IH+]


      -5.98 (lO)""6  C  =  0                                             (139)



                   2 + 2C - E | [H+]2  +  J2.559(10)~5  + 7.2(10)~3C
                                            5.98(10)~6C  -  5.98(10)~6E} =  0


assuming  [HT]3 =  0.


          -2.559(10)"5
            +3 =                                                   (140)
           +  5.9(10)~3E                /J2.559(10)~5+7.2(10)~3C


                                         - 5.9(10)~3E(2


                                   +  /  - 4 { 1.106(10)"2+ 2C - E)


                                         {-1.896 (10)~8 +  5.98(10)~6C
                                         -5.98(10)~6E
                                  2
                       2.212(10)    + 4C - 2E
                                                                    (141)
           -2.559(10) 5 - 7.2(10)  3C
          +  5.9(10)"3E                / 4(10)"6C2+2.556(10)"7C


                                    +  /  +1.489(10)~9 +1.089(10)~5E2


    ,                                 \  -1.132(10)"7E  -1.32(10)"5CE
   j"*"!  =  	I	

                       2.212(10)~2  + 4C  - 2E                      (142)



                                    183

-------
The family of curves obtained by application of Equation 142 are
shown in Table 23 and Figure 79.

Although alkalinity has been shown to be linear with respect to
pH when no free chlorine is present, it is important to note
that the alkalinity monitor reads potential in mv, not pH.  To
convert [H ]  in the theoretical results to potential, the Nernst
equation is applied:

          E = E° + 0.059 log [H+]                            (143)

or

          E = E° -  (0.059)pH                                 (144)

where:

          E° = a function of reference electrode potential
               and E is the measured potential

E° cannot be predicted theoretically.  If it is assumed to be
equal to zero, then:

          E = -(0.059)pH                                     (145)

Data from Table 23 are expressed as potentials in Table 24 and
Figure 80.  As shown in Figure  80, experimental and theoretical
results for alkalinity solutions with no free chlorine residual
vary only about three to seven  mv at alkalinity equal to 30 to
70 mg/1 as CaC03.  This slight  variation is likely attributable
to experimental error and instrument limitations.

In Table 25, data from Table 23 for C12 = 1.0 and 2.0(10)~5F
 (approximately 0.7 and 1.4 mg/1 free residual chlorine) are ex-
pressed as potentials and compared with theoretical results for
C12 = 0.  The greatest variation in potential associated with a
specific level of alkalinity and a change from C12 = 0 to
C12 = 1.4 mg/1 is 0.3 mv.  This small change in potential cannot
be observed with the alkalinity monitor.


Experimental Results

Experimental data obtained by NSF closely approximated the theore-
tical results.  NaOCl was substituted for gaseous C12 in the pro-
cedure described on page 181.   Alkalinity standards included 30,
40, 50, 60, 70, and 80 mg/1 as  CaCO3, prepared from Na2C03•
Changes in volume with addition of NaOCl to the standards were
negligible and disregarded.   (The greatest addition was 0.1 ml
to 60 ml of standard, or 0.167  percent change in volume.)  The
result of simple regression applied  to the data is shown  in
Figure  81.


                                184

-------
                                                                           o
                                                                           o
                                                                           o
                                                                           00
                                                                           o
                                                                           CN
                                                                                    •P
                                                                                    •H
                                                                    o
                                                                   4-1


                                                                    CM

                                                                   H
                                                                   U
                                                               O
                                                               O  I
                                                                nj  
                                                                                    O
                                                                                    m
                                                                                    0)
LO
 •

in
o
 •
m
                                m
                                                                 m
                                                                      ro
                                                      ro
                     OT
                              185

-------
                                                      o
                                                      oo
                                                      o
                                                      in
                                                      o
                                                      O
                                                      n
                                                           O
                                                           U
                                                           
               (d
              rH
               0)
 u
•H
4->
 OJ
 !-l
 O
 (D
X!
                                                                  rd  tn
                                                                  td -H
                                                                  4-1 4-)
                                                                  G  C
                                                                  CU  (D
                                                                  6 4J
                                                                 -H  O
                                                                  !H  O<
                                                                  0)
                                                                     m
                                                                 o
                                                                 00
     O
     O
     CM
O
cn
(AUI)
          186

-------
                                      co
                                                                          0s-
                                                                                   a      E

                                                                                  •     r-,.
                                                                           s       s
                                                                          o-.      o
                                                                                                      c
                                                                                                      o
                                                                                                     •H
                                                                                                  C -P
                                                                                                  o  3
                                                                                                  •H rH
                                                                                                  -P  O
                                                                                                  3  Cfi
                                                                                                  O
                                                                                                  W
                                                                                                      C
                                                                                                     •H
                                                                                                     m
 a
ro
01


 0)
rH
X!
 fO
EH
                        in     in
                                *)     o
                        U;     ii ;     ii;
                                               e      8      £
                                                                                                  •H  O
                                                                                                      O
                                                                                                   OJ  OJ
                                                                                                  rH  fd
                                                                                                  U  !3
                                                                                                   o  o
                        OOC=OOO
                                                                          O     O     O
                                                                           e       e       s

                                                                          CO     0--     O
fd  fc
e  e
M  S-l
O  O
pr i  Pr i
                                                   187

-------
o
 •
o
                                 CO U
                 CN
                 I
00
 I
                                                       C
                                                       O
                                                      -H
                                                      -P
                                            o
                                            IT)
                                            o
                                            ro
                                                 
-------
Table 24.  EXPERIMENTAL AND THEORETICAL  RELATIONSHIPS
           BETWEEN ALKALINITY AND  POTENTIALS
Alkalinity
(mg/1 as CaC03)

24
36
48
60
72
84
Potentials (mv)
Experimental

-194.9
-197.5
-200.1
-202.6
-205.3
-207.6
Theoretical

-193.1
-194.4
-195.7
-197.1
-198.4
-199.8
   Table 25.  THEORETICAL RELATIONSHIP BETWEEN  ALKALINITY AND
          POTENTIAL FOR SOLUTIONS WITH AND WITHOUT FREE
                          RESIDUAL C12
Alkalinity
(mg/1 as CaC03)

24
36
48
50
72
84
Potential (mv)
No C12

-193.1
-194.4
-195.7
-197.1
-198.4
-199.8
0.7 mg/1

-193.0
-194.3
-195.6
-196.9
-198.3
-199.7
1.4 mg/1

-192.9
-194.2
-195.5
-196.8
-198.2
-199.5
                             189

-------
Assuming that the NaOCl was made stoichiometrically according to
the reaction:
2NaOH
C1
2Na
Cl
OC1
                                            H2O
(146)
i.e., there was no excess NaOH in the NaOCl, the net effect of
substituting NaOCl for gaseous C12 is a change in sign of the
quantity representing C12 in Equation 146; i.e., a cation (Na )
is added in lieu of an anion (Cl ).  The predicted slope is
approximately -0.18 mv per mg/1 free C12.  This value is slightly
larger than the experimental value obtained in the NSF laboratory
(.07/.18), but it is 28 times smaller than the value obtained
by EPA.   (See Table 26.)
          Table 26.  CHANGES IN POTENTIAL FROM ADDITION
                     OF 1.0 mg/1 FREE RESIDUAL C12


Slope =(as mv/mg/1)
Experimental
NSF

-0.07
EPA

-5
Theoretical

-0.18
Despite the theoretical relationships which were established and
closely related data acquired experimentally by NSF, anomalous
results continued to occur with Cincinnati tap water monitoring.
Although it does not seem likely that the observed effect can be
attributed only to the presence of free chlorine residual,
chlorine is very likely involved, perhaps synergistically with
a complex organic material.

Additional experiments were undertaken to determine the effect
of adding a reducing agent to the alkalinity system.  Thiosulfate
(five mg/1) added to the buffer solution did not interfere with
potentiometric measurement of alkalinity and it eliminated the
interference effect.   (Data are shown in Table 27 and Figure 82.)
As a result, it is now recommended as standard procedure that
S20a2  be added to pH 3.1 phthalate buffer solution used in all
applications of the potentiometric alkalinity monitoring system.
                               190

-------
 M
 Q)
-P
 ttf  O
-P  (0
    U
-P
•H
 rd
                                                               DO
 0)
-P
 fll
 H
-P
•H
                                       O
                                       in
                                             O
                                             O
                                             (0
                                             u

                                             w
                                             to
                                             tn
                                             g
                          -P
                          •H
                          C
                          -H
                          rH
                                                               o
                                                               n
     O
     rH
     (N
IT)
O
Cvl
O
O
CM
                                                    en
                                                    (U
                                                   -H
                                                    M
                                                    (D
                                                    O
                                                   -H
                                                   -P
 O

 >i
4J
•H
 C
-H
rH
 fO
y,
rH
 (0

 W
 3
 en
 H
 (D
                                                                          •rH
                                                                          -P
                                                                           c
                                                                           0)
                                                                          -P
                                                                           o
                                                                          CN
                                                                          oo

                                                                           CD
                                                                           M
               (AUI)
                            191

-------
LLJ


<
CO
O
LL

O
O
LU

LL
LL

LU
CM















"*>
j=

CO
'•*-«
c
£
O
Q.












CN
CN
co
O
CM
to
-f-
«—

CM
O

CD
Q.
E
CO
to
CM
O
+
CD
Q.
E
CO
to
CN
CO
O
CM
to
+

CD
Q.
E
&
CD
Q.
E
c3
"co
>8
lo
CO VI
S/ CO
—
~^
O>
_§






in
| CO 1 | 1
8








O
1 g 1 1 1
CN







CD O CN
CM O3 | | 00
r- 0 O
CM CM CN


CM O CM O If)
CN 00 CM 00 r^
«- o o en o
CM CN CM ^ CM
O O O O % O5
CN 
                                                               JN

                                                               O
                                                                     o>
                                                                     E
CM

 CO

O
 CM
CO
                                                                           to
                                                                           c
                                                                           O
                                                                          '5
                                                                          Tl
                                                                          -a
                                                                           CD
                                                                           i_
                                                                           CO
                                                                           O.
                                                                           CD
                                                                           C
                                                                           O
                                                                          O
                                      192

-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO. 2.
EPA-600/2-77-074
..TITLE AND SUBTITLE
WATER QUALITY MONITORING IN
DISTRIBUTION SYSTEMS
/. AUTHOR(S)
Nina I. McClelland
K . H . Mancy
P. PERFORMING ORGANIZATION NAME AND ADDRESS
National Sanitation Foundation
Ann Arbor, Michigan 48106
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory — Cin. ,OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
3. RECIPIENT'S ACCESSION-NO.
5. REPORT DATE
March 1977 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
ICB047;ROAP 21AQB ;TASK011
11. CONTRACT/GRANT NO.
68-03-0043
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
16. ABSTRACT

 A mobile laboratory with  18  integrated, computer controlled  parametric
 systems for monitoring  potable water quality in distribution systems
 was developed and  field evaluated at ten locations in four United
 States cities:  Chicago,  Illinois;  Ann Arbor and Detroit, Michigan;
 and Philadelphia,  Pennsylvania.   Temperature, conductivity,  pH,
 chloride, dissolved oxygen,  free and total residual chlorine,  turbidity
 corrosion rate, free and  total fluorides, alkalinity, hardness,
 nitrate, copper, cadmium,  lead,  and calcium carbonate deposition rate
 are measured using commercially available and newly developed  sensor
 systems.
7. KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Water quality
Chemical analysis
Monitors
Data acquisition
Distribution systems
Potable Water
8. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b. IDENTIFIERS/OPEN ENDED TERMS

19. SECURITY CLASS (This Report)
UNCLASSIFIED
20. SECURITY CLASS (This page)
UNCLASSIFIED
c. COSATI Field/Group
7 B
13 B
21. NO. OF PAGES
207
22. PRICE
PA'Form 2220-1 (9-73)
                                     193
                                                      *US GOVERNMENT PRINTING OFFICE 1977-757-056/5576

-------
  HAY -. 1  1977
          . ,,'!  V
0) in ~-~ •-» .^
O- .
  8 I 8- A
    ?• 
                                             o
                                             o
                                                   CD  0
                                                   o  c
                                                   o  3.
                                                   ?r  -y
                                                      \
                                                      O

                                                      0)
                                                      2.
                                                      (0
                                                                 o ^
                                                                 CD -J3
                                                   Ooo
                                                                 o O
                                                           II-
                                                           ^  ?L
                                                           w  	
                                                              -+>
                                                           o  q
                                                           ?  B
                                                           o'  &

                                                           sf

                                                           %  
                                                     Q
                                                     m
                                                     Z
                                                     n
                                                        X 2
                                                        £  2

                                                        2  3
                                                        -(  f>
                                                        o  j-

                                                        *  •£
                                                        lo
                                                        D m
                                                        s  q

                                                        Is
                                                        •D  Z
                                                        1  >
                                                        3  o
                                                        2.  m
                                                           z
                                                           o
c
(fl

m
Z
<
33  -0
o  o
z
2
m
                                                        Q
                                                   m -(  m

                                                   ^ >  >
                                                   > r  z
                                                   co TJ  D
                                                   S ^  -n
                                                   01 O  m
                                                     H  m
                                                     m  (/)
                                                     O
                                                                 TJ
                                                                 >

-------