PB81-172280
Reaction Products from the
Chlorination of Seawater
Rosenstiel School of Marine and Atmospheric
Science, Miami, FL
Prepared for
Environmental Research Lab.
Gulf Breeze, FL
Mar 81
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
EPA-
600/4-81-010
-------
PB81-172280
EPA 600/4-81-010
March 1981
REACTION PRODUCTS FROM THE CHLORINATION OF SEAWATER
BY
James H. Carpenter
Carroll A. Smith
Rodney
University of Miami
Rosenstiel School of Marine and Atmospheric Science
4600 Rickenbacker Causeway
Miami, Florida 33149
Final Report
Grant No. R 803893
Project Officer
Dr. William P- Davis
Bears Bluff Field Station
U.S. Environmental Protection Agency
Gulf Breeze, Florida 32561
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
GULF BREEZE, FLORIDA 32561
REPRODUCED 8V
NATIONAL TECHNICAL
INFORMATION SERVICE
U.S. DEPARTMENT OF COMMERCE
SPRINGFIELD, VA. 22161
-------
NOTICE
THIS DOCUMENT HAS BEEN REPRODUCED
FROM THE BEST COPY FURNISHED US BY
THE SPONSORING AGENCY. ALTHOUGH IT
IS RECOGNIZED THAT CERTAIN PORTIONS
ARE ILLEGIBLE, IT IS BEING RELEASED
IN THE INTEREST OF MAKING AVAILABLE
AS MUCH INFORMATION AS POSSIBLE.
-------
ERL,GB 0100
TECHNICAL REPORT DATA
(Please read Instructions an the reverse before completing)
REPORT NO.
;pA-600/4-81-010
3. REC
TITLE AND SUBTITLE
Reaction Products from the Chlorination of Seawater
5. REPORT DATE
MARCH 1981 TSSUTNG DATE,
S. PERFORMING ORGANIZATION COOS
7. AUTHOR(S)
J. H. Carpenter, C. A.. Smith and R. G. Zika
3. PERFORMING ORGANIZATION
UM RSMAS 80006
IT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
University of Miami
Rosenstiel School of Marine and Atmospheric Science
4600 Rickenbacker Causeway
Miami, Florida 33149
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
R 803893
12. SPONSORING AGENCY NAME AND ADDRESS
Envrionmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Gulf Breeze, FL 32561
13. TYPE OF REPORT AND PERIOD COVERED
Final- 7/15/75 to 7/14/80
14. SPONSORING AGENCY CODE
EPA/600/4
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Current methods underestimate the residual oxidants in chlorinated seawater by as
much as 70% depending upon the details of the procedures.
Chlorination of seawater in the presence of light produces bromate ions which can
influence standard analytical procedures and represent an unknown factor in estuarine
and coastal waters. Toxicity of bromate ion and persistence in coastal waters has not
been'determined.
The copper complexing capacity of Biscayne Bay, Florida, water was reduced with
the addition of chlorine. Analysis was by anodic stripping voltammetry on water samples
after successive additions of copper sulfate solutions. Chlorination of seawater may
produce toxicity and growth reduction through the indirect mechanism of copper release
and/or reduced binding capacity.
Laboratory Chlorination of water from the intake of the Port Everglades, Florida,
power plant produces bromoform levels comparable to that found in the plant.
Chloroform extracts of chlorinated Biscayne Bay water are found to contain
halogenated compounds which are new and different, and which pose unusual analytical
problems. Studies using GC/EC, GC/MS, HPLC, 1E NMR, differential pulse polarography
and other techniques on natural extracts and synthesized compounds are reported.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATl Field/Croup
Chlorination
Seawater
Halogenated organics
Power plant cooling
Copper complexing
Bromate formation
07/C
06/F
'8. DISTRIBUTION STATEMENT
/
Release unlimited
19. SECURITY CLASS (This Report)
None
21. NO. OF PAGES
52
20. SECURITY CLASS (This page>
None
12. PRICE
?*"* 2220-1 (R.v. 4-771 PREVIOUS e
-------
CONTENTS
PAGE
Foreword . . i
Abstract . . ii
List of Figures iii
List of Tables . v
Acknowledgements vi
1. Introduction 11
2. Conclusions 2
3. Recommendations 3
4. Analytical Procedures for the Determination of Residual
Oxidants 4
5. Sunlight-Induced Bromate Formation in Chlorinated
Seawater 11
6. Chlorination and the Copper Complexing Capacity of
Seawater 17
7. A Power Plant Study: Chlorination at the Port Everglades,
Florida Power Plant 23
8. Lipophylic Halogenated Compounds: The Liquid/Liquid
Extractables 26
8. a Introduction 26
8.b Hexane and chloroform extractions 27
S.b.l ECD-capillary column results 27
8.b.2 GC/MS results 27
8.c Preparation and identification of unknowns found in
chloroform extracts 37
8.c.l Synthesis of unknown compounds 37
8.C.2 Elemental analysis 39
8.c.3 NMR spectra 3Q
8.C.4 Fluorescence and electrochemical analyses of
chlorinated proline solutions ^Q
8.C.5 Possible artifacts ^
8.C.6 Limited search for molecular ions 45
8.d Electrochemical analyses of chlorinated baywater . . .46
8.e High performance liquid chromatography (HPLC)
on chlorinated amino acid solutions ^g
8.f Summation ^g
References . 5Q
-------
FOREWORD
The protection of our estuarine and coastal areas from damage caused by
toxic organic pollutants requires that regulations restricting the
introduction of these compounds into the environment be formulated on a sound
scientific basis. Accurate information describing dose-response relationships
for organisms and ecosystems under varying conditions is required. The
Environmental Research Laboratory, Gulf Breeze, contributes to this
information through research programs aimed at determining
the effects of toxic organic pollutants on individual species and
communities of organisms
- the effects of toxic organics on ecosystem processes and components;
the significance of chemical biocide reaction products in the
estuarine and marine environments.
Chemical treatment of natural waters, in particular the use of chlorine as
a biocide s modifies the chemistry of these waters in ways that are not fully
understood. The research described in this report examined both inorganic and
organic reaction products from the chlorination of seawater using a variety of
ana-lytrcarlappr-oachesr Conventional methods for the determination of
"residual chlorine" were found to underestimate levels in seawater by as much
as 70%, depending upon the details of the procedure. It was found that the
chlorination of seawater in the presence of light produces substantial
quantities of bromate ions which influence standard analytical procedures and
represents an unknown toxicity factor in estuarine and coastal waters.
Chlorination of Biscayne Bay (Florida) water was found to reduce the copper-
complexing capacity which raises the question of whether chlorination may also
create toxic effects through the indirect mechanism of copper release or
reduced organic binding capacity for ionic copper in addition to direct
oxidation and halogen addition reactions. Bromoform was found to be the
dominant haloform produced by chlorination of seawater and bromoform was found
in the cooling water discharge of the one power plant effluent examined during
chlorination. Solvent extracts of chlorinated Biscayne Bay water were found to
contain halogenated compounds, primarily brominated, which were resolvable by
gas chromatography and detectable by mass spectrometry. Some of these
compounds appear to be new and probably are a significant part of "chlorine-
produced oxidants" which can be expected to be biologically active. These and
other aspects of.the determination of reaction products of the chlorination of
seawater are discussed in the body of this report.
Hefrrylfc? Enos, Director
Environmental Research Laboratory
Gulf Breeze, Florida
-------
ABSTRACT
A general study of the reaction products from the chlorination of seawater
is reported. The results include the following:
Some analytical methods in widespread current use underestimate the
residual oxidants in chlorinated seawater by as much as 70% depending upon the
detail of the procedures.
The chlorination of seawater in the presence of light produces substantial
quantities of bromate ions which can influence standard analytical procedures
and represents an unknown factor in estuarine and coastal waters. The toxicity
of bromate ion and the possibility of bromate as a persistent source of
brominated compounds in coastal waters needs to be determined.
The copper complexing capacity of Biscayne Bay, Florida water was found to
be substantially reduced with the addition of chlorine. Analysis was made by
anodic stripping voltammetry on water samples after successive additions of
copper sulfate solution. The chlorination of seawater may, therefore, produce
toxicity and growth reduction through the indirect mechanism of copper release
and/or reduced binding capacity.
Laboratory ch-lorination of water from the intake of the Port Everglades,
Florida power plant produces bromoform levels comparable to that found in the
plant discharge. These results are in contrast to results reported in the
literature for a power plant on the Patuxent estuary in Maryland, so that
bromoform production appears to be site-specific.
Chloroform extracts of chlorinated Biscayne Bay water are found to contain
halogenated compounds which are new and different, and. which pose unusual
analytical problems. Studies using GC/ECD, GC/MS, HPLC, H NMR, differential
pulsed polarography and other techniques on natural extracts and synthesized
compounds are reported.
11
-------
LIST OF FIGURES
Page
Variations of triiodide ion absorbance with time after addition of
bromine 7
2 Variation of triiodide ion absorbance with time after addition of ~~"
iodine, potassium iodate, chlorine and potassium bromate . . 8
3 Differential pulse polarographic verification of sunlight-induced
bromate production in chlorinated seawater 14
4 Disappearance with time of residual oxidants with concomitant
appearance of bromate in chlorinated seawater 15
5 Anodic stripping current variation with copper added to Biscayne Bay
water showing copper complexing capacity of 12 ppb . . . .18
6 Anodic stripping current variation with copper added to chlorinated
Biscayne Bay water 19
7 Anodic stripping current variation with copper added to Biscayne Bay
water showing copper complexing of 6 ppb 21
8 Anodic stripping current variation with copper added to chlorinated
sample of Biscayne Bay water, showing a loss of complexing capacity 22
9 GLC chromatograms of hexane and chloroform extracts of chlorinated
and unchlorinated Biscayne Bay water 29
10 Total ion chromatogram of chloroform extract of chlorinated Biscayne
Bay water 33
11 Mass spectra of major peaks shown in Figure 10 34
12 Mass spectra of peak at 5.9 minutes 35
13 Mass spectra of peak at 4.3 minutes 36
14 Proton NMR spectrum of unknown compound having m/e 69,149,151 mass
spectrum 41
15 The oxidation of 5 mM proline in pH 8.1 borate buffered chloride /
iii
-------
42
bromide solution
16 Differential pulsed polarogram of chlroinated and unchlorinated ^
Biscayne Bay water
IV
-------
LIST OF TABLES
Page
Apparent residual oxidant concentrations in distilled water and
seawater with different analysis procedures for two different
chlorine additions 10
2 Bromoform in Port Everglades (Florida) power plant intake and
discharge water 24
3 Laboratory chlorination of Port Everglades (Florida) power plant
intake and discharge waters 25
4 Tabulated GC/MS chromatograms of chloroform extracts of chlorinated
Biscayne Bay water 28
5 Elemental analysis of brominated unknown 40
6 Chlorinated and unchlorinated HPLC separations of dansyl derivatives 49
-------
ACKNOWLEDGEMENTS
Dr. William P. Davis (Bears Bluff Field Station) provided stimulating
discussion and suggestions throughout the project.
Ms. Cynthia A. Moore capably executed much of the experimental work.
Dr. Donald L. Macaldy made major contributions to the work on bromates.
Ms. Mary Jo Spencer and Dr. Carol AuBuchon performed the work on copper
complexation and assisted generally in voltammetric techniques.
Support from the Department of Energy also contributed to this work.
VI
-------
SECTION 1
INTRODUCTION
The toxicant of choice for the control of fouling organisms in power plant
coolants and the disinfection of sewage has been chlorine for many years. Most
power plants rely upon the continuous or periodic addition of chlorine added as
gas or hypochlorite to maintain efficient heat transfer. We estimate that over
100,000 tons of chlorine are consumed annually in a manner that constitutes a
probable input to saline water resources. The saline waters that receive these
effluents are usually productive near shore estuarine waters that are used
extensively by man for recreation and for food.
Chlorine is a desirable pesticide in many respects. Application can be
precise, it (or its reaction products) is toxic to many of the organisms that
man seeks to control, and it has a limited persistence. It is this very
desirable property of short residence time that causes us concern, however.
Unlike many pesticides, chlorine is highly reactive in the broadest sense. The
variety of possible reaction products is immense and almost entirely unexplored
in saline waters, especially for organic reactions. In the research reported
herein, we first examined the analytical procedures available for determining
"residual oxidants" in seawater, then looked for and found the- formation of
bromate in sunlight, but not in dark. We then found the most abundant
"volatile" product (i.e., removable by gas purge) was bromoforin->. Subsequently
we examined "lipophylic" reaction products using gas chromatography with a
variety of detectors including mass spectrometry.
Administratively, this is a "final report," however, in terms of our
understanding of the use of chlorine as a pesticide in saline waters, it is an
initial contribution to the work that remains to be done as described herein.
-------
SECTION 2
CONCLUSIONS
. Analytical methods for the determination of "residual chlorine" (more
correctly termed "chlorine produced oxidants") developed for freshwater use
cannot be applied to saline waters without critical examination. A technique
is described that is shown to be applicable to seawater.
. Bromate ion results when seawater is chlorinated in the presence of
sunlight. Photochemical reactions, both inorganic and organic, can be expected
to play a significant role in chlorination reactions.
. Chlorination of Biscayne Bay: Florida, waters reduces the capability of
that water to bind free copper, measured electrochemically. It is probable
that the reduction in complexing capacity is due to general oxidation of
organic constituents in the water.
. Observations at several coastal plants are needed for a meaningful
assessment of the haloform discharge rate that are related to laboratory
chlorination studies before prediction of power plant discharges can be made.
. Chloroform-soluble brominated compounds of moderate volatility and
molecular weight have been observed in extracts of chlorinated Biscayne Bay
water. These compounds, which pose unusual analytical problems, are believed
to be an organic fraction of "residual oxidants" that participate in iodometric
titration.
-------
SECTION 3
RECOMMENDATIONS
Determine and observe the uptake of compounds found to result from the
chlorination of seawater using selected organisms in EPA controlled-expos_ure
facilities.
Develop techniques for determining selected series of haloamines in
seawater at trace levels and examine chlorinated coastal waters for these
compounds using HPLC with fluorescence detection at sub-picomolar levels.
Utilize polarography (e.g., differential pulse) to determine electro-
reducible compounds in natural saline waters before and after chlorination in
conjunction with HPLC/fluorescence measurements.
Study the formation, decomposition and reactivity kinetics of haloamines
utilizing sophisticated stop-flow reaction techniques. Equipment for these
studies is available at no capital cost to EPA.
Study the reactivity of haloamines with other substrates in living
organisms in vitro and vivo. Mutagenesis is of particular concern and can be
explored using the Ames test. Coupling reactions with RNA would be
particularly interesting and could be sought using C-14 tagged amines.
-------
SECTION 4
ANALYTICAL PROCEDURES FOR THE DETERMINATION OF RESIDUAL OXIDANTS
With the increasing recognition of the responses of aquatic organisms to
low levels of "residual chlorine" (1, 2), measurement of the residual oxidants
in the waters discharged from wastewater treatment plants and electricity-
generating plants has become important in order that the environmental impact
of the discharges can be properly assessed and regulated. Operation of
electricity-generating plants frequently involves the use of chlorine as an
antifouling agent in the cooling water system, and the large number of plants
that have been built at estuarine and coastal sites during the past decade has
led to much greater input of chlorinated waters. The wastewater collection
systems of many coastal communities contain some seawater as a result of
infiltration with brackish ground waters, and chlorine added as a disinfectant
during treatment reacts with the seawater constituents during treatment and in
receiving water. We find that the analytical methods in widespread current use
underestimate the residual oxidants in chlorinated seawater by as much as 70%,
depending on the details of the procedures (3).
The addition of chlorine to waters containing sea salts lead to reaction
with the natural bromide ion (65 mg/L in ocean water) to produce hypobromous
acid and hypobromite ion (4). If ammonia is present, a mixture of
monobromamine and monochloramine may be formed (5). In addition, reaction with
organic compounds may produce a variety of brominated substances. Thus, the
determination of "residual chlorine" actually corresponds to the estimation of
the sum of this complex mixture and is better termed "residual oxidant deter-
mination. "
The toxicity of chlorinated waters has been reported in terms of the
combined residual chlorine concentration (1), and the results of iodometric
amperometric titration measurements appear to be the most closely correlated
with biologically active chlorine residuals (6). Since responses by fish have
been found at the low concentrations of 0.001-0.01 mg/L, the high sensitivity
of amperometric titration equiment with large electrodes has been attractive.
The high sensitivity is attained by vigorous agitation of the sample with the
danger that volatile halogen compounds may be lost, so that rapid titration has
been recommended. The basic procedure (7) consists of adjusting the pH of the
sample to 4 with acetate buffer, adding KI solution (final concentration 0.001-
0.003 M depending on sample size), and rapidly titrating the liberated iodine
with either sodium thiosulfate or phenylarsine oxide solutions. Similarly, the
direct solid electrode amperometry (8) involves the continuous addition of an
acetate buffer solution containing KI to the sample stream and the monitoring
4
-------
of the current produced by electrochemical reduction of the resulting iodine,
with the refinement of using coulometrically generated iodine for periodic
calibration. These procedures involving reaction of KI with the residual
oxidants during a brief reaction time do not respond to all the residual
oxidants in chlorinated seawater.
EXPERIMENTAL
The determination of residual oxidants were carried out using reagent
solutions and procedures as described in "Standard Methods" (7), except that
the titration of iodine was followed by photometrically using the apparatus and
procedure outlined by Carpenter (9). Stock solutions of bromine was prepared
by dilution of distilled water saturated with liquid bromine.
The triiodide ion concentration in experimental soltuions was monitored
at 350 nm with a Beckman Model 24 spectrophotometer with a recorder.
RESULTS AND DISCUSSION
The results of titrations for residual oxidants in chlorinated seawater
showed that Gulf Stream water filtered thorugh 0.22-ji Fluoropore filters had a
large apparent chlorine demand, even though such waters have a very low organic
carbon content. In addition, the end point of the titrations was not
persistent, and triiodide ion was slowly generated after the intial end point
has been reached, with 18-24 h required for cessation of additional appearance
of triiodide -ion. For examplej for a chlorine dosage that produced an apparent
residual oxidant concentration in distilled water of 4.0 x 10 N in two
separate trials. The pH 4 buffer and KI results (final solution 0.0024 M KI)
were added in less than 1 min after the chlorine dose to minimize reaction with
seawater organics or decomposition. The residual oxidants reacted completely
with KI in 1 min or less in solutions of pH 2 and 0.024 M KI in distilled water,
and the test seawater showed a residual oxidant concentration of 3.6 x 10 N
under these conditions.
The high "chlorine demand" of the Gulf Stream seawater appears to be due
to the rapid formation of chemical species that react slowly with 0.0024 M KI
at pH 4.
The addition of chlorine to seawater results in the rapid production of
hypobromous acid and hypobromite ion. Based on observation of the 330-nm peak
of .hypobromite ion, the reaction is complete in less than 1 min. It seemed
probable that the slow reacting species was formed from the bromine rather than
the chlorine. This possibility was confirmed by use of the simple system of KI
solutions in distilled water at various pH values to which an aqueous bromine
solution was added and the absorbance to 350 nm (the triiodide peak was
monitored with time). The rate of appearance of the triiodide ion depended on
the pH of the solutions and KI concentration. For the "Standard Methods" (7)
conditions of pH 4 and 0.0024 M KI, 70% of the total potential triiodide ion
concentrations appeared in less than 20 s, and further production was very
slow. The slow formation of triiodide ion observed photometrically is
-------
analogous to the fading end point in the iodometric titrations. The inference
is that the formation of the slow reacting species is not peculiar to seawater
The initial formation of tiiodide ion was less (30-50%) if the stock KI and
buffer solutions were added to brominated distilled water, as in the case of
the procedures for analysis of residual oxidants, but was difficult to
reproduce with precision. We used the bromination of the diluted KI solution
to illustrate the formation and kinetics of the species that reacts slowly with
iodide ion.
The rate of production of triiodide ion by the species formed when bromine
was added depended strongly on the pH of the solutions (Figure 1). Slow
reaction at pH 4 was found even in the presence of a tenfold higher
concentration of KI than is normally used in the amperometric procedures for
residual chlorine.
We sought to identify the slow reacting species by observing the triiodide
ion concentration variation with time after the addition of iodine, chlorine,
iodate, and bromate to 0.025 M KI solutions at various pH values (Figure 2).
Addition of iodine produced a triiodide ion absorbance nearly instantaneously
that did not vary with time or pH, showing that hydrolysis or disproportiona-
tion reactions were not responsible for the variations shown in Figure 1.
Similarly, the addition of chlorine caused an absorbance that did not vary with
time or pH, which supports the notion that the slow appearance of iodine in
chlorinated seawater is not caused by reactions involving chlorine.
An obvious species that could cause the observed behavior is bromate ion,
since the rate of reaction between iodide and bromate is strongly dependent on
acidity (10), and the formation of bromate from bromide by hypochlorite is
favored by a high concentration of chloride (11). As shown in Figure 2,
bromate reacts with 0.24 M KI only slowly at Ph 2 and does not appear to be the
species that causes the results shown in Figure 1. Furthermore, we were unable
to detect bromate in chlorinated seawater polarographically, unless the
solutions were exposed to sunlight (12)-
Addition of iodate to the 0.24 M KI solutions produced patterns of
triiodide ion appearance with time at various pH values (Figure 2) that are
similar to the results hown in Figure 1, ranging from rapid at pH 2_t_o very slow
at pH 5. We infer that hypobromous acid or bromine (1.4 x 10 N) reacts
rapidly with iodide (0.024 M) to produce a mixture of iodine and iodate;
subsequently, the iodate reacts with the excess iodide to produce additional
iodine at rates that depend on the pH.
Confirmation of the formation of iodate from iodide by added bromine at
pH 4 was found from differential pulse polarograms. Differential pulse
polarography was used in order that the iodate wave could be resolved on top of
the large current due to the reduction of the iodine that is also present as a
result of the reaction of the iodide with the added bromine.
-------
1.0
.8
CD
0 6
c.o
CO
JD.4
<
.2
pH 2.0
pH 3.5
pH 4.0
pH 5.0
Time (min)
8
Figure 1. Variation of triiodide ion absorbance (350 nm, 5 cm path)
with time after addition of bromine (final solution 14 yN) to 0.024
M KI distilled water solutions adjusted to various pH values with
sulfuric acid.
-------
.8
c .6
o
.2
KIO
.8
S.6
o
2-4
_o
.2-
CI
- pH 2.0
pH 3.5
pH 4.0
pH 5.0
NaBrO
g
6 80 2
Time (min)
6
8
Figure 2. Variation of triiodide ion absorbance (350 nm, 5 cm path) with
time after addition of either iodine, potassium iodate, chlorine, or
potassium bromate (final solutions 14 uN) to 0.024 M KI distilled
water solutions adjusted to various pH values with sulfuric acid.
Variation with added iodate similar to variation with added bromine
in Figure 1.
-------
COMPARISON OF ALTERNATE PROCEDURES
These results indicate that the amperometric standard method badly
underestimates the residual oxidants in chlorinated seawater because the
resulting bromine oxidizes part of the iodide to iodate. In our experiments
the residual oxidant was primarily hypdbromous acid because Gulf Stream
seawater contains very low concentrations of ammonia and organic matter. We
considered the ferrous ammonium sulfate titrimetric method with N, N-diethyl-
p-phenylenediamine (DPD) end point indicator (13) as a possible aternate
procedure for these unpolluted waters. Application of this method to Gulf
Stream seawater gave only 80-85% recovery of added chlorine. The reasons for
this discrepancy have not been identified. The DPD ferrous titrimetric method
does not appear to be suitable for low (0.1-0.01 ppm) concentrations of
residual oxidants because the visual end point is not sharp in these dilute
solutions.
Another possible procedure would be a modification of the iodometric
procedure at lower pH and higher potassium iodide concentrations, so that
iodate would react rapidly. Greater acidity and iodide ion concentration
should be limited to conditions that do not produce significant air oxidation
of the iodide, and pH 2 with ca 0.03 M KI appears to be a realistic limit. The
high KI concentration has the additional virtue or reducing the volatilization
of the iodine by the formation of the triiodide ion complex, but has the
drawback of reducing the amperometric response since the electrode senses
primarily the free iodine (14). In the photometric titration procedure that we
used, the triiodide ion concetration is measured, and the high KI concentration
is desirable.
The results of titrations for residual oxidants in chlorinated seawater,
by use of variations of the iodometric procedure, are shown in Table 1. The pH
was adjusted and the KI solutions were added within 1 min after the addition of
the chlorine solution to minimize the decomposition of the resulting oxidants.
Use of pH 4 and low KI concentration produced values that are substantial
underestimates, as delineated above. The chlorine disappearance in seawater
has been described as involving an initial rapid decline followed by a much
slower decline (15), and our results suggest that the apparent rapid initial
decline may be partially an artifact of the analytical method.
The use of greater acidity and KI produced values that correspond to a
loss of consumption of chlorine in the Gulf Stream seawater of roughly 1 to
2 yeq/L, which is possibly a reasonable result. However, the use of the back
titration procedure (6, p. 382), in which the sample was added to an excess of
phenylarsine oxide solution that had been mixed with the pH 4 buffer and the KI
solution and the excess phenylarsine oxide titrated with a standardized iodine
solution, gave results corresponding to 100% recovery of the oxidizing capacity
of the added chlorine. The reverse tiration has been recommended for
wastewaters to avoid reaction of the liberated iodine with the organic material
in such solutions. A possible explanation for the difference between the
direct titration results and the back titration results would be the reaction
of the iodine with organic matter during the, 10 min or so required for the
photometric titration that was used. If so, this effect would be greater in
-------
estuarine and coastal waters that contain much more organic matter than the
Gulf Stream seawater.
The quantitative recovery of the added chlorine oxidizing capacity with
the back titration procedure suggests this procedure is suitable for residual
oxidant determinations in chlorinated seawater. The procedure has the
following advantages:
A low KI concentration may be used without the danger of
iodine volatilization, and the response of the amperometric
electrodes is not reduced, as it would be if a high KI
concentration were required.
The pH 4 acetate buffer is convenient for adjusting the
pH of samples.
The high pH and low KI concentration should reduce the
possible effects of interference from air oxidation and
ferric, manganic, or nitrite ions.
TABLE 1. APPARENT RESIDUAL OXIDANT CONCENTRATIONS IN DISTILLED WATER
AND SEAWATER WITH DIFFERENT ANALYSIS PROCEDURES
FOR TWO DIFFERENT CHLORINE ADDITIONS3
Procedure
a
b
c
d
a
b
c
d
Residual oxidants,
M
21.7
6.7
18.7
21.4
40.6
21.2
38.8
41.3
Recovery
&f
/o
31
86
99
...
52
95
102
a(a) Distilled water, direct iodometric at pH 2 in 0.024 M KI; (b) seawater,
direct iodometric at pH 4 in 0.0024 M KI; (c) seawater, direct iodometric at pH
in 0.024 M KI; and (d) seawater; back titration with standardized iodine
solution after excess phenylarsine oxide solution added at pH 4 and 0.0024 M
KI.
The large errors that we find with current procedures make it difficult to
evaluate and compare various toxicity studies involving chlorine since the
actual exposure levels probably have been underestimated substantially.
Simple correction for the errors does not appear to be practical because the
magnitude of the errors depends on the particulars in the analyses, such as
whether pH 4 or 3.5 was used and the rate of titration. A need for careful
evaluation of the analytical procedures seems obvious, particularly for water
containing higher concentrations of ammonia and organic compounds than Gulf
10
-------
Stream water.
SECTION 5
SUNLIGHT-INDUCED BROMATE FORMATION IN CHLORINATED SEAWATER
Chlorine and its compounds have been used for water disinfection and as
general aqueous biocides in increasing quantities since the turn of the
century. The popularity of these materials stems partly from the remarkable
apparent tolerance of mammals to them (16) at concentrations that produce
mortality of organisms ranging from bacteria to fish; that is, it kills them,
not us. Recent estimates (17) indicate that more than 100,000 tons of chlorine
are used annually for the partial disinfection of effluents from wastewater
treatment plants, and such may be expected to increase substantially as the
secondary treatment systems mandated by Congress in Public Law 92-500 begin
operation. An additional major use of these compounds is as antifouling agents
in the cooling waters of electric generating plants. Somewhat more chlorine is
used for this purpose than for wastewater treatment, based on a cooling water
flow of 300,000 cubic feet per second (8400 m /sec) (18) and a dose of 0.5 mg of
C12 per liter.
The release of chlorinated waters is producing effects that are slowly
being better documented as a result of continuing research. Summaries of
current knowledge (19) show avoidance behavior and reproductive failure in many
freshwater invertebrates and fish at chlorine concentrations of 0.003 to
0.005 mg/liter. Federal and state regulations have been used on measurements
of "residual chlorine" for both control of wastewater treatment (in the state
of Virginia, chlorine is added until the concentration is effluent is
0.2 mg/liter) and effluent limitations on power plants. Considering the strong
sensitivity of aquatic organisms to "residual chlorine" and the present levels
of chlorine use, substantial damage to aquatic resources may occur. For
example, the present releases of chlorine to Chesapeake Bay and its tributaries
would sterilize the whole system if there were not environmental degradation of
the added chlorine. However, transformation of chlorine to persistent, but
less acutely toxic, compounds may be hypothesized to produce slow changes in
the abundance and diversity of aquatic species in such situations.
Degradation is operationally defined as the disappearance of the
analytical signal for "residual chlorine." As pointed out by Eppley e_t al.
(15), different analytical methods produce very different estimates of
"residual chlorine." In fact, the products from chlorination of wastewaters
and natural waters are a mixture of chlorine, hypochlorous acid, hypochlorite
ion, inorganic and organic chloroamines, and other compounds. A better term is
"residual oxidants," as noted elsewhere in this report.
Since a large fraction of the U.S. population resides in coastal areas,
11
-------
much of the chlorine is discharged to saline natural waters. There is an
extensive literature (19) on chlorination of freshwater systems, but coastal
and estuarine waters have not been studied extensively. Research programs are
underway at several federal and university marine laboratories to alleviate
this situation. The work on freshwaters, unfortunately, does not have much
application to marine environments, because seawater has a bromide ion concen-
tration of 65 mg/liter and the added chlorine reacts with it to produce
hypobromous acid and hypobromite ion. Bromoamines and chloroamines may be
formed in the presence of ammonium ion (20).
For normal seawater of pH 8, the initial products of chlorinations are a
mixture of hypobromous acid and hypobromite ion. Both of these compounds are
unstable with respect to decomposition and disproportionation (23).
2HOBr -ป 2H+ + 2Br~ + 00;
"
20Br -> 2Br + 0.
3HOBR -ป 3H+ + 2Br~ + BrO ~;
_ (2)
30Br~ -ป- 2Br + BrO.,
The rates of these reactions have not been measured in seawater. The
decomposition of HOBr-OBr solutions has been considered most recently by Lewin
and Avrahami (21) and by Engel et al. (22). Both groups conclude that the
decomposition is to bromate plus bromide, with the disappearance of HOBr-OBr
or bromide ion concentration and to decrease strongly with increasing pH (22).
It is observed to be independent of the chloride ion concentration up to 0.5M
(21). No photolytic effects were investigated.
Previous investigators (15, 19) considered only the rate of disappearance
of residual oxidants in chlorinated seawater and did not identify the products.
The initial rapid decline was ascribed to reactions with organic compounds and
the ensuring slower decline to "decomposition." No attention was given to
photolysis by laboratory lighting or, more important, by natural sunlight. We
report here our observations of residual oxidant disappearance and bromate
formation, with particular reference to the significance of photolysis.
Chlorinated seawater was exposing to sunlight in open beakers placed in a
bath of running seawater. In each experiment, six 400 ml beakers, each
containing 300 ml of filtered (Millepore, 0.22 vim) Florida Current water, were
placed in a batch. After temperature equilibrium was reached sufficient NaOCl
solution (buffered to pH 8.1 with Na CO ) was added to each beaker to give an
initial OC1 normality equivalent to approximately 4.5 mg Cl per liter.
Actual initial OC1 concentrations varied somewhat (4.2 to 4.9 mg of Cl per
liter) among experiments because of varying OC1 concentrations in the stock
solution. Florida Current water contains less than 1 urn NH_, so formation of
haloamines cannot take up more than 5 percent of the added Cl .
Light intensities were estimated with a Yellow Springs Instrument -
Kettering Model 65A radiometer, operated with the focusing head removed from
12
-------
the sensing element. This makes it possible to measure relative light
intensities from day to day without the extreme angular dependence caused by
the focusing head. Experiments were conducted under full sun, partial sun, and
heavily overcast conditions.
After chlorination, beakers were removed from the sunlight at regular
(usually 30 minute) intervals, placed in a dark box, and analyzed for bromate
and residual oxidants without delay. Residual oxidants analyses were performed
by the I,~ spectrophotometric titration procedure described by Carpenter (9)
with a pH of 2 and a KI concentration of 4 g/liter. Bromate analyses were made
by differential pulse polarography at 25ฐC and a pH of 8.35 (after 02 stripping
with N ), using a Princeton Applied Research Model 174A polarographic analyzer.
A typical polarographic recording is shown in Fig. 3. Curve "a" is the
polarogram obtained for chlorinated seawater analyzed immediately after
chlorination. Identical traces were observed for nonchlorinated seawater and
for chlorinated seawater kept in the dark for periods up to 24 hours at
temperatures up to 40^C> which indicates a lack of bromate formation under
these conditions (BrO, _< 10 M, less-|han 0.5 percent conversion of Cl-).
Addition of copper sulrate to give a Cu concentation in the seawater of 100
parts per billion did not induce measurable bromate production in the dark.
Curve "b" was obtained from a chlorinated (4.9 parts per million (ppm) )
seawater solution that was exposed to full sunlight for 70 minutes. Curve "ฃ,-"
which is offset by 0.4 ya with respect to curves "a" and "b." shows 1.0 x 10 M
sodium bromate in seawater.
Figure 4 illustrates kinetic data for the appearance of bromate (Fig. 4A)
and disappearance of residual oxidants (Fig. 4B) in chlorinated seawater
exposed to sunlight. Curves "a" were obtained from solutions exposed to full
midday sunlight for the duration of the experiment; curves "b" are for exposure
to partial sunlight (the average light intensity was approximately 65 percent
of full sunlight); and curves "c" are for overcast conditions (average light
intensity, 20 percent of full sunlight). Curve "d" in Figure 4A shows the
disappearance of residual oxidants with time at 40ฐC in the dark. The
ordinates are calibrated as the percentage of the added chlorine recovered as
residual oxidants (Fig. 4A) or as bromate formed according to Equation 2
(Fig. 4B).
The lack of observable bromate production in' the dark is not inconsistent
with the report of Lewin and Avrahami (21) that substantial bromate was formed
in their 0.05M hypobromite solutions. Our solutions, which correspond to
chlorine use, were 1000 times more dilute. Using their rate constants, we
calculate in our solutions a conversion to bromate of less than 1 percent after
24 hours.
The loss of residual oxidants does not correspond exclusively to bromate
formation, and other reactions including oxidation of organic matter and
perhaps those in Equation 1, also take place. The rate and extent of bromate
formation depend on the intensity of sunlight.
In another experiment, 1.0 x 10 M solutions of sodium bromate in seawater
were exposed to full midday sunlight for periods up to 4 hours and the residual
13
-------
8 -
-1.25
1.40 --
Volts vs S.C.E
1 .70
Figure 3. Differential, pulse polarographic verification of sunlight-induced
bromate production in chlorinated seawater. (Curve a) Polarogram from
untreated seawater, seawater immediately after chlorination to 4.9 ppm,
or chlorinated seawater kept in the dark for 4 hours at 40ฐC. (Curve b)
Polarogram from chlarinated seawater exposed to full sunlight for 70
minutes. (Curve c) Standard: 1.0 x 10~5M sodium bromate in seawater,
offset with respect to curves a and b. Polarogram were recorded at
25ฐC and pH 8.35; SCE, saturated calomel electrode.
14
-------
TOO
2 3
Time (hr)
Figure 4. (A) Disappearance with time of residual oxidants and (B) con-
comitant appearance of bromate (Eq. 2) in chlorinated seawater (4.2 to
4.9 ppm of Cl2) as a function of exposure to sunlight. The conditions
were: (curve a) full midday sunlight, (curve b) 65 percent of full sun-
light, and (curve c) overcast, 20 percent of full sunlight. Curve d
shows residual oxidant disappearance in the dark at 40ฐC. No bromate
production was observed in the dark.
15
-------
oxidant and bromate concentrations were monitored. No measurable decline in
bromate concentration or increase in residual oxidant was found.
Thus, the production of substantial amounts of bromate ion will cause
erroneous results when standard analytical procedures are used for residual
oxidants, especially procedures involving reaction of the oxidants with iodide
ion. Bromate reacts sluggishly with iodide ion and the rate is dependent on
factors such as reactant concentrations, pH, temperature, light, and content of
transition metals. More important, it appears that large amounts of bromate
have already been produced in estuarine and coastal waters with unknown
effects. Extremely limited information is available on the direct toxicity of
bromate ion (19). Further, the formation of bromate may provide a persistent
source of low levels of known toxicants (such as hypobromite and bromoamines)
and brominated organics through the reverse of the formation reactions. In
summary, present knowledge is totally inadequate to assess the environmental
impacts of our discharge of chlorine to saline waters.
16
-------
SECTION 6
CHLORINATION AND THE COPPER COMPLEXING CAPACITY OF SEAWATER (24)
While copper has long been recognized as an essential micronutrient for
the growth of aquatic plants, recent research has drawn attention to the
extreme sensitivity of microalgae to free ion in seawater (25) and freshwater
(26). Reduction in growth was reported at the nanomolar level for several
species. Addition of chlorine has the potential for modifying the organic
compounds that nearly completely complex copper in natural waters and thus
increase the toxicity of the copper presence. The results of some initial
experiments to explore this possibility are reported here.
PROCEDURE
Anodic stripping voltammetry (27) provides a technique for measuring
copper with a minimum of sample manipulation or addition of reagents. Biscayne
Bay water samples were placed in an electrolysis cell and dissolved oxygen was
removed by bubbling argon through the sample for 20 min. The electroactive
copper was plated out on a hanging mercury drop electrode (PAR Model 9323)
during 10 min with an applied potential of minus 0.8 V vs SCE, using a
Princeton Applied Research Model 174 Polarographic Analyzer (Princeton Applied
Research Corp., Princeton, N.J.). Then the potential was made anodic at the
rate of 2 mV/sec and the resulting stripping current peak due to copper at
minus 0.185 V was recorded. Aliquots of a standard copper sulfate solution
were added to the electrolysis cell and the copper stripping current measured
after each addition to produce a titration of the copper complexing capacity of
the sample.
RESULTS
Figure 5 is a plot of the copper stripping currents in a sample of
Biscayne Bay water with copper added in increments corresponding to 1 ppb in
the sample solution. The initial increments of added copper produced some
increase in the copper stripping current and, then, additional increments
produced a more rapid increase in the copper stripping current. The initial
increase in the copper stripping current may be ascribed to electroactive
(labile) copper complexes. The sharpness of the transition between the two
slopes suggests that a complex with a very large formation constant is present.
Figure 6 shows the titration of a sample of the same water after
chlorination (1.5 ppm Cl_ added). The copper complexing capacity is reduced.
17
-------
300
200-
E
E
100 -
10
Cu Cone ppb
Figure 5. Anodic stripping current variation with copper added to a sample
of Biscayne Bay water, showing copper complexing capacity of 12 ppb.
18
-------
300
e
s
E
200 .
100
10
Cu Cone ppb
20
Figure 6. Anodic stripping current variation with copper added to a
chlorinated (1.5 ppm C12) sample of Biscayne Bay water, showing decreased
copper complexing capacity and changed character of the complexes (same
water as in Figure 5).
19
-------
Also, Che copper complex that remains is altered, as shown by the increased
slope of the curve before the endpoint.
Figure 7 shows the titration of another sample of Biscayne Bay water which
had a copper complexing capacity of 6 ppb. An aliquot of this sample was
chlorinated (5 ppra ! added). This dosage did not produce a measurable,
persistent residual oxidant concentration. This sample was titrated with
results that are plotted in Figure 8. The addition of chlorine removed all the
measurable copper complexing capacity from this water.
These results suggest that chlorination of seawater may produce toxicity
and growth reduction through the indirect mechanism of modifying the copper
coraplexing capacity of the treated water.
20
-------
300
200 I
E
E
100 4.
10
Cu Cone ppb
20
Figure 7. Anodic stripping current variation with copper added to a
sample of Biscayne Bay water, showing copper complexing capacity of
6 ppb.
21
-------
300 ป
200 -
E
E
JZ
O)
S
100 ..
0
0
10
Cu Cone ppb
Figure 8. Anodic stripping current variation with copper added to a
chlorinated (5 ppm 0X2) sample of Biscayne Bay water - showing a loss
of complexing capacity.
22
-------
SECTION 7
A POWER PLANT STUDY:
CHLORINATION AT THE PORT EVERGLADES, FLORIDA, POWER PLANT (28)
Much of the present information on the products formed when seawater^is
chlorinated is based on observations of laboratory experiments in which
chlorine was added to seawater to simulate conditions of electricity generating
plants. The work of Helz e_t al. (29) is the only comparison of laboratory
simulation and direct sampling at an operating power plant that we have found
in the literature. They found that chlorine-produced oxidants disappeared to a
much greater extent (factor of ten) in the water that passed through the Chalk
Point plant than did an equal oxidant dose under laboratory conditions or in
samples withdrawn from the plumbing upstream of the condensers inside the power
plant. Also, they measured only traces of halofonns in the discharge canal
waters, even though 10-100 yg/1 CHBr_ was formed in chlorinated intake water.
On some of their sampling dates, ammonia nitrogen was found to be higher in the
discharge canal than in the surface water at the intake, but this feature was
not observed under laboratory conditions. These discrepancies led Helz et al.
to the view that "laboratory dosing of water from the intake canal with 1-
2 mg/1 Cl as NaOCl provided a poor model for what was observed in the field."
Sampling at the Port Everglades power plant of the Florida Power and Light
Company was undertaken for comparison with the observations at the Chalk Point
plant on the Patuxent estuary in Maryland by Helz e_t al. The Port Everglades
power plant consists of four oil-fired units with four condensers per unit.
Cooling water is pumped from the harbor and during observation had salinities
of 29.5-31.1 parts per thousand (ppt)3 in contrast to the average salinity at
the Chalk Point plant of 5 ppt. Microorganism fouling of the condensers is not
a 'substantial problem at Port Everglades, and chlorination for 15 min/day on
each condenser sequentially at a dose of 0.6 ppm is used. Samples were
collected from the pipe just downstream of the circulating water pump (chlorine
is injected through a manifold in the intake structure and mixing takes place
in the pump) and from the discharge sluiceway. Mixing of chlorinated and
unchlorinated water takes place in the discharge sluiceway and the resulting
nonuniformity produces variability in replicate samples. Additional samples
at the circulating pump were held in glass flasks for 2.5 min and then
stabilized for comparison with the water collected at the discharge that passes
through the plant in approximately 2.5 min.
METHODS
The samples for residual oxidant determinations were stabilized
23
-------
imnnned lately after collection by adding phenylarsine oxide (PAD) solution, and
the unreacted PAD was titrated with standard iodine solution using photometric
endpoint detection (3) within 24 hr of collection. Bromoform concentrations
were determined using a purge-and-trap procedure (30) with a Hewlett-Packard
5730A chromatograph and electron-capture detection (ECD). The samples for the
ammonia nitrogen determinations were stabilized at collection by adding phenol
solution and analyzed using the Solorzano procedure (31).
RESULTS
Comparison of the concentrations of residual oxidants in the discharge
water and in water from the circulating water intake pump held for 2.5 min did
not show a remarkable difference in the rate of disappearance of the residual
oxidants. The decrease in residual oxidant concentration in either case was
typically 50%. As shown in Table 2, bromoform remained essentially constant
within a factor of two during travel through the power plant. (Intake water
was held for 2.5 min before quenching to allow for passage time through the
plant.) Due to physical factors in the plant, there was both some periodic and
uncontrolled mixing at the intake and discharge of the plant, but these
analyses were made on water taken as close to intake and discharge as possible.
Laboratory chlorination of water taken at the plant site produced varying
levels of bromoform, but at 4 ppm added oxidant, we did not find the 10X excess
found by Helz as shown in Table 3. (We do not have a measure of chlorine added
at the plant.)
TABLE 2. BROMOFORM IN PORT EVERGLADES, FLORIDA,
POWER PLANT INTAKE AND DISCHARGE WATERS
*
Unchlorinated Chlorinated
Date
12 Sep 1980
20 Oct 1980
Intake Water
1 ppb
1 ppb
Intake
75 ppb
78 ppb
Discharge
86 ppb
32 ppb
Chlorine residual was 0.5-0.6 ppb.
Ammonia nitrogen in the intake water samples average 25 vg/1. Similar
values were observed in the discharged waters and in the intake samples that
were held for 2.5 min before addition of the phenol solution.
Observations show no major effects associated with passage of the cooling
water through the Port Everglades plant. Comparison with the results of Helz
ej: al. for the Chalk Point plant suggests there must be some site-specific
considerations that have not been identified. The total organic carbon (TOC)
at both sites was approximately 5 mg/1 and the ammonia nitrogen levels were
comparable; thus, gross water quality was not strikingly different. The
condensers at Chalk Point have copper-nickel tubing and those at Port
Everglades have aluminum-brass tubing, but there is no information on the
24
-------
reactivity of chlorinated water with these materials. Observations at several
coastal plants are needed for assessment of the haloform discharge rates and
usefulness of laboratory chlorinated studies for anticipating the nature of
power plant discharges.
TABLE 3. LABORATORY CHLORINATION OF POWER EVERGLADES
POWER PLANT INTAKE AND DISCHARGE WATERS
Chlorine
Added Detected
1 ppm 6 . 5 ppb
2 ppm 107 ppb
4 ppm 272 ppb
25
-------
SECTION 8
LIPOPHYLIC HALOGENATED COMPOUNDS: THE LIQUID/LIQUID EXTRACTABLES
8.a. INTRODUCTION
Bromoform was found to be the major "volatile" halogenated s-ompound
produced by the chlorination of seawater, measured using the purge-and-trap
method. However, bromofonn (and its one and two carbon halogenated analogues)
cannot account for more than a small percentage of the total organic carbon
consumed. It was postulated that halogenated reaction products which would not
extract by sparging with inert gas but would partition favorably into solvents
such as ethyl ether or chloroform (i.e., lipophylic in nature) were present in
the chlorinated saline waters studied. Lipophylic halogenated compounds are of
particular interest due to their capacity for incorporation into fatty tissue
and their transport across cell membranes. Fish, for example, would be
expected to assimilate such compounds through the gut and across gill membranes
with possible accumulation in fatty tissue being transferred up the food chain.
Such compounds could have long-term environmental effects while not
necessarily demonstrating large LD/50 values for mature or perhaps even larval
stage organisms, in a food chain sequence. Such compounds would be expected to
partition favorably into halogenated hydrocarbons such as chloroform. In
addition, we would expect that some of these compounds would behave as mild
oxidants making them susceptible to reducing agents such as potassium
thiosulfate. We chose, therefore, to look for such compounds using large
volume extracts of chlorinated Biscayne Bay water, first using chloroform and
later using ethyl ether as the extraction solvents. Initially, the study of
lipophylic halogenated compounds centered on GC/MS analysis of chloroform
extracts of chlorinated Biscayne Bay water. Mass spectral analysis showed that
the compounds observed were not similar to compounds in mass spectral
libraries, but definitely were brominated and possibly contained nitrogen. One
interpretation suggested a compound such as bromopyrrolidine although such a
compound is not recorded in the literature. Subsequently, attempts to
synthesize a compound with the same mass spectrum were performed using proline
which was considered to be a potential precursor. This resulted in
intermittent low yields of compounds that matched the mass spectra and
retention times of those found in natural extracts. Suspecting cyclic amino
acids to be possible precursors, a variety of amino acids were chlorinated in a
seawater matrix and chromatographed using HPLC with fluorometric detection.
The unknown compounds proved to be stable in solvent solutions but virtually
impossible to purify by the usual methods such as adsorption chromatography,
distillation, etc. Milligram quantities of the compounds were obtained by a
small capacity preparative gas chromatograph and this set up was used to
generate enough sample for proton NMR and elemental analysis.
26
-------
8.b. HEXANE AND CHLOROFORM EXTRACTIONS
ECD-Capillary Column Results
Early in the study of liquid/liquid extractables, a comparison was made
between hexane(s) soluble and chloroform soluble fractions. As shown in
Figure 9, chloroform extracts of Biscayne Bay water, both chlorinated and
unchlorinated, contained a much larger number and quantity of compounds
detectable by the electron capture detector (ECD) than did hexane extracts
obtained in a similar manner. In this case, the experiment consisted of the
collection of 4 carboys of 16 liters each of Biscayne Bay water; two of which
were chlorinated to ca_ 5 ppm. Extraction was done in tandem, i.e. the
raffinate from the hexane extraction was fed directly into the chloroform
extractor and the extracts treated equally. The process took several days to
complete so the chlorine (as chlorine saturated water) was added 30 mio. from
the time the extraction was started. The water was not filtered but was
acidified to pH 5 just before extraction to reduce foaming. The hexane
fraction contained few compounds detectable by the ECD and a much more
sensitive gain level (X128) was required, compared with the chloroform extract
(X2048). GC/MS on the hexane extract produced no usable spectra as would be
expected from lower sensitivity of that instrument (in the broad scan mode)
compared with the ECD. GC/MS on the chloroform extract did, however, show
compounds that are of great interest.
GC/MS Results
The first chloroform extracts of chlorinated Biscayne Bay water were made
"on T4~ December 1977 and on 30 January 1978; both contained compounds having
unusual mass spectra. The 14 December 1977 experiment was similar in execution
to the 30 January 1978 run described below except that only CHC1, was used as
solvent and that more attention was given to the compounds more volatile than
bromoform. As shown in Table 4, eights peaks were observed to have usable mass
spectra from the GC/MS of whch five were tentatively identifiable by comparison
with published mass spectra. The halogenated methane and ethane compounds were
expected based on previous volatiles analyses and not considered particularly
noteworthy. The compounds that eluted at 4.1-4.2 and 5.8-5.9 min had mass
spectra that did not match known published spectra and were major in abundance.
These two compounds, and in particular the one that eluted at 5.8-5.9 min, were
studied extensively.
The total ion chromatogram of the CHC1., extract of chlorinated 30 January
1978 Biscayne Bay water (Figure 10) contains 18 discernable peaks, of which 8
are too small to give usable mass spectra. Bromoform, at 3.7 min, overlaps
with another smaller peak at 3.8 min and a contaminant, probably methylisobutyl
ketone is recorded at 2.7 min. The remaining 7 peaks are all halogenated
compounds, as shown by the characteristic M, M+2 couplets in the mass spectra
shown in Figures 11-13. Even though the mass spectrometer was carefully
calibrated by using perfluorotributylamine (PFTBA) and decafluorotriphenyl-
phosphine (DFTPP) in the manner described by Eichelberger e_t al_. (32), none of
these spectra are recorded in the Aldermaston Eight-Peak Index (33) or the
Cornell-McLafferty mass spectral libraries.
27
-------
TABLE 4. TABULATED GC/MS CHROMATOGRAMS OF CHLOROFORM EXTRACTS
OF CHLORINATED BISCAYNE BAY WATER
Retention
Time , min
1.7
1.8
2.3
2.4
2.7
3.7
4.1
5.8
2.9
3.7
3.8
4.2
5.9
9.6
10.0
10.5
11.9
12.2
Compound Spectrum, Mass (abundance)
Run 5076 Chlorinated Baywater , 14 December 1977
CC14 119(100), 117(99). 121(31), 47(22)
CHCl2Br 83(100), 85(66), 47(34), 48(19)
45(100), 69(99), 91(79), 53(32), 148(10), 150(10)
CHClBr2 129(100), 127(83), 131(25), 81(19)
C2H2C14 166(100). 164(82), 168(49), 129(42), 131(41)
CHBr3 173(100), 171(55), 175(51)
105(100), 69(69). 107(37), 53(20), 149(20), 151(20)
151(100), 149(99)^69(54), 53(11)
Run 5078 Chlorinated Baywater, 30 January 1978
C2H2C14 166(100), 164(69), 168(46), 129(38), 131(35)
CHBr3 173(100), 171(55), 175(51)
43(100), 59(95), 153(37), 151(37)
105(100), 69(65), 107(36), 53(19), 149(15), 151(12)
149(100), 151(99), 69(44), 53(10)
43(100), 99(57), 55(28), 42(9), 135(8), 137(7)
137(100), 139(95), 43(48), 151(39), 153(36)
139(100), 137(96), 153(73), 151(72), 43(72)
185(100), 183(75), 67(40), 103(35), 147(29), 149(31)
227(8), 229(10), 231(8)
185(100), 182(81), 67(42), 103(38), 147(32), 149(38)
28
-------
00
CVJ
V)
c
o
Q.
ir
u
0)
Q
unchlorinated Biscayne Bay water
hexane extract
t
If me
Figure 9a. GLC chromatogram of hexane extract of unchlorinated Biscayne
Bay water.
29
-------
GO
Sf
O
CM
C
O
Q.
ซrt
CD
OC
O
O
0>
Q
unchlorinated Biscayne Bay water
chloroform extract
t
Time
Figure 9b. GLC chromatogram of chloroform extract of unchlorinated
Biscayne Bay water.
30
-------
CD
CM
o>
to
C
o
ex
in
-------
00
SJ-
o
CVJ
W5
c
o
Q.
(O
o
o>
Q
t
chlorinated Biscayne Bay water
chloroform extract
Time
Figure 9d. GLC chromatograph of chloroform extract of chlorinated Biscayne
Bay water.
32
-------
TI
Figure 10. Total ion chromatogram of chloroform extract of chlorinated
Biscayne Bay water, total ion abundance vs. elution time, minutes.
33
-------
5073
*7S0
5072
*739
5078
$6-43
507S
t606
5078
J576
5078
t3!3
5078
*192
40
,f J
J.1 ll
II
,. L
1 ^ r^ r
,,V
T | 1 , 1
, ..I
' 1 1
, , f i j ( ,
60
"V ** r''~i i
i_
i i i ' i
i.
i i : i
i ' i ' i ' i
l T i T i i
_,
1 i T i i
80 100
1 T ' T ' I r
L i I
i ' 1 ' 1 '
1 1 1 ' 1 '
' 1 ' I ' 1 '
I ' I ' ! '
1 ' 1 ' 1 '
L
1 1 ' 1 ' 1 '
120 1
l|
1 '
([
i T i
7 i
1 i
t ' i
r ' i
!
<;0
r i ' r ''i
1 i i i
i ' i i
It
i i i i
i ' i i
i - i i
i i i l i i i
160 180
12.2 min.
* I ' I ' ' I i i
11.9 min.
r i i i i T
10.5 min.
,
i i i T i i
10.0 min.
i ' i i i i
9.6 min.
i i ' i ' i i
5 . 9 min .
i ' i ' i ' i i '
4 . 3 min .
i ' i ' i ' i ' r
200 230 34
-
-
.
-
-
*
-
-
0
100
100
100
100
100
100
100
Figure 11. Mass spectra of major peaks shown in Figure 10, relative
abundance vs. atomic mass units.
34
-------
FRN 5078 SPECTRUn 313 RETENTION TIP1E 5.9
LnROST 4- 149.0.100.0 151.
LAST 4' 232.0. .1 233.
100
SO
SO
40
100
30
60
40
SO
f)
!ni llli' mi I
" ""i i11 i"" ""i"" ""(!ป ....(iiii .1.
50 40 60
"" ""1" 1" 1" 1"" ""1"" ""inTT-rm,
180 200 220
0, 99.4 69.1. 44.2 53.1, 9.9
1, .1 234.9. .1 237.9. .1
PAGE 1 V 1 . 00
!ui illlii in i iiiitiiluii iiiililr M iii-i i i i 1 lii ii
80 100 1S0 140 16O
240 260 289 300 32ซ
Figure 12. Mass spectrum of peak at 5.9 minutes, relative ion abundance
vs. atomic mass units.
35
-------
FRN 5978 SPECTRUM 192 RETENTION TIP1E 4 3
LAROST 4 105.1.100 0 S9.1, S3 -4 107 1 , 33 6 53 2, 18 4
Lซ5T 4 2236. .1 2278. .1 230.2. 1 234.6. .1
100
30
60
40
20
0
100
80
60
40.
20
0
, ,l|l!|n,1|i|
,., ,,, .,,,,, ,,..|,.,. ,.,, ,,.,, ,,,,,,
20 40
""i"
60
i i
- A
PAGE 1 V-- 1 00
, i, H
' "'I | | | |"" "! -|" | f 1
80 100 130 140 16O
130 200 220
|nll llll|iin iui|im iiii|iiii mijini M.I,..,. ini|i ,..i. ......... ....,
240 260 2He 3Q& 32ฎ
Figure 13. Mass spectrum of peak at 4.3 minutes, relative ion abundance
vs. atomic mass units.
36
-------
Compounds represented by the two major peaks at 4.2 and 5.9 min were
studied further due to their abundance. GC/MS runs at low energy (20 EV)
electron impact mode (El) and in the chemical ionization mode (CI) (methane
reactant gas) did not reduce the m/e 149/151 couplet, nor were there any new
ions observed at higher masses. Lacking any guidance from the literature to
the contrary, the hypothesis was taken that m/e 149 represented a true
molecular ion accepting that confirmation was not obtained by CI mass spectro-
metry (i.e., M + 1 and M + 15 ions were not observed with methane). Accepting
this assumption, the 5.9 min peak of the 30 January 1978 chloroform extract can
be represented by an elemental composition of C.HgNBr. The m/e 69 fragment
suggested a neutral loss of HBr, a common neutral fragment in El MS. It appears
that the compound could be an isomer of bromopyrrolidine:
Br
Further, the 4.3 min peak of the 30 January 1980 was given a hypothetical
structure of CcHgNCvCl, or chloroproline on the basis that the m/e 44 neutral
(149-150 m/e) was COO and that the m/e isotope ratio of 100/30 for m/e 105/107
is chlorine.
This rationale suggested the investigation of the halogenation of
naturally occurring amino acids, especially proline, as a direct step toward
the synthesis and chemical characterization of the major compound found in the
chloroform extract of chlorinated Biscayne Bay water.
8.c. PREPARATION AND IDENTIFICATION OF UNKNOWNS FOUND IN CHLOROFORM EXTRACTS
8.c.l. Synthesis of Unknown Compound
Early efforts to synthesize a compound having a mass spectrum matching
that of the major compound found by GC/MS in the chloroform extract of the
chlorinated 14 December 1977 Biscayne Bay water (i.e., m/e 69, 149, 151) were
founded on the premise that m/e 149 was a true molecular ion. Interpretation
based upon that premise led to the prediction that the compound was
bromopyrrolidine (with the location of the bromine left unstated) and that the
precursor of the compound was a simple five membered cyclic secondary amine.
Proline and pyrrolidine were considered as possible candidates, but only the
former has been observed in natural waters. An early experiment in which
proline-spiked Gulf Stream water was chlorinated and extracted with chloroform
produced a compound which matched the mass spectra of the unknown served to
support that direction. The most direct approach to the synthesis of the
compound seemed to be the bromination of pyrrolidine. This was done
37
-------
successfully by adding bromine in sodium hydroxide solution directly to
pyrrolidine in 0. 1M phosphate buffer with subsequent extraction using
chloroform. This method was successful, but it provided very low yields.
Subsequently, a variety of approaches were taken including the halpgenation of
pyrrolidine by light-induced free radicals from N-halosuccimide (Cl and Br),
the Hunsdiecker rearrangement of the silver salt of proline, and the
halogenation by N-halosuccimide in various solvent and temperature
combinations. Halogenation of aqueous solutions of proline confirmed that
decarboxylation occurs as a first or early step in the oxidation of that amino
acid. In general, however, aqueous oxidation of proline by HOBr followed by
extraction with chloroform did not prove to be a reliable synthesis method.
Chloroform has a number of undesirable properties as an extraction solvent,
i.e. probably toxicity, relatively high boiling point, possible reactivity to
aqueous free radical attack, etc. As a result, an alternate solvent was
sought. Hexane was not successful in baywater extracts (reported in section
8b); neither were saturated halogenated solvents such as carbon tetrachloride,
and 1,1,2,trichlorotrifluoroethane in laboratory extracts of halogenated
proline. Ethyl ether (ethanol stabilized or unstabilized) appeared to have
sufficient advantages to make it the solvent of choice. These advantages
included higher extraction efficiency, high volatility (lower distillation
temperatures, very rapid sample preparation), and much improved
chromatographic properties.
Concurrent postulation of reaction mechanisms suggested that an important
intermediate in the formation of the haloamine from proline might be the cyclic
imine, 1-pyrroline:
,COOH + HOBr
HOBr, X-
RX
(X = Cl or Br)
Supporting evidence for the the imine's equilibrium product, 4-
aminobutryaldehyde, is shown in section 8.c.4. Since increasing the
equilibrium concentration of 1-pyrroline is favored by removing the water, it
was considered advantageous to conduct the reaction in non-aqueous media.
Also, it has been reported (34) that HOC1 (and presumable HOBr) can be
extracted into ether. Consequently, the following procedure was devised and
was found to be consistently successful:
38
-------
Step 1: Dissolve proline _to 0.005 jnolar in 200ml of buffered
"matrix" (0.3M Cl , 0.3M Br , O.lM borate buffer to 8.1
pH).
Add sufficient NaOCl solution to ca 1.5X molar excess.
Step 2: Dissolve sufficient NaOCl in 200 ml of the "matrix" to give
ca 5-10X molar excess to proline. Adjust pH to 8.1.
Step 3: Extract solutions from Steps 1 and 2 separately with HPLC
grade ethyl ether, then add extract from Step 2 dropwise to
extract from Step 1.
Step 4: Dry the combined ether extract over anhydrous sodium
sulfate and concentrate to 1 ml using Kuderna-Danish and
micro-Snyder glassware.
Step 5: Isolate and purify the compound using gas chromatography
(GLC) with a 3 mm ID, 2 meter OV-101 packed column with a
temperature program of 55 C isothermal for 4 minutes and
then increasing at 16 C per minute to 90 C hold, and a
helium carrier gas flow rate of 60 ml/min. Inject
30 microliters of the concentrated solution from step 4.
Monitor the discharge with a thermal conductivity detector
and manually trap the fraction containing the peak at 7.4
minutes in a dry-ice cooled U-tube microtrap. Transfer the
collected liquid material to a cone-shaped reaction vial
with a septum cap.
Rechromatograph the collected material isothermally at 80 C
and collect by trapping the material from the peak at 5.0
minutes. Use 4 microliter aliquots of the material that was
collected in the reaction vial. Transfer the material to a
1 ml scalable glass ampule and dry over P^O,- to constant
weight before sealing. Commonly, yields of 5 milligrams
were found for this overall and rather tedious procedure.
8.C.2. Elemental Analysis of Synthesized Unknown Compound
Elemental analyses of the synthesized unknown compound were performed by
Galbraith Laboratories, Inc. on two samples of the compound that matched both
the retention time and mass spectrum of the major unknown in
chloroformanalytical schemes, normally the minimum sample is roughly 30 to 50
milligrams. Several months of effort were devoted to preparing the required
amount of material through the use of the above outlined preparative scale gas
chromatography. The results are shown in Table 5. No chlorine was found as
might be expected since bromide ion oxidation in chlorinated seawater is rapid
as described above. The absence of nitrogen indicates that the cyclic amine
structure in the proline is not present in the reaction product. Possible
interpretation of these results is discussed in conjunction with additional
mass spectral data below.
39
-------
TABLE 5. ELEMENTAL ANALYSIS OF BROMINATED UNKNOWN
Sample Elemental Analysis
* 2C %R %N %Br %C1 %0
M2-146
M3-12
22.83
19.90
4.02
3.97
* * *
0.58 64.16 0.5
*
* 11.4
*
Not determined due to insufficient sample.
Performed by Galbraith Laboratories, Inc., 2323 Sycamore Drive,
Knoxville, TN 37921.
8.c.3. NMR Spectra
NMR spectra of the synthesized unknown suggest a compound having three
methyl groups, i.e., nine protons, one methyl being split by a methine group at
low field. A typical 60MHz H NMR spectrum is shown in Figure 14. The peaks at
3.7-3.9 appear to be composed of two singlets and a doublet. The doublet is
not obvious until the quartet at that value is decoupled (not shown) causing a
disappearance of the doublet and the appearance of two singlets of about equal
magnitude. The interpretation of the spectra is not straightforward because an
isomeric mixture may' be present. This possibility is indicated because
repetition of synthesis and chromatography produced materials that had
somewhat-different NMR spectra. The probable presence of isomers was indicated
also in the GC/MS analyses in that several of the small peaks in the
chromatogranis gave the same major ions as did the primary large peak.
8.c.4. Fluorescence and Electrochemical Analyses of Chlorinated Proline
The chlorination of proline in pH 8.1 borate buffered chloride/bromide
matrix solution was studied using two different polarographic procedures
systems in the differential pulse mode (DDP), and two fluorescence procedures
one with dansyl chloride, the other fluorescamine (Fluram). As shown in
Figure 15, proline concentration was essentially reduced to zero at a molar
ratio of 3.0 (Cl/proline) as measured by dansyl chloride and at this same molar
ratio, an unkown product appeared to be at its maximum. Fluorescamine
(Figure 15B) which reacts with primary amines only, indicated a maximum yield
at a molar ratio of 1.5-2.0 (Cl/proline) of an unknown product which is
probably the same as observed with the dansyl chloride. Theory suggests that
this is (as shown in 8.C.2.) 4-amino-butryaldehyde. Differential pulse
polarography shown in Figure 15 C and D show reduction waves for three unknown
species with maxima at molar ratios of 1-2 (-1.1 to -1.2 volts SCE) and 3 (-1.5
volts SCE). An easily reducible species (-0.2 to -0.3 volts SCE) appears at a
molar ratio of ca 2 with no maximum observed. Interpretation of these results
40
-------
/H,
ppm (S)
0
Figure 14. Proton NMR spectrum of unknown compound having m/e 69,149,151 mass spectrum (60mhz, H,
CDC13, Varian 360A, single scan, TMS internal reference).
-------
40
ซ 30 -
c
3
Q
O
O
i_
o
O
O>
Q.
Cl/proline ratio
Figure 15a. The oxidation of 5 mM proline in pH 8.1 borate buffered
chloride/bromide solution - dansyl chloride derivative separated by HPLC,
42
-------
t_
a
o
c
0)
o
W
0)
L_
o
3
5mM proline
Cl/proline ratio
Figure 15b. The oxidation of 5 mM proline in pH 8.1 borate buffered
chloride/bromide solution - measured by fluorescence of fluorescamine
(Fluram) derivatives.
43
-------
a
o>
Q.
40 -
30 -
20 -
10 -
-1.12 VSCE
-I.49VSCE
-0.24 VSCE
C!/proline ratio
Figure 15c. The oxidation of 5 mM proline in pH 8.1 borate buffered
chloride/bromide solution - measured by differential pulse polarography,
PAR model 384.
44
-------
40
-L20VSCE
JC
o>
1C
a
Q>
a.
Cl/pro!ine ratio
Figure 15d. The oxidation of 5 mM proline in pH 8.1 borate buffered
chloride/bromide solution - measured by differential pulse polarography,
PAR model 174.
45
-------
is presently limited to the recognition that proline is oxidized by increased
amounts of halogen and that new chemical species appear and in some cases
disappear with increased halogen addition. The use of DPP and fluorescamine
and dansyl chloride derivatization shows promise for future experimental
schemes.
8.C.5. Possible Artifacts
Chloroform, or some contaminant or stabilizer (usually ethanol), was
considered as a possible source of the unknown compounds observed. Chloroform
is known to participate in free radical reactions, and the ingredients for such
reactions (heat, light, aqueous interphase) were present. The probability of
artifact compounds from chloroform appears low since chloroform extracts of
Gulf Stream water did not produce the unknown compounds. Further; chloroform
treated directly with aqueous HOBr also produced negative results. Co-
extraction of HOBr and the organic precursor(s) with subsequent reaction in the
non-aqueous phase cannot be totally ruled out, but no evidence for such
reactions has been found. Alternate analytical schemes that eliminate the
extraction step such as differential pulse polarography (DPP) are attractive in
this regard and could provide an approach to the study of these "as yet to be
identified" chlorination products.
8.C.6. Limited Search for Molecular Ions
While the chemical ionization results and the stability of the m/e 149-151
couplet as the ionization potential was varied suggested that the m/e 149-151
couplet was possibly a molecular ion, the failure to find nitrogen in the
elemental analyses denies the idea that this odd-numbered couplet was a
molecular ion. Based on the results of the elemental analyses, alternate
empirical formulas that had oxygen in place of the previously hypothesized
nitrogen and the presence of two bromine atoms were considered. The GC/MS
system was operated in the more sensitive selection ion monitoring (SIM) mode
and a peak at m/e 228 and smaller peaks at 230 and 232 were observed. Further
work on elucidating the structure of the compound should be based on a more
effective synthesis scheme to provide larger quantities of the material, which
should improve the accuracy of the elemental analysis and NMR spectra and allow
a conventional molecular weight determination. In particular, lanthanide shift
analysis with the NMR would be extremely valuable.
8.d. ELECTROCHEMICAL ANALYSES OF CHLORINATED BAYWATER
An exploratory study was made of the use of differential pulse
polarography (DPP) as an alternate method of detecting reaction products of the
chlorination of seawater, by passing the solvent extraction step. In a very
limited series of experiments, a DPP reduction wave was obtained at ca -0.9 v
(SCE) upon chlorination that was reduced with time, as shown in Figure 16. The
observed response did not appear in the blank (unchlorinated) nor does it
correspond to analyses on chlorinated organic-free seawatsr or proline-spiked
seawater. DPP is sufficiently sensitive to detect reducible species in situ
46
-------
200 r-
- 100
3
0
0
Baywater, 4ppmCI
initial
Baywoter blank
-.27 -.39 -.51 -.63 -.75 -.86 -.98 -1,10 -1.22 -|.34 -1.46 -1.56
Volts vs SCE
Figure 16. Differential pulse^ polaroeram'of chlorinated (1 ppm added) Biscayne Bay seawater showing
blank (unchlorinated), immediately, after chlorination, and after 30 minutes to 5 hours.
-------
that are generated by chlorination, and are probably organic. The promise that
DPP offers is that of a rapid, low cost assay for reducible species resulting
from chlorination. Much work remains, however, before the reducible species
can be identified.
8.e. HIGH PERFORMANCE LIQUID CHROMATOGRAPHY ON CHLORINATED AMINO ACIDS
Chlorination of amino acids in seawater cause a reduction in the
concentration of the amino acid and in some cases produces reaction products
that form dansyl derivatives, based upon a brief study using high performance
liquid chromatography (HPLC) with fluorescence detection. As shown in Table 6
of the eleven amino acids studied, all were reduced in concentration to some
degree, from 29 percent for hydroxy-proline to 90 percent for lysine after
reaction with a 3X molar excess of HOC1 in pH 8.1 borate buffered Gulf Stream
sea water. Of the amino acids studied, three produced unambiguous new peaks in
the fluorescence chromatograms. Separations were made on reverse phase RP-18
columns with methanol-water gradient elution. Amino acids are attacked by free
halogen in seawater and in some cases form unknown derivatives that may be of
interest. Variables such as time, concentration, light, temperature, and pH
remain to be studied.
8.f. SUMMATION
Our work shows that GC/MS analysis of chloroform extracts of chlorinated
Biscayne Bay water reveals new halogenated compounds that cannot be dismissed
as artifacts, nor have they been amenable to laboratory synthesis on a scale
sufficient to obtain a satisfactory chemical description. It is reasonable to
conclude that these compounds are "residual oxidants" with finite lifetimes in
natural waters with unknown toxicities and environmental impacts. The
precursors of these compounds, and their ubiquity in natural waters, remains to
be defined. Each analytical method provides clues to the chemical processes
that chlorination incurs, but a clear picture has not yet emerged. The
elucidation of the exact chemical structures of "residual oxidants" presents a
tantalizing challenge that could be developed using the discoveries described
above as a foundation. Another approach in which fractions of natural extracts
are characterized according to their halogen content, redox properties, and
biological response may be productive also from an environmental viewpoint.
48
-------
TABLE 6. CHLORINATED AND UNCHLORINATED AMINO ACIDS -
HPLC SEPARATIONS OF DANSYL DERIVATIVES
Amino
Acid
hydroxy-proline
glycine
alanine
phenylalanine
tryosine
lysine
methionine
cysteine
tryptophan
leucine
isoleucine
Percent Reduction
with Chlorination
29
43
35
58
78
90
92
33
60-
66
76
Number of
New Compounds
2
0
0
0
0
3
0
0
1
0
0
49
-------
REFERENCES
1. Brungs, W.A., J. Water Pollut. Control Fed. 45, 2180-2193 (1973).
2. Proceedings of Conf. on Environ. Impact of Water Chlorination, NTIS, Oak
Ridge National Lab., Oak Ridge, Tenn., Oct. 22-24, 1975.
3. Carpenter, J.A., C.A. Moore and D.L. Macalady, Environmental Science and
Technology, _U, 992 (1977).
4. Farkas, L., Lewin, M., Anal. Chetn. , ^9, 665-666 (1947).
5. Joahnnesson, J.K., Analyst, 83_, 155-159 (1958).
6. Rosenberger, D.R., Thesis, Michigan State Univ. East Lansing, Mich.,
1971.
7. American Public Health Assoc., "Standard Methods for the Examination of
Water and Wastewater", 13th Ed., pp. 112-116, 1971.
8. Marinenki, G., Huggett, R.J., Friend, D.G., J. Fish. Res. Board Can., 33,
822-826 (1976).
9. Carpenter, J.H., Limnol. Oceanogr., H), 135-140 (1965).
10. Kolthoff, I.M., J. Anal. Chem., 60., 348-352 (1921).
11. Van Der Meulen, J.H., Chem. Weekbl., 28_, 82-85 (1931).
12. Macalady, D.L., Carpenter, J.H., Moore, C.A., Science, 195, 1335-1337
91977).
13. Palin, A.T., Water Sewage Works, 108, 461-462 (1961).
14. Bradbury, J.H., Bambly, A.N. , Aust. J. Sci. Res. Ser. A, 5_, 541-554
(1952).
15. Eppley, R.W., Renger, E.H., Williams, P.H., Estuarine Coastal Mar. Sci.,
4, 147-161 (1976).
16. Blabaum, C.J. and Nichols, M.S., J. Am. Water Works Assoc., 4B, 1503
(1956).
17. Jolley, R.L., Oak Ridge Natl. Lab. Tech. Memo, (1973).
50
-------
18. Federal Power Commission, The 1970 National Power Survey (Government
Printing Office, Washington, D.C., 1971), Part 1.
19. Environmental Studies Board. National Academy of Sciences-National
Academy of Engineering. Water Quality Criteria 1972 (Government
Printing Office, Washington, D.C., 1972); W.A. Brunge, J. Water
Pollut. Control Fed., 45, 2180 (1973); R.L. Jolley in Proceedings of
a Conference on the Environmental Impact of Water Chlorination
(CONF-751096, Oak Ridge National Laboratory. -Oak Ridge, Tenn.,
1976).
20. Johnson, J.D. and Overby, R<> J. Sanit. Eng. Div. Am. Soc. Civ. Eng., 97,
8425 (1971).
21. Lewin, M. and Avrahami, M., J. Am. Chem. Soc., 77, 4491 (1955).
22. Engel, P., Oplatka, A., Perlmutter-Hayman, B., ibid, J76_, 2010 (1954).
23. Eppley, R.W., Renger, E.H., Williams, P.M., Estuarine Coastal Mar. Sci.,
4, 147 (1976).
24. Downs, A.J., Adams, C.J., in Comprehensive Inorganic Chemistry,
A,F. Trotman-Dickenson, Ed. (Pergamon, Oxford, 1973), Vol. 2,
-pt 1404.
25. Carpenter, J.H., Limnol. Oceanogr., 10, 135 (1965).
26. Carpenter, J.H. and Smith, C.A., Chapter 15 Water Chlorination,
Environmental Impact and Health Effects, Vol. 2, Ann Arbor Science,
Michigan, 1978.
27. Sunda, W. and Guillard, R., "The Relationship Between Cupric Ion Activity
and the Toxicity of Copper to Phytoplankton," J. Mar. Res., 34, 511-
529 (1976).
28. Chau, Y.K. and Lum-She-Chan, K., "Determination of Labile and Strongly
Bound Metals in Lake Water," Water Res., 8, 383-388 (1974).
29. Batley, G.E. and Florence, T.M., "The Effect of Dissolved Organics on the
Stripping Voltammetry of Seawater," J. Electroanal. Chem.
Interfacial. Electrochem., 72_, 121-126 (1976).
30. Carpenter, J.H., Smith, C.A., Zika, R.G., Chapter 34, Water Chlorination
Environmental Impact and Health Effects, Vol. 3, in press. Ann Arbor
Science, Michigan, 1980.
31. Helz,;G.R., Sugman, R. and Hsu, R.Y., "Chlorine Degradation and
Halocarbon Production in Estuarine Waters," In Water Chlorination:
Environmental Impact and Health Effects, Vol. 2, R.L. Jolley,
H. Gorchev and D.H. Hamilton, Jr., Eds. (Ann Arbor Science
Publishers, Inc., 1978), p. 209.
51
-------
32. Carpenter, J.H., Moore, C.A. and Macalady, D.L., "Errors in Determination
of Residual Oxidants in Chlorinated Seawater;" Environ. Sci.
Technol. IKIO), 992-994 (1977).
33. U.S. Environmental Protection Agency, "Sampling and Analysis Procedure
for Survey of Industrial Effluents for Priority Pollutants," IDE No.
WA-77-B133, Appendix B (March 1977).
34. Solorzano, L. "Determination of Ammonia in Natural Waters by the Phenol-
hypochlorite Method," Limnol. Oceanogr. 14, 799-801 (1969).
35. Eichelberger, J.W., Harris, L.E. and Budde, W.L., Anal. Chem., 46_, No. 7,
June 1975, pp. 995-1000.
36. EightPeak Index of Mass Spectra, Second Edition (1974), Mass Spectrometry
Data Centre, AWRE, Aldermaston, Reading, RG7 4PR, UK.
37. Fieser; L.F. and Fieser, M., "Reagents for Organic Synthesis," Wiley and
Sons, New York, 1967, p. 487.
52
------- |