EPA-R2-72-077
October 1972 Environmental Protection Technology Series
Control of
Mercury Contamination
in Freshwater Sediments
Ill
Office of Research and Monitoring
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and
Monitoring, Environmental Protection Agency, have
been grouped into five series.. These five broad
categories -were"established to facilitate further
development and application of environmental
technology. Elimination of traditional grouping
was consciously planned to foster technology
transfer and a maximum interface in related
fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL
PROTECTION TECHNOLOGY series. Thisf series
describes research performed to develop and
demonstrate instrumentation, equipment and
methodology to repair or prevent environmental
degradation from point and non-point sources of
pollution. This work provides the new or improved
technology required for the control and treatment
of pollution sources to meet environmental quality
standards..
-------
OPT3-tSCH3IClL IKJFORJUIIOH C**Titt EPA-R2-72-077
October 1972
CONTROL OF MERCURY CONTAMINATION
IN FRESHWATER SEDIMENTS
By
George Feick
Edward E. Johanson
Donald S. Yeaple
Contract No. 68-01-0060
Project 16080 GWU
Project Officer
Dr. Curtis C. Harlin, Jr.
National Water Quality Control Research Program
Robert S. Kerr Water Research Center
Ada, Oklahoma 74820
Prepared for
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, B.C., 20402 - Price $2
-------
EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents necessarily
reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names
or commercial products constitute endorsement or
recommendation for use.
11
-------
ABSTRACT
Methods for controlling the release of mercury from sediments have
been developed, and the effects of dredging on the redistribution of
mercury have been evaluated. A program of laboratory studies was
conducted concurrently with a field survey where the extent of
mercury contamination at a typical site was evaluated.
Laboratory studies consisted of both partitioning and aquarium exper-
iments using artificially contaminated sediments as well as sediments
from the polluted field site. Inorganic sulfides and long-chain alkyl
thiols with suitable modifications were found to be the most effective
binding agents. A number of factors were identified which affect the
decision to decontaminate a polluted sediment or to remove the
material by dredging. If the material is to be dredged, precautions
must be taken when land disposal methods are used. The field survey
consisted of determining both the horizontal and vertical extent of the
mercury contamination as well as pertinent hydraulic parameters.
From results of the laboratory and field work, a pilot field project is
described whereby techniques for controlling mercury contamination
can be evaluated at a site where the field conditions have been fully
established.
This report was submitted in fulfillment of Project Number 16080 GWU,
Contract 68-01-0060, under the sponsorship of the Office of Research
and Monitoring, Environmental Protection Agency.
-------
CONTENTS
SECTION PAGE
I CONCLUSIONS 1
II RECOMMENDATIONS 5
III INTRODUCTION
Scope and Purpose 7
Approach 7
IV SEALING OR CHEMICALLY BINDING
MERCURY IN PLACE
Criteria for Evaluating Mercury-
Complexing Agents 15
Measurement of Partition Coefficients 16
Aquarium Studies 21
Cost of Materials 26
Natural Organic Soils 27
Inorganic Sulfides 27
Long-Chain Alkyl Thiols 28
Natural Proteins 29
V DREDGING OF MERCURY-CONTAMINATED
SEDIMENTS 31
VI FIELD STUDIES
Ashland, Massachusetts Site Description 35
Process Description--Nyanza Chemical
Corporation 41
Mercury Disposal Prior to June, 1970 42
Reservoir Description 42
Extent of Mercury Contamination 43
Discussion 48
-------
SECTION PAGE
VII PROPOSED PILOT FIELD PROGRAM
Test Site Description 51
Test Structure ~*2
Dredging Tests 54
Tests of Mercury Bonding and
Sealing Agents 55
Test Procedures 56
Schedule for Field Pilot Program 59
Work Summary 61
VIII REFERENCES 63
IX ACKNOWLEDGEMENTS 65
X APPENDICES
Appendix A - Partition Coefficients 67
Appendix B - Aquarium Experiments 99
Appendix C - Dredging of Mercury-
Contaminated Sediments 115
Appendix D - Physiological Effects of
Organic Thiols 129
Appendix E - Analytical Procedures
and Method Development 133
Appendix F - Field Survey Sample
Collection 141
VI
-------
FIGURES
FIGURE PAGE
1 Decision Sequences for Control of Mercury-
Polluted Water Bodies 8
2 Decision Sequences for Dredging and Treatment
of Mercury-Contaminated Sediments 9
3 Decision Sequences for Sealing or Treating
Sediments in Place 10
4 Plan View- -Framingham Reservoir No. 2 . 39
5 Plan View--Nyanza Chemical Corporation
Relative to the Sudbury River and Framingham
Reservoir No. 2 40
6 Mercury Contour Mappings, 0-2 inches, Ashland
Test Site 44
7 Mercury Contour Mappings, 2-4 inches, Ashland
Test Site 45
8 Mercury Contour Mappings, 4-6 inches, Ashland
Test Site 46
9 Mercury Contour Mappings, 6-8 inches, Ashland
Test Site 47
10 Suggested Frame Construction--Test Structure 53
11 Field Evaluation Test Plan Outline 57
12 Test Structure Layout 58
13 Schedule for Field Pilot Program 60
A-1 Approach to Equilibrium for a Sandy Clay with
Continuous Agitation ?3
A-2 Approach to Equilibrium for a Sandy Clay •with
Continuous Agitation 74
C-l Decrease of Total and Dissolved Mercury as a
Function of Time After Initial Dredging Disturbance 1 19
C-2 Settling Chamber 120
C-3 Settling Velocities Characteristic of Various
Particle Groups 123
C-4 Settling Velocities vs Particle Group 124
C-5 Mercury Content vs Turbidity Levels 126
E-l Microcell for Sampling Output of Gas
Chromatograph 139
vn
-------
TABLES
TABLE PAGE
1 Some Representative Distribution Data for
Mercuric Chloride at 24-25°C 17
2 Some Representative Distribution Data for
Methylmercuric Chloride at 24-25°C 20
3 Summary of Aquarium Data 23
4 Summary of Site Data 36
A-l Partition Coefficients for Acton Sediments
with Mercuric Chloride at 24-25°C 76
A-2 Partition Coefficients for Minerals and Sediment
from Ashland, Mass., at 24-25°C 80
A-3 Partition Coefficients for Pyrite Additives with
Mercuric Chloride at 24-25°C 82
A-4 Partition Coefficients for Various Inorganic
Sulfide Additions with HgCl2 at 24-25°C 86
A-5 Partition Coefficients for Miscellaneous Materials
with HgCl2 at 24-25°C 88
A-6 Partition Coefficients for Long-Chain Alkyl Thiols
with HgCl2 at 24-25°C 90
A-7 Partition Coefficients for Methylmercuric Chloride
with Acton Sediments at 24-25°C 94
A-8 Partition Coefficients for Methylmercuric Chloride
with Various Additives at 24-25°C 95
A-9 Effect of Soluble Chlorides on Partition Coefficient
at24-25°C 97
B-l Extraction of Mercury from Sediment under
Static Conditions 100
B-2 Summary of Aquarium Experiments 103
C-l Simulated Dredging Experiments 116
C-2 Simulated Dredging Experiments 117
C-3 Elapsed Times to Reach Various Turbidity Levels 122
C-4 Settling Velocity as a Function of Height, Turbidity,
and Elapsed Time 122
C-5 Settling Velocities with Revised Elapsed Time 127
D-1 Some Solubilities of Normal Mercaptans and
Normal Alkanes at 20-30°C 131
Vlll
-------
TABLE PAGE
E-l Characteristic Frequencies (cm ) of Some
Mercury Compounds 137
F-l Grab-Sample Analyses--Ashland, Massachusetts
Area 142
F-2 Core-Sample Analyses--Ashland, Massachusetts
Area 145
F-3 Water-Sample Analyses--Ashland, Massachusetts
Area 148
F-4 Fish-Sample Analyses--Ashland, Massachusetts
Area 153
F-5 Water Quality Parameters--Framingham Reservoir
Watershed 154
F-6 Average Flow Volumes, Framingham Reservoir
No. 2 (1968) I55
IX
-------
SECTION I
CONCLUSIONS
1. The behavior and mobility of mercury in natural water and soil
systems are governed mainly by the strong binding of the mer-
curic ion to sediments, suspended particles, and soils. The
binding capacity of solids is conveniently measured by the par-
tition coefficient, which may be defined as the equilibrium ratio
of mercury concentration in solution to that in the solid. The
lower the value of this ratio, the more effective is the mercury-
binding action.
2. The mercury-binding capacity of natural sediments varies widely
and increases with the content of organic matter. A highly or-
ganic peaty sediment may give a partition coefficient on the order
of 10~°, which is about the limit measurable by present analytical
methods. The partition coefficient of a sandy sediment may be
around 10 and of a pure kaolin or silica sediment from 0. 1 to
1. 0.
3. When natural sediments are oxidized (as by mixing, with oxygen-
rich water or exposure of dredge spoil to air), the mercury-
binding capacity is decreased. The capacity is also decreased
by the presence of salt in concentrations similar to that of sea
water.
4. The binding capacity of a given sediment for methylmercurie ion
is several orders of magnitude less than for the mercuric ion.
Since more than 99% of the mercury in most natural sediments is
in the mercuric form, however, the main problem is to bind the
inorganic mercury in a form which is resistant to methylation.
5. The mercury-binding capacity of sediments may be increased by
the addition of sulfur compounds, such as long-chain alkyl thiols,
inorganic sulfides, or natural proteins. Of these, the long-chain
alkyl thiols most nearly meet all the requirements for useful and
practical mercury-complexing agents. These thiols are capable
of producing partition coefficients on the order of 10-8, which is
comparable to the best natural organic sediments we have meas-
ured. The thiols are also useful in binding me thylmer curie ion.
6. The long-chain alkyl thiols can readily be applied to bottom sedi-
ments or to dredge spoils by the use of appropriate surface-active
agents. The sediments so treated are less readily affected by
oxidation than are the natural sediments or sediments treated
with inorganic sulfides. The effectiveness of the thiols in pre-
venting mercury uptake by fish and their lack of toxicity to the
fish have been confirmed by aquarium experiments.
-------
7. The major drawback of the thiols is that they may impart an ob-
jectionable odor or taste to the water. We believe that this
objection can be overcome by the proper choice of materials
and by chemical modification of the thiol group in such a way
that its reactivity toward mercury is retained. The modified
thiols will probably also be useful for complexing other heavy
metals.
8. Plastic films (e. g. , of polyethylene) do not appear to provide an
effective barrier against methylmercuric ion. In conjunction
with chemical sealants, however, films may be useful for re-
tarding oxidation of the complexing agent and retaining it in
place.
9. The uptake of mercury by goldfish in aquariums with contaminated
sediments is less than that of fish in a natural environment with
comparable sediments. The observed difference may be due to
short time of exposure, to greater variability of the natural en-
vironment, or to greater uptake through the food chain. Large-
scale tests will be needed to evaluate any dredging or sealing
technique.
10. Mechanical dredging of mercury-contaminated sediments may
increase local concentrations of waterborne mercury from less
than 1 ppb to values on the order of 0. 1 to 1.0 ppm. Of this in-
crease, less than 1% is in the form of water-soluble mercury.
The remaining 99% represents mercury bound to particulate
matter, which will be redistributed by settling. The sediment
so redistributed will be readily ingested by bottom-feeding fish.
On the basis of laboratory experiments, we estimate that the
amount of mercury resuspended in the water may be on the order
of 10% of that removed with the dredge spoil. Hydraulic dredging
may reduce the amount of material resuspended but will result
in a higher percentage of water in the spoil. The mercury con-
centrations in the runoff water will probably require some reduc-
tion.
11. Mercury-contaminated dredge spoil placed on a landfill may re-
lease mercury due to oxidation and leaching. Release of mercury
may be prevented by proper landfill design to prevent percolation
and infiltration of oxygen-rich water, and by adding long-chain
alkyl thiols to the spoil as it is put into place.
12. If corrective action is contemplated at a mercury-contaminated
site, mercury concentrations in both the horizontal and vertical
distributions should be mapped and the basic hydraulic parameters,
such as velocity and flow volume, determined. This action is
necessary in order to plan a dredging operation which will either
-------
result in the removal of all mercury-contaminated sediment or
will provide quantitative information on the amount of mercury
to be complexed if the contamination is to be treated instream.
13. When dredging of a mercury-contaminated site is required by
navigational considerations, provisions should be made prior
to the operation for adequate land disposal.
-------
SECTION II
RECOMMENDATIONS
1. The site survey conducted in Phase II at the Framingham Reser-
voir in Ashland should be extended over a longer period of time
to account for seasonal variations in water concentration and in
mercury input. The contribution of mercury from the adjacent
landfill should be assessed and monitored over a meaningful
period of time. If possible, mercury balances should be made
in the reservoir system to determine how its mercury content is
changing with time.
2. Analytical studies should be undertaken to determine the specific
form of soluble organic compound believed to be present in the
waters of the Ashland test site. Special attention should be given
to the possible presence of mercurated anthraquinone derivatives.
3. A large-scale test should be conducted at the Framingham Reser-
voir in Ashland to determine the effectiveness of dredging and
sealing methods under field conditions. The natural environment
should be simulated as closely as possible with respect to the food
chain, seasonal variations of mercury concentration, and time of
exposure. The redistribution of mercury during dredging opera-
tions should be measured.
4. The leaching of contaminated dredge spoil from land disposal areas
should be monitored and the effect of added complexing agents
measured.
5. A laboratory program should be undertaken for the development of
chemically modified thiols which will be free of the objectionable
taste and odor of most of the presently available materials. This
should involve the synthesis of new organic sulfur compounds and
laboratory screening for effectiveness as mercury-binding agents.
The effect of these thiols and similar complexing agents on the
rate of methylation in naturally contaminated sediments should be
investigated.
6. A laboratory program to determine the binding action of saltwater
sediments should be undertaken. The effects of oil pollution and
of salt concentration should be measured.
-------
SECTION III
INTRODUCTION
Scope and Purpose
The mercury contamination of some fish in fresh waters of the United
States has been well documented since 1970. In many cases, the source
of mercury has been found to be industrial discharges of various mer-
curic compounds which have accumulated in the sediments. There have
also been cases where the specific source is unknown but may be the
result of a general fallout of mercury from the air. The release of
mercury from the burning of fossil fuels has recently been documented
and is a possible source, both through direct fallout and runoff.
The specific purpose of the JBF mercury program has been to develop
and evaluate both physical and chemical methods of binding the mercury
in the sediments to prevent its release to the overlying water. We have
also investigated the feasibility of removing mercury-contaminated
sediments by dredging and have evaluated in the laboratory the possible
effects of dredging. Realizing that laboratory methods may not always
be a true indication of what may happen in the field, we have in addition
conducted a field investigation of mercury-contaminated sites. One of
these sites was selected for extensive testing, including vertical and
horizontal mapping of mercury concentrations, hydraulic parameters,
and other water quality indicators. Sediments from this site have also
been used in the laboratory evaluation of physical and chemical binding
techniques. As part of the field investigation, a test plan has been pre-
pared for a pilot-scale field evaluation of various binding techniques
and an evaluation of the effects of dredging.
Approach
The program thus far has been divided into two concurrent phases, one
being a laboratory investigation and evaluation of binding and dredging
techniques, the second being a survey of mercury-contaminated sites
and the conduct of an extensive mapping survey at one of the sites. A
third phase, the conduct of a field pilot project, has not yet been per-
formed, although a test plan for this phase has been proposed as part
of Phase II.
In preparing our laboratory and field investigations, we have been
guided by a set of decision sequences which define the mercury con-
tamination problem and show how various actions are related. When
a mercury pollution problem is suspected, it becomes necessary to
identify the nature and magnitude of the problem and to decide on an
appropriate course of action. The steps involved in arriving at such a
decision are outlined in Figures 1, 2, and 3.
-------
1. Known mercury sources--natural, industrial,
Identify landfill.
Polluted 2. Analysis of water.
Areas 3. Analysis of fish and other biota.
4. Analysis of sediments
survey
of
Site
1. Estimate history of input-output.
2. Identify continuing sources of mercury--
landfill, mine dumps, etc.
3. Map classification of sediments and concentra-
tion of Hg in three dimensions.
4. Estimate limnographic parameters and probable
future conditions.
Identify
Environ-
mental
Effects
1. Drinking -water standards- -5 ppb Hg.
2. Limits on edible fish--0. 5 ppm Hg.
3. Effects on shellfish and bird life.
4. Other effects on biota, including plant life.
Evaluate
Alterna-
tives
V
1. Estimate cost benefits for various courses of
action.
2. Estimate side effects- -both environmental and
for Action economic.
i
(
1. No
action
required
>
i
2. Con-
tinue moni -
toring.
V
> 1
3. Dredge
(See Figure
2.)
i
4. Chemically
treat in place.
(See Figure 3. )
Figure 1. Decision Sequences for Control of
Mercury-Polluted Water Bodies
-------
Consider special needs for dredging (e.g., navigability),
and estimate environmental impact.
X
Effects of dredging on aquatic
environment: oxygen demand,
mercury release, turbidity.
I
Effects of spoil disposal: re-
lease of mercury to ground or
surface waters.
Chemically treat before
dredging (Figure 3)
Cover disturbed or silted
areas after dredging.
1
No special treat-
ment required.
Landfill:
estimate en-
vironmental
effects.
Dump at sea
if accessible
and if per-
mitted.
Process for
mercury re-
moval before
dumping.
Chemically
immobilize
mercury after
placing (see
Figure 3).
Chemically treat
before dumping
to immobilize
mercury ( see
Figure 3).
Seal in place
with imper-
vious top and
bottom layers
or membranes.
Control and
chemically
treat runoff
water.
Combination
of the fore-
going.
Figure 2. Decision Sequences for Dredging and Treatment
of Mercury-Contaminated Sediments.
-------
Evaluate Cost Effectiveness of Various Treatments
1. Cost of materials at site.
2. Methods and costs of deployment.
3. Side effects on biota and environment.
Mineral Coverings:
1. Sand, clay, ground quartz, mine tailings, etc.
2. Combinations of the above with chemical treatments.
Natural'Organic Coverings:
1. Peaty sediments, sawdust, protein, hair, feathers, etc.
2. Combinations of the above with mineral covers and/or
added sulfides.
3. Evaluate biochemical and water quality problems.
Inorganic Sulfides: FeS^, FeS, ZnS
1. Prevent oxidation by inert cover or by organic additives.
2. Effects of chlorides if present.
3. Effects on water quality and environment.
Organic Sulfides
1. Choose molecular weight and structure.
2. Chemically modified thiols.
3. Evaluate costs, deployment, sinking agents.
4. Water quality and environmental effects.
Figure 3. Decision Sequences for Sealing or
Treating Sediments in Place
10
-------
Referring to Figure 1, we begin with the assumption that a polluted
area has been identified and confirmed by analysis. For this purpose,
the analysis of mercury in predatory fish and organic sediments is
most useful, since they concentrate mercury entering the water. We
also assume that the sources of mercury have been stopped as far as
possible. Even after the discharge of mercury has ostensibly been
checked, it is still possible for mercury to enter the system from such
sources as landfills, mine tailing dumps, and sediments in tributary
streams.
A site survey should then be performed and should include a study of
the distribution in depth of mercury in various classes of sediment.
The hydraulic parameters should be defined sufficiently to permit an
estimate of the future conditions in the water and sediments if the sys-
tem is left to itself.
The environmental effects of the mercury should also be considered,
including not only established standards for water quality and edible
fish and shellfish, but also such additional effects as those on bird life
and other biota, including plant life.
With this background we can evaluate the cost benefits of various courses
of action and make an estimate of the probable side effects, including
both environmental and economic considerations. The major alternatives
appear to be: (1) take no action; (2) continue monitoring for future pre-
dictive information; (3) dredge; and (4) chemically treat the sediments
in place. These last two alternatives are considered in Figures 2 and
3, respectively.
Figure 2 shows some of the decision sequences involved in dredging a
contaminated site. In some cases the navigability of the waters will be
the overriding consideration, and it will be necessary to dredge to main-
tain water depth.
If dredging is decided upon, it will be necessary to consider the effects
of disturbance and possible mercury release on the water and on the
biota. An alternative here is to chemically treat the sediments before
dredging to minimize mercury release. After dredging, the undisturbed
and/or silted areas may be further treated with chemicals or sealants
to minimize the effects of freshly exposed mercury.
The handling and disposal of the contaminated dredge spoil present an
especially severe set of problems. It may be necessary to impound or
treat the runoff water before returning it to the source. If the spoil
can be dumped at sea, a considerable economic advantage may be ex-
pected. Feasibility will depend on location and on regulations govern-
ing disposal at sea. Recent observations indicate that biological
activity is reduced by a factor of 10 to 100 in the ocean depths [l],
and it can be inferred that biological methylation and oxidation will
probably be minimal. Transportation of the spoil to deep water, how-
ever, would increase the costs.
11
-------
If the spoil is to be disposed of in a landfill, the question of mercury
release becomes crucial. Our work indicates that mercury is best
kept in insoluble form under anoxic conditions. In a landfill, unless
the cover material is impervious and well drained, oxidizing conditions
will eventually prevail as oxygenated surface waters percolate through.
Mercury may then be released in soluble form in the leachate. Careful
design of the landfill can prevent this.
One alternative solution would be to remove the mercury from the spoil
before dumping. The low concentration of mercury (typically a few
hundred parts per million) and the colloidal character of the spoil
make this alternative unattractive.
A second alternative is to treat the spoil with chemicals before dumping,
in order to immobilize the mercury. This affords a convenient oppor-
tunity to secure good mixing with the treating agents. The various types
of chemical treatment are listed in Figure 3.
A third alternative is to dump the spoil without treatment and to try to
contain the mercury by proper design of the landfill. This might involve
chemical treatment of selected areas, sealing with impervious layers,
control and treatment of runoff waters, or a combination of these.
Containment of mercury will be easier in arid regions than in areas
of high rainfall.
The types of chemical treatment involved in sealing mercury in place
are shown in Figure 3. Cost-effectiveness considerations include cost
of materials delivered at site, methods and cost of deployment, and
possible side effects on aquatic biota and water quality.
At present we have considered four types of treatment, which can be
used alone or in combination. The first alternative is to cover the
sediment with mineral coverings, such as sand, clay, or other fine
mineral material. Although these materials have little mercury-
binding capacity in themselves, they can prevent disturbance of or-
ganic sediments and aid in maintaining anoxic conditions. If the con-
taminated sediments to be covered contain enough sulfide, the mercury
will be adequately immobilized.
In cases where little or no natural organic matter is present, it may be
desirable to add such materials in the form of natural peaty sediments
or proteinaceous materials. Because of their low density, the organic
coverings are easily disturbed and slow to settle. It may prove neces-
sary to cover them with a denser material.
Among the inorganic sulfides which may be considered as mercury-
binding agents are pyrite (FeS2), ferrous sulfide (FeS), and sphalerite
or zinc sulfide (ZnS). Pyrite is a cheap by-product of ore-dressing
operations, but our experiments indicate that it is less effective than
FeS or ZnS.
12
-------
The organic thiols, in the form of high-molecular-weight alkyl mercap-
tans, are among the most effective mercury-binding materials -we have
observed. The thiols of interest are oily liquids which float on water.
For these reasons we have given careful consideration to methods
deployment of these materials. One promising method is to absorb
them on hydrophobic oil-sinking agents, of which a large variety are
known.
Another possibility is to chemically modify the thiols in order to tem-
porarily inactivate the sulfhydryl group, thus making the odor less
offensive. In the presence of mercury, the thiol will be regenerated
and will bind the mercury. Preliminary investigations along this line
have produced encouraging results.
An attractive feature of the thiols is that they are potentially effective
in very low dose rates. We estimate that about 200 ppm of thiol in the
sediment will be sufficient to bind 100 ppm of mercury. Thus it will
usually be possible to effectively bind the mercury at levels of thiol
treatment which will not exceed EPA guidelines for oil-polluted sedi-
ments. If the thiol is used as a barrier layer, even less may be re-
quired. In any case, the effects of these materials on the environment
and on water quality must be carefully evaluated.
The decision processes discussed above show that many of the possible
cures to a mercury problem involve either binding the mercury in place
by chemical or physical means or removing it by dredging. Each of
these actions has been shown to have some other consequence which
must also be evaluated. We have considered these relationships in the
design of our laboratory and field program.
13
-------
SECTION IV
SEALING OR CHEMICALLY BINDING MERCURY IN PLACE
This section presents a discussion and comparison of various physical
and chemical means of immobilizing mercury at the bottom of contam-
inated water bodies. Since no method of treatment can produce perfect
immobilization, a realistic objective will be to reduce the rate of mer-
cury release to values which will not appreciably affect the biota in the
overlying or downstream waters.
Many sediments in the natural states are found to bind mercury very
strongly. For this reason, it is necessary that any method of treatment
be highly effective in order to reduce the natural rate of mercury release.
This section covers first the requirements for effective mercury-binding
agents suitable for application on a large scale. This is followed by a
discussion of laboratory experiments designed to evaluate various mater-
ials in terms of these requirements. Finally, the cost of several materials
for a typical application is estimated.
Criteria for Evaluating Mercury-Complexing Agents
The following list is proposed as covering the main requirements for
mercury-complexing agents to be distributed in contaminated waters:
1. The equilibrium constant for the formation of the
mercury complex should be as high as possible.
2. The resulting mercury complex should be ex-
tremely insoluble in water.
3. The complex should be stable toward oxidation,
reduction, hydrolysis, biological action, and
the presence of dissolved salts such as chlorides.
4. The rate of reaction with mercury at very low
concentrations should be reasonably high.
Preferably the reaction should proceed substan-
tially to completion in a few days or weeks.
5. The material used should not adversely affect
the quality of the water or the bottom sediment
for its intended uses. This includes effects on
fish and bottom biota in areas where such con-
siderations are important.
6. The material should be readily convertible into
a dense, granular form that will sink quickly and
will not readily be dispersed into the water.
15
-------
7. The cost of the material, in place at the bottom
of the water and per unit of mercury complexed,
should be as low as possible.
The first four of the above requirements were evaluated by measuring
the partition coefficients for mercury between water and various treated
sediments. The practical application of mercury-binding agents and
their possible side effects on biota were studied by means of aquarium
experiments with goldfish.
Measurement of Partition Coefficients
Although it is known that the transport of mercury in natural water
systems is sharply limited by its strong absorption on soils and sedi-
ments, little quantitative data on the absorption has heretofore been
available [2], Such data is needed to predict and control the move-
ment of mercury in water and soil systems as well as to evaluate the
effects of mercury- complexing agents. A useful and quantitative
measure of mercury absorption is provided by the partition coefficient,
which, for purposes of this report, is defined as the equilibrium ratio
of mercury concentration in solution to the concentration in the solid.
The lower the numerical value of this ratio, the more effective is the
mercury-binding action of the solid.
The partition coefficients of mercuric chloride and of methylmercuric
chloride have been measured for a variety of natural sediments and
minerals with and without mercury- complexing additives. Details of
this work are given in Appendix A; the main results and conclusions
are discussed below. Some typical data for mercuric chloride are
shown in Table 1.
Runs A- 16 and B-4 were made with fresh Acton sand and fresh Acton
peat, respectively. The mercury- binding action of the sand is relatively
low, while that of the peat is one of the best we have measured. The
effectiveness of the peat may be due to the presence of sulfides and to
the highly reducing conditions in the organic sediment.
Runs A- 36 and B-18 show the results obtained when the same materials
were aged in air for five to eight weeks before testing. In both cases,
the binding capacity is diminished, probably because of oxidation. This
result indicates that mercury-containing dredge spoils may release
mercury if they are placed on a landfill where they can become per-
meated with oxygen-rich surface waters.
Runs C- 1 and C-7 were made with Georgia kaolin. The addition of
to raise the pH in the latter run made some improvement, but the binding
capacity is low in both cases.
16
-------
Table 1
Some Representative Distribution Data for Mercuric Chloride at 24-25 C
Run
No.
A- 16
A- 36
B-4
B-18
C-l
C-7
S-l
C-5
C-13
C-25
C-19
C-51
C-27
C-28
C-2
B-ll
CF-6
Time
(days)
7
7
7
7
8
7
7
7
7
7
7
7
7
7
7
7
7
Description
Fresh Acton
sand
Aged Acton
sand
Fresh Acton
peat
Aged Acton
peat
Clay (Georgia
kaolin)
Clay plus 5%
CaCO3
Ground silica,
about 240 mesh
Clay plus 3%
coarse pyrite
Clay plus 5%
milled pyrite
(-325 mesh)
Clay plus 1%
FeS (fired pyr. )
Clay plus 1%
ppt. FeS
Clay plus 5%
ppt. ZnS
Clay plus 1% n-
dodecyl mercpt
+ 5% CaCO3
Same as C-27
+ 3. 5% NaCl
Ppt. FeS +
3.5% NaCl
Fresh Acton
peat + 3. 5% JfeCl
Chickn. feathrs.
Mercury Cone, (ppm)
Dry
Sediment
412
258
1430
1335
82
314
33
193
300
321
378
300
1000
300
86
800
1780
Water
0.52
10.0
<0. 00002
0. 0031
40. 1
11.5
51.5
31.8
0. 0025
0.154
0. 168
0.00053
0.00002
0.00006
26
0.004
0. 140
&^H20
K"^++JSed.
1. 3 x 10"3
0.037
cl.4 x 10"b
2. 32 x 10-6
0.49
0.037
1.56
0. 165
8. 3 x lO-6
4. 8 x 10-4
4. 5 x 10-4
1.77 x 10-6
2.0 x 10-8
2.0 x 10-7
0.30
5. 0 x 10-6
7.87 x 10-5
PH
6.2
.5-7
5.2
4.8
5.2
7.4
6.8
5.0
4.5
5.0
4.3
5. 1
6.8
7.2
4. 5
4.8
'6. 5
Dissolved
Oxygen
(ppm)
6.5
0.0
0.6
5.0
7.5
6.0
10.0
7.1
9.5
4.0
7.0
11.5
8.5
0.0
0.9
17
-------
The use of ground silica to cover mercury- contaminated sediments in
pond bottoms has been proposed. Run S-l shows, however, that this
material has practically no binding capacity.
Because of the low natural mercury-binding ability of kaolin, we have
used this material as a substrate for testing a number of additives.
Run C-5 shows the effect of adding 3% of coarse (84% over 200 mesh)
pyrite to kaolin. Only a slight improvement over straight kaolin (run
C-l) is observed. Since mercuric sulfide is known to have a low
solubility, it appears that very little reaction has taken place in seven
days and that the reaction is probably kinetically limited.
A number of experiments were made in the attempt to increase the
reaction rate of pyrite. One of the easiest and most effective methods
was to mechanically mill the pyrite to a very fine powder (100% through
325 mesh). This produces a major improvement in partition coefficient,
as shown by run C-63.
Another way of modifying pyrite is to heat it in the absence of air to a
temperature above about 700 °C, when one atom of sulfur is lost, ac-
cording to the reaction:
FeS2 — - FeS + S
A sample of calcined pyrite was made by this method which was estimated
by weight loss to be about 35% converted to FeS. Run C-25 shows that
this treatment produces some improvement over coarse pyrite but not as
much as fine grinding.
The effect of particle size on the reactivity of FeS is shown by run C-19»
in •which the FeS was precipitated in situ by reaction of FeSO^. with CaS.
This material is somewhat less active than fired pyrite, even in the
presence of somewhat less dissolved oxygen. If the oxygen is reduced
to 1 ppm, however, this material is greatly improved (Appendix A,
Table A-4).
Run C-51 shows that precipitated ZnS (laboratory reagent) is somewhat
more effective than milled pyrite or precipitated FeS.
The solubility of mercuric sulfide in near-neutral solutions is controlled
by its hydrolysis according to the equation:
HgS + 2H2O — - Hg + 2OH" + H2S
Since H2S is very slightly ionized, the concentration of Hg in the
presence of HgS is greater than its solubility product (10~53. 5) -would
indicate. Hydrogen sulfide is both soluble and reactive and is therefore
readily lost from the reactive zone by diffusion and oxidation. In this
manner, the hydrolysis of HgS can progress until an appreciable con-
centration of mercuric ion is reached.
18
-------
As an alternative to the inorganic sulfides, we considered the long-
chain alkyl thiols. These thiols form insoluble mercury compounds
according to the reaction:
Hg++ + 2RSH —^ Hg(SR)2 + 2H+
Like H2S, the thiols are very weak acids, and their mercury compounds
are subject to hydrolysis. Unlike f^S, however, the long-chain thiols
are highly insoluble in water and tend to remain in the reaction zone,
where they continue to be effective in preventing progressive hydrolysis.
Several long-chain alkyl thiols (mercaptans) are commercially available.
Normal dodecyl mercaptan was chosen for most of this work because its
cost is moderate and its odor is relatively low. Run Cr27 shows the ef-
fect of 1% n-dodecyl mercaptan buffered with calcium carbonate. This
is better than the inorganic sulfides by almost two orders of magnitude
and is almost as effective as the Acton peat.
An important feature of mercury-complexing agents in some environ-
ments is their ability to function in the presence of salt or brackish
water. Run C-28 shows that the effectiveness of the mercaptan de-
creases by about a factor of 10 in the presence of 3. 5% NaCl. Under
the same conditions, the precipitated FeS is five orders of magnitude
less effective, and the Acton peat is two orders of magnitude less ef-
fective, as shown by runs C-21 and B-ll. The mercaptan is by far the
best material we have found for use in a saltwater environment.
Another approach to the bonding of mercury is the use of natural protein-
aceous materials, such as wool, which has been studied by Friedman
et al. [3], Run CF-6 shows the results of reacting mercuric chloride
solution with chicken feathers, which are a cheaper source of protein
than wool. The distribution ratio obtained agrees well with the data of
for wool but is not as good as the data for sulfides or mercaptans. As
will be shown in a later section, the low capacity of feathers renders
them uneconomic in comparison with the sulfides or mercaptans.
Table 2 gives the results of a variety of runs made with methylmercuric
chloride. The materials which gave the best results with HgCl? gener-
ally yield the best results with CH^HgCl, but the distribution ratios are
less favorable by several orders of magnitude. Fortunately, the methyl-
mercury content of most contaminated sediments is less than 1% of the
total mercury (see below for example). The main problem, therefore,
appears to be to immobilize the inorganic mercury. To do so effective-
ly, it is desirable to bind the inorganic mercury in a form which will
not be appreciably methylated. We recommend that the effect of thiols
and similar complexing agents on the rate of methylation of mercury be
investigated.
Further measurements have been made on a sediment from Framingham
Reservoir No. 2 in Ashland, Massachusetts, which is believed to have
19
-------
Table 2
Some Representative Distribution Data for
Methylmercuric Chloride at 24-25 C
Run
No.
B-20
B-14
B-15
C-32
C-33
C-60
C-62
C-64
Time
(days)
7
7
7
7
7
7
7
7
Description
Fresh Acton
peat
Fresh Acton
peat
Aged Acton
peat
Kaolin clay-
Kaolin clay
Clay plus
5% ZnS
Clay plus 1%
n-dodecyl
mercaptan,
plus 5% CaCOj
!
Clay plus 5%
pyrite (-325
mesh)
Mercury Cone, (ppm)
Dry
Sediment
1470
1
1
2630
2860
382
842
300
300
300
Water
1.0
2.76
6. 5
470
1665
0.45
0.24
37. 5
&**\20
^^sed.
6.8 x 10"4
1. 05 x 10~3
2. 27 x 10"3
1. 23
1.98
1. 5 x 10"3
8. 0 x 10"4
. 125
PH
5.2
5. 1
5.3
5. 1
5.0
5.4
7.6
4. 1
Dissolved
Oxygen
(ppm)
0.2
0.4
0.4
9.0
9. 1
2.8
20
-------
been contaminated with mercury by discharges from a dye-manufacturing
plant. One sample of this sediment contained about 100 ppm of total
mercury and 0.428 ppm of methylmercury. The partition coefficient
for total mercury in this sediment is about 2. 7 x 10~5.
Recent experiments indicate that a major part (on the order of 80%) of
the mercury in this sediment may be organically bound, although less
than 1% is in the methylated form. We now believe that much of the
mercury in this sediment may be in the form of mercurated anthraquin-
one derivatives, which are a probable by-product of the dye manufactur-
ing process.
We recommend that Phase III of this program include analytical studies
aimed at identifying the specific forms of mercury present in the sedi-
ment and in brackish-water sediments frem the vicinity of a dye plant
in Dighton, Massachusetts. Such information will be needed to study
the binding of mercury in these sediments and to evaluate the results
of the pilot-scale experiments. The problem of identifying these chem-
ical species is further discussed in Appendix E, which covers work on
analytical methods.
Aquarium Studies
The aquarium studies were intended to supplement the results of the
distribution experiments under conditions more closely approximating
field conditions. The behavior of mercury in an actual water body will
be governed not only by equilibrium conditions but by rates of diffusion
and reaction and by flows of mercury and of natural mercury-complexing
materials through the system.
The aquarium experiments may also serve to screen out materials which
are toxic to fish. Procedures and results are summarized below and
discussed in detail in Appendix B.
The experiments were conducted in five-gallon glass aquariums, eight
inches by 14 inches by 10 inches deep. A one- or two-inch layer of
sediment containing HgCl? or CH^HgCl plus any required mercury-
binding agents was added and allowed to stand about a week. A cover
layer was then added, and the aquarium was carefully filled with water.
The aquarium was allowed to stand one or two days before the fish •were
added. The experiment was started by adding three or four (depending
on size) goldfish about two inches long. The aquariums were aerated
with bubblers during the tests.
The fish were fed about every other day with a commercial fish food
containing about 20% protein. Our analysis showed negligible amounts
of mercury in the food.
21
-------
After nine day's exposure, the fish were killed, gutted, and the heads
and tails removed. The remaining portion was then analyzed for mer-
cury. New fish were then added to the tank, exposed for about 30 days,
and analyzed in the same way. Some runs were made with different
periods of exposure, as noted in Appendix B. The water was period-
ically analyzed for soluble mercury while the fish were being exposed.
The effectiveness of mercury binding by the sediment was judged by the
mercury uptake of the fish and by the concentration of soluble mercury
in the water.
In all the cases the dissolved-mercury concentration decreased markedly
with time, and the uptake of mercury by the fish was usually less during
the second period of 30 days than in the first exposure of nine days. The
loss of mercury from the water was approximately equal to the uptake by
the fish
The data obtained for various sediments and additives is summarized
in Table 3.
The Acton peat was the best natural mercury-binding sediment found,
as might be expected from the low value of the partition coefficient.
With 185 ppm of Hg in the sediment, the fish took up only about 1/2 ppm
during the 30-day exposure (run C). This appeared to be due mainly to
ingestion of the sediment by the fish. When the Acton peat was covered
with 1/2 inch of sand, the fish lost mercury during both the nine-day
and 30-day exposures (run D).
The Acton sand showed low mercury-binding capacity (runs A and S),
and the fish took up mercury rapidly. One inch of clean sand cover
(run B) lowered the concentration of mercury in the water and produced
a loss from the fish during the 30-day test. Runs E and F showed that
1/2 inch of Georgia kaolin or of 240-mesh silica were relatively inef-
fective as covers. These materials were readily stirred up by the fish,
and both tanks were turbid for the duration of the test.
Run G shows the effect of a thin layer of precipitated zinc sulfide.
Although the fish lost mercury during the first nine days, there was
a larger gain in the 30-day test. This may have been due to the oxida-
tion of sulfides by long contact with aerated water.
Runs I, K, and L show that milled pyrite, fired FeS-ZnS mixture, and
fired FeS were less effective than ZnS.
The effect of the long-chain alkyl mercaptans is shown by runs H, J,
and V. Although the concentrations of mercury in the water are not
particularly low, the uptake by the fish is generally less than with the
inorganic sulfides. Run V showed low mercury levels in the water
after the first run of 21 days. Time did not permit a second run to be
made in this aquarium.
22
-------
Table 3. Summary of Aquarium Data
Run
No.
C
D
A
S
B
E
F
G
I
K
L
H
Bottom
Sediment
Acton peat
Acton peat
Acton sand
Acton sand
Acton sand
Acton sand
Acton sand
Acton sand
Acton sand
Acton sand
Acton sand
Acton sand
Hg Content
ppm
185
100
100
100
100
100
100
100
100
100
100
100
Added as
HgCl2
HgCl2
HgCl2
HgCl2
HgCl2
HgCl2
HgCl2
HgCl2
HgCl2
HgCl2
HgCl2
HgCl2
Cover Layer
None
1/2" clean sand
None
None
1 " clean sand
1/2" Ga. kaolin
1/2" 240-mesh
silica
. 015 lb/ft2 ZnS
. 0291 lb/ft2
milled pyrite
.015 lb/ft2
ZnS-FeS
.015 lb/ft2 FeS
.0051 lb/ft2 (b)
MTM on carrier
Hg Concentration in
Water (ppm)
Initial
. 0004
.048
0. 18
.00055
. 032
.074
.0018
3 days)
. 0407
. 0204
. 078
.0035
9 Days
. 00037
. 000077
. 0049
. 00025
.--
. 0096
. 0063
.0076
.001
Final
.000056
.000055
(18 days
. 0002
.00012
. 0003
.0008
. 0008
(28dys
. 0075
. 0049
. 0008
. 0036
;23dys
Mercury Uptake by Fish
(ppm, wet basis)
First Set of
Fish(9 days)
+0. 144
-0. 155
+ 29.9
+6. 0 (2 days)
+ 1. 72
+13.96
+ 10.83
-0. 06
+ 1. 81
+ 11.7
+ 16. 3
+0. 53
Second Set of
Fish (9 days)
+0.47
-0.48
+ 1.98
+0. 064
+0.42
+ 1. 30
+ 3.41 (19 days)
+ 5.78
+ 14. 4
+ 20. 2
+ 0. 83 (19 days)
Remarks
Sediment in-
gested by fish.
Sand cover
prevented in-
gestion.
All fish died
in 2 days.
N)
-------
Table 3 (continued)
lun
No.
J
V
N
M
R
T
U
Q
Sedinient
Acton sand
Acton sand +
NDM, CaCO3
Acton sand
Acton sand
Acton sand
Acton sand
Acton sand
Ashland
12/71
Hg Content
ppm
100
100
30
30
30
30
30
100.5
Added as
HgCl2
HgCl2
CH3HgCl
CH3HgCl
CH3HgCl
CH3HgCl
CH3HgC
(c)
Cover Layer
.0247 lb/ft2 ,, .
NDM on sand1 '
None
None
. 0247 lb/ft2
NDM on sand
Polyethylene
film, . 001 "thck
Polyeth. film
over milled py-
rite .0291 lb/ft2
Polyeth. film
over NDM on
sand , . 0247
lb/ft2
None
Hg Concentration in
Water (ppm)
Initial
.0045
. 0016
4.6
.048
0.45
. 046
.021
(d
. 0009
9 Days
.0035
00045
21 days)
4. 0
.035
.012
(lOdays
.010
(10 days
.0003
Final
.001
2.9
.024
.002
.003
. 0003
Mercury Uptake by Fish
(ppm, wet basis)
First Set of
Fish (9 days)
+ 0.78
0.90(21 days)
(16.7(4 hrs)
+ 11. 1
+6. 0 (6 hrs)
+ 7. 0(10 days
+ 3.8 (lOdays
+ 0. 18
Second Set of
Fish (9 days)
+0.83
+ 12.7
+2.0
+ 1.4
-0.05
Remarks
1st exposure
was 21 days.
All fish died
in 4 hours.
All fish died
in 6 hours.
ro
NOTES: (a) See Appendix B for more detail on these runs.
(b) MTM = mixed tertiary mercaptans
NDM = normal dodecyl mercaptan
(c) Contaminated sediment, no additional mercury added.
(d) These analyses may be somewhat low; we later found that much of the Hg
is organically bound.
-------
Runs N through U were made with 30 ppm of methylmercuric chloride
in Acton sand. This level is considerably higher than found in any-
natural sediment of which we are aware, but the data afford rapid
comparison of various materials.
Run N was made with no cover, and the fish all died within four hours.
The use of 0. 0247 lb/ft^ of n-dodecyl mercaptan (run M) lowered the
initial concentration of mercury in the water about 100 fold. The fish
survived both the nine-day and the 30-day tests, although the uptake of
mercury was large.
A cover of 1-mil polyethylene film (run R) sealed at the edges with a
little clean sand was less effective than the mercaptan; all the fish died
within six hours.
The bests results with methylmercury were obtained with polyethylene
film over milled pyrite or n-dodecyl mercaptan, as shown in runs T
and U.
Run Q shows the results obtained with a contaminated sediment from
the Framingham reservoir in Ashland, Massachusetts. This sediment
probably contains both inorganic and organically bound mercury in the
form of mercurated anthraquinone derivatives. During the 30-day run
with this sediment, the fish lost mercury, although fish caught in the
reservoir itself have analyzed from 0. 5 to over 7. 0 ppm of mercury.
This indicates that the aquarium test does not adequately duplicate the
conditions in the reservoir. The difference may be due to the reservoir
fish ingesting sediment over long periods of time or to other mercury-
contaminated food. There is also a continuing and variable mercury
input to the reservoir from a landfill upstream, which may contribute
to the problem.
In terms of the requirements for practical mercury-complexing agents,
the aquarium results indicate that both the inorganic sulfides and the
long-chain thiols are capable of markedly reducing the concentration of
water-soluble mercury. The thiols are generally more effective than
the inorganic sulfides, although neither has produced results equivalent
to Acton peat in these short-term tests. Neither class of material
showed any toxic effect on the fish in these tests. The toxicology of the
thiols is further discussed in Appendix D.
The thiols, especially the mixed tertiary mercaptan, impart an objection-
able odor and taste to the water. We believe this problem may be over-
come by carefully selecting the type and purity of the thiol used. In
addition, the odor may be eliminated by temporary chemical masking
of the thiol group in such a way that its reactivity toward mercury is
not impaired. We recommend that this approach be further investigated.
Under recent guidelines issued to regional representatives by the En-
Environmental Protection Agency in February, 1971, a concentration
25
-------
of up to 1500 ppm of oily material is permitted in a dredge sediment
before it is classified as "polluted with organic matter. " If 1000 ppm
of a mercaptan was added to an otherwise oil-free and unpolluted sedi-
ment, the resulting mixture would be well within EPA guidelines, and
the mercaptan would have the theoretical capacity to bind about 500 ppm
of mercury.
The inorganic sulfides are most effective when they are in finely divided
form, such as milled pyrite (-325 mesh) or as precipitated ZnS. In
this form the sulfides do not sink rapidly, and they are readily dispersed
into the water. It is not clear how these problems can be overcome
-while still maintaining the required degree of reactivity.
The alkyl thiols, being liquid at ordinary temperatures, can readily be
coated onto sand by means of cationic surface-active agents. In this
form they sink rapidly and are not readily redispersed or released from
the treated sand. Thus, the thiols appear to be preferable to the inor-
ganic sulfides from the standpoint of deployment.
Cost of Materials
The final requirement for a useful mercury-binding agent is that its
cost should be moderate. In this section we discuss some preliminary
estimates of the cost of materials for a typical situation. No attempt
is made to estimate heavily site-dependent costs, such as for dredging
or for moving large amounts of sand or earth cover. For purposes of
these estimates, we consider four typical classes of mercury-binding
materials:
1. Natural organic soils
2. Inorganic sulfides
3. Long-chain alkyl thiols
4. Natural proteins
As a basis for these estimates, we will consider the upper basin of the
Framingham Reservoir, which is estimated (see Section VI) to contain
about 250 Ibs of mercury in seven acres of bottom, or an average of
36 Ibs per acre. We assume that we wish to lower the mercury content
of the water in the upper basin to a level which will permit raising edible
fish. Since the maximum level of mercury permitted by the Food and
Drug Administration is 0. 5 ppm, and the fish, in general, will concen-
trate mercury by a factor of about 3000, the maximum permissible
concentration in the water is taken as 0. 167 ppb. No mercury input
from upstream is assumed, and the mercury content of the water column
is considered negligible. The mercury will be considered as mercuric
ion, although recent work with sediments in this reservoir indicates
that this is probably not the actual case.
26
-------
Natural Organic Soils
The Acton peat will be considered as a typical organic soil (44. 3%loss
on ignition--see Appendix A). If we take a partitio'n coefficient of
5.3x10"' for this material (run B- 5, Table A- 1), we obtain a figure
of 315 ppm of mercury in the peat in equilibrium with 0. 167 ppb in the
water. At this level it •will require 1. 14 x 10^ Ibs of dry peat per acre
to bind the 36 Ibs of mercury. At an estimated moisture content of
79%, this is equivalent to 270 short tons of wet peat per acre. At a.
wet density of 65 lbs/ft-3, this is equivalent to a layer 2. 3 in. thick.
The above figure represents a worse case, since it takes no account
of the absorptive capacity of the layer of organic sediment already
existing in the upper basin.
This layer of peat should be covered with a layer of sand to prevent
re suspension of the peat in the water and to prevent ingestion by fish,
as well as to maintain anoxic conditions in the bottom. If we use 1/2 in.
of a material similar to Acton sand (wet density about 130 lbs/ft^), we
•will require 118 short tons per acre.
The cost of these covering materials will depend heavily on their avail-
ability at the site and on the means used for deployment. If the reser-
voir could be drained, the cover could probably be spread with road-
grading equipment. Otherwise it would have to be dropped into the
water. This latter operation would be much more difficult to control.
In view of these factors, no detailed estimate of costs will be attempted.
Because of the large tonnages of materials involved, however, the cost
is expected to be high.
Inorganic Sulfides
For estimating purposes, we consider the case of precipitated ferrous
sulfide, FeS. Run C-20 (Appendix A, Table A-4) gives data for FeS
formed in situ by reaction of CaS and FeSO^- 7H2O in the presence of
clay. The concentration of Hg in the water was less than 0. 2 ppb, and
we assume it meets the requirement of 0. 167 ppb. The amount of FeS
present in this experiment is estimated to be 0. 316 g, and this has re-
moved 0. 0378 g of mercury from solution. The experimental ratio of
FeS/Hg is, therefore, 8. 36 Ibs per Ib, or about a 19-fold excess over
the theoretical. There is, therefore, reason to expect that this ratio
may be improved with further work, but we will use this conservative
figure for the present estimate. Thirty-six pounds of mercury will
require 300 Ibs of ferrous sulfide.
The cost of sufficiently reactive ferrous sulfide has not been adequately
explored at present, but, if we assume a reasonable figure on the order
of $l/lb, the cost of $300 should be relatively minor as compared to the
cost of moving 270 tons of peat. If milled pyrite could be used in place
27
-------
of ferrous sulfide (run C-82, Table A-4), a cost of $200 per ton, or
about 10£ per pound, might reasonably be projected. This would reduce
the material cost to about $30 per acre.
The data of runs C-20 and C-82 were obtained under anoxic conditions,
and it is anticipated that any sulfide will have to be covered to prevent
oxication. The cost of 118 tons of covering sand will, therefore, be the
same as in the case of the Acton peat. Some economies might be
achieved if the reservoir could be drained and the sulfide harrowed or
plowed into the bottom sediments. For the Framingham Reservoir this
would seem to be a relatively low-cost operation as compared with try-
ing to place and cover the sulfide under water.
Long-Chain Alkyl Thiols
The data obtained on the organic thiols as mercaptans (Tables A-6 and
A-9) show that these materials are easily capable of reducing the con-
centration of dissolved mercury to levels below 0. 167 ppb, even in the
presence of dissolved oxygen and of much chloride. Aquarium experi-
ments to date show no toxic effect on goldfish, and a review of the liter-
ature indicates that, because of their extreme insolubility, no toxic
effects are to be expected (see Appendix D). Under anaerobic conditions
no biological degradation is anticipated.
A major drawback of the thiols is their objectionable odor, which may
be imparted to the overlying water. We believe that this problem may
be overcome by using certain chemically modified compounds, in which
the thiol group is temporarily masked but is available for reaction with
mercury under appropriate conditions. Some of these masked thiols
have very little or no objectionable odor.
A second drawback of the thiols (and the modified thiols) is that they
are oily liquids with a density less than that of water. In order to dis-
tribute them at the bottom of the water, they must, therefore, be ab-
sorbed on porous or oleophilic materials, which will carry them to the
bottom. A number of oil-sinking agents have been developed for treat-
ing oil spills and should be readily adaptable for this purpose. In
particular, long-chain amines have been developed for rendering wet
sand oleophilic. It should, therefore, be possible to dredge sandy
bottom sediments from a water body, treat them on a barge with
amines and thiol derivatives, and return them to the water. By this
means, the transportation of large tonnages of material will be avoided.
Alternatively, if a reservoir could be drained, the surf ace-active and
complexing agents might be plowed or harrowed into the bottom sedi-
ments.
It is not yet clear whether a cover will be needed in the case of the
modified thiols. Although a cover is probably desirable, the bottom
in many cases may be sufficiently anoxic to prevent excessive biode-
gradation.
28
-------
If we consider the results of run C-26 (Table A-6), we note that 0. 844 g
(1 ml) of n-dodecyl mercaptan has complexed 0. 100 g of mercury,
leaving 0. 150 ppb in solution. This is a ratio of 8.44 Ib of thiol per
pound of mercury complexed, or slightly more than a fourfold excess
over the theoretical- Even better results could be obtained by the addi-
tion of small amounts of calcium carbonate. The quoted price of n-
dodecyl mercaptan is about $. 80/lb in drum lots. Other mercaptans
are available in volume at prices as low as about $. 30/lb. No prices
are available on modified thiols. Taking the higher value of $. 80/lb
for n-dodecyl mercaptan, we arrive at a cost of $244 per acre for the
306 Ibs of n-dodecyl mercaptan required for 36 Ibs of mercury. This
cost would be reduced to about $100 if a thiol at $. 30/lb could be used.
Further assuming that we emplace a mixture containing 5% of mercaptan
and 0. 1% of surface-active amine on wet sand, we will require a little
over three tons of sand per acre and six Ibs of surface-active agent. At
$. 40/Ib the cost of the latter will be negligible.
Natural Proteins
The absorption of mercury by wool has been studied by Friedman and
Waiss [3], who find that mercuric chloride approximately follows a
Freundlich isotherm, given by the equation:
log x ~ 0. 33 log C + 1.94
where x is the mercury absorbed by the wool in mg per gram, and C is
the concentration of mercury in the water in grams per liter. We have
made some preliminary experiments with the absorption of HgCl£ by
chicken feathers and find that the results agree well with those obtained
for wool by Friedman and Waiss. Since chicken feathers are a cheap
by-product (estimated cost $. 04/Ib), we will base the present cost
estimates on the use of feathers but will use the distribution data for
•wool.
From the equation of Friedman and Waiss, we estimate that feathers
in equilibrium with water containing 0. 167 ppb of Hg++ will contain
about 500 ppm of mercury. This is a ratio of 2000 Ibs of feathers per
Ib of Hg, or 36 short tons of feathers to complex 36 Ibs of mercury.
At $. 04/Ib of dry feathers, the cost of material will be $2, 880 to cover
an acre of bottom. If the dry feathers are compressed to a density of
5 Ibs/ft3, this will be equivalent to a layer of feathers 4-1/4 in. thick
over the area of the upper basin. Collection, transportation, and em-
placement of such a large quantity of low-density material will certainly
be difficult and costly. As in the case of the Acton peat, about 825 tons
of sand will be required to provide a cover 1/2 in. thick.
The use of such a large quantity of proteinaceous organic matter in a
small area will almost certainly have an adverse effect on the taste and
29
-------
odor of the water. In addition, biodegradation of the feather protein
may release soluble mercury back into the water. If the feathers are
treated with a reducing agent to convert the disulfide linkages into
thiol groups, the absorptive capacity may be increased by a factor of
about two. The costs for such treatment have not been worked out.
By destroying the natural cross-linking of the feather keratin, however,
the reducing treatment may render the complexed mercury still more
soluble and increase the dangers of its release into the water column.
Other agricultural by-products, such as walnut expeller meal, may be
an order of magnitude more effective than wool, but large tonnages
would still be required.
In summary, we find that highly active mercury-binding agents such as
organic thiols or inorganic sulfides are likely to provide greater over-
all economy than natural materials such as peat or proteinaceous sub-
stances. The principal saving is in reducing the need to transport and
emplace large tonnages of material.
If the thiols can be used without a cover layer of sand, they will be
more economical than the sulfides.
30
-------
SECTION V
DREDGING OF MERCURY-CONTAMINATED SEDIMENTS
The preceding section of this report indicates that chemical treatment
of mercury in place is potentially less costly than physically covering
the sediments with large tonnages of cover materials. For the same
reason, chemical treatment, where applicable, will probably be much
less costly than dredging. There will be some situations, however,
where dredging of mercury-contaminated sediments will be required
to maintain navigable water depths.
We have also found that in many cases mercury will be found in local-
ized areas which may be relatively shallow. The material can be
easily removed by dredging once the extent of mercury contamination
is defined. This situation occurs quite often near an outfall, where a
sludgelike material accumulates along the banks of a brook or river.
Dredging presents two main problems of environmental impact: disper-
sal of mercury throughout the water column and disposal of contaminated
spoil.
During the course of this project we have gathered data pertinent to the
analysis of these environmental effects, especially with regard to dis-
persal of mercury in the water column. Details of this laboratory work
are given in Appendix C. The decisions involved in dredging and spoil
disposal are discussed in Section III.
Experiments in aquariums with simulated mechanical dredging have
indicated that the amount of mercury dispersed in the water column is
on the order of 2-10% of that removed. With 100 ppm of mercury (as
HgCl^) in the sediment, total mercury concentrations in the water on
the order of 1 ppm were observed after dredging.
The dissolved mercury fraction increased from fifteen- to thirtyfold
after dredging. The highest value observed was 5.6 ppb, which exceeds
the permissible standard for drinking water. In this case, however,
the amount of mercury in solution was less than 1% of the total water-
borne mercury. This indicates that the major redistribution of mer-
cury will take place in the form of suspended particles. Measurements
of sedimentation rates and their application to the prediction of mercury
redistribution are discussed in Appendix C.
Since the bulk of the mercury is in the suspended form rather than in
solution, it may be concluded that treatment of the bottom with mercury-
complexing agents before dredging will have little effect on the total
waterborne mercury. A more effective method of controlling the dis-
persal of mercury lies in the possible use of vertical cloth or screenlike
31
-------
barriers in the water to limit the travel of fine silt. The results
obtained -with such barriers by the Florida Department of Transporta-
tion have recently been described by Hutt [4]. Suitable barriers might
also tend to limit the zone of oxygen depletion in the water, which is
caused by the dredging of reduced sediments that become a localized
source of high oxygen demand.
In some cases a more effective way to prevent redistribution of mer-
cury is to use a suction dredge in place of a mechanical dredge.
Suction dredging has been successfully used for several years in a
demonstration project for the restoration of Lake Trummen in southern
Sweden [5j. A cutter head is required for roots and consolidated sedi-
ments, but the cutter is not necessary for recent sediments which have
not yet consolidated.
A major drawback of suction dredging is the problem of spoil handling,
since the spoil contains a high percentage of water. At Lake Trummen,
the spoil and water are pumped ashore to two settling ponds, which are
filled alternately. The overflow from the ponds is clarified with alum-
inum sulfate, resettled, and returned to the lake. Such facilities for
spoil handling are not available in many cases.
In some isolated cases, such as the Framingham Reservoir, mercury
contamination may occur in impounded areas where •water level is con-
trolled both for flood control and water resource purposes. In these
cases, it may be possible to lower the water level and expose much of
the contaminated sediment. This material could then be removed with
conventional earth removal equipment rather than the more expensive
dredging systems.
Regardless of the method used to remove the contaminated sediment,
disposal of the spoil will present a potential hazard. Sediments con-
taining more than 1 ppm of mercury are classed as "polluted with heavy
metals" under EPA guidelines of February, 1971. Such sediments may
be disposed of only in the ocean at depths greater than 100 fathoms or
on land disposal sites.
For purposes of this report, we are concerned with how contaminated
sediments can be disposed of in a landfill without risking potential re-
lease of the mercury to air or groundwater or back to the water from
which it was removed. The decisions involved are shown in Figure 2.
Partitioning experiments (Appendix A) have shown that mercury is
firmly bound to organic sediments and is only partially removed by
such powerful complexing agents as cysteine hydrochloride. Further,
these sediments are colloidal in nature and difficult to separate from
the interstitial water. For these reasons, we do not consider it feas-
ible to remove mercury from the sediment before dumping.
32
-------
When placed on a landfill, contaminated sediment can lose mercury to
the water which runs off or percolates through it. Our partitioning ex-
periments indicate that more mercury tends to be released as the sedi-
ment becomes oxidized. Mercury release can be prevented by complex-
ing it in an insoluble form. Either the inorganic sulfides or the long-
chain thiols appear suitable for this purpose. As discussed in Section IV,
the thiols are somewhat more effective than the sulfides, especially under
oxidizing conditions. The conversion of mercury to sulfide has been
found to reduce the rate of methylation [6], and it is probable that the
thiols may have a similar effect.
It appears advantageous to allow the treating agent to mix with the dredge
spoil as it is being moved to a landfill site. Inorganic sulfides can be
added as dry powder or as a slurry in water. The thiols can be emul-
sified in water with the aid of a cationic agent for easy mixing with the
spoil. Alternatively, they can be coated on sand which is mixed with the
contaminated material.
In some cases it may be more advantageous to apply the treating agents
to the spoil after it has been drained and placed on the landfill. The
solid treating agents can be plowed or harrowed into the surface, or
the emulsion can be sprayed on. In any event, it will be advantageous
to provide a well sealed landfill to minimize oxidation and prevent
leaching by oxygenated surface waters.
In summary, we can now enumerate a number of actions that should be
taken when it is known that the dredge spoil material will contain an ex-
cessive concentration of mercury.
First, the extent of the mercury contamination, both horizontally and
vertically, should be surveyed. In some cases the contamination can
be localized to a small area, with a consequent decrease in handling
effort. If the vertical concentration is known, the vertical cut can be
planned so that all of the contamination is removed. If it is not pos-
sible to remove all of the material, it may be necessary to add a bind-
ing material in order to prevent release of mercury to the water from
the freshly exposed sediment surface.
The amount of turbidity which results from the dredging should then be
estimated. If the sediment contains organic or other natural mercury-
binding material, one could expect that some mercury would be released
with the turbidity and that there would be both an increase of total mer-
cury in the water column and a redistribution in the sediment. Methods
of controlling this turbidity, such as the screening material discussed
above, can then be investigated.
The disposal of the spoil material should be planned in advance. If a
diked disposal area of landfill is to be used, the overflow and drainage
patterns should be checked. If it appears that mercury could be released
in the overflow, the spoil material should be treated to bind the mercury.
33
-------
A decision should then be made as to whether or not the spoils area
should be covered, preferably with an impervious fill material. If
there is little possibility of leakage to groundwater and if the overflow
will not contain much mercury (this might be true if the spoils contain
a high percentage of organic material), then it might not be necessary
to add a binding agent. However, it might be desirable to cover the
area with an impervious seal to prevent volatilization of mercury to
the atmosphere and to prevent the penetration of oxygen-rich surface
water.
If the dredge spoil is relatively free of contaminants other than mercury,
it may be possible to dispose of the material at an approved open-water
disposal site if the mercury can be effectively bound. A binding agent
in this case should be resistant to oxidation, reduction, hydrolysis,
biological action, and dissolved salts, such as chlorides. The binding
agent would be mixed with the spoil material en route to the disposal
area. A relaxation of the present EPA guidelines would be required.
34
-------
SECTION VI
FIELD STUDIES
The field studies constituted Phase II of the program and commenced
approximately three months after the start of Phase I. The primary-
purpose of the field investigation was to gain familiarity with site con-
ditions which might influence the effectiveness of physical and chemical
binding techniques and dredging operations to remove the contaminated
sediment. A second purpose was to obtain sufficient field data at a
selected site to design a field project for testing the binding and dredging
techniques investigated during Phase I.
The field investigations were divided into two parts, the first being a
survey of the regional offices of EPA to obtain data on known mercury-
polluted sites.
The information was then compiled and reviewed in order to select one
site for an intensive on-site survey. A summary of site information
received is given in Table 4. A number of the sites were ruled out for
on-site investigations because the environment involved saltwater and
our laboratory program was limited to the control of mercury in fresh-
water environments. In order to hold travel costs to a reasonable level,
we also sought a New England area site, if a suitable one could be found.
Three mercury-contaminated areas were reported in New England.
The sites were located near Orrington, Maine, on the Penobscot River;
near North Dighton, Massachusetts, on the Taunton River; and at
Ashland, Massachusetts, near the Sudbury River and Framingham
Reservoir No. 2. The Penobscot River and Taunton River sites in-
volved brackish water and, except for a preliminary reconnaissance,
were not considered further. The third site in Ashland, Massachusetts
was selected.
Ashland, Massachusetts Site Description
The site selected for intensive investigation during Phase II was the
Framingham Reservoir No. 2 located in Ashland and Framingham,
Massachusetts. The source of mercury to the reservoir has been the
Nyanza Chemical Corporation, located approximately 1 mile away in
Ashland. Until mid-1970, mercury was discharged to a swampy area
near the company, and thence to a small brook which joined the Sudbury
River about one-half mile from the company site. The reservoir is
formed by impounding the Sudbury River in the Town of Framingham.
A general layout showing Nyanza, the brook, the Sudbury River, and
the reservoir system is shown in Figures 4 and 5.
Direct releases of mercury to the swamp were discontinued by Nyanza
in June, 1970, however, a large quantity of mercury has been found in
35
-------
Table 4
Summary of Site Data
Site Location
Whitewood
Creek, South
Dakota
Belle Fourche
River, South
Dakota
Berry's Crk.,
New Jersey
Arthur Kill
R iver, New
Jersey
North Fork,
Holston R iver
Saltville,
Virginia
Savannah
River, Augus
ta, Georgia
Type of Site
Stream, Hg
from mine
waste
River, Hg
from mine
waste
Stream,
tributary of
Hackensack
R i ve r
R iver,
small creek
feeds river
from GAF
plant
R iver, dis -
charge
Olin Math-
ieson Corp.
River, dis-
charge from
Olin Mathie
son Corp.
Other Pollution
Problems at Site
Cyanide, Arsenic
Silt
Cyanide, Arsenic,
Silt
Industrial area,
other pollution
not identified
BOD, Color,
Silt
Calcium chloride
High coliforms
(25, 000 per 100
ml)
Hydraulic Conditions- -
Flow, Seasonal Effects, etc.
Average discharge = 25 cfs
Maximum " - 100 cfs
Minimum " - 18 cfs
Average discharge - 245 cfs
Maximum " - 1 500 cfs
Minimum " -T 0
Not reported
Waste effluent from GAF
10-14 million gallons per
day (mgd).
Not reported
Plant flow, 1 -3 mgd.
Extent of Total Hg
in Sediments and Water
In water, 2-8 ppb with
normal runoff. In bot-
tom sediment, <1 ppm.
In Homestake mining
effluent, < 57 ppb.
In water,
-------
Table 4 (continued)
Site Location
Brunswick
Estuary,
Brunswick,
Georgia
French Broad
River, Ashe-
ville, North
Carolina
Cold Creek,
Alabama
Tombigbee
River, Me -
Intosh, Ala-
bama
Lower Tenn.
River, Mus-
cle Shoals,
Alabama
Type of Site
Tidal estu-
ary, dis-
charge from
Allied
Chem. Corp
River, dis-
charge from
sewage
treatment
plant, in-
cludes some
mercury
Creek, trib-
utary of
Mobile R ivr .
Mercury
from Stauf-
fe r Chemica!
Corp.
Settling ba-
sin, dis-
charge to
river from
Olin Corp.
River, dis-
charge from
Diamond
Shamrock
Corp.
Other Pollution
Problems at Site
High BOD, sludge
from pulp and
paper mill 1 /4
mi. south of
Allied outfall
None reported
Other chemicals
BOD discharge
from Geigy
Chemical Co.
Swamp drainage
Hydraulic Conditions- -
Flow, Seasonal Effects, etc.
Plant flow, 6-8 mgd
Treatment plant discharge,
0. 2 mgd
Flow from chlor -alkali plant
40 gpm.
Variable
Variable
Extent of Total Hg
in Sediments and Water
Concentration in sedi-
ment not known. Mer-
cury discharge not
reported.
Not reported
Discharged 0. 15 Ibs Hg
per day prior to July
1970. Now apparently
reduced to 0. 07 Ib per
day.
Mercury discharge has
been reduced to 0. 12
Ibs per day.
Discharged^ 8. 0 Ibs
per day prior to May
1970. Reduced to v3. 0
Ibs per day after July
1970.
Depth of
Water at Site
3-10 feet
Tidal varia-
tion
Not reported
Cold Creek,
1-2 feet;
Mobile
River, 5-40
feet.
Settling ba-
sin outlet,
2-10 feet;
river, 20
feet.
6-8 inches
at discharge
point
-------
Table 4 (continued)
Site Location
Penobscot
River, Or-
rington,
Maine
Androscoggin
River, Rum-
ford, Maine
Taunton
River, Digh-
ton, Mass.
Sudbury
River, Ash-
land, Mass.
Detroit River
Wyandotte,
Michigan
Type of Site
River, dis-
charge
from Sobin
Chem. Corp
River, dis-
charge from
Oxford Pa-
per Co.
River, dis-
charge from
ICI, Inc.
Settling
lagoon
Brook, dis-
charging to
Sudbury Rvr
Swamp dr.
from Nyanza
Chem. Corp.
River, dis-
charge from
Wyandotte
Chemical
Corp.
Other Pollution
Problems at Site
High BOD in
river
High BOD
Dye waste, other
heavy metals
Dye waste.
Color
Not reported
Hydraulic Conditions--
Flow, Seasonal Effects, etc.
Tidal
J. i-UCL-L
Not reported
Flow from ICI to lagoon,
8 mgd.
River flow, 0-200 mgd
depending on season.
Flow controlled at dam.
Discharge at brook influ-
enced by runoff after heavy
rain.
Major river
Extent of Total Hg
in Sediments and Water
Mercury discharge re-
duced to •+• 0. 2 Ibs per
day. Sediment concen-
trations near discharge
up to 200 ppm.
Mercury concentrations
in sediment up to 20
ppm. Location has been
shifting in pockets down-
stream. Plant closed on
August 15, 1970.
Mercury in sediment of
upper lagoon, 120-820
ppm; lower lagoon, 10-
70 ppm; in mouth of la-
goonat river, 10- 15 ppm.
Sediment in brook had
up to 1000 ppm. Con-
centration in sediments
of Sudbury River and
Framingham Reservoir
5-160 ppm. Levels in
water up to 5 ppb.
Mercury discharge re-
duced to 0. 2-0. 5 Ibs
per day from over 10
Ibs per day prior to
July 1970. Concentra-
trations in sediment 5-
85 ppm within one mile
downstream.
Depth of
Water at Site
1 C f_->f
i - -) le et
Variable
Less than 1
foot in brook;
0. 5 to 4 feet
in river; 4-
25 feet in
reservoir
1-5 feet near
shore
-------
TO
C
>-i
(D
4>.
<5
H"
(fi
TO
tr
O
H"
Lower Section
-------
Upper Section
Framingham Reservoir No. 2
Figure 5. Plan View Showing Nyanza Chemical Corporation Relative
to Sudbury River and Framingham Reservoir No. 2
40
-------
the swamp sediments. Sampling in the brook has indicated that mercury-
is being released in runoff-water from the swamp.
Process Description--Nyan2a Chemical Corporation
Nyanza is located at the end of Magunco Road in the Town of Ashland.
Mercury is used by the company as a catalyst in the production of
anthraquinone compounds, which are used primarily for polyester
dyeing. In the process of producing the anthraquinone dyes, approx-
imately 37 Ibs of mercury has been added per production batch, with
a total of 2400 Ibs purchased and consumed in 1970. The amount of
mercury used per batch has recently been reduced to about 25 Ibs.
Nyanza is now employing other measures to remove mercury from
the process wastes [7].
In the present production process, mercury, sulfuric acid, and
anthraquinone compound are added to a reactor vessel, heated, and
stirred. Benzoic acid is added, and the mixture is heated for about
three hours. The mixture is then sulfonated by additional heating.
After 24 hours, the contents of the reactor vessel are blown to a water
tank, where sodium chloride is added, and the contents are boiled for
24 hours. The mixture at this point is soluble in water. Up to now the
mercury has been complexed with the sulfonated anthraquinone, and
this step breaks the complex. After boiling, sodium sulfide and carbon
black are added. This addition is made in order to remove the mercury
at this point as a mercuric sulfide. Until 1970 the mercury was carried
through the process with no apparent removal.
The entire mixture is pumped to a filter press, where the liquor is
drawn off and pumped back to the reactor vessel. The mercuric sulfide
scraped from the presses is stored in drums. Nyanza is presently in-
vestigating methods for recovering the mercury from the mercuric
sulfide. The mercury concentration in the sulfide cake is about 7000
mg/kg on a dry-weight basis.
At the reactor vessel salt is added to the liquor to precipitate the di-
sulfoanthraquinone. The liquor at this point contains approximately
5 to 15 ppm of mercury. The mixture is then pumped to a filter press,
where the liquor is drawn off and pumped to the plant's sewer system.
The disulfoanthraquinone is scraped from the filter press and removed
for additional processing.
The liquid waste in the sewer system is blended with other liquid wastes
from the plant and is treated with either lime or sulfuric acid for pH
control. The treated waste is then discharged to a series of four settling
basins, with a detention time of approximately 12 hours. After settling,
the waste is discharged to the Ashland town sewer system and eventually
through the MDC system to the Nut Island treatment plant in Boston
Harbor.
41
-------
Mercury Disposal Prior to June, 1970
Prior to June, 1970, the mercury folio-wed the liquor to the settling
basins, receiving only lime treatment for sulfate removal. A small
part of the mercury probably settled -with the calcium sulfate. The
sludge at the bottom of the lagoon was periodically removed and dis-
posed of in a landfill site on the Nyanza property. The liquor was dis-
charged from the settling basin to a small brook which runs through a
swampy area near the Nyanza plant.
The brook has been traced from Nyanza through the Town of Ashland
to where it joins the Sudbury River. Sediments from the brook bed have
been analyzed and show high concentrations of mercury. From measure-
ments in the area around Nyanza it appears that the swampy area has
very high concentrations of mercury and that some mercury continues
to enter the brook in the drainage water from the swamp.
An accurate accounting has not been made of how much mercury was
bought and consumed by Nyanza prior to 1970. If the amount was close
to the 2400 Ibs consumed in 1970 and if operation had been carried out
for even a 10-year period, one would have to assume that up to 24, 000
Ibs of mercury was deposited in the brook or removed in the sludge to
be deposited in a landfill site. Measurements made in the reservoir
system indicate that a significant amount of mercury has reached the
re servoir.
Reservoir Description
Several large reservoirs are operated by the Boston Area Metropolitan
District Commission in the Framingham, Massachusetts region.
Reservoir No. 2, encompassing approximately 130 acres, is formed
by impounding the Sudbury River in Framingham. Mercury contam-
ination of the sediments has been found in all parts of the reservoir
and in Reservoir No. 1 on the other side of the impoundment dam from
Reservoir No. 2.
The reservoirs in this area are not presently used for water supply,
although they do constitute part of the long-range water supply plan for
the Boston area. Water flow in this particular drainage basin is markedly
seasonal, varying from a high monthly average of over 250 million gallons
per day from March to April, down to 5 to 10 million gallons per day
from August through October.
A plan view of Reservoir No. 2 is shown in Figure 4. At the southern
end of the reservoir in the Town of Ashland is a small section isolated
between a railroad bridge and the Union Street bridge. Mercury levels
in the sediments from this section have been found to be as high as
164 ppm (dry-weight basis). Water depth in this seven-acre section
42
-------
is 4-7 feet. The water current depends on the flow volume and may be
less than 0. 1 knot from August through September and up to 2 knots in
the early spring. In this section of the reservoir, which is approximately
750 feet long by 325 feet wide at the maximum points, the velocity profile
is greatest under the Union Street bridge, with a large area of relatively
quiescent water on the near left bank area and on the far right bank. The
sediment sampling program has shown that these are also the areas of
highest mercury concentration. The land area surrounding this section
is owned by the Metropolitan District Commission.
Extent of Mercury Contamination
In order to determine the extent of mercury contamination in the reser-
voir, and also in the area between the reservoir and Nyanza, a field
sampling program was designed and executed. Initially, a series of
grab samples were taken, commencing in the swamp area near Nyanza
property, progressing down the brook through Ashland to the Sudbury
River and finally to the reservoir. The grab samples were analyzed
for total mercury, and, using the results, a plan was developed for
taking sediment cores in the areas of high mercury concentration.
Core samples were taken with 2-foot-long by 1. 5-inch-diameter plastic
tubes, which were quick-frozen after sampling. Before analyzing, the
cores were cut into 2-inch sections and each section was analyzed for
total mercury and percentage of moisture in the sample. Knowledge of
the sampling location and the vertical section position permitted mapping
of the mercury concentrations in both the horizontal and vertical planes.
From this data a series of contour maps for the 7-acre section was de-
veloped. Each map shows the horizontal distribution of mercury within
a 2-inch vertical section. Sufficient data were available for mappings
of the 0-2, 2-4, 4-6, and 6-8 inch sections. Several 18-inch cores
were also examined to determine the depth of mercury penetration.
The contour mappings are presented in Figures 6, 7, 8, and 9- Results
of the mercury analyses on grab samples and cores are tabulated in
Appendix F.
In addition to the sediment samples analyzed for total mercury, several
samples were also analyzed for methylmercury. The methylmercury
fraction of the total is on the order of 0. 4% (see Appendix A). In addition
to the sediment analyses, a number of water samples were taken. Levels
of dissolved mercury in the reservoir water are generally lower than the
5 ppb standard for drinking water supplies, although there appear to be
seasonal excursions above the limit. As a result of analyzing the water
samples in two different ways, we believe that over 50% of the total mer-
cury found in the water is in the form of a soluble organic compound.
This possibility is discussed in more detail in Appendix A.
Water samples taken closer to the source of mercury generally have
higher concentrations of both total and dissolved mercury than do samples
43
-------
OQ
C
>-t
ft)
s-
t— ,?
o
CO O
a'S
fD "<
OP
01
o
i
ro
n
tr
n>
CO
Framingham Reservoir No. 2
Between Union Street and
Railroad Bridge, 2/16/72
-------
OQ
c
o
c
l-j
13
QfQ
CO
CO
n
tr
(D
CO
Framingham Reservoir No. 2
Between Union Street and
Railroad Bridge, 2/16/72
-------
TO
C
H
a
oo
en o
£ c
a
oq
en
n>
CO
Framingham Reservoir No.
Between Union Street and
Railroad Bridge
-------
TO
e
H
(I>
sO
01 0)
V *
£* n
g c
0) o
tl
TO
CO
i
00
I—I
tJ
n
cr
n>
01
Frarningham Reservoir No. 2
Between Union Street and
Railroad Bridge, Z/16/7Z
-------
taken in the Sudbury River and the reservoir. When runoff is high in
the spring season, mercury levels in the brook appear to increase,
probably because of increased erosion and leaching in the swamp and
brook sediments. Levels in the reservoir tend to decrease after a
heavy runoff period, probably because of the increased dilution from
the river.
The results of the water analyses are given in Table F-3 of Appendix F.
In order to determine whether or not aquatic life in the reservoir was
affected by the mercury concentration in the sediment, we requested
through the Massachusetts Division of Water Pollution Control that fish
samples be taken from the reservoir by the Massachusetts Division of
Fish and Game. The results of the first set of analyses are given in
Appendix F. In this sampling all fish analyzed had mercury concentra-
tions in excess of 1 ppm.
A later sampling of largemouth bass indicated that fish of this species
in excess of 12 inches in length would probably have greater than 6 ppm
concentrations of mercury in their tissue. Although water levels of
mercury were uniformly low, the fish were accumulating relatively
high levels. This latter work was performed by Mr. Thomas Palermo
of the Massachusetts Division of Fisheries and Game.
Discussion
From the contour maps it is possible to determine the approximate
quantity of mercury in the upper 7-acre section of the reservoir. From
grab sample analyses in the remainder of the 130-acre reservoir it is
also possible in a much less precise manner to estimate the quantity of
mercury in the entire reservoir system.
We have determined the density of the sediment in the upper two inches
of the 7-acre section to be about 78. 5 Ibs per ft^. The bottom area in
O •*•
that section is 305, 000 ft^. The average Hg concentration on a wet-weight
basis is about ZO ppm. From this data we have determined that there is
about 80 Ibs of Hg in the top two inches. For the two-to-four-inch layer
we have estimated a density of about 90 Ibs per ft-^ and an average Hg
content of about 15 ppm. This gives a result of about 70 Ibs of Hg. In
the four-to-six-inch layer the sediment density is about 100 Ibs per ft^
and the Hg concentration approximately 8 ppm, giving a mercury con-
tent of 40 Ibs. Further calculations for the six-to-eight-, eight-to-ten-,
and 1 0-to-l 2-inch layers give an approximate total quantity of 250 Ibs
in the 7-acre section.
Grab samples in the remaining 125 acres have indicated that the mercury
concentrations in the zero-to-two-inch layer range between nine and 80
ppm by dry weight. Although we have not performed core sampling in
this portion of the reservoir, we can estimate on the basis of samples
in the upper layer that the overall mercury concentration in this area
48
-------
is on the order of one-half that in the upper 7-acre section. On this
basis , the total quantity of mercury in the 125-acre lower section would
be about 2250 Ibs and about 2500 Ibs in the total 132 acres.
Although we do not know over how long a period Nyanza has been re-
leasing mercury, we do know that in 1970 the company consumed
2400 Ibs. If in the previous 10 years a similar annual amount was
consumed, a total of 24, 000 Ibs could have been released to the environ-
ment. If only about 2500 Ibs can be accounted for in the reservoir sedi-
ments, some may have travelled further downstream and some may still
be in the swamp adjacent to the company. We have, in fact, found con-
centrations in this swamp of up to 3500 ppm by dry weight. No analyses
of lower river sediments have yet been made.
From our measurements of the mercury concentrations in the brook
between Nyanza and the Sudbury River, we have evidence that mercury
is continuing to be transported to the reservoir. An examination of
Table F-3 of Appendix F shows that turbid samples taken from the
drainage area near the Nyanza plant have a high level of mercury asso-
ciated with the particulate matter in suspension. In the clear water
samples from the brook, between 20% and 50% of the total mercury has
been in a dissolved form. However, our water samples do not include
the sediment particles along the bottom, which by visual observation
appear during heavy runoff periods to be moving almost continually
downstream. We believe that this bottom shifting may be responsible
for a large share of the mercury transport to the reservoir.
We have discussed in Appendix A the conclusion that much of the dis-
solved mercury coming from the swamp area around Nyanza is probably
in the form of a soluble organic compound. Under a separate contract
to the Commonwealth of Massachusetts, we are investigating in more
detail the circumstances of mercury leaching from the swamp. One
additional point can be made about mercury in the brook between Nyanza
and the reservoir, i.e., that the dissolved levels of mercury in the
water are well diluted when they reach the reservoir. The highest
levels of dissolved mercury we have found in the reservoir have been
between 5 and 6 ppb during our October measurements. This is a per-
iod of low flow, thus we might expect that the total quantity of mercury
reaching the reservoir would be highest in the spring, when the flow
volume is greater.
Although mercury concentrations in the water of the reservoir are
between 1 and 6 ppb, the fish have accumulated a significant amount,
as witnessed by the data presented in Table F-4. It is of note that the
small bluegill have over 2 ppm of mercury. These fish are bottom
feeders, and this indicates that the bluegill may be accumulating mer-
cury from ingestion of bottom material. The high levels of mercury in
the largemouth bass, which are predators, indicate that the food chain
may also be responsible for increasing mercury concentrations. The
contribution of the mercury in the water column to the fish is unknown
at this point.
49
-------
SECTION VII
PROPOSED PILOT FIELD PROGRAM
One of our tasks under Phase II has been to develop a detailed pilot
field program at the surveyed site, whereby- the techniques of binding
and dredging investigated during Phase I could be tested under real-
istic conditions. The importance of testing these techniques on a pilot
scale before attempting a large-scale decontamination cannot be over-
stated. The laboratory work using aquariums has indicated that several
possible methods may be effective in controlling the release of mercury
from sediments. However, we have found that it is very difficult to dup-
licate in the laboratory the physical and chemical conditions existing at
the field site.
We believe that the validity of the laboratory results must be substan-
tiated in the field. The proposed pilot program will allow for testing
of hydraulic and mechanical dredging techniques, will provide data on
treatment of dredge spoils prior to disposal in a landfill site, and will
test the effectiveness of complexing agents added to the bottom sedi-
ments.
The proposed pilot program should be conducted at the Ashland, Mas-
sachusetts site, since this site has been extensively surveyed and
mapped over the past year. There is good access to the site over land
owned by the Boston Metropolitan District Commission. This agency,
which controls the water resources of the area, is quite agreeable to
the proposed task.
Test Site Description
At the southern end of Framingham Reservoir No. 2 in the Town of
Ashland (see Figure 4), a small section of the reservoir is isolated
between a railroad bridge and the Union Street highway bridge. Mer-
cury levels in this section have been mapped during Phase II and are
shown in Figures 6, 7, 8, and 9. The water depth in this section,
which encompasses an area of 7 acres, is 4-7 feet. Water currents
are low, ranging from less than 0. 5 knot to about 2 knots during heavy
runoff periods. The current is confined generally to a fairly well de-
fined course, leaving several large quiescent areas near the left and
right banks.
The area in which the tests would be conducted is approximately 750
feet long by 325 feet wide at the maximum points. The land area sur-
rounding the water, except for the two bridges at either end, is owned
by the Boston Metropolitan District Commission. There is good access
to the water from several locations along the shore. The bottom sedi-
ment material is primarily organic detritus and fine silt to a depth of
51
-------
10-12 inches, except in the area of higher water velocity extending out
from the Union Street bridge, where scouring has uncovered a gravelly-
base material.
Test Structure
We believe the decontamination tests should be conducted with as little
effect as possible on the sediments or water of the reservoir. For this
reason, the tests have been designed to be performed in a wooden struc-
ture similar to a cofferdam. The components of the wooden structure
can be assembled on shore, with final assembly in the water. The test
box, when completed, would be a rectangular structure with no bottom
and could be floated into place. When ready for testing, it would sit on
the bottom, with sides extending about 18 inches above the water surface.
Access catwalks can be installed and flow gates provided at both ends to
control the flow of water during the tests.
Figure 10 shows a suggested type of construction. The frame would be
about 20 feet long and would be assembled using 2 inch x 6 inch uprights
every 2 feet. After the frame is assembled, prepainted sheets of marine
plywood could be nailed to the frame on both sides with staggered seams.
The finished section would be approximately 20 feet long by 7 inches in
width. The ends of each section would have a full-length rubber gasket,
so that sections could be bolted together and still maintain a good degree
of watertight integrity. The seal is not too critical, as there would be
little pressure head across the section, and there is little water current
in the test areas.
As each section is needed, it can be picked up with a truck crane or
rolled into the reservoir and floated to the test location. After the
structure is approximately in the right position 1/4-inch sheet steel
or aluminum would be bolted to one side of the section to act as a keel
to penetrate into the muddy bottom of the reservoir. If needed, bricks
or sandbags could be used as added ballast above the keel in order to
facilitate maintaining a vertical position in the water. With the proper
choice of wood, metal keel, and bricks or sandbags, each section can
be made almost neutrally buoyant. After all of the required sections
for a test cell (40 x 20 feet) have been floated into position, they can be
bolted together and cross supports installed as necessary. The result
would be a floating, bottomless box that can be towed about the reser-
voir as required. Once it is in position, sandbags could be piled on the
upper structure, and the cell would sink into place.
Based on the site survey, we would expect the keel plates to sink into
the unconsolidated sediments. However, if this does not happen in some
locations, the appropriate area can be "jetted" out, using a jet pump
along the outside of the keels until the structure sinks into the bottom
sediments approximately 12 inches.
52
-------
in
M?
OP"
c
ft
I—'
o
C/3
C
CTQ
en
rt-
ft
3
ft
O
o
C
O
H
ft
en
C
o
rt-
C
hi
ft
8'
4' x 8' /
Marine Plywood
I I
I I
! I
' L
I /
I /
V
20' —
Z
z
-------
An alternative assembly mode would involve the pre-assembly of a
section, including keel plate and ballast, on land and then simply
picking the section up with a truck crane and lowering it into place.
Since the bottom drops off very quickly, the truck crane can remain
on shore and, piece by piece, lower sections into place in 6 feet of
•water. The entire assembly can then be floated anywhere in the
reservoir.
We anticipate that there may be some difficulty experienced in refloat-
ing the test structure, thus provisions should be made for installing
inflatable floats to help break it loose prior to floating the structure
to another location.
Dredging Tests
There are two dredging problems to consider in conducting the test
program. In some cases we will want to clean the bottom area inside
a test cell prior to conducting dredging simulations in an adjacent cell.
This would be required in order to determine whether or not there is
a noticeable effect on the clean area by adjacent dredging activities.
Then there is also the problem of how to conduct the actual dredging
simulation.
In most cases of dredging a polluted area, conventional dredging equip-
ment will not be suitable. Unless dredging is required in a navigable
waterway, most of the cases we have observed involved lakes, small
rivers, and streams where the depth of cut would be confined to less
than 2 feet. This will require equipment not normally used for dredging,
such as suction trash pumps or equipment used to pump out disposal
lagoons or septic tanks. In some cases it may also be possible to use
specially designed dragline equipment, although this may cause exces-
sive turbidity.
In this test program the hydraulic removal of sediments can be demon-
strated using typical trash pumps of the diaphragm type. Pumps are
available using electric, internal combustion, or air drive and can be
rented or purchased. Rental costs for a gasoline-driven trash pump are
on the order of $350 per month, including hoses. Complete units can be
purchased for about $1,000.
The pump would be mounted on a flotation platform (a -wooden raft with
polystyrene floats), which could be positioned in different locations
within the test cell. The discharge lines would be run ashore to several
large storage containers. Several dredging conditions could be simulated
by allowing some of the discharge to return to the test cell. The test
program should include several experiments whereby the dredged mater-
ial is allowed to settle in the storage containers and the supernatant
treated and returned to the reservoir. The temporal behavior of mer-
cury concentrations in the liquid would be monitored during this treat-
54
-------
merit. Several effective techniques have been developed recently for
the treatment of liquid streams from chlor-alkali plants, and these
could be tested at this point if desired by EPA.
Tests should also be conducted on the treatment of the settled spoil
material to determine the most desirable form of ultimate disposal.
It may be necessary to add complexing agents to the spoil before send-
ing it to a. landfill site, or it may be possible to dispose of it in an un-
treated form if the landfill is designed to prevent percolation and
leaching.
Tests of Mercury Bonding and Sealing Agents
In the course of our laboratory program, we have evaluated a number
of materials to determine their effectiveness in decreasing the rate of
release of mercury from the sediments to the water. For purposes of
the pilot program, we have considered the following materials:
1. Natural sediments--organic or sandy
2. Inorganic sulfides
3. Organic thiols (mercaptans)
4. Proteinaceous materials, such as hair or feathers
Our findings have indicated that the organic and inorganic sulfides are
likely to provide greater overall economy and effectiveness than the
natural materials such as peat or proteinaceous substances. In some
cases, two materials may be required, such as the use of a layer of
sand to stabilize a complexing agent in the sediment. The details of the
laboratory evaluations are discussed in Section IV above.
At the field test site we are proposing that tests be conducted on the
effectiveness of materials mentioned above in reducing the rate of re-
lease of mercury to the water. We expect that the use of fish and pos-
sibly freshwater mussels will be required in these tests to indicate the
effectiveness of the materials. The freshwater mussels have been
used as indicators of pesticide pollution in tests conducted by the Mas-
sachusetts Division of Fish and Game, and we feel that they may also be
useful as indicators of mercury and other heavy metals in the water.
They can be suspended in mesh bags, both on the bottom and above, to
indicate the amount of mercury taken up from the water column and
from bottom sediments. We expect that tests could be conducted on the
uptake rates of these organisms and a statistical base for use in the
field programs could be developed. The fish would be needed to indicate
if mercury continues to be concentrated through the food chain.
The test fixtures provide four 40 x 20 foot basins for the experimental
program. Since the upper sediment surface (4-6 inches) is highly organic
55
-------
and has a high binding capacity for mercury, one of the test fixtures
will be used as a control. Indicator organisms will be added to the test
basin. The control test will be run for one month with no flow through
the basin. Three samples of each indicator organism will then be with-
drawn and tested for mercury. The remaining organisms will then be
removed and a net set added. The flow gates will be opened and for
one month the basin will be open to the flowthrough of the reservoir.
At the end of one month, samples of each indicator will again be taken.
The flowthrough test will be run several times at different flow rates.
The total reservoir flow can vary from about 5 million gallons per day
in late summer to as much as 200 million gallons per day in early
spring. We will also be sampling the mercury concentration in the
water throughout the test period.
After the control conditions have been established, tests will be run on
up to five combinations of sealing and complexing agents. An outline of
the proposed plan is shown in Figure 11.
Test Procedures
Using the test structure layout shown in Figure 12, we believe the
dredging tests and the sealing and binding tests can be conducted as
follows. The procedures are tentative and indicate our present think-
ing. A final test plan should be drafted during the first two months of
the program.
1. Remove by dredging the organic and underlying contam-
inated sand material in Section B. All removed material
is to be deposited in one section of the shore container.
2. Remove by dredging only the organic material in Sec-
tion A. Deposit material in section of shore container.
This will be a simulated dredging operation with gates
between Sections A and B open. Effects of the dredging
in A will be monitored in both A and B, which is free of
contaminated sediment.
3. Concurrent with step 1, control tests will be started in
Section C with both inflow and outflow gates closed.
The control tests have been described previously. After
one month, open inflow and outflow gates in both Sections
C and D, and continue control tests.
4. Concurrent with step 3, conduct first sealing and binding
agent test in Section D (gates closed). After one month,
the control test in Section C will require gates to be open.
This will be compatible with test of the binding agent
with gates open.
56
-------
TESTS
J t [
2 0 - 1
[ 1
to *• &
c .2 •"
0 1 £ < K 0 3
•O i C 5 c C
c U3 a o .X .2 °
c
-------
80'
Gate
H 1-
40'
D
B
-20'
-40'
Figure 12. Test Structure Layout
58
-------
5. Section A, which has had the contaminated organic
material removed, can now be used for tests on the
effects of dredging a sandy material. The bottom of
Section B will be cleaned by skimming to insure the
bottom and water column are relatively free of mer-
cury, and then the sandy material in Section A will
be removed by dredging and deposited in a section of
shore container. Effects of dredging will be monitored
in both Sections A and B.
6. Refloat test structure, move to new location, and
repeat sequence of tests. Tests will take about two
months at this location.
7. Refloat test structure and move to new location.
Tests in Sections A and B will be sealing and binding
tests on sandy sediment rather than dredging. The
organic sediment will be removed prior to commencing
the tests.
The above procedure allows for two tests of dredging organic sediments,
two tests of dredging sandy sediments, and five tests of sealing and
complexing agents: three on organic sediments and two on sandy sedi-
ments. The test series will require a period of about six to eight
months. If additional tests are required, it may be advisable to add
two additional 20 x 40 foot sections rather than extend the time period.
The following list of cover materials to be tested is suggested. The list
may be augmented by EPA if desired.
1. One to two inches of clean sand over organic sediment.
This material was tested and described as the "Tank D"
part of the aquarium experiments (Appendix B).
2. A cover of ferrous sulfide or milled pyrite.
3. A material consisting of a long-chain mercaptan on
treated sand with and without ground limestone.
4. An organically modified mercaptan or other organic
sulfur compound.
Schedule for Field Pilot Program
The field pilot program should be conducted over a 12 to 14 month
period. Figure 13 shows a 12-month schedule, which would not
allow for any contingencies. We suggest that two additional months
should be allowed for this purpose.
59
-------
1. Continue water and
sediment monitoring
of the site.
2. Design and construct
the wooden cofferdam
3. Assemble and posi-
tion the cofferdam in
the site.
4. Develop test plan for
the site experiments.
5. Conduct site tests
in situ.
6. Conduct dredge spoil
tests.
7. Evaluate data and
perform cost analysis.
8. Write final report.
1 2 3 4 5 6 7 8 9 10 11 12 - 14
1 2 3 4 5 (
*
'
7 8 9 10 11 12-14
Months
*A two-month contingency is allowed at the end of these tests,
if required.
Figure 13. Schedule for Field Pilot Program
60
-------
Monitoring of conditions at the test site should continue throughout the
field program. We have been aware of seasonal changes in the mer-
cury levels in the water, and these should be monitored. Although a
preliminary design of the cofferdam structure has been presented, this
will require finalizing. A time period of 1. 5 months is allowed for
design and construction. The final assembling and positioning of the
cofferdam will take about one month.
While the test structure is being constructed, the final test plans can
be prepared; this will take about two months. The actual tests will take
about six months, although the two-month contingency period would not
likely be required at the end of this period. Data evaluation, cost
analysis, and report writing consume the remainder of the time.
Work Summary
1. Continue monitoring mercury concentrations in the
sediment and water of Framingham Reservoir No. 2.
Establish three stations where bi-weekly samples will
be taken over an 11-month period.
2. Design and construct a 40 x 80 foot wood test structure
divided into four 20 x 40 foot sections. Each section
will have an input and output gate located on the 20-
foot side.
3. Assemble and position the test structure at the site.
4. Develop a final test plan for the site experiments,
based on the preliminary plans in this proposal.
5. Conduct dredging, sealing, and bonding tests within
the test structure.
6. Conduct tests on the binding of mercury in the dredge
spoils. These tests will be conducted in a sectioned
container located on shore near the test site.
7. Evaluate data from the site experiments and write
final report.
61
-------
SECTION VIII
REFERENCES
1. Jannasch, H. W. Einhjellen, K., and Wirsen, C. O.,
Science, 171, 672-675 (1971).
2. Jenne, E. A. , U. S. Geological Survey Professional Paper
713, U. S. Government Printing Office, Washington, D. C.
(1970).
3. Friedman, M., and Waiss, A. C. , Environmental Science and
Technology, 6_, 457-458(1972).
4. Hutt, A., Sea Frontiers, 18_, 86 (1972).
5. Bjork, S. , University of Lund, Sweden (personal
communication).
6. Fagerstrom, T. , and Jernelov, A. (unpublished communi-
cation).
7. Devine, T. , et al. , Nyanza, Inc., Ashland, Massachusetts,
1971 Mercury Surveys, Region 3, Environmental Protection
Agency.
8. Tuttle, J. H. , et al. , Journal of Bacteriology, 97, 594-602
(February, 1969).
9. Sutton, L. E. , U. S. Patent 1, 926, 797 (September 12, 1933).
10. Chick, H. , Journal of Hygiene, 8, 92 (1908).
11. Bidstrup, P. L. , Toxicity of Mercury and its Compounds,
Elsevier, New York (1964), p. 30.
12. Pagnotto, L. D. , Brugsch, H. G. , and Elkins, H. B. ,
American Industrial Hygiene Association Journal, 21, 419
(I960).
13. Bidstrup, P. L. , loc. cit. , pp 54 and 62.
14. Takahashi, H., and Hirayama, K. , Nature, 232, 201 (1971).
15. Ganther, H. E. , et al. , Science, 175, 1124(1972).
16. Sillen, L. G. , and Martell, A. E. , Stability Constants of
Metal-Ion Complexes, The Chemical Society, London (1964).
63
-------
17. Van Horn, W. M. , Anderson, J. B. , and Katz, M. , Trans-
actions of the American Fisheries Society, 79, 55-63 (1949).
18. Shugaev, B. B. , Khim. Seraorg. Soedin. , Soderzh. Neftyakh
Nefteprod. , 8, 681-86 (1968) (U.S.S.R. ); see C. A. 71, 99956 s
(1969).
19. Turnbull, H. , DeMann, J. G. , and Weston, R. F. , Industrial
Engineering Chemistry, 46, 324-33 (1954).
20. Reid, E. E. , Organic Chemistry of Bivalent Sulfur, Vol. 1,
P 49, Chemical Pub. Co. , Inc. , New York (1958).
21. American Petroleum Institute, Manual on Disposal of Refinery
Wastes, Volume on Liquid Wastes, Chapter 20, New York
(1969).
22. Hatch, W. R. , and Ott, W. L. , Analytical Chemistry, 40,
2085-87 (1968).
64
-------
SECTION IX
ACKNOWLEDGEMENTS
The support and assistance of the Project Officer, Dr. Curtis C.
Harlin, Jr., of tlie Robert S. Kerr Water Research Center, Environ-
mental Protection Agency, is acknowledged with sincere thanks. In
addition, the assistance of Mr. Charles Myers, of the EPA, has been
greatly appreciated.
The authors also wish to acknowledge the assistance of Ms. Birgit Foley,
Mr. Stephen Greene, and Mr. Thomas Hall for their patience and skill
in performing both laboratory and field analytical work.
65
-------
APPENDIX A
PARTITION COEFFICIENTS
The partition coefficient provides a quantitative measure of the mercury-
binding capacity of sediment in the presence of overlying or percolating
water. This kind of information is necessary to understand and control
the movement of mercury in natural water and soil systems. For the
purposes of this report, the partition coefficient will be defined as the
equilibrium ratio of mercury concentration in solution to the concentra-
tion in the solid. The lower the numerical value of this ratio, the more
effective is the mercury-binding action of the sediment.
In this appendix, we describe the experimental methods used for meas-
uring the partition coefficient, the materials studied, and the results
obtained both with natural sediments and with chemical additives.
Experimental Procedure
Partition coefficients were measured by placing a few hundred grams
of sediment in a quart glass jar and covering it with several hundred
milliliters of distilled water. A known amount of mercury was added
as a standard solution of HgCl2 or CH-^HgCl, together with any required
additives or complexing agents. The jars were tightly covered and
placed in an agitator, where they were slowly tumbled (about 10 rpm)
for periods of 1 to 7 days at room temperature (24-25°C). Preliminary
experiments had shown that such continuous agitation was necessary to
approach equilibrium within a reasonable period of time.
Most of the runs were made with 200-300 ml of air in the sample bottle.
As a result, most of these equilibrations were made with oxygen-
saturated liquid (7-9 ppm of dissolved oxygen). In some runs, espe-
cially those made with highly organic sediments, enough reducing
material was present to consume all the oxygen in the bottle and reduce
the dissolved oxygen to a low value. Dissolved-oxygen measurements
identify this situation. In other cases it was desirable to conduct the
equilibration with a minimum of oxygen present. For these runs, the
water used was freshly boiled and cooled to room temperature. The
sample bottle was then filled to the brim with this oxygen-free water
(the inclusion of 5-10 ml of air was unavoidable with the type of screw-
cap we used), and the bottle was sealed. Runs made in this way are
referred to as "low oxygen" in the discussion of the data.
After equilibration, dissolved oxygen and pH were measured with immer-
sion electrodes, and the samples were roughly filtered through paper on
a Buchner funnel to remove the bulk of the sediment. The filtrates •were
again filtered through a 0.45 micron membrane filter in order to remove
fine particles and to ensure that only mercury in true solution was
67
-------
measured. The final filtrates were acidified with 1 ml of HNOo in
order to hold the mercury in solution while awaiting analysis.
The samples were analyzed for mercury with a Coleman Model 50
flameless atomic absorption analyzer, using the Hatch and Ott pro-
cedure. We found that solutions containing less than about 0. 001 ppm
of mercury were difficult to analyze precisely with this instrument.
Most of the samples below this concentration were therefore analyzed
by Jarrell-Ash, using a high-sensitivity atomic absorption apparatus
with a hydrogen flame. Analyses made by this high-sensitivity method
are marked with an asterisk in the following tables.
All sediment samples and all solutions containing methylmercuric
chloride were refluxed -with a mixture of nitric and sulfuric acids prior
to analysis in order to destroy organic matter and bring the mercury
into solution as Hg+"*~. Details of the analytical methods are given in
Appendix E.
The accuracy of the analyses was checked by adding the total mercury
found in the sediment to that in the filtrate and comparing this figure to
the known amount of mercury originally added. In most cases, the
mercury balance checked within 4- 15%. At relatively low concentra-
tions (~10 ppm in the sediment) a check of 4- 25% was considered ac-
ceptable. If the mercury balance was outside these limits, the analysis
was repeated.
The mercury content found by analysis of the wet sediments was con-
verted to the dry basis by measuring and correcting for the moisture
content. In a few cases, the analysis was also corrected by subtracting
the mercury content of the solution contained in the pores of the wet
solid. In all but a few runs, however, this latter correction was neg-
ligible in comparison to experimental error.
The results were expressed as the partition coefficient:
_ ppm Hg in solution
ppm Hg in dry solid
- 8
The values obtained in this work ranged from about 1 to about 10
The latter value represents the limit of sensitivity of the analytical
method, using the Jarrell-Ash high-sensitivity atomic absorption ap-
paratus.
Description of Materials Used
Acton Sand
This sample was obtained from Nagog Pond, a municipal reservoir in
Acton, Massachusetts. It consists largely of a siliceous sand with a
minor proportion of very fine clay or silt. The material retains about
68
-------
27% of water when allowed to settle and drain, and 9. 3% of moisture
when dewatered on a suction filter. The loss on ignition is 0.8%, which
places an upper limit on the content of organic matter. A part of this
loss may be due to removal of bound water from the clay fraction.
The sample has a pH of about 6.5 and contains no measurable mercury.
It appears to contain some ferrous iron, as indicated by the fact that it
becomes covered with a yellowish layer (probably ferric hydroxide)
when a water-covered sample is allowed to stand in air.
Runs made with this sediment are identified by the prefix "A" to the run
number.
Acton Peat
This material is a black, fibrous sediment obtained from a different
part of Nagog Pond in the vicinity of a wooded shoreline. It is thought
to consist largely of decomposed leaves from deciduous trees, and it
contains 79% of water in the drained condition and 74. 5% when vacuum
filtered. The loss on ignition is 44. 3%. It has a pH of 5. 4 and a mer-
cury content of 0. 342 ppm (dry basis).
Runs made with this sediment are identified by the prefix "B" to the
run number.
Georgia Kaolin
This was a commercial pure kaolin clay (Pioneer Brand, Georgia Kaolin
Co. , Dry Branch, Georgia), sold for use in ceramic work. The pH was
about 4. 6 and the loss on ignition about 16%. This loss is considered to
be mainly the combined water of the clay material.
This clay was considered to be a relatively pure representative of a
mineral commonly found in bottom sediments. Because of its low natural
affinity for mercury, it 'was used mainly as a substrate for testing various
chemical additives.
Runs made with this clay are identified by the prefix "C. "
Ground Silica
A sample of ground silica (about 240 mesh) was obtained from Fisher
Scientific Company. This material may be representative of another
common constituent of sediments. In Sweden, ground silica has been
proposed as a sealant for contaminated sediments.
Runs made with this material are identified by the prefix "S. "
69
-------
Chicken Feathers
A sample of wet chicken feathers was obtained from a local farm and
stored frozen until used. Feathers are similar in composition to various
animal proteins, such as wool and hair, which contain relatively large
amounts of the thiol amino acid cysteine.
Runs made with these feathers are identified by the prefix "CF. "
Ashland Sediment
This sediment was collected from the upper basin of Framingham Reser-
voir No. 2, located in Ashland, Massachusetts. The reservoir is down-
stream from a dye manufacturing plant which uses mercury catalysts in
the production of anthraquinone sulfonic acids. Until June, 1971, the
mercury-containing waste solutions from this process were discharged
into a swamp and a tributary stream. Although the discharge has been
stopped, the sediments in the reservoir contain mercury in amounts from
4 to over 100 ppm on the dry basis.
Two samples were used for partitioning and aquarium studies — one col-
lected in October, 1971 and one collected in December of the same year.
Both were black, highly organic, and gave evidence of chemical and
industrial contamination, as shown by free oily material and colored
water-extractable materials. Through the cooperation of Mr. James
Longbottom of the Environmental Protection Agency, Cincinnati, we
have obtained analyses of these sediments for methylmercury. The
results of these and other analyses are as follows:
October, 1971 December, 1972
Total Hg (dry basis) 31. 8 ppm 100. 5 ppm
Methyl Hg (dry basis) 0.125 ppm 0.428 ppm
Percentage of total Hg
as methyl 0.39% 0.43%
pH --- 7.2
Moisture 65%
Loss on ignition 16%
When acidified with dilute H^SO f these sediments gave off a strong odor
of hydrogen sulfide. This is probably due to the presence of FeS.
Runs made with these sediments are identified by the prefix "ASH. "
70
-------
Climax Pyrite
This material was obtained from the Climax Molybdenum Co. and is
produced as a by-product of their milling operations at Climax,
Colorado. A typical analysis is as follows:
Chemical Analysis In Percentage Weight (dry basis)
Sulfur 51.89
Iron 44.94
Insoluble s 2.80
Copper 0.05
Lead 0.05
Zinc 0.23
Arsenic 0. 01
Selenium 0.002
Tellurium 0. 001
Phosphorus 0.006
Screen Size Weight (percent)
Plus 35 mesh 0. 1
Plus 100 mesh 38. 0
Plus 200 mesh 46. 0
Plus 325 mesh 12. 0
Minus 325 mesh 4. 0
This pyrite was shipped to us with about 8% moisture. Before use, it
was washed with strong HC1, followed by acetone, and then it was dried.
This treatment was intended to remove possible oxidized layers (Fe(OH)o)
and possible residues of flotation reagents from the surface.
Since the coarse powder as received was found to be relatively unreactive,
some later runs were made with the above material, which was hand
ground in a mortar. Other experiments were conducted with pyrite
which had been fired in a crucible to partially decompose it to FeS.
This material was also hand ground. The screen analyses of these
hand-ground materials were:
Fired
Screen Size Pyrite Pyrite
Plus 120 mesh 24.2% 6.4%
Minus 120 plus 200 mesh 32.4% 27.5%
Minus 200 mesh 43.7% 65.7%
Some later tests were made with pyrite which had been mechanically
milled with alumina balls until it all passed through a 325-mesh screen.
The cost of this pyrite is $3. 80 per ton at Climax, Colorado, before
loading. The cost of shipping to eastern points, however, is expected
71
-------
to be in the range of $70 to $100 per ton. It -would therefore be advisable
to locate a source of pyrite as near to the point of use as possible.
Miscellaneous Materials
Other materials used in this work -were commercial products or labora-
tory- reagent chemicals obtained from local sources.
Rate of Equilibration
The rate of equilibration was checked by agitating samples of Acton
sand with various concentrations of HgC^ for periods of 1, 2, 4, and
7 days. When equilibrium is attained, the concentration of mercury in
solution should no longer decrease with time. Figure A-l shows the
results obtained with a mercury concentration of 412 ppm based on the
dry sediment. The concentration of mercury in solution changes rapidly
for 2 days and then rather slowly up to 7 days. We have considered the
7-day point to represent substantial equilibrium, since the change after
the second day is probably within the limits of analytical accuracy. In
order to check this point, we performed a reverse experiment in v/hich
a mercury-saturated sediment was equilibrated with pure water. If
equilibrium is being attained, the concentration of mercury in solution
should reach the same level in this experiment as in the previous exper-
iment. The point marked "x" in Figure A-l shows that after 11 days
the reverse equilibration attained only about 40% of the concentration
of the 7-day point. Again, this observation must be tempered by con-
siderations of analytical accuracy. It appears, however, that the true
equilibrium is reached very slowly in this case and lies between 0. 2
and 0. 5 ppm of Hg++ in solution. The slow approach to equilibrium is
in agreement with the results of Malcolm and Kennedy [2], who find that
ion-exchange equilibria in coarse sediments may take several weeks to
reach substantial completion.
Some results obtained at lower concentrations of mercury are shown in
Figure A-2. At 137 ppm of mercury, the results are similar to those
at 412 ppm, except that the 7-day point is somewhat higher than the
2-day point. (The 4-day points on these samples were rejected because
of procedural problems. ) The difference, however, is comparable to
analytical error.
At 41. 2 ppm, the equilibrium appears to have been reached by the first
day, and the readings increase regularly thereafter. The increase, how-
ever, is again comparable to analytical error.
The two reverse equilibration runs at 11 days (points marked with "R"
on Figure A-2), show an increase in mercury concentration. We sub-
sequently found that this sediment is very sensitive to aging in the
presence of air. When aged, the sediment loses a part of its mercury-
binding capacity, probably due to the oxidation of sulfides contained in
72
-------
O T3
g-
w ct-
OQ
H
O
H
03
13
(X
O
h-1
(U
3
H"
rt-
tr
o
"3
r-H
O
c
O
I)
O
a
o
U
3
u
1. 5
1.0
0. 5
Concentration: 412 ppm Hg on dry sediment
Temperature: 24-25°C
pH: 6.6^0.2
© Mercury added with water
x Mercury added with sediment
5 6
Time (days)
10 11
-------
OQ
t-j
(D
I
INJ
O >
O "0
o
c
en
>
TO
C
3
p)
C/3
13
o
i—•
p>
3
ff
tr
G
O
O
CO
c
O
u
C
o
>H
U
-------
the freshly dug sediment. The rise in mercury concentration at 11
days is probably due to oxidation of the sediment by prolonged agita-
tation with air-saturated solution.
Similar experiments with Acton peat and with Georgia kaolin indicated
that equilibrium was rapidly attained and was substantially complete
in 7 days. This period of time was therefore adopted as a standard
length of run for all materials.
There is some evidence, however, that a 7-day run does not produce
equilibrium in all cases. In the case of coarse pyrite, the concentra-
tion appears to be limited by reaction rate rather than by equilibrium.
This question is further discussed below.
Discussion of Results
Acton Sediments
Partition coefficients measured for Acton sand and Acton peat are
summarized in Table A-l. The first five runs of this table show that
the partition coefficient increases from around 4 x 10" at the lower
concentrations to 1.3 x 10~3 at 412 ppm in the dry sediment. The
increase in value with concentration shows that several different
absorption mechanisms are active in this complex mixture, and the
stronger binding sites are the first to become saturated.
Run A-42 was made with low oxygen and shows little difference
from run A-20.
Further work •with this sediment showed that its mercury-binding
capacity is diminished on storage, as shown by run A-36 with sand
which had been stored in an open tub in the laboratory for 5 weeks
after digging. The effect is even more marked if the aged sand is
allowed to become completely dry, as shown in runs A-37 through
A-39. In these runs, also, there is an increase in partition
coefficient with increasing mercury concentration.
Further experiment showed that, -when the fresh sand was acidified
with dilute H2SO4, a distinct odor of H2 was given off. This was
probably due to the decomposition of a trace of FeS, since the sample
had already been observed to contain considerable iron. When the dry
sand was acidified, however, no trace of P^S was detectable. This
indicates that the FeS had been lost on standing, probably by oxidation
to an iron sulfate. This loss of FeS would account for the loss of
mercury-binding ability in the aged or dried sand.
Following these experiments, all sediment samples were stored in
closed barrels and covered with several inches of water in order to
retard oxidation as much as possible.
75
-------
Table A-l
Partition Coefficients for Acton Sediments with Mercuric Chloride at 24-25°C
Run
No.
A~25
A-27
A-24
A-20
A-16
A-42
A-36
A-37
A-38
A- 39
B-3
B-4
B-5
B-6
B-7
B-8
Time
j(days)
6
6
7
(
7
6
7
7
7
7
4
7
7
7
4
2
Description
Acton sand
(fresh)
Acton sand
(fresh)
Acton sand
(fresh)
Acton sand
(fresh)
Acton sand
(fresh)
Fresh sand
low C>2
Aged sand
( 5 weeks)
Aged and
dried sand
Aged and
dried sand
Aged and
dried sand
Acton peat
(fresh)
Acton peat
(fresh)
Acton peat
(fresh)
Acton peat
(fresh)
Acton peat
(fresh)
Acton peat
(fresh)
Mercury Cone, (ppm)' *• ^ ^H£O
Dry ' K~(W+1
Sediment Water ' L s Jsed
13.7 . 0052 : 3. 8 x 10"4
13.7 .004
41.2 . .024
137
412
137
258
244
10.0
30
1430
1430
476
2670
2670
2670
. 048
0. 52
. 045
10.0
33.8
0.42
. 046
<. 0002*
< .00002*
<. 0002*
. 0044*
. 074
. 062
2.9 x 10'4
5.8 x 10-4
3. 5 x ID"4
1. 3 x 10-3
3. 3 x 10-4
.037
0. 14
.042
.014
<1. 4 x 10"7
< 1. 4 x 10-8
<5. 3 x 10-7
1. 65 x ID'6
2. 8 x 10-5
2. 3 x 10"5
PH
6.6
6.6
6.6
6.6
6.2
6.2
5.7
6.0
6.2
6.4
4.9
5. 2
5. 3
5. 1
4.8
5. 0
!
Dissolved
Oxygen
(ppm)
4
5
3
0. 0
6.5
5
4
3.5
0.0
0. Of
0.0
0.0
1.0
1.0
76
-------
Table A-1 (continued)
Run
No.
B-9
B-10
B-18
B-21
B-22
B-23
Time
(days)
7
7
7
7
7
7
Description
Acton peat
(fresh)
Acton peat
(fresh) low O2
Acton peat
Aged 2 months
Fresh peat, 3 g
cysteine HC1
Fresh peat
3 g thiourea
Fresh peat
3 g Na2S203
Mercury Cone, (ppm)
Dry
Sediment
8000
2670
1335
1320
1365
890
Water
. 192
.0258
.0031
8.8
. 044
.09
^+\0
v -
FH^+I
L"r Jsed.
2.4 x 10-5
9. 7 x 10~6
2. 3 x ID'6
6. 7 x lO-3
3. 2 x 10-5
1. 0 x 10-4
pH
4.7
5. 1
4.8
4. 1
5.4
5.8
Dissolved
Oxygen
(ppm)
0.0
0.0
0.6
0.6
0.4
0. 5
>-High-sensitivity analyses by Jarrell-Ash Division.
77
-------
Runs B-3 through B-5 were made •with Acton peat at mercury levels up
to 1430 ppm. In no case was any mercury detectable in the filtrates,
although the detection limit for run B-4 (as reported by Jarrell-Ash)
was . 00002 ppm. These results lead to a partition coefficient of less
than 1.4 x 10~° for this run. In the case of run B-3, a low value was
reached in only 4 days, indicating that this sediment approaches
equilibrium even more rapidly than the Acton sand.
Runs B-6 through B-9 indicate that the partition coefficient increases
as the mercury content is raised to 8000 ppm. This again indicates
that this sediment contains several types of binding sites.
It is interesting to note that, although no special precautions were
taken to exclude air, little or no dissolved oxygen was found in the
overlying liquid in the above runs. We attribute this to the chemical
oxygen demand of this highly reduced sediment. The oxygen is
probably consumed by reduced iron species, such as Fe(OH)2 and FeS.
This question is further discussed in Appendix C, where it is shown
that the dissolved oxygen can be reduced to low values by this sedi-
ment within 10 or 15 minutes.
Run B-10, with low oxygen, was substantially equivalent to runs B-6
and B-7 made in the presence of 2-300 ml or air. This indicates that
a considerable amount of oxygen can be consumed by this sediment
without impairing its mercury-binding capacity. When a sample of
this peat was aged in air for two months, however, the partition
coefficient increased by about two orders of magnitude, as shown by
run B-18. If such a sediment were dredged up and placed in a land-
fill, we would expect mercury to be released as the spoil became
permeated •with oxygen-rich surface waters.
Further experiments showed that the Acton peat gave off a strong
odor of H2S when acidified with dilute H2SO4, indicating an even
higher FeS content than the sandy sediment from the same site. The
presence of iron was confirmed by precipitating Fe(OH)3 from the
acid extract. The excellent mercury-binding capacity of this sediment
is probably due to its sulfide content, together with the anoxic condi-
tions maintained by its high biochemical oxygen demand.
The ferrous sulfide in these sediment probably originates from the
biochemical reduction of sulfate ions in the presence of iron by the
organic materials. This conclusion is in agreement with the results
of Tuttle et al. [8], who have found that acid mine drainage (essentially
ferric sulfate) can be reduced to FeS by heterotrophic bacteria with
sawdust as the only nutrient. This natural scavenging mechanism may
provide a powerful tool for the control of mercury in contaminated
waters.
We were unsuccessful in obtaining quantitative analyses for sulfur as
sulfide in the Acton sediments, but such an analysis is obviously of
78
-------
major importance in assessing the natural binding capacity of sedi-
ments. Sulfide analyses should be obtained as a part of any large-
scale operation.
Runs B-21 through B-23 of Table A-1 show the results of dding
water-soluble mercury-complexing agents to contaminate' Acton peat.
The object of these experiments was to bring the mercury into solution
so that is could be removed from the sediment. These results relate
to the problem of trying to decontaminate dredge spoil before damping.
Of the additives tries, cysteine hydrochloride (run B-21) was the most
effective in solubilizing mercury. Even in this case, however, over
99% of the mercury would be bound to the sediment in equilibrium with
an equal weight of water. Thiourea and sodium thiosulfate were even
less effective. The colloidal nature of these peaty sediments is such
that they are not readily amenable to washing with water. These
results indicate that it is probably not practical to remove mercury
from spoil by this method, even if cysteine or similar" complexing
agents were available in sufficient quantities.
Kaolin Clay and Silica
Results of these materials are shown in Table A-2. Runs C-l through
C-14 were made -with straight clay containing no additives. The
mercury analyses on the sediment were corrected where necessary
for the mercury contained in the pore water of the moist filter cake.
The results show that the Hg concentration in the clay increases with
liquid concentrations up to a. value of 80 or 90 ppm. This maximum
value is reached at a. liquid concentration of 1 to 6 ppm, and further
increase of the liquid concentration to about 166 ppm produces no
further increase in the sediment concentration. This kind of result
would be expected if the clay had a limited ion-exchange capacity
which become saturated at these relatively low concentrations of
mercury.
The decreasing values of K_ at liquid concentrations of 1 ppm and
below indicate that a variety of types of binding site are active and that
the binding may become very effective at low concentrations of mercury
in the solid.
Run C-l shows the effect of controlling the pH by addition of CaCC>3.
A marked improvement over straight clay is found.
Runs S-l and S-2 show that the ground silica (240 mesh) has even less
mercury-binding ability than the kaolin. This binding capacity is
probably due mainly to surface adsorption.
Ashland Sediments
Our first measurements of the partition coefficients of these sediments
were based on mercury analyses of the filtrates made by the usual
79
-------
Table A-2
Partition Coefficients for Minerals and Sediment from Ashland, Mass, at 24-25 C
Run
No.
C-l
C-2
C-3
C-4
C-13
C-14
C-7
S-l
S-2
ASH-
1A
ASH-
IB
ASH-
2A
ASH-
2B
Time
(days)
8
4
4
7
6
6
7
7
7
7
7
7
7
Description
Georgia
kaolin
Georgia
kaolin
Georgia
kaolin
Georgia
kaolin
Georgia
kaolin
Georgia
kaolin
5 g CaCO3
Silica,
240 mesh
Silica,
240 mesh
Ashland sed.
October
Ashland sed.
October '71
Ashland sed.
December '71
Ashland sed.
December '71
Mercury C
Dry
Sediment
82
83
88
31.6
90
10
314
33
29.6
31.8
31.8
102. 5
98. 5
one. (ppm)
Water
40. 1
166.4
38.8
1. 08
6.2
0. 0175
11.5
49
31.6
0.0016
0.0016
0.0028
0.0028
,. ++,
1Y — r, T -1-4-1
fR§ led.
0. 49
2. 0
0. 44
0. 034
0. 069
1.8 x lO-2
0. 037
1. 5
1. 1
5.0 x 10-5
5. 0 x lO-5
2.7 x 10-5
2.6 x lO-5
PH
5. 2
5.2
5.2
5. 1
5.2
5.4
7.4
6.8
7.4
6.4
6.4
6.3
6.3
Dissolved
Oxygen
(ppm)
5. 0
6.0
5. 0
10. 0
.
7. 5
6. 0
8.0
0. 5
0. 5
1.9
1.9
80
-------
room-temperature oxidation with permanganate. We subsequently-
found that, if these filtrates were digested with nitric-sulfuric acids
under reflux, the measured mercury content was increased by a factor
of 5 to 10. This result indicates that most of the mercury was organ-
ically bound, yet less than 1/2% of the total mercury in these samples
was in the methylated form.
We therefore postulated that most of the mercury in these sediments
was in the form of mercurated anthraquinone sulfonic acids or similar
derivatives. Such mercurated species are a probable by-product of
the mercury-catalyzed alpha-sulfonation of anthraquinone. In order
to check this hypothesis, we dissolved 5 g of the sodium salt of
commercial 1-anthraquinone sulfonic acid in 250 ml of boiling water
and allowed the bulk of the dissolved salt to crystallize out. The
mother liquor was considered to simulate the by-products of the
commercial sulfonation operation. It was found to contain 7.-5 ppm
of mercury when analyzed by reflux digestion, but only 1. 5 ppm or
20% of the total when analyzed by room-temperature digestion.
This result is consistent with the hypothesis that most of the mercury
in the Ashland sediments is organically bound to anthraquinone
derivatives.
As a result of this finding, we have revised some of our earlier work
with the Ashland sediments. The last four runs of Table A-2 show
some revised values of the distribution coefficient. It should be noted
that these values of 2 to 5 x 10"^ are not directly comparable to
coefficients measured with HgCl2, since the Ashland sediments con-
tain different species of mercury compounds.
The distribution coefficient of this material after aging is further
discussed below in Apendix C under treatment of dredge spoil.
Natural and Fired Pyrite
The results obtained with Climax pyrite are summarized in Table
A-3. Runs C-5 and C-6 show the results of adding 3% of pyrite (as
received, except far washing and drying) to Georgia kaolin.
Comparison with runs C-l and C-2 of Table A-2 shows that the amount
of mercury jbound to the solid is about doubled but that the concentra-
tions in solution are still much higher than would be expected from the
formation of mercuric sulfide. It therefore appears that the reaction
of the pyrite is very slow and is probably limited by insufficient
surface area. Alternatively, the reaction may be limited by a layer
of highly insoluble ferric hydroxide on the surface of the pyrite
particles.
Runs C-37 through C-55 were designed to test these hypotheses. For
these runs the pyrite was hand-ground to reduce its particle size
(see section on materials), and various iron complexing agents were
81
-------
Table A-3
Partition Coefficients for Pyrite Additives with Mercuric Chloride at 24-25°C
Run
No.
C-5
C-6
C-37
C-38
C-39
C-40
C-45
C-46
C-55
C-56
A
C-63
C-82
Time
(days]
7
7
7
7
7
7
7
7
7
7
7
7
(a)
Description
3 g pyrite, as
received
3 g pyrite, as
received
5 g ground py-
rite, 10 ml
acetic acid
Same as C-37
+ 0.6g BHA(b)
Same as C-37
+ 0. 85 g KF •
2H20
Same as C-37
+ 0. 5 g .oxalic
acid
Same as C-37
+ 10. 0 g oxal-
ic acid
Same as C-37
+ 15. 8 g KF-
2H2O
5 g ground py-
rite, low C^,
200 ml saw-
dust ext.
Same as C-55
+ 5g Na2S03
5 g milled py-
rite -325 mesh
Same as C-63
5gCaC03, 5g
Fe, low O2
Mercury Cone, (ppm)
Dry
Sediment
193
176
300
300
173
171
210
236
300
59
300
299
Water
31.8
20.8
.003
. 044
24.2
24. 5
17. 1
12.2
0.78
46.4
. 0025*
<. 00004*
t»^H20
*~IH«++].ed.
0. 16
0. 12
1. 0 x lO-5
1. 5 x 10-4
0. 14
0. 14
0. 08
0. 05
2.6 x 10-3
0. 79
8. 3 x lO-6
<1. 3 x 1Q-7
PH
5.0
5. 1
3.4
3.4
4.0
3.0
2.4
5.8
4.6
7.9
4.5
7.0
Dissolved
Oxygen
(ppm)
10.0
9.0
2.0
6.2
0.8
7. 1
1. 1
82
-------
Table A-3 (continued)
Run
No.
C-10
C-Z5
C-41
C-4Z
C-43
C-44
C-56
Time
(days)
7
7
•7
7
7
7
7
(a)
Description
5 g fired py-
rite
5 g fired py-
rite
5 g ground,
fired pyrite,
10 ml acetic
acid
Same as C-41
+ 0. 6g BHA(b)
Same as C-41
+ 0.85 g KF-
2H2O
Same as C-41
+ 0. 5 g oxalic
acid
5 g ground,
fired pyrite;
low O2, 200
ml sawdust
ext.
Mercury Cone, (ppm)
Dry
Sediment
377
321
300
300
300
210
300
Water
0.21
. 154
. 0006
.0005
0. 09
16.9
.027
^++^2o
Y
[Hg+ + ] ,
L 6 Jsed.
5. 6 x 10-4
4.8 x 10~4
2. 0 x 10-6
1. 7 x 10-6
3. 0 x 10"4
0.08
9.0 x 10-5
PH
5.4
5. 0
3. 5
3. 3
3.8
2.4
4.6
Dissolved
Oxygen
(ppm)
4.0
9. 5
2.6
:;cHigh- sensitivity analysis.
(a) All runs made with 100 grams oven-dried Georgia kaolin.
(b) BHA = Benzohydroxamic acid
83
-------
added in order to try to dissolve any layer of Fe(OH)3. The best runs
of this series were runs C-37 and C-58 with and without the addition
of acetic acid. The acid appears to produce a slight improvement,
but neither run gives the low values of soluble mercury expected from
a sulfide. The remaining runs of this series show that the three iron-
complexing agents, benzohydroxamic acid, potassium fluoride, and
oxalic acid, have a deleterious rather than a beneficial effect.
Run C-55 was intended to be low in oxygen, but the oxygen concen-
tration of 6. 2 ppm shows some air was inadvertently dissolved. This
run also contained 200 ml of the liquid extract from decomposing
sawdust, in the hope that this would simulate the bioreducing action of
natural organic sediments. Not only was this attempt unsuccessful,
but the extract appears to have solubilized some mercury.
To test the possibility that the reaction may be inhibited by a surface
layer of free sulfur, run 56A was made with the addition of 5 grams of
sodium sulfite. This reagent is known to dissolve elemental sulfur
to form sodium thiosulfate. The concentration of mercury in solution,
however, was increased by this treatment.
The best results with pyrite were obtained by mechanically milling
the material to -325 mesh, as shown in runs C-63 and C-82. The use
of CaCO^ and low-oxygen conditions in run C-82 gave a slight improve-
ment. In neither of these runs, however, is the mercury concentration
reduced to a value comparable to the Acton peat.
Since pyrite is structurally a disulfide (i. e. , it contains the 82" ion)
rather than a simple sulfide, we considered that a simple iron
sulfide (FeS) should be evaluated. One way of obtaining this material
is to heat pyrite in the absence of air to a temperature in excess of
about 700°C, when one atom of sulfur is lost, according to the
equation:
FeS2 —- FeS+ S
A sample of calcined pyrite was prepared by this method and, accord-
ing to weight loss measurements, it was about 35% converted to FeS.
Runs C-10 and C-25, made with this material, show a considerable
improvement over straight pyrite, but the mercury in solution is still
much higher than expected from theoretical considerations. Some
of the observed improvement may be due to increase of surface area
during firing.
Runs C-41 through C-44 were made with hand-ground (see section on
materials), fired pyrite in combination with various iron-complexing
agents. Benzohydroxamic acid (run C-42) appears to produce a
slight improvement over straight acetic acid, but the difference is
probably •within experimental error. None of these materials is
significantly better than the -325 mesh pyrite.
84
-------
Run C-56 shows a deleterious action of the sawdust extract, similar
to that found in run C-55.
Other Inorganic Sulfides
Further results with various inorganic sulfides are shown in Table
A-4. The effect of adding 5 grams of calcium sulfide to Acton sand
is shown by run A-33. The high concentration of mercury in solution
is probably due to the formation of soluble HgS2= or similar species
at the high pH produced by this excess of sulfide. If the excess of
sulfide is reduced, as in run C-ll, a considerable improvement is
effected. Still better results (K <1. 6 x 10"^) were obtained by
restricting the supply of oxygen. The actual concentration of oxygen
could not be measured because of interference by sulfide. The addi-
tion of acetic acid (run C-31) also produces an improvement over run
C-ll, possibly by preventing formation of
A very finely divided form of FeS can be prepared in situ by reaction
of CaS with FeSO4. Run C-19 is comparable to runs C-10 and C-25
of Table A- 3, which were made with fired pyrite. Runs C-20 and C-58
show that an improvement of two or three orders of magnitude is
achieved by restricting the oxygen. Run C-58 contained 200 ml of the
liquid extract from a decomposing sawdust slurry which had been
inoculated with Acton peat. It was hoped that this would simulate the
biochemical reducing effect of the peat. The small residual oxygen
shows that this was not completely effective.
Run C-51 shows the results obtained with zinc sulfide (precipitated
laboratory reagent). This material is comparable to ground pyrite.
Run C-54 contained 200 ml of sawdust extract and the oxygen was
restricted. The lack of improvement over run C-51 may be due to
some mercury- solubilizing effect of the sawdust extract (compare run
C-55, Table A-3).
Run C-52 was made with free sulfur in the form of reagent- grade
flowers of sulfur (sublimed). No appreciable mercury -bind ing action
is observed (compare run C-l, Table A-2).
In seeking ways to improve the utilization of pyrite, we heated a
sample with an equimolar amount of powdered electrolytic iron. The
reaction:
FeS2 + Fe — *- 2FeS
took place smoothly at a low red heat. After cooling, the product was
easily disintegrated into a coarse powder.
A similar reaction was carried out using an equimolar quantity of
zinc dust in place of iron. The reaction:
85
-------
Table A-4
Partition Coefficients for Various Inorganic Sulfide Additions with HgCl., at 24-25°C
Run
No.
A- 33
C-ll
C-12
C-31
C-19
C-20
C-58
C-51
C-59
C-52
C-69
C-70
C-71
C-72
Time
(days)
6
7
7
7
7
7
7
7
7
7
7
7
7
7
(a)
Description
Acton sand
5 g CaS
. 180 g CaS
. 180 g CaS
low C"2
. HOg CaS, 25
ml acetic acid
. 11 g CaS, 1. 0 g
FeSO4- 7H2O
Same as C-19
low C>2
Same as C-20
200 ml saw-
dust ext.
5 g pptd. ZnS
Same as C-51
low O2, 200 ml
sawdust ext.
5 g flowers of
sulfur
5 g fired FeS
5 g fired FeS
5 g fired
FeS • ZnS
5 g fired
FeS- ZnS
Mercury Cone, (ppm)
Dry
Sediment
412
1260
1260
300
378
378
300
300
300
79
268
101
263
85
Water
10. 1
.248
< . 0002*
.001
1 .68
< .0002*
.0092*
. 00053*
.0013*
32. 5
.0175
.0125
.0152
.0007
[Hg++lH20
K>
W+Jsed.
.024
2.0 x 10"4
<1.6 x 10"7
2. 2 x 10"6
4. 5 x 10"4
<5.3 x 10"7
3.0 x 10'5
1.8 x ID'6
4.0 x 10'6
0.41
6. 5 x 10"5
1.2 x 10'4
5.8 x 10-5
8. 2 x 10"6
PH
9.0
8.4
8.4
3.4
4. 3
4. 5
5.9
5. 1
5. 2
5. 4
5.6
5.8
6.2
6.8
Dis solved
Oxygen
(ppm)
4.0
1.0
0.6
7.0
0.6
8.2
5.2
5.3
4.2
4.6
* High-sensitivity analysis.
(a) All runs prefixed with "C " were made with
100 grams oven-dried Georgia kaolin.
86
-------
FeS2 -f Zn —*- FeS + ZnS
was not violent but was somewhat more energetic than in the previous
case. The product may be considered approximately equivalent to
natural sphalerite (a natural ZnS, some of which contains much iron).
No fine grinding was performed on either of these materials.
Runs C-69 through C-72 show that these two materials offer no appre-
ciable advantage over milled pyrite as far as the partition coefficient
is concerned. They are, however, less finely divided than the milled
pyrite and would therefore have less tendency to become re suspended
in the water column. If they could be made economically and utilized
efficiently, they might compete with milled pyrite.
Miscellaneous Additives
Results obtained with various additives to kaolin clay and with chicken
feathers are shown in Table A-5.
Run C-22 was made with calcium carbonate and ferrous sulfate, which
at this pH is oxidized in situ to Fe(OH)3. An improvement over
straight clay of about 100-fold is found.
In runs C-23 and C-50 the Fe(OH)3 was formed by reaction of calcium
carbonate with ferric sulfate. In this case the precipitation involves
no oxidation but only hydrolysis of the ferric salt. Run C-23 contained
insufficient CaCC^, as evidenced by the pH of 2.9. At this hydrogen
ion concentration the absorption of mercury is poor. Run C-20 con-
tained more CaCC>3 and had a pH of 6. 6. The distribution ratio is
considerably higher than in the case of run C-22, where the Fe(OH)3
was formed by oxidation. In all these cases the binding of mercury by
Fe(OH)3 is much less effective than the binding by sulfides.
Runs C-48 and C-49 were intended to learn if sawdust has an appre-
ciable reducing effect under these experimental conditions. These
experiments were prompted by literature reports that ferric sulfate
in acid mine drainage could be biologically reduced to FeS in the
presence of sawdust. The sawdust used in these experiments was
inoculated with turbid water from the Acton sediments in the hope of
introducing suitable bacteria. At the end of a week, however, run
C-48 still contained 5 ppm of dissolved oxygen, indicating no appre-
ciable biochemical oxygen demand. Run C-49, with ferric sulfate,
also showed no evidence of biochemical reduction. It appears likely
that the duration of these runs was too short to produce appreciable
biochemical reduction. Bacterial action may also have been inhibited
by the high concentrations of mercury in solution in these runs.
Sodium thiosulfate has occasionally been used as a precipitant for
mercury. When heated, the solutions deposit HgS. Run C-57,
however, shows that this reaction is not effective at room temperature
within 7 days. In fact, the thiosulfate appears to have a solubilizing
87
-------
Table A-5
Partition Coefficients for Miscellaneous Materials with HgCl2 at 24-25 C
R ' in
No.
C-22
C-23
C- 50
C-48
C-49
C-57
C-83
C-84
CF-1
CF-2
CF-3
CF-4
CF-5
CF-6
CF-7
CF-8
!
('days
j
7
7
7
7
<
7
7
1
1
1
7
3
7
7
1
:
< (a)
' Description
1 g CaCO?
2. 1 g FeSO4
1 1 g CaCG3, 3g
| Fe2(SO4)3-nH2O
5 g CaCOj, 3 g
Fe2(SO4)3-nH2O
200 g sawdust
22 g sawdust, 3g
Fe2(SO4)3-nH2O
5 g Na2S04-
5H2°
5 g Dow ex A- 1
5 g Dowex 1x8
5 ml HNO3
'-> g CaC03
No addition
No addition
No addition
No addition
No addition
No addition
Mercury C
D r \r
Sediment
321
12.8
251
.
300
172
59
135
87
1630
1360
1380
1785
1780
1780
1780
1680
[HP++!
one. (ppm) l K JH?O
Y ~~
Water [Hg + +] PH
Jsed.
0.91 2.8 x 10"3 6. 5
57 4.4 ! 2.9
9. 5 0. 04 6. 6
4.9 0.016 5.0
26. 5 o. 15 ; 2. 8
46.4 0.79 7.9
0. 022 1. 7 x lO-4 6.9
0. 008 9. 3 x 10"5 ; 7. 2
J '
0. 68 \ 4. 17 x 10-4 < 2. 0
0. 55 ; 4. 04 x 10-4 ; 7. 0
0. 43 3. 12 x 10-4 ; 6. 4
0. 062 3. 48 x ID'5 '. 6. 2
0.245 1.38 x 1C'4 6.2
0. 140 7.87 x 10-5 j 6. 5
0. 164 9.2 x 10-5 | 6.6
2.9 j 1.73 x 10'3 ; ---
(Dissolved
o
(ppm)
i 3.0
I
5.0
5.0
5.0
0.8
4. 4
4.6
1.9
0.7
0.9
0.5,
(a.) Buns prefixed with "C " made with 100 grams dry clay;
Runs prefixed with "CF" made with 14 grams of chicken
feathers (dry basis).
88
-------
effect on the mercury, probably by formation of a soluble thiosulfate
complex.
Runs C-83 and C-84 were made to test the mercury- binding capacity
of two commercial ion-exchange resins. Dowex A-l is a chelating
resin consisting of a styrene-divinyl benzene matrix to which are
attached iminodiacetate groups. Dowex 1x8 is a strongly basic anion-
exchange resin which may function by attaching HgCl4= or similar
anionic species. The results show that neither resin is as effective
as the sulfides with inorganic mercury. The cost of these resins
(about $400 per ft3 for A-l and $60 to $80 per ft3 for 1x8) will probably
be too high for expendable use on a large scale. One cubic foot of these
resins •weighs about 50 Ib.
The remaining runs of Table A- 5 show the results obtained by exposing
chicken feathers to mercuric chloride solutions under various condi-
tions. Similar work with wool, which is chemically similar to
feathers, has been reported by M. Friedman et al. [3]. Runs CF-1
through CF-3 show that pH has little effect on the sorption of mercury
by feathers. The remaining runs, equilibrated for various times, show
show the best ratios obtained at 7 days. It is possible that some small
improvement would result from longer equilibration times. The dis- •
tribution ratio of 3.48 x 10-5 obtained in run CF-4 agrees well with
the results of Friedman et al. for wool.
* >
It should be noted that the feathers produced a turbid solution contain-
ing much colloidal matter, which was difficult to remove on a mem-
brane filter. In practice, this colloid would probably become suspend-
ed in the water and would increase its total mercury content.
Long- Chain Alkyl Thiols
Table A- 6 summarizes the partition data we have obtained with long-
chain alkyl thiols (mercaptans).
Run C-7 is a control run made with calcium carbonate only and gives
a partition coefficient of . 037. The addition of n-dodecyl mercaptan
with calcium carbonate lowers the partition coefficient to the order of
10~8, as shown by runs C-15, C-16, and C-27. The observed con-
centrations of dissolved mercury are on the order of .00002 ppm
(. 02 ppb), which equals the best results obtained with Acton peat.
As shown by run C-27, these low mercury concentrations are obtained
even in the presence of 11. 5 ppm of dissolved oxygen. The molecular
weight of n-dodecyl mercaptan is around 202. Therefore, about
2 pounds of mercaptan will theoretically be required to complex
1 pound of mercury in the form of the mercaptide,
Runs C-24 and C-26 show that, in the absence of the calcium carbonate
buffer, the partition coefficients are higher. Run C-24 appears to be
somewhat out of line and is probably in error. Run C-26 shows that
even in the absence of CaCGs the concentration of mercury in solution
89
-------
Table A-6
Partition Coefficients for Long-Chain Alkyl Thiols with HgCl, at 24-25°C
R un
No.
C-7
C-15
C-16
C-24
C-26
C-27
C-53
A -47
A- 48
A- 57
A- 58
A- 59
A-60
A-61
C-17
C-18
Time
(days
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
8
(a)
Description
5 g CaCO3
1 g CaCCu,
1 ml NDM (b)
Same as C-15
low O2
1 ml NDM
1 ml NDM
5 g CaCO
1 ml NDMT^
1 ml NDM,
aged 36 days
Sand + Armac T
1 ml NDM
Same as A- 47
Same as A-47
fresh batch
Same as A- 57
•f 5 g CaCO3
Same as A- 57 +
5 g Fe, low 02
Same as A- 48 +
5 g Fe, low Q£
Same as A-60
old batch NDM
1 g CaCO
1 ml MTM^ (b)
1 g CaCO,,
1 ml THM (b)
Mercury Cone, (ppm)
Dry
Sediment
Water
314 11.5
378 . 00003*
378
321
1000
1000
300
307
108
112
125
93
93
92
378
378
<. 00002=:=
. 05
.00015*
. 00002*
.0154
. 0005
<. 00004*
<. 00004*
<. 00004*
.0016
.012
. 0094
< . 00002*
< . 00002*
fH^H20
¥ -
[Hg++l „
sed.
. 037
7. 9 x 10-8
<5. 3 x 10'8
1.6 x 10-4
1. 5 x 10"7
2.0 x 10'8
5. 2 x 10"5
1. 6 x 10-6
3. 7 x 10~7
3. 6 x 10~7
< 3.2 x 10~7
1. 7 x 10-5
1. 3 x 10"4
1. 0 x ID'4
< 5. 3 x 10-8
< 5. 3 x 10'8
Dissolved
| Oxygen
PH i (ppm)
7. 4 ; 7. 5
7. 5 i
'7.7 3.0
5. 1 10.0
4.4' 12.0
6. 8 11. 5
5. 0 15. 0
6.41 7.6
6.2 6.1
8.5 4.9
7.2 5.0
7. 3 1.1
6.8. 1.2
i
6-9! i.!
7.6
7.6 5.0
I
i
90
-------
Table A-6 (continued)
Run
No.
C-65
C-66
C-67
C-68
Time
(days)
7
7
7
7
(a)
Description
1 ml DDD (4)
1 ml DDD
5 g Zn dust
1 ml DDD
5 g Fe powder
1 ml DDD
100 ml saw-
dust ext.
Mercury Cone, (ppm)
Dry
Sediment
300
300
300
300
Water
6.2
.0114
.0254
3.8
[Hg++]H20
Y ~
[«g++]sed.
. 017
3.8 x 10-5
8. 5 x 10-5
.013
PH
5.4
5.4
7.6
6.0
Dissolved
Oxygen
(ppm)
6.2
0.8
0.8
6.4
#High-sensitivity analysis
(a) All runs prefixed with "C" made with 100 grams
dry kaolin.
(b) NDM = n-dodecyl mercaptan
MTM = mixed tertiary mercaptans
• THM = t-hexadecyl mercaptan
DDD = di-t-dodecyl disulfide
91
-------
is reduced to well below 1 ppb with 1000 ppm in the solid.
Run C-53 was made with 1 ml of n-dodecyl mercaptan, which had
been added to 100 grams of dry clay and aged in air for about 2 months.
During this time the mixture lost most of its odor, and it was thought
that the mercaptan was oxidized to a disulfide by the reaction
2RSH + 1/2 O2 -*~ RSSR + H2O
Surprisingly, this preparation retained a substantial effectiveness as
a mercury scavenger, as shown by the partition coefficient of 5.2 x
10" . It is probable that the mercaptan was not completely oxidized
during this period and that the loss of odor was due to selective oxida-
tion of some volatile impurity.
Since the long-chain thiols are oily liquids which float on water, it is
necessary to combine them with some denser material in order to
deploy them at the bottom of the water. As a sinking agent we chose
Acton sand, which was treated with a cationic surf ace-active agent to
render it preferentially wettable by oil. A number of suitable fatty
amines and their derivatives are commercially available for this
purpose, and we used Armac-T, which is described as a tallowamine
acetate and is made by Armour Industrial Chemical Co. These cationic
agents have the advantage that they will displace water from the surface
of the wet sand, thus avoiding the need to dry it before applying the
mercaptan. Since the Armac-T functioned satisfactorily, we made no
search for an optimum cationic agent.
We made a mixture of 550 g of wet sand, 1/2 g of Armac-T (about
1/10% on the dry sand), and 50 ml of n-dodecyl mercaptan with enough
water to permit stirring. After mixing, the mercaptan was found to
be well absorbed and to be held by the sand even after long periods of
submersion in water. Ten grams of this mixture was equivalent to
about one ml of pure mercaptan.
Runs A-47 through A-6l were made with the above mixture, which was
added to fresh Acton sand in amounts sufficient to give 1 ml of mer-
captan to 2-300 grams of sand. Runs A-47, A-48, and A-57 show that
the mercaptan is highly effective when applied in this way, although
the partition coefficients are not quite as low as those previously
obtained with clay. Run A-58, with CaCC^, produced a mercury con-
centration in solution of less than 0. 04 ppb, which is in the range of
our best previous results.
Runs A-59 through A-6l were made with low oxygen and with 5 g of
powdered iron added to provide additional reducing action. Although
the dissolved oxygen was reduced to about 1 ppm in these runs, the
partition coefficients are less favorable than those obtained in the
presence of oxygen. These results are unexpected, and we are not
yet able to offer an explanation.
92
-------
Run C-17 was made with a mixed tertiary mercaptan (MTM) which had
an average molecular weight of 212 and an average of 13. 3 carbon atoms
in the chain. In addition to having the thiol group in a tertiary carbon
atom, this material probably has a more-or-less branched hydrocarbon
chain. Its biodegradability may therefore be less than that of a straight-
chain primary mercaptan. The partition coefficient of
-------
Table A-7
Partition Coefficients for Methylmercuric Chloride with Acton Sediments at 24-25°C
Run
No.
A-43
A- 44
A-45
A-46
A-49
A-50
A-51
A-52
A-53
A-54
A-55
A-56
A-62
B-17
B-19
B-20
Time
(days'
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
Description
Acton sand
Acton sand
Acton sand
Acton sand
Sand + Armac
T, 1 ml NDM
Sand + Armac
T, 1 ml NDM
Sand + Armac
T, 1 ml NDM
Sand + Armac
T, 1 ml NDM
(fresh)
Same as A-52
+ 5 g CaCO3
Same as A- 53
low Q£
Same as A- 53 -r
5 g Fe, low C>2
Same as A-52 r
5 g Fe, low C>2
Same as A- 55
old batch NDM
Acton peat,
aged 2 months
Acton peat,
aged 2 months
Acton peat,
fresh
Acton peat,
fresh
Mercury Cone, (ppm)
Dry
Sediment
40. 5
»
40.7
11.9
12.0
100
30
19
106
96.5
119
126
132
111
2860
143
2630
1470
Water
4.6
3.9
1. 03
0.91
8.0
1. 21
0. 65
6. 35
4. 25
0. 02
0. 11
0. 35
0. 10
6. 5
0. 048
2.76
1.0
[Hg++]H20
* " ^++Jsed.
. 12
.096
.086
.076
.08
.04
.034
.06
. 044
1.7 x 10-4
8.7 x 10-4
2.7 x 10-3
9.0 x 10"4
2.3 x 10-3
3.4 x 10'4
1.0 x 10-3
6.8 x 10-4
PH
6.9
6.9
7.2
7. 1
6.3
6.4
6.4
6.7
6.9
7.0
7.6
8.4
7.2
5.3
5.4
5. 1
5.2
Dissolved
Oxygen
(ppm)
8. 1
7.7
6'! 5
7.6
6.0
5.7
4.9
9.4
7.8
1.2
1.2
1.2
1. 1
0.4
0.2
0.4
0.2
94
-------
Table A-8
Partition Coefficients for Methylmercuric Chloride with Various Additives at 24-25 C
Run
No.
C-32
C-33
C-60
C-61
C-62
C-64
C-73
C-74
C-75
C-79
C-80
C-81
C-86
Time
(days)
7
7
7
7
7
7
7
7
7
7
7
7
7
(a)
Description
Georgia kao-
lin, no additive
Georgia kao-
lin, no additive
5 g ZnS
5 g ZnS, low
O2, 200 ml
sawdust ext.
5 g CaCO3
1 ml NDM
5 g milled py-
rite -325 mesh
5g FeS-ZnS
5g FeS-ZnS
5g FeS-ZnS
5g CaC03, 5g
- 325 mesh py-
rite
5 g -325 mesh
pyrite, 5 g Fe,
low O2
Same as C-80
+ 5 g CaC03,
low C>2
5 g Dowex A- 1
Mercury Cone, (ppm)
Dry
Sediment
382
842
300
300
300
300
68
35
17
32
162
242
104
Water
470
1665
0. 45
0. 68
0. 24
37. 5
3.6
1.96
1. 11
51.5
13.6
8.8
1. 16
[Hg++JH20
K"^++]sed.
1.23
1.98
1. 5 x 10~3
2. 3 x 10"3
8.0 x 10"4
0. 125
. 054
.056
.065
1. 60
.084
.036
.011
PH
5. 1
5.0
5.4
5.3
7.0
4. 1
7.2
7.2
6.4
6.9
6.8
7.0
6.9
Dissolved
Oxygen
(ppm)
9.0
0.4
9.1
2.8
4.6
4.2
3.8
2.9
0.9
1.2
7.6
(a) All runs made with 100 grams oven-dried Georgia kaolin.
95
-------
strongly bound that is Hg , the partition coefficients being several
orders of magnitude greater in the case of methylmercury-. Compari-
son with the pervious tables shows that the materials which bind Hg++
most strongly are also the best binding agents for CH^Rg^. The
lowest partition coefficient we have found is the value of 1.7 x 10 for
n-dodecyl mercaptan coated on Acton sand (run A- 54).
Effect of Dissolved Chlorides on the Partition Coefficient
The effect of dissolved chlorides is important to the study of mercury-
sediment interactions in marine and estuarine environments and also
in estimating possible effects of runoff of road deicing salts. Table
A-9 gives the results of a series of runs made with various sediments
and additives in the presence of NaCl and of CaCl2-
Runs A- 24 and A- 26 show that 35 g per liter of NaCl (about the con-
centration of sea water) will increase the distribution ratio by almost
2 orders of magnitude. This corresponds to an increase in the
concentration of dissolved mercury by a factor of about 70.
The same general type of result is shown by runs B-4 through B- 14.
The effect is more severe at the higher mercury concentrations and
at the higher concentrations of chloride obtainable with CaCl..,.
Runs C-20 and C-21 give some results showing the effect of NaCl on
the precipitation of mercury as a sulfide. A very large increase of
solubility is produced, which is quite unexpected from the known
equilibrium constants of mercury with sulfide ion and with chloride ion.
Runs C-27 through C-30 show the effect of chlorides on the trapping
of mercury by n-dodecyl mercaptan. An increase of only 1 order of
magnitude in dissolved mercury is observed. Run C-29 shows an
increase of about 2 orders of magnitude in the presence of a very high
concentration of CaC^. Run C-30 confirms the results of C-28 at a
higher mercury level. With the exception of run C-29, these very
high mercury removals were obtained in the presence of 8. 5 to 12 ppm
of dissolved oxygen.
Runs C-74 and C-78 show that 3. 5% NaCl has little effect on the parti-
tion coefficient of methylmercuric chloride, probably because the
CHoHg ion is less strongly complexed by the chloride ion than is
Runs C-84 and C-85 show the effect of chloride on binding by the anion
exchange resin Dowex 1x8. About a sixfold increase in partition
coefficient was produced by NaCl.
Runs ASH-9A and 9B show the effect of 35 g per liter of NaCl on the
partition coefficient of the Ashland sediment. It is about an order of
magnitude greater in the presence of salt than in its absence (compare
runs ASH-1A, IB, 2A, and 2B of Table A-2).
96
-------
Table A-9
Effect of Soluble Chlorides on Partition Coefficient at 24-25 C
Run
No.
A- 24
A-26
B-4
B-6
B-ll
B-12
B-13
B-14
C-20
C-21
C-27
C-28
C-29
C-30
C-74
C-78
C-84
Time
(days)
7
6
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
(a)
Description
Fresh Acton
sand, no salt
Same as A-24
+ 35 g/1 NaCl
Acton peat
no salt
Same as C-37
Acton peat
35 g/1 NaCl
Same as B-ll
Acton peat
165 g/1 CaCl2
Same as B-13
Low O2, 0. 11
g CaS, 1.0 g
FeSO4-7H2O
Same as C-20
+ 35 g/1 NaCl
5 g CaCO
1 ml NDM (b)
Same as C-21
T 35 g/1 NaCl
Same as C-27
+ !65g/lCaCl2
Same as C-27
+ 35 g/1 NaCl
5 g FeS-ZnS
CH3HgCl
Same as C-74
+ 35 g/1 NaCl
5 g Dowex 1x8
Mercury Cone, (ppm)
Dry
Sediment
41.2
45.4
1430
2670
800
2670
800
1535
378
86
1000
300
300
1000
35
54
87
Water
. 024
1. 70
<. 00002*
. 0044
. 004*
25.0
1.58
215
. 0002*
26
. 00002*
. 00006*
.0025*
. 00024*
1.96
1.24
^^H20
^["^sed.
5.8 x 10-4
3.8 x 10'2
<1.4 x 10-8
1.6 x 10-6
5.0 x 10-6
9.4 x lO-3
1.9 x 10-3
0. 14
5. 3 x 10-7
0. 30
2. 0 x 10-8
2. 0 x 10-7
8. 3 x lO-6
2. 4 x lO'7
. 056
| .023
. 0081 9. 3 x 10'5
PH
6.7
6.6
5. 2
5. 1
4. 8
4.6
3.7
'3.6
4. 5
4. 5
6.8
7. 2
5.4
7. 2
7. 2
3.4
7. 2
Dissolved
Oxygen
(ppm)
3.0
7.0
0.0
0.0
0.0
0.0
0.0
0.0
1.0
—
11.5
8.5
2.0
12.0
4.2
6.2
4.6
97
-------
Table A-9 (continued)
Run
C-85
ASH-
9A
ASH-
9B
Time
(days)
7
7
7
(a)
Description
Same as C-84
+ 35 g/1 NaCl
Ashland, De-
cember, 1971
35 g/1 NaCl
Same as ASH-
9A
Mercury Cone, (ppm)
Dry
Sediment
105
100
100
Water
.062
.017
.019
[Hg++]H 0
K ~
[Hg++]sed.
5.9 x 10-4
1. 7 x 10"4
1.9 x 10"4
pH
5. 2
6.0
6.0
Dissolved
Oxygen
(ppm)
5.6
3.6
3.6
*High-sensitivity analysis
(a) Runs made with HgCl, except as noted.
98
-------
APPENDIX B
AQUARIUM EXPERIMENTS
This appendix gives the results of measurements of the release of
mercury from contaminated sediments and its uptake by goldfish.
The first section discusses some preliminary experiments, in which
the rate of extraction of mercury was measured for a period of several
weeks with no fish present. The following sections describe the results
obtained in aquariums with goldfish. As in the partition measurements,
we have chosen to work at moderate-to-high mercury levels in order
to study the effects of various addition agents.
Static Extraction Experiments
The results of some static extraction experiments with three different
types of sediment are shown in Table B-l. The mercury-laden sedi-
ments used in these experiments -were those filtered off from the
correspondingly numbered equilibrium runs. In making a static run,
the bottles containing the sediment •were carefully refilled with
distilled water (about 600 ml) and allowed to stand quietly for a number
of -weeks at room temperature. Water samples -were periodically with-
drawn, filtered through a 0.45 micron membrane filter, and analyzed
for mercury.
In the case of the clay sediments (runs C-l, C-2, and C-3), much of
the original mercury was lost with the filtrate during the original
equilibration run. In these cases, we have estimated the true mercury
content of the sediment at the beginning of the static experiment.
Table B-l shows the mercury concentrations in the water obtained in
the original equilibrium run and at the end of one, two, and three
weeks, in the static experiments. The "A" and "C" runs show a lower
concentration of mercury during the static runs than at equilibrium.
Runs A-30 and A-31, which contain n-dodecyl mercaptan, not only
show a low initial concentration, but the mercury analysis diminishes
with time.
The "B" runs (Acton black peat), however, show a higher concentration
of mercury in the static experiments than during equilibrium. This
appears to indicate that disturbing these sediments can release
mercury, possibly by oxidation of mercuric sulfide to more soluble
species. Runs B-3 and B-4 appear to indicate that mercury is
reabsorbed by this sediment between 7 and 14 days. This may be due
to the re-establishment of strongly reducing conditions at the bottom
of the sediment layer during standing.
99
-------
Table B-l
Extraction of Mercury from Sediment
under Static Conditions (all runs
made with HgCl-,)
No. (a)
A-16
A-20
A-24
A-30
A-31
A-36
B-3
B-4
B-5
C-l
C-2
C-3
Mercury
Content
in ppm
(dry basis)
412
137
41. 2
137
137
258
1430
1430
476
93.5
207
80.0
Mercury in Water (ppm)
Equilibrium (b)
0. 52
. 048
. 024
.0048
. 0024
10. 0
<. 0002
<. 0002
<.0002
40. 1
166.4
38.8
7 days
. 0424
.0138
.008
.0014
. 0010
. 036
.00785
0. 154
nil
.85
7.4
.78
14 days
.0327
.0180
.0213
. 0017
.00133
.0150
.0019
.057
.0046
.875
8.9
.68
21 days
. 0585
. 036
. 0154
. 00087
. 0003
. 026
. 057
. 75
8.0
.65
Remarks
Cont. n-dodecyl
mercptn+CaCO^
Same as above
Aged sediment
(a) Run numbers beginning with "A" denote Acton
sandy sediment, "B'1 denotes Acton black peat,
and "C " denotes kaolin clay.
(b) This column gives equilibrium concentrations
obtained with the same sediment by continuous
mixing for 7 days.
100
-------
Procedure for Aquarium. Experiments
These experiments were made in 5-gallon glass aquariums, 8 inches
by 14 inches by 10 inches deep. A one- or two-inch layer of the
sediment to be tested was placed in the bottom (after removing excess
water), and a solution of the required amount of mercuric chloride was
distributed over the sediment and stirred in. Methylmercuric chloride
was added as dry solid, which was well mixed with several hundred
grams of dry sand in order to facilitate uniform distribution. About
3. 5 kg (dry basis) of sand was used in one tank or about half that
weight of peat. The sediment was allowed to stand about a week with
daily stirring in order to equilibrate the mercury. The sediment was
then leveled and a layer of covering material added if required. The
aquarium was filled by carefully pouring water onto a floating wooden
board in order to minimize disturbance of the sediment.
The water used was from the Burlington, Massachusetts, municipal
water supply. It is obtained from local wells and has a pH of 5.9 to
6.4 and a hardness of 62-149 mg/1 CaCO^ equivalent.
The aquarium was allowed to stand one or two days before the fish
•were added. The experiment was started by adding three or four
goldfish about 2 inches long. The aquariums were aerated with
bubblers during the test.
This fish were fed about every other day with a commercial fish food
having the folio-wing reported analysis:
Crude protein--not less than 20%
Crude fat--not less than 2%
Crude fiber--not more than 5%
Ash--not more than 12%
Moisture--not more than 12%
and the following reported ingredients:
Wheat flour, meat meal, cornmeal, 2. 5% steamed
bone meal, 5% ground malt flour, 2% alfalfa leaf meal,
fish liver oil, fish meal, 0. 5% irradiated dried yeast,
0. 5% salt.
Our analysis of the food showed a mercury content of 0. 0405 ppm.
The amount fed -was not accurately measured, but it is estimated to
be on the order of 0. 05 g per day for each aquarium. This gives an
estimated mercury input of 0. 08 microgram over the 40-day duration
of an experiment.
After 9 days exposure, the fish were killed, gutted, and the heads and
tails removed. The remaining portion was then analyzed for mercury.
101
-------
New fish -were then added to the tank, exposed for 30 days, and
analyzed in the same way.
It should be noted that the mercury content of the fish is reported on
a wet basis in order to be comparable with the FDA guidelines on edible
fish (0. 5 ppm). Since the fish filets contain about 83% moisture, the
reported mercury content may be multiplied by 5 to give the approximate
analysis on the dry basis.
Results of Aquarium Experiments
The results of the aquarium tests with goldfish are summarized in
Table B-2. Runs A and B were made with Acton sand with 100 ppm
of mercury as HgC^. Tank A was uncovered and tank B was covered
with 1 inch of clean sand over the mercury-contaminated sand. The
initial water concentration in tank A of . 048 ppm is about the equili-
brium value obtained by tumbling the sand -with mercury solution for
7 days. (Compare run A-20 of Table A-l.) The final value was
.0002 ppm, as shown in Table B-2.
It is of interest to compare the mercury loss of the water with the
mercury uptake of the fish. If we estimate that aquarium A contained
6 liters of water, then the change in concentration over the first 9 days
indicates a loss of 259 micrograms of mercury. If we estimate the
total weight of the whole fish to be about 8. 4 grams and assume that the
increase in concentration of the filets is about the same as that of the
whole fish, we find a total uptake of 251 micrograms in 9 days. The
close agreement of these two figures indicates that most of the mercury
lost from the water was taken up by the fish. If anything, a slight
amount of mercury was taken up by the sediment, but there is no
evidence for a release of mercury. The 0. 08 microgram of mercury
added with the food is negligible in the above estimate.
For the following 30-day period we estimate that 28 micrograms of
mercury was lost from the water and 17 micrograms gained by the
fish. Again a slight absorption, rather than a release, of mercury
is indicated.
Applying the same estimates to the first 9 days of tank B, we find a
loss of 1. 8 micrograms in the water and a gain of 1.45 micrograms in
the fish, again a slight absorption of mercury by the sediment. The
30-day run in tank B showed a loss of mercury by both the water and
the fish: 0. 78 and 0. 54 micrograms, respectively, or a total of 1. 3
micrograms of mercury which must have been abosrbed by the
sediment.
In order to explain these results, we must postulate not only that the
bulk of the sediment is releasing mercury either not at all or at most
very slowly and that some part of it is actively absorbing mercury.
A reasonable postulate is that the excreta of the fish are taking up
the mercury. If we estimate that about 3. 0 g of fish food was placed
102
-------
Table B-2
Summary of Aquarium Experiments
O
oo
Tank
No.
A
B
C
D
E
Bottom Sediment
Type
2" Acton
sand
(HgCl2)
2 " Acton
sand
(HgCl2)
2" Acton
peat
(HgCl2)
2" Acton
peat
(HgCl2)
2" Acton
sand
(HgCl2)
Hg
(ppm)
100
100
185
100
100
Cover Layer
None
1 " clean sand
None
1/2" clean
sand
1/2" kaolin
Total Hg in Water
Time
(days)
0
3
9
11
30
41
0
3
9
11
30
41
0
3
9
11
30
41
6
9
18
39
0
9
24
39
ppm
.048
.017
. 0049
. 00136
.00020
. 00055
. 00040
.00025
.000066
.000122
.0004
.0003
.00037
.00020
.000056
.000077
.000055
.032
.001
.0003
Total Hg in Fish
Wet Basis (about 83% Water)
Exposure
(ppm)
. 176
. 24
. 176
.24
. 176
. 240
. 240
. 202
. 190
.077
Exposure
(ppm)
31. 1
2.22
. 348
. 176
. 32
.71
. 085
. 154
14. 15
. 50
Change
+29.9
+ 1.98
+ 172
-.064
+. 144
+.47
-. 155
-. 048
+ 13.96
+ .42
Remarks
-4 fish added.
-After 9 days
-New fish added.
-After 30 days
-4 fish added.
-After 9 days.
-New fish added.
-After 30 days.
-4 fish added.
-After 9 days.
-New fish added.
-After 30 days.
-4 fish added.
-After 9 days.
-After 30 days.
-9-day exposure, new
fish added.
-30-day exposure.
-------
Table B-2 (continued)
Tank
No.
F
G
H
Ashland
I
J
Bottom Sediment
Type
2" Acton
sand
(HgCl2)
2" Acton
sand
(HgCl2)
2" Acton
sand
(HgCl2)
1" Ash-
land Res-
ervoir
Sediment
(Oct. '71
sample)
2" Acton
sand
(HgCl2)
2" Acton
Hg
(ppm)
100
100
100
26
100
100
Cover Layer
1 /2 " ground
silica
ZnS, 5 g =
.015 lb/ft2 on
carrier
Mixed tertiary
me re apt an
. 0051 lb/ft2 on
carrier
None
Milled pyrite
.0291 lb/ft2
n-dodecyl mer-
captan(on sand
. 0247 lb/ft2
Total Hg in Water
Time
(days)
0
9
24
39
3
9
23
28
3
9
23
28
19
33
38
0
5
9
24
39
0
5
9
24
39
ppm
. 074
.0061
. 0008
.0018
.0008
.0035
. 0036
.0002
.0407
. 0144
.0096
. 0056
. 0075
. 0045
.0048
.0035
.00175
.0010
Total Hg in Fish
Wet Basis (about 83% Water)
Exposure
(ppm)
. 190
. 077
. 190
.143
. 190
. 143
. 190
. 190
. 143
. 143
. 143
. 143
Exposure
(ppm)
11. 02
1. 38
. 128
3. 55
. 724
.968
. 330
. 256
1.95
5.92
0.92
0.97
Change
+10. 83
+ 1. 30
- .06
+ 3.41
+ . 53
i . 83
+ . 14
+ . 07
+ 1. 81
+ 5.78
+ . 78
+ .83
Remarks
-9- day exposure, new
fish added.
-30-day exposure.
-9-day exposure, new
fish added.
-19 -day exposure.
-9-day exposure, new
fish added.
-19 -day exposure.
-19-day exposure.
-38-day exposure.
-9-day exposure, new
fish added.
-30-day exposure.
-9-day exposure, new
fish added.
-30-day exposure.
-------
Table B-2 (continued)
Tank
No.
K
L
M
N
P
Bottom Sediment
Type
2" Acton
sand
(HgCl )
2" Acton
sand
(HgCl,)
2" Acton
sand
(CH3HgCl)
2" Acton
sand
(CH3HgCl)
2" Acton
peat
Hg
(ppm)
100
100
30
30
0
Cover Layer
ZnS-FeS,
.015 lb/ft2
FeS
.015 lb/ft2
n-dodecyl mer-
captan ( on sand
. 0247 Ib/ft2
None
ZnS-FeS added
. 015 lb/ft2
Polyethylene
film
Total Hg in Water
Time
(days)
0
5
9
33
39
0
5
9
33
39
0
7
9
26
40
0
1
5
7
9
12
13
14
15
21
26
70
ppm
.0204
.0086
.0063
.0022
.0049
.048
.042
.035
.024
.024
4.6
3.2
3.4
3.8
4.0
4. 1
2. 3
2.2
2.0
2.9
Total Hg in Fish
Wet Basis (about 83% Water)
Exposure
(ppm)
. 143
. 143
. 143
. 143
. 143
. 143
.143
. 143
Exposure
(ppm)
11.8
14.5
16. 4
20. 3
11.2
12.8
16.8
19. 1
Change
+ 11.7
+ 14.4
+ 16. 3
+20. 2
+ 11. 1
+ 12. 7
+ 16.7
+ 19.0
Remarks
-9-day exposure, new
fish added.
-30-day exposure.
-9-day exposure, new
fish added.
-30-day exposure
-9-day exposure, new
fish added.
-30-day exposure.
-All fish died within
4 hours.
-All fish died within
4 hours. ZnS-FeS
added after water
sample
-No gas bubbles
formed under film
in 70 days.
-------
Table B-2 (continued)
Tank
No.
Q
R
S
T
U
Bottom Sediment
Type
Ashland
sediment
(Dec. '71
sample)
2" Acton
sand
(CH3HgCl)
2" Acton
sand
(Repeat of
Tank A)
2" Acton
sand
(CH,HgCl)
2" Acton
sand
(CH3HgCl)
Hg
(ppm
100. 5
30
100
30
30
Cover Layer
None
Polyethylene
film (.001")
None
Polyethylene
film over
milled pyrite
. 0291 lb/ft2
Polyethylene
film over
NDM-coated
sand
. 0247 Ib/ft^
Total Hg in Water
Time
(days)
0
6
9
27
30
0
6
0
1
2
7
0
7
10
25
40
0
7
10
25
40
ppm
. 0009
. 0007
. 0003
. 0005
. 0003
0.45
0. 33
0. 18
. 0056
.046
.012
. 012
. 009
. 002
.021
. 010
. 010
. 015
. 003
Total Hg in Fish
Wet Basis (about 83% Water)
Exposure
(ppm)
. 143
. 158
0. 16
0. 21
.21
.21
. 21
.21
.21
Exposure
(ppm)
. 327
. 109
6. 15
1. 7
9.3
7. 2
•
2. 25
4.04
1.63
Change
+ . 18
- .05
+6. 0
+ 1. 5
+9. 1
+7. 0
+ 2. 0
+ 3.8
+ 1.4
Remarks
-9-day exposure, new
fish added.
-30-day exposure.
-All fish died in 4-6
hours. Tank drained
and refilled.
-All new fish died in
4-6 hours. Experi-
ment terminated.
-4 of 5 fish died.
-Last fish died.
-Experiment termin.
-New fish added.
-30-day exposure.
-New fish added.
-30-day exposure.
-------
Table B-2 (continued)
Tank
No.
V
Bottom Sediment
Type
2" Acton
sand
(HgCl )
Mixea
with
CaCO,
.I66lb7ft2
+ NDM-
coated
sand. 0247
lb/ft2
NDM.
Hg
(ppm)
112. 5
Cover Layer
None
Total Hg in Water
Time
(days)
0
19
21
ppm
. 0016
. 00055
. 00045
Total Hg in Fish
Wet Basis (about 83% Water)
Exposure
(ppm)
.21
Exposure
(ppm)
1. 10
Change
+ .90
Remarks
-21-day exposure.
-------
in tank A over the 41-day run and that 19 micrograms of mercury was
lost during the same period, this would require a mercury concentra-
tion of 10 ppm in the total excreta, or about 50 ppm based on the
protein content only. The data of Friedman et al. for wool [4] (see
Appendix A) indicate that these values are consistent with very low
concentrations (less than 0.01 ppb) in the water.
The salient result of these experiments appears to be that in no case
were we able to observe any evidence of mercury release by the sedi-
ments after the fish were added.
The results obtained with Acton peat are shown by runs C and D. Note
that tank C contained 185 ppm of mercury in the sediment (dry basis)
rather than 100 as in the other cases. The initial concentration of
0. 0004 ppm Hg in the water is somewhat higher than would be expected
from the equilibrium experiments. (Compare run B-5, Table A-l.)
This higher value may be due to the fact that the aquarium was aerated,
while the equilibrium experiment was essentially anoxic. As before,
the mercury concentration in the water decreased steadily with time.
During the first 9 days, the water lost 0. 18 micrograms of mercury,
while the fish gained 1.21 micrograms. This is in contrast to the
results with Acton sand, where a net loss of mercury was observed.
The same type of result was observed during the 30-day run on tank C,
where the water lost 1.88 micrograms, but the fish gained 3.94 micro-
grams. In both cases we believe that the net gain in mercury was due
to ingestion of the organic bottom sediment by the fish. Th'is position
is supported by the fact that, when these fish were killed, the intestines
were found to be full of black sediment. These results indicate that,
in order to prevent the uptake of mercury by fish, not only must the
mercury content of the water be low, but they must also be denied
access to mercury-laden sediments high in organic matter.
Tank D of Table B-2 shows the result of covering the Acton peat with
a layer of clean sand. The mercury content of the water was so low
that some difficulty was experienced with the analysis and only two
acceptable results were obtained. The initial mercury content of the
water was 0.46 micrograms, while the total mercury loss by both sets
of fish was 1.7 micrograms. Thus, the system shows a net loss of
mercury to the sediment. This indicates that the 1/2-in. sand layer
was effective in preventing ingestion of the mercury-laden peat by the
fish. Another factor may be that the mercury-laden sediment was
held under anoxic conditions by the cover layer of sand, and the
mercury was thus prevented from being returned to solution by
oxidation or by methylation.
Tanks E and F show the results of covering mercury-contaminated
Acton sand with 1/2-in. of kaolin clay in 1/2 in. of ground (about
240 mesh) silica, respectively. The effect of these coverings is
shown by comparison with runs A and B.
108
-------
Although the results are somewhat better than those obtained with
uncovered Acton sand, they are less favorable than those found with
1 inch of clean sand cover. The cost of clay or silica would probably
preclude the use of a thickness much greater than 1/2 in. , while at
least about 1/2 in. is needed to obtain a reasonable coverage.
The finely divided nature of these materials permitted them to be
easily stirred up by the fish. Both tanks, but especially the one
containing kaolin, were turbid for the duration of the experiment.
The use of these materials is not recommended.
Tank G was covered with a thin layer of precipitated zinc sulfide on a
granular ceramic material (oil-absorbent). The mixture contained
about 5% of ZnS by weight and was applied at the rate of only 0. 3 Ib
per square foot. This formed a layer about 1/8 in. thick. Since the
ZnS was not adhered to the granules, it tended to become suspended
in the water and formed a turbid tank. This may have promoted
intereaction of the ZnS with dissolved mercury but would be undesirable
in a Irage-scale experiment. The results indicate that the ZnS was
very effective during the period of the first 9 days but that the fish
gained excess mercury during the second test period of 30 days. Any
conclusions based on these tests must be regarded as tentative, since
none of the results have been confirmed by repetition. We may
postulate, however, that the precipitated HgS may have been re-oxidi-
zed to a soluble form during the prolonged exposure to aerated water.
Tank H was covered with a mixture of 100 g of a porous-ceramic oil
absorbent (treated to render it oleophilic), 5 g of CaCO3, and 2 ml of
mixed tertiary mercaptans. This mixture was applied at the rate of
about 0. 3 Ib per square foot, forming a layer about 1/8 in. thick.
While this material was not entirely effective in preventing mercury
uptake by the fish, it was more effective than a much thicker layer
of clay or ground silica. Unlike the ZnS, it did not appear to lose
its effectiveness during the 30-day test. We believe that, since this
mercaptan was absorbed on the interior of the porous granules, it
was largely inaccessible to the dissolved mercury in the water.
Further experiments, in which a mercaptan is absorbed on the
external surface of sand particles, are discussed below.
The run marked "Ashland" (following run H) was made with a sample
of sediment obtained in October, 1971 from Framingham Reservoir
No. 2 in Ashland, Mass. This sample contained about 32 ppm of
mercury (dry basis) as the result of industrial pollution of a tributary
stream. The very low mercury content of the water appears to be
related to a high sulfide content of the sediment. No cover was used
in this aquarium, and the small mercury uptake of the fish is probably
due to the ingestion of the sediment by them. These results may be
contrasted with the results of analyses of fish taken from the actual
reservoir, most of which show over 1 ppm of mercury, with some in
excess of 10 ppm. It appears that aquarium tests do not simulate the
109
-------
actual environment as far as mercury is concerned. The difference
may be due to uptake of mercury in the food chain, to duration of
exposure, or to continuing or intermittent mercury input to the
reservoir from the tributary stream.
Runs I through L were made with mercuric chloride in Acton sand to
compare the effects of various mercury-complexing agents. Run I
shows the effect of milled pyrite, which was applied at the rate of
about 0. 03 Ibs per square foot of bottom area and was mixed into the
sand layer to a depth of about 1/2 in. The results can be compared to
run A, which is the control run for this group. Several points of
difference arise from this comparison. In the first place, the concen-
tration of dissolved mercury is higher after the first 9 days in run I
than in the control run. The final value for run I (0. 0075 ppm) can
also be compared to the value of 0. 0025 ppm obtained with a higher
total mercury concentration in the distribution experiment C-63,
Table A-3. On the basis of the distribution experiment, a lower
concentration would have been expected in run I if equilibrium had
been approached. Possibly the mercury was initially concentrated
by the pyrite and then oxidized to soluble form by the aerated
aquarium water.
The uptake of mercury by the fish (1.81 ppm) during the first nine days
of run I appears to be less than expected from the corresponding period
of the control run. During the final 30-day period, however, the
uptake was greater in run I than in the control, as expected from the
higher concentration of dissolved mercury.
Run J was covered with about 0. 024 lb/ft^ of n-dodecyl mercaptan.
The mercaptan was applied to 100 g of sand with the aid of the surface-
active agent Armac T. The mercaptan-coated sand was then stirred
into the top half-inch of the mercury-containing bottom sediment.
The concentration of dissolved mercury in run J shows a considerable
improvement over run I and over the initial nine days of run A. The
very low values of dissolved mercury obtained in run A after 40 days
suggest that this control run should be repeated (see run S below).
With respect to mercury uptake by the fish, run J shows a consider-
able improvement over both the control run and run I. This again
illustrates the superiority of the mercaptans over the inorganic
sulfides for complexing mercury.
Runs K and L, made with lower dosages of pyrogenic sulfides formed
by heating pyrite with powdered zinc or iron, respectively. It was
hoped that these sulfides would be more reactive than pyrite, but the
uptake of mercury by the fish shows that neither was highly effective.
The dissolves oxygen in these runs was in the range of 7.0 to 7.4 ppm,
and the pH was 7. 2, which is typical for these experiments. At the
end of the run the dissolved iron was 0.05 and 0. 06 ppm for K and L,
110
-------
respectively, and there was no yellow precipitate of ferric hydroxide
in either tank. Thus, there is no evidence of excessive oxidation of
these sulfides.
Runs M and N were made with 50 ppm of mercury as methylmercuric
chloride in Acton sand. This concentration is much higher than any
naturally occurring level of which we are aware. The values of 0. 1
to 0.4 ppm previously reported for Ashland sediment are probably
typical. The high values used in the aquariums, however, permit
comparative data to be rapidly acquired.
Tank M was treated with about 0. 025 Ib/ft2 of n-dodecyl mercaptan,
while tank N was an untreated control.
The water analyses show that the treatment lowered the concentration
of soluble mercury about 100-fold throughout the duration of the
experiment. The fish in the treated tank survived for the full test
period but picked up a considerable concentration of mercury. Under
actual field concentrations the pickup would have been much less.
In the control tank N, however, all the fish died within about four
hours, and they picked up more mercury in this time than in 30 days
in tank M.
Tank N was allowed to stand for 13 days with periodic water analyses,
and a new set of fish was added. These fish again died within four
hours. On the 14th day the tank was treated with 0. 15 lb/ft^ of ZnS-
FeS mixture, but this did not appreciably reduce the mercury in
solution, either immediately or on subsequent standing.
Tank P was covered with a polyethylene film over Acton peat. The
object wa-s to see if the peat would give off gas bubbles which would
gather under the film and tend to lift it. No bubbles were observed
during the 70-day test period at'room temperature.
Tank Q was made with the sample of contaminated sediment from the
Framingham reservoir, in Ashland, Massachusetts which was collected
in December, 1971. The results are quite similar to those obtained
with the earlier sample (October, 1971) of Ashland sediment, as
reported above. Again the uptake of mercury by the fish was small
compared to the values reported for fish from the reservoir. No
carp have been taken from the reservoir, but 4- to 6-in. bluegills
are reported to contain from 1. 5 to 3. 5 ppm of mercury. Since
bluegills are also a foraging fish, the data may be comparable to that
for carp. The higher levels in the reservoir fish indicate that their
mercury uptake is probably through the food chain rather than directly
from the water.
As discussed in Appendix A, the mercury content of the Ashland
sediments may be mainly bound as anthraquinone derivatives. The
identity and physiological action of these substances should be
111
-------
characterized in more detail. Runs R, T, and U were made to learn
the effect of a plastic film cover on Acton sand with Cf^HgCl and
various chemical treatments. For these runs, the polyethylene film
(1 mil) •was cut to fit the aquarium and laid on the surface of the
sediment before the tank was filled with water. The edges of the
film were weighted and sealed with a little clean sand.
Run R was made with no chemical treatment and only the plastic film.
Comparison with run N (without film) shows that the concentration of
mercury had been reduced by about an order of magnitude, both
initially and at the end of six days. Two sets of fish were used, one
at the start and one at the end of six days. Both sets of fish died with-
in four hours and the experiment was terminated.
Tank T was treated with 0. 291 Ibs/ft^ of milled pyrite, which reduced
the initial mercury concentration by an order of magnitude as compared
to run R. The mercury continued to fall off with time. The fish in this
tank survived for both the 9-day and 30-day test periods, and the
mercury gained during the first nine days was comparable to that
gained in 4-6 hours in the untreated tank R.
Tank U was treated with 0. 0247 Ibs/ft^ of n-dodecyl mercaptan under
the plastic film. The results may be compared with those of run M
(mercaptan with no film) and run R (film with no mercaptan). The
initial mercury concentration in tank U was about half that of tank M
and 1/20 that of tank R. The major part of the improvement is thus
attributable to the mercaptan treatment rather than to the film. Tank
U may have suffered some disturbance at about 25 days, which caused
a slight increase in mercury concentration. Despite this, however,
the pickup of mercury by the fish is appreciably lower than in tank M
or in tank T, which was treated with milled pyrite.
Tank S was a duplicate of run A with uncovered Acton sand and
The initial mercury concentration of tank S was about four times that
of tank A, and the fish survived only one or two days. The cause of
this difference in behavior is not known but may be due to differences
in age, organic content, or oxidation of the two sediment samples.
The result indicates that in making comparisons between different
treatments care should be taken that the sediments being compared
are as nearly identical as possible.
Tank V was run by mixing NDM- coated sand and calcium carbonate
•with the entire mass of mercury-laden sand. Only one run of 21 days
was made with this aquarium because of lack of time. Run V may be
compared with run J, in which the NDM- coated sand was used as a
cover layer only and to which no CaCOo was added. Although the
mercury concentrations in the water are lower in run V than in run J,
the uptake by the fish is about the same.
112
-------
It is estimated that the water lost about 9. 1 micrograms of mercury,
while the fish gained 7. 5 micrograms--a slight net loss. It is
probable that, if run V had been continued with new fish, the mercury-
uptake would have been much less, since only about 3.8 micrograms
of mercury remained in solution. A cover layer of clean sand would
probably have greatly improved the performance of this tank (compare
runs A and B).
113
-------
APPENDIX C
DREDGING OF MERCURY-CONTAMINATED SEDIMENTS
The dredging of mercury-laden sediments presents two major
problems of environmental impact: first, the dispersal of mercury
throughout the water column and, second, the disposal of the contam-
inated spoil. This appendix gives the results of laboratory experi-
ments aimed at providing some of the data needed to analyze these
problems.
Simulated Dredging Experiments
Dredging experiments were conducted in aquariums A, B, and C after
the fish had been removed and after the water had been allowed to
stand for several days. Dredging was simulated by removing about
a liter of bottom sediment by repeated dipping of a small spoon into
the tank. Mercury content, turbidity, dissolved oxygen, and pH were
measured before and after the disturbance, as shown in Table C-1.
As expected, both dissolved and total mercury are increased by the
dredging. In the case of the sand tanks, total mercury appears to
increase with turbidity. Because of the black color of the Acton
peat, the turbidity reading is high, although the actual weight-of
suspended material may be less than in the case of the sand.
Since the bulk of the mercury is in the suspended form rather than in
solution, it may be concluded that the treatment of the bottom with
mercury-complexing agents before dredging will have little effect on
the total waterborne mercury.
From the volume of water in the tanks (about 11 liters), the amount
of mercury resuspended in the water can be estimated as a fraction
of the mercury removed. These values range from about 2 to 10
per cent, as shown in Table C-l.
Dissolved oxygen and pH were little affected by the dredging except
in the case of the Acton peat, where the dissolved oxygen dropped
from 7. 1 ppm to 1. 8 ppm in the course of a few minutes. Further
experiments showed that the result could be repeated at will simply
by stirring Acton peat into oxygen-rich water. We believe that the
disappearance of oxygen is caused by rapid reaction •with reduced iron
compounds, such as Fe(OH)9 or FeS, which may be contained in the
sediment in finely divided form. This view is supported by the fact
that we obtained a similar result by stirring a slurry of precipitated
Fe(OH)2 (from FeSO4 anc^ NaOH) into oxygen-rich water. In small
ponds or in confined areas, this depletion of dissolved oxygen may be
detrimental to biota.
Table C-2 shows the results of dredging tanks I and J, which were
treated with milled pyrite and n-dodecyl mercaptan, respectively. It
115
-------
Table C-l
Simulated Dredging Experiments
Sediment
Hg Content (ppm)
Cover
Mercury Content
of Water (ppm)
Filtered
( Before Dredging)
Filtered
(After Dredging)
Turbid
(After Dredging)
Total Hg Removed
in Spoil (mg)
Total Hg Suspended
in Water (mg)
Percent of Hg
Suspended in Water
Turbidity (JTU)
Before Dredging
( With Fish Present)
Before Dredging
(No Fish)
After Dredging
(10 min. )
Dissolved Oxygen
(ppm)
Before Dredging
(10-15 min. )
After Dredging
10 Days
After Dredging
oH Before Dredging
pH After Dredging
Aquarium No.
A
Acton sand
100
None
. 00020
. 0056
0. 58
166
6.4
3.8
20
^
280
6. 0
5.4
6.9
7. 4
7. 2
B
Acton sand
100
1 " Clean sand
. 000122
.002
1. 66
193
18.3
9. 5
15
4
680
6.9
5. 1
7. 3
7. 4
6. 2
C
Acton peat
185
None
. 000056
. 0008
.208
100
2. 3
2. 3
2*40
6
1050
7. 1
1. 8
7. 0
5. 7
5. 4
116
-------
Table C-2
Simulated Dredging Experiments
Sediment
Hg Content (ppm)
Cover
Before Dredging
Total Hg (ppm)
Hg in Solution (ppm)
Turbidity (JTU)
Dissolved Oxygen (ppm)
PH
After Dredging
15 minutes
Total Hg (ppm)
Hg in Solution (ppm)
Turbidity (JTU)
Dissolved Oxygen (ppm)
pH
6 hours
Total Hg (ppm)
Hg in Solution (ppm)
Turbidity (JTU)
Dissolved Oxygen (ppm)
PH
24 hours
Total Hg (ppm)
Hg in Solution (ppm)
Turbidity (JTU)
Dissolved Oxygen (ppm)
PH
Aquarium No.
I
Acton sand
100
Milled pyrite
. 0291 Ib/ft2
0. 17
0.0.35
40
7.0
6.7
0.785
0.033
110
8.4
7. 1
0.445
0. 042
45
7.4
7.0
0. 170
0. 040
20
7.2
6.5
J
Acton sand
100
Milled pyrite
.0247 lb/ft2
0. 15
0.027
27
6.6
6.6
0.948
0.07
80
8.5
7.0
0.65
0.05
30
7.6
7.0
0. 181
0.014
20
6.0
6.7
117
-------
should be noted that the concentration of dissolved mercury in these
tanks is higher than that shown in Table B-2 at the conclusion of the
fish tests. This is because these two tanks were disturbed and a
allowed to settle for a day before the simulated dredging was started.
The total mercury in the water after dredging is due mostly to
suspended matter rather than to mercury in true solution. As
previously suggested, the use of mercury-binding agents has made
little difference in the total mercury concentration. The data at
6 hours and at 24 hours show that the initial conditions have been
largely restored at the end of the latter period. This is shown
graphically in Figure C-l, in which total and dissolved mercury are
plotted as a function of time.
Prediction of Mercury Redistribution
As a direct consequence of physically removing sediment by mechan-
ical means, a certain amount of benthal deposits will become dispersed
in the water column. This is primarily a result of washoff and over-
flow as the sediment is lifted. The use of vacuum dredging techniques
will significantly reduce the amount of material that becomes entrained
in the water column, since most of the particulate matter, when
disturbed, will be drawn by the vacuum system. From a knowledge
of the length of time that the material from either type of removal
operation is suspended, one can predict its redistribution by super-
imposing the effects of a current velocity. Previous tests have
indicated that the majority of mercury becomes adsorbed to organic
particles, hence, an estimate of the redistribution of mercury asso-
ciated with "lost sediment" from dredging can be obtained.
Sediments may be broadly classified as noncohesive or cohesive.
Noncohesive sediments consist of discrete particles whose movement
depends on their physical properties, such as size, shape, density,
and relative location with respect to other particles. In cohesive
sediments significant forces exist between the particles, and these
forces may inhibit the individual particle behavior. In the case of
dredging, sediment may be initially cohesive in character, but once
the bond is broken they may behave noncohesively as far as transport
is concerned. It is also possible that a reverse transformation may
occur; sediment initially noncohesive in nature may, through chemical
or physical reactions, coalesce.
Because natural sediment is irregular in shape, settling velocities
cannot be accurately predicted by the application of hydrodynamic
theories such as Stokes Law, which holds for spherical particles.
Hence, it was felt that, in order to get settling velocities character-
istic of the Ashland sediment, a quiescent settling test should be
conducted. The test apparatus basically consisted of a cylindrical
container which had sampling taps at various heights (see Figure C-2).
118
-------
H"
TO
C
H
a
n
4 U
C CD
0 n
o 4
8
3
(D
(3 O
TO C
en p>
s1
C
OH
OH
1. 00
0. 80
0. 60
0. 40
0. ZO
x Tank I
O Tank J
(D
-------
28'
)OOOOOOOOOOOOQ[
A
~ T
4"
1
B
C
D
T
4"
1
, i
4"
- -
~ T
4"
_ 1
E ~
2"
Figure C-2. Settling Chamber
120
-------
The transparent cylinder -was 29 in. high and 5-1/2 in. in diameter
(I. D. ). Five sampling positions were used, spaced 4 in. apart, with
the first being 2 in. from the bottom. The mean water height was
28 in. , with variations from 28-1/2 in. to 27-1/2 in. during the
course of the test.
At various time intervals after the start of the test, samples were
drawn from different heights, and a measure of relative amounts of
suspended solids was determined by turbidity tests. The settling
velocities were then determined by dividing the depth of water above
the sampling point by the period of settling. Extreme care was taken
to ensure uniform temperature in the water and uniform ambient
temperature (_+ 1/2°C). Subsequent tests were made to determine
mercury level associated with turbidity reading. Turbidity values
were measured in JTU and mercury concentrations in ppm.
From the test data elapsed times required for turbidity levels to drop
to predetermined levels were determined. The elapsed times required
to reach turbidity levels of 400, 300, 200, and 100 JTU at various
heights in the settling chamber are given in Table C-3.
If the assumption is made that the same collection of particles is
responsible for a given turbidity level as settling continues, the
settling velocities can be estimated by dividing the traversed height
by elapsed time. For example, for the 400-JTU particles, 67 minutes
(73-6 = 67 minutes) was required to traverse 12 inches (18-6 = 12 inches)
which results in an estimated characteristic settling velocity of
, -, . 1 ft.
12 in. x
v 12 in' -3
V400 = TH = 0.248x10 ^ft/sec.
/_ . 60 sec.
67 mm. x :
mm.
Corresponding velocities associated with other particle groups are
listed in Table C-4.
Figure C-3 illustrates the settling velocities for each group. From
Table C-4 it can be seen that larger velocities are associated with the
higher turbidity groups, with the exception of the 100, group. The 400,
300, and 200 groups appear to be made up of particles of progress-
ively smaller dimensions (the smaller the particle, the higher the
drag force and, consequently, the lower the settling velocity). For
the 100 group, one would normally extrapolate that the particles would
be smaller and that the velocity would be lower. However, as is clear
from Table C-4, the 100 group's behavior could be the result of a
system of fine particles coalescing to form larger particles after a
period of time, resulting in turbidity levels of 100 JTU in the upper
layer and then proceeding to settle at a velocity characteristic of the
larger size. Such an occurrence would account for the long delay
121
-------
Table C-3
Elapsed Time in Minutes to Reach Various Turbidity Levels
Height
(in.)
18
14
10
6
2
400 JTU
6
8
18
73
360
300 ITU
21
24
36
107
200 JTU
69
81
93
172
100 JTU
270
255
280
350
Table C-4
Settling Velocity as a Function of Height, Turbidity, and Elapsed Time
Particle Group
400
300
200
100
Elapsed Time
(min. )
67
86
103
80
Transversed
Height (in. )
12
12
12
12
V
(ft/sec)
0.248 x 10'3
0. 196 x 10-3
0. 162 x 10'3
0. 238 x 10-3
122
-------
ts)
OP
c
H
n
O
*
0 0
C 0
en
fD
CO
o
V
hi
f)
rt-
(B
H
M-
0)
rt-
o'
O
P>
H
i-1*
O
C
CO
u
0)
(n
o
O
QC
C
0. 30
0. 20
0. 10
(0.211
Av.
100
200
Particle Group
300
400
-------
O
0.26-
O
IS)
cn
OQ
o
o
fO
o
o
O
H
O
u
o
be
G
CO
0. 10
v = 6 x 10"6 (JTU)
100
200 300
Particle Group
400
-------
time of 270 minutes at the 18-in. level (see Tables C-3 and C-4),
followed by an 80-minute transit time (350-270 = 80) to reach the
6-in. level. However, to estimate transport distances of suspended
particles, it is necessary to know the amount of time in suspension,
which would include formation time. Hence, a more representative
velocity can be obtained by using the total time from the start of the
test to determine the settling velocities. Settling velocities deter-
mined on this basis are presented in Table C-5.
Figure C-4 contains the results of Table C-5, indicating higher settling
velocities associated with higher-number turbidity groups.
Figure C-5 contains a plot of mercury concentration versus turbidity
readings. Examination of this group shows that the majority of the
adsorbed mercury is associated with turbidity levels of over 100, thus
useful settling velocities will be those in the vicinity of 100 to 400 JTU.
The horizontal distance traveled by a particle group is given by
d = v t
c
V
s
/ hv
d = 1-66xl°i_%r
where d = transport distance
v = stream current velocity (ft/sec)
h = height (ft)
v = settling velocity (ft/sec)
S
JTU = turbidity level
For a given current and height of disturbance, the redistribution of
particles within a particular turbidity grouping can be predicted. In
order to determine the approximate quantity of mercury deposited at
a distance d from the dredging site, the time of deposition and the
mercury distribution as a function of turbidity are required. For the
Ashland test site, Figure C-5 shows mercury concentration versus
turbidity.
125
-------
ro
00
c
>-j
0)
o
s
(D
o
>-t
O
o
rt-
0>
H
c.
a*
t->'
a
300
p
H
>N
4J
• rH
^H
In
200
100
a>
01
0.010 0.020
Mercury Concentration (ppm)
0. 030
-------
Table C-5
Settling Velocities with Revised Elapsed Time
Group
400
300
200
100
Test Time
(min. )
73
107
172
350
V
ft/sec
0.238
0.163
0.097
0.048
127
-------
Treatment of Dredge Spoil
In the course of working with the Acton sediments (Appendix A —
Table A-l) it was found that the mercury-binding capacity decreased
on aging and exposure to air. This leads to the possibility that
contaminated dredge spoil may release mercury if it is placed on a
landfill and exposed to air and oxygen-rich surface waters.
We made several experiments in which samples of Ashland sediment
were exposed to air and alternately moistened and dried for a period
of two weeks. These results were inconclusive in that only a slight
and variable decrease in partition coefficient was observed after the
above aging treatment.
We now beliver that, because of the high organic content of the
Ashland sediments, two weeks was insufficient to produce appreciable
oxidation and that longer term exposures are needed. We recommend
that such long-term experiments be conducted as part of Phase III of
this program.
128
-------
APPENDIX D
PHYSIOLOGICAL EFFECTS OF ORGANIC THIOLS
If the long-chain alkyl thiols (mercaptans) are to be used for complex-
ing mercury on a large scale, it is necessary to make an estimate of
their possible impact on the aquatic environment. It is well known
that the thiol (-SH) group is an essential constituent of animal protein
and that many organic thiols are physiologically active. As with other
physiologically active materials, an excess may be harmful.
In this appendix we review first some of the beneficial properties and
medicinal uses of thiols and related materials. Second, we review
the recent literature on the toxicity of the simple alkyl thiols and
estimate their probable effect on fish life.
Beneficial Properties and Medicinal Uses of Thiols and Related
Substances
It has long been recognized that thiols are beneficial to cell repro-
duction. An early patent by Sutton [9] discusses the use of alpha thio-
glycerol as a cell stimulant to decrease the healing time of wounds.
Much of the recent literature deals with the use of thiols as antidotes
for poisoning by heavy metals. In this review we will confine the
discussion to the effects of thiols on mercury poisoning.
It is known that certain sulfides will interfere with the disinfecting
action of mercuric chloride. As early as 1908, Chick [10] found that
a culture of B. paratyphosus, which had apparently been killed •with
solutions of HgCl2, could be revived within a limited period of time
by exposure to solutions of hydrogen or ammonium sulfide. These
sulfides are believed to act by removing the mercuric ion from its
combination with the organism, through precipitation of the very
insoluble mercuric sulfide.
The principle of removing mercury from combination with the organ-
ism has been widely applied in the development of antidotes for
mercury poisoning in humans. Bidstrup [ll] discusses the use of the
dithiol BAL (2, 3-dimercaptopropanol) in the treatment of acute
poisoning by mercuric chloride. The use of BAL has greatly improved
the prospects for recovery in such cases.
In cases of chronic mercury poisoning, BAL is less useful, but the
thiol derivative N-acetyl penicillamine has been found effective in
relieving the symptoms and in increasing the elimination of mercury
[12, 13]. More recently, Takahashi and Hirayama [14] have suggested
the use of indigestible and unabsorbable thiol compounds to accelerate
the elimination of methylmercury from animals. Reduced human hair
129
-------
powder contains thiol groups originating from its cysteine content
and was found effective in aiding the elimination of methylmercury
from mice. Synthetic resins containing thiol groups were also
suggested.
Another recent report by Ganther et al. [15] indicates that the presence
of selenium in diets fed to Japanese quail is effective in decreasing the
toxicity of methylmercury compounds. It is known [16] that selenium
forms an even less soluble compound with mercury than does sulfur,
and the authors suggest that the protective action of the selenium
derivatives is probably due to its mercury-binding capacity. Thus,
even an element which by itself may be highly toxic has a beneficial
effect in the prevention of mercury toxicity.
Toxicity of Alkyl Thiols
The short-chain alkyl thiols are known to be toxic to fish and other
biota. Methyl mercaptan, found in kraft paper mill effluents, is com-
parable to hydrogen sulfide in its toxic effects. Van Horn et al. [17]
report that the safe concentration (no mortality in 120 hours) of
methyl mercaptan for minnows is 0. 5 ppm.
As the length of the hydrocarbon chain is increased, the solubility,
volatility, and sulfur content of the mercaptans decrease. Shugaev
[18] concludes that the toxicity of the long-chain mercaptans is more
or less equal to that of the hydrocarbons of the same chain length. We
have found no data on the toxicity to fish of mercaptans having more
than four carbons in the alkyl group. Turnbull et al. [19] have per-
formed experiments on the toxicity of butyl mercaptan (in the form of
the sodium salt) to bluegill sunfish. Their results (calculated as free
mercaptan) show that 50% of the fish will survive for 24 hours at a
concentration of 20. 2 ppm and 50% will survive for 48 hours at a con-
centration of 15 ppm. From these figures they estimate a safe concen-
tration of 2. 5 ppm or about 5 times that for methyl mercaptan.
As the length of the alkyl chain is further increased, the toxic effects
of the mercaptans become limited by their rapidly decreasing solubility
in water. Reid [20] points out that the solubility of the longer-chain
normal mercaptans is about the same as that of the normal alkane
containing one more carbon atom. The solubility in a given hydro-
carbon series decreases rapidly with increasing molar volume [21],
These facts are illustrated by the data shown in Table D-l. Although
no experimental solubilities are available for the higher mercaptans,
their solubility may be taken as equal to that of the corresponding
hydrocarbon, as shown in the adjacent column of Table D-l. The
solubility of n-dodecyl mercaptan thus estimated is 0. 013 ppm, •which
is far below the safe limit of 2. 5 ppm discussed above for butyl
mercaptan.
130
-------
Table D-l
Some Solubilities of Normal Mercaptans
and Normal Alkanes at 20-30°C
n-Alkyl
Radical R
C6H11
C?H13
C8H15
C11H21
C12H23
C17H33
Solubility in mg/1 (ppm)
Mercaptan (R-SH)
43
14
Alkane (R-CH3)
52
15
6
0.2
0.013
0.007
131
-------
The solubility considerations thus indicate that no toxic effect on fish
is to be expected from treating the bottom sediments with long-chain
thiols. This conclusion is supported by our own aquarium experiments,
in which no toxic effect was apparent with goldfish exposed for 30 days
to water in contact with sediments containing either n-dodecyl or
t-dodecyl mercaptan.
132
-------
APPENDIX E
ANALYTICAL PROCEDURES AND METHOD DEVELOPMENT
Most of the analytical work performed on this project was concerned
with the determination of total mercury in samples containing inorganic
mercury or organomercury compounds or both. We also undertook,
in cooperation with the Jarrell-Ash Division of Fisher Scientific
Company, to establish a procedure by which methylmercury and other
organomercury compounds could be separated from their naturally
occurring mixtures and unambiguously identified. The method chosen
was to separate the individual mercury species by gas chromatography
and collect them in a specially designed microcell. The collected
samples were identified by laser Raman spectroscopy. A successful
microcell has been demonstrated, and Raman spectra have been
obtained with sub-microgram quantities of a number or organomercury
derivatives. Considerable difficulty has been experienced with the gas
chromatographic separations, however, and further work on separa-
tion techniques will be required to obtain an operational system.
We are indebted to Mr. James Longbottom, of the Environmental
Protection Agency, Cincinnati, Ohio, for analyzing sediments from
the Framingham Reservoir for methylmercury.
Inorganic and Total Mercury Analyses
These analyses were made by the procedure of Hatch and Ott [22]
using atomic absorption spectrophotometry. In preparing samples
for this analysis it is important that the final solutions be freed from
organic matter and that all the mercury be in inorganic form. Samples
of solutions known to contain only inorganic mercury and containing
only traces of organic matter were analyzed by the following procedure:
1. The sample was acidified with a solution of 2. 5% HNO,
and 2. 5%
2. A few drops of 5% KMnO4 solution were added and the
mixture allowed to stand Tor a few minutes. The color
of permanganate should persist on standing.
3. The excess permanganate was destroyed with
hydroxylamine hydrochloride solution.
4. The mercury was reduced to the metallic state with
stannous chloride.
5. A stream of air was passed through the solution to
vaporize the mercury and carry it into the spectro-
photometer.
133
-------
All samples containing organic matter or organically bound mercury-
were digested prior to analysis by boiling under reflux with a mixture
of H2SC>4' HNC>3. Such samples include all sediments, fish, and all
samples known to contain methylmercury or other organically bound
mercury. The procedure is as follows:
1. The sample (1 to 5 g of solid or 10 ml of solution) was
weighed or pippetted into a 250-ml boiling flask.
2. A few boiling chips, 10 ml cone. HNOs, and 10 ml 50%
were added.
3. The flask was fitted with a water-cooled reflux condenser
and boiled under reflux for two hours.
4. The flasks were cooled for 15 minutes, and 5 ml of a
mixture of two parts cone. HNC>3 and one part 50%
^2 C>2 was added.
5. The samples were again refluxed for 45 minutes.
6. The samples were cooled and the condensers washed
down with 25 ml deionized water.
7. Twenty ml of 5% KMnC>4 were added and the mixture
allowed to stand 1 /2 hour. (If the permanganate color
does not persist, small amounts of KMnC>4 crystals are
added until it does. Some samples low in organic matter
require only a few drops of KMnC>4 solution at this point.
The minimum amount required to give a permanent color
should be used, since KMnO^. contributes appreciably to
the blank reading.
8. The excess KMnO^ is destroyed with hydroxylamine
hydrochloride, and the solution is diluted to 100 ml.
9. The digested solution is analyzed as above.
Most of the samples were analyzed with a Coleman Model MAS- 50
atomic absorption spectrophotometer. Some samples containing
less than 1 ppb of mercury were sent to Jarrell-Ash Division for
analysis with a specially equipped absorption spectrophotometer,
using a hydrogen flame. The detection limit of this device is in the
range of 0. 02 to 0. 04 ppb, and the calibration curve is stated to be
linear down to this limit. Most of the samples analyzed by both
instruments gave somewaht lower values with the Jarrell-Ash
equipment. This may indicate some low level of interference in
the flameless Coleman instrument.
134
-------
Gas Chromatographic Separations
T"he gas chromatograph which was used for this study was a Fisher/
Victoreen Catalog No. 11-104-11, which embodied a M&3 electron-
capture detector and was equipped with a selection of columns, the
stationary phases of which were Chromosorb W and Carbowax, as
recommended in the literature for the separation of mercury
compounds. However, columns with Chromosorb W alone proved
satisfactory for the separation. The gas chromatograph was modified
to allow for collection of fractions by splitting the effluent stream from
the column in a ratio of 1:1 to the detector and to an output collector
port. The effluent collector port was maintained at the column tem-
perature, and, although the chromatograph was equipped with variable
temperature programming, capabilities, no use was made of this.
Although gas chromatography is a technique not difficult to practice,
in the course of this study it was found difficult to obtain a column
with the required properties to give sharp effluent peaks. This was
possibly due to the circumstance that the columns were of stainless
steel, and this might have had a deleterious effect through Hg
adsorption. Within the time scale of the experiment, it was not
possible to investigate the potentials of glass columns. The experi-
mental evidence, however, disclosed that the detector was not respon-
sible for the difficulties encountered, and improvements to the column
undoubtedly are the primary requirements.
An attempt was made to develop a method by which both inorganic and
methylmercury could be simultaneously determined on the same gas
chromatogram. We therefore tried to isolate the two forms of
mercury as mercaptides and to separate and estimate both forms by
gas chromatograph. In theory, the methylmercury should form a
mercaptide of the type CHsHgSR, while inorganic mercury should
form Hg(SR2), where R^ is a hydrocarbon radical. If R_ is sufficiently
large, there should be an appreciable difference in molecular weight
between the two species. We chose dodecyl mercaptan (C^H^SH)
because of its molecular weight and because its low volatility renders
it relatively inoffensive to work with.
Preliminary extraction experiments were made in which 300 ml of
aqueous solutions of HgCl£ or CH^HgCl, containing 100 ppm of Hg,
were extracted with 1 ml (849 mg) of n-dodecyl mercaptan dissolved
in 100 ml of petroleum ether. After four days of agitation, the
aqueous solutions were separated and both were found to contain less
than 1 ppb of Hg (Hg++ = 0. 234 ppb and CH^Hg+ = 0. 832 ppb), indicating
substantially complete extraction of the mercury. The petroleum ether
extracts were evaporated down until only the mercury-containing
mercaptan remained.
These concentrates were then gas chromatographed on a 5% SE-30
column at 195°C. The initial chromatogram showed a characteristic
peak at six minutes, which appeared with CH^Hg"^ and in a mixture of
135
-------
^ with Hg++, but not in Hg++ alone. Other peaks were present,
but their interpretation was less clear.
It was then found that the heater on the electron-capture detector of the
chromatograph was not functioning reliably and that the detector was
not operating at the desired temperature of about 220°C. When the
chromatograms were repeated after the heater was repaired, however,
the six-minute peak had disappeared and only minor differences were
found between the two forms of mercury.
We then prepared pure samples of Hg^C^H^)? and of
by reaction of the mercaptan with HgO or witn CrlsHgOH, respectively.
These materials were examined by laser Raman spectroscopy and
found to produce spectra which correlated well with the various inter-
atomic bonds assumed to be present (see Table E- 1 below).
When these compounds were heated for 5 minutes at 200°C, however,
extensive decomposition was found to take place. The Raman spectra
were altered, with disappearance of peaks attributed to the C-Hg
bond and to the Hg-S bond. Visual examination of the heated samples
revealed the presence of droplets of free mercury, together with dark
material which may have been HgS. A brief examination of the
literature indicated that the mercury mercaptides are known to be
readily decomposable by heat by at least two mechanisms:
Hg(SR)? — *- Hg + R-S-S-R
Hg(SR)2 -^ HgS + R-S-R
By analogy with the decomposition of methylmercuric sulfide, we infer
that the methylmercury mercaptide may decompose according to some
reaction, such as:
2CH3HgSR -*- Hg(CH3)2 + Hg(SR)2
These products may further decompose according to the reactions:
Hg(CH3)2 — Hg + C2H6
Hg(SR)2 — - Hg + R-S-S-R
Hg(SR)2 — HgS + R-S-R
Thus, except for the possible production of ethane by the methyl-
mercury derivative, the decomposition products of the two types of
mercaptide are much the same. This would account for the general
similarity and for the multiple peaks obtained on the gas chromato-
grams. The extent of decomposition of Hg(CH3)2 at 200°C is not yet
clear, but it is evident that any analytical scheme involving separation
of the mercaptides at high temperatures will be subject to difficulties.
136
-------
Table E-l
Characteristic Frequencies (cm ) of Some Mercury Compounds
Compound
(CH3)2Hg
(C2H5)2Hg
(C6H5)2Hg
(C6H5CH2)2Hg
CH3HgCl
C6H5HgCl
(C12H25S)2Hg
CH3HgSC12H25
C12H25SH
HgCl2
HgS (red)
S-H
2580
C-S
725
730
660
C-Hg
515
488
(660)
(445)
560
662
535
S-Hg
425,
330
425
300
Hg-Cl
295
315
280
137
-------
We made a further attempt to obtain thermally stable derivatives of
Hg and CH^Hg"*" by reacting them •with ortho mercapto aniline. It
was hoped that the formation of chelate rings by the nitrogen atom of
this ligand would confer sufficient thermal stability on the complex to
permit it to be chromatographed as such. Both the Hg++ and the
CH3Hg+ complex, however, were found to liberate mercury when
heated for five minutes at 200°C.
It may be possible to separate the mercaptides in solution at room
temperature by column or thin-layer chromatography. We recommend
that these techniques be investigated. For the present program,
however, we elected to try to separate methylmercury as CH^HgCl by
gas chromatography according to known methods.
The level of detection of methylmercurie chloride was found to be
much higher (in excess of 1500 nanograms in benzene solution) than
expected. Mud samples of known methylmercuric chloride concentra-
tion were analyzed on the gas chromatograph, but no methylmercuric
chloride could be detected. The concentrations •were as high as
1500 ppm Hg as CH-^HgCl (dry basis). A Raman analysis on these
samples was not undertaken due to the strong Raman lines of the
solvent benzene. The water solutions are known to be too weak in
concentration for Raman analysis.
Solutions to the problems of the detector and the low sensitivity
observed in an instrument capable of much better performance were
being sought when the mercury program was terminated. It is highly
probable that the solutions are quite simple ones involving an increased
detector temperature and switching from a stainles s-steel column to a
glass column.
Design of the Microcell
Several cells were designed to permit collection of sub-microgram
quantities of sample from the gas chromatograph. The filled cell
could then be mounted in the sample compartment of the Raman system
and aligned with the laser beam for excitation. Each successive design
permitted the use of smaller quantities of collected material. The
final design, shown in Figure E-l, permitted handling of samples with
a volume of less than 200 nanoliters. The cell was joined to the effluent
part of the gas chromatograph via a hollow septum, and the U-shaped
portion was immersed in a reservoir maintained at -50°C. When the
elution from the gas chromatograph was completed, the cell was
transferred to another cold reservoir maintained at -190°C to freeze
the sample onto the walls of the cell. Under this condition, the air
and other gasses in the cell were evacuated with assurance of minimal
sample losses. Subsequent to evacuation, the liquid nitrogen was
transferred from the U-tube section to the microcapillary portion of
the cell, and by the process of sublimation the sample was released
from the walls of the U-tube and condensed and trapped in the micro-
capillary cell. Raman spectra were obtained from less than
138
-------
Figure E-l. Microcell for Sampling
Output of Gas Chromatograph. (Capillary
into which sample is distilled for Raman
spectroscopy is at lower right. )
200 nanoliters of dimethylmercury and diethylmercury from an
injection of 250 nanoliters of each compound into the gas chromato-
graph, following the above procedure.
Raman Spectroscopy of Mercury Compounds
Organomercury compounds are good candidates for analysis by Raman
spectroscopy because the carbon-mercury bond produces a strong
polarizability change in the molecule. This in turn leads to strong
Raman scattering. Nevertheless, Raman spectroscopy is not a
technique that lends itself to trace analysis, and some means of
preliminary concentration and isolation of the compounds of interest
is needed.
The instrument used for the present work was a Jarrell-Ash Model
25-500 Raman Spectrometer equipped with a CRL Model 52 Organ-
Krypton ion laser and an f/0.95 collection lens assembly.
139
-------
Raman spectra were obtained on pure samples of available organo-
metallic compounds, using volumes of 10 microliters of the pure
material. Each sample was contained in a melting-point capillary-
tube positioned in a focused laser beam directed onto the sample at
90° with respect to the Raman radiation imaged on the entrance slit
of a 0. 5-m focal length double monochromator. Typical data obtained
are shown in Table E-l. For dimethylmercury, the band at 515 Acm~^
is produced by the symmetrical stretch of C-Hg-C; a shift in frequency
occurs for this mode of vibration in other compounds. Thus it is
present at 488 Acm- 1 in the diethyl form and at 445 Acm~l in the
dibenzyl compound.
Similarly, the characteristic frequencies due to the S-Hg and Hg-Cl
bonds can be identified by comparison with HgS and HgC^, respectively.
The Raman technique should be readily adaptable to all classes of
compounds. Dependent only on the ability to isolate the components,
it is capable of providing a positive identification of sub-microgram
quantities of material. It thus provides an excellent complement to
the gas chromatograph, which can separate components but provides
no positive identification. We recommend that efforts to interface the
Raman spectrometer with the gas chromatograph be continued.
140
-------
APPENDIX F
FIELD SURVEY SAMPLE COLLECTION
As discussed in Section VI, a field survey was performed in the
Ashland, Massachusetts area. Grab sediment samples, core sediment
samples, and water samples were collected in Framingham Reservoir
No. 2 and in the brook and Sudbury River between Nyanza Chemical
Corporation and the reservoir. The results of the analyses on these
samples are presented in Tables F-l, F-2, and F-3.
Fish samples were collected by the Massachusetts Division of Fish
and Game'at the request of the Massachusetts Division of Water
Pollution Control. The analyses were performed by the Lawrence
Laboratory of the Massachusetts Department of Public Health. Results
of the fish analyses are presented in Table F-4. Analyses on the
largemouth bass samples were also performed by the Westboro
Laboratory of the Division of Fish and Game, with results showing
somewhat higher concentrations of mercury. Since most of the fish
had been analyzed by the Lawrence Laboratory, their results are
reported herein.
Analyses of other water-quality parameters in the Framingham water-
shed area were obtained from the Boston Metropolitan District
Commission, as were the flow volume data. These results are
reported in Tables F-5 and F-6, respectively.
Water samples were preserved in the field by the addition of 3 ml of
concentrated HNO3 per 100 ml of sample. In some cases water
samples were filtered in the field through a 0.45 micron Millipore .
filter prior to acidification. These samples were used for determina-
tion of the dissolved-mercury fraction.
Grab samples were obtained using a small scoop at the end of an
extendable pole. Samples were refrigerated in the laboratory until
analysis. The core samples were obtained by forcing a 2 ft x 1. 5 in.
plastic coring tube into the reservoir or river bottom. Cores could
be taken in water depths up to 15 feet. In the upper seven-acre section,
the maximum water depth is eight feet. In some areas of the lower
section, water depth reaches a maximum of about 30 feet. Core
samples were frozen until analysis. The sample was retained in the
plastic tube, which was then cut into two-inch sections. The thawed
samples were analyzed for total mercury and per centage of moisture.
Results of the core analyses in the upper two-inch sections agreed
closely with the nearby grab-sample analyses.
Water samples were analyzed in two different ways. Near the end of
the program, we discovered that a large fraction of the dissolved
mercury in the reservoir was in the form of a s^'uble organic
141
-------
TABLE F-l
Grab Sample Analyses--Ashland, Massachusetts Area
Sample Number
E-l (Sediment)
E-2 (Sediment)
E-3 (Sediment)
E-4 (Sediment)
E-5 (Sediment)
E-6 (Sediment)
E-7 (Sediment)
E-8 (Sediment)
E-l 2 (Sediment)
E-l 3 (Sediment)
E-14 (Sediment)
E-l 5 (Sediment)
E-16 (Sediment)
E-l 7 (Sediment)
E-18 (Sediment)
Hg(mg/kg,
wet weight)
2.2
39. 0
19.9
3.38
12.3
315
4.6
54.8
36. 3
15.7
40. 8
13.9
49. 1
31. 7
74. 0
% Moisture
21
53
66
11
38
70
50
14
59.5
27
64.6
42. 0
52. 1
35
55
Hg(mg/kg,
dry weight)
2.79
83.0
58. 5
3.8
20. 0
1050
9.25
64.0
89.5
21.5
115.0
24.0
100.2
Location
Chestnut Street,
Sudbury River
Union St. Bridge,
Sudbury River
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Brook near
Nyanza Plant
Reservoir No. 2,
Upper Section
Cherry Street,
Brook
Reservoir No. 2,
Lower Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
48.8 Reservoir No. 2,
[Upper Section
164.0 |Reservoir No. 2,
Upper Section
142
-------
TABLE F-l (Continued)
Sample Number
E-19 (Sediment)
E-20 (Sediment)
E-21 (Sediment)
E-22 (Sediment)
E-23 (Sediment)
E-24 (Sediment)
E-25 (Sediment)
E-26 (Sediment)
E-27 (Sediment)
E-28 (Water lily)
E-29 (Sediment)
E-31 (Sediment)
E-33 (Sediment)
E-35 (Sediment)
E-37 (Sediment)
E-38 (Sediment)
E-39 (Sediment)
Hg(mg/kg,
wet weight)
18.0
12. 5
11.3
18.2
10.6
30.6
20.0
14.5
5.96
0.556
18.3
7.28
14.4
3.30
3.23
2. 18
3.42
% Moisture
35
33.7
12.7
51. 8
43
62
67. 2
62.5
36.2
85.3
36.6
38.6
28.3
30.0
38.0
20.0
33.2
Hg(mg/kg,
dry weight)
27.7
18.9
12.9
37.7
18.6
80.5
61. 0
38.7
9.35
3.78
28.9
11.9
20.5
4.7
5.2
2.72
5. 1
Location
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Lower Section
Reservoir No. 2,
Lower Section
Reservoir No. 2,
Lower Section
Reservoir No. 2,
Lower Section
Reservoir No. 2,
Lower Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 1
Reservoir No. 1
Reservoir No. 1
Reservoir No. 1
143
-------
TABLE F-l (Continued)
Sample Number
E-41 (Sediment)
E-43 (Sediment)
E-44 (Sediment)
E-45 (Moss)
E-46 (Moss)
E-47
E-48
E-49
E-50
Hg(mg/kg,
wet weight)
27.6
41.4
39. 0
1. 268
0. 246
582. 0
43.7
64.4
3. 37
% Moisture
48. 5
36.4
34. 7
40.0
40.0
83. 5
89. 3
44.4
24. 2
Hg(mg/kg,
dry weight!
53. 6
65. 1
59. 8
2. 1
0.41
3504. 0
408. 0
116. 0
4.44
Location
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Lower Section
Reservoir No. 2,
Lower Section
Cherry Street
near Nyanza
Concord St. near
unpolluted brook
Nyanza swamp
area
Brook before rr
culvert (1. brnch)
Brook before rr
culvert (r. brnch)
Gravel material
near railroad
culvert
144
-------
TABLE F-2
Core-Sample Analyses--Ashland, Massachusetts Area
Sample Number
C-l (0-2 inches)
C-l (2-4 inches)
C-l (4-6 inches)
C-2 (0-2 inches)
C-2 (2-4 inches)
C-3 (0-2 inches)
C-3 (2-4 inches)
C-3 (4-6 inches)
C-4 (0-2 inches)
C-4 (2-4 inches)
C-4 (4-6 inches)
C-5 (0-2 inches)
C-5 (2-4 inches
C-5 (4-6 inches)
C-6 (0-2 inches)
C-6 (2-4 inches)
C-6 (4-6 inches)
C-6 (6-8 inches)
C-7 (0-2 inches)
C-l (2-4 inches)
C-8 (0-2 inches)
C-8 (2-4 inches)
C-8 (4-6 inches)
C-8 (6-8 inches)
C-8 (8-10inches)
C-9 (0-2 inches)
C-9 (2-4 inches)
C-9 (4-6 inches)
C-9 (6-8 inches)
C-9 (8-10inches)
C-10 (0-2 inches)
C-10 (2-4 inches)
C-10 (4-6 inches)
Hg (mg/kg,
wet weight)
30.4
17. 7
8.6
10.6
0.81
20.0
17.4
11.6
9.55
7.21
10. 3
14.9
10. 2
8.95
21.9
22.8
9. 72
7.7
1. 35
0.49
14. 1
6. 35
17.7
5.4
4.36
15.8
7.75
7.42
9.50
9.85
16.3
14. 0
3.5
% Moisture
67. 0
53.0
33.3
53.0
54. 0
75.0
53.6
54. 1
83. 0
42.2
34. 8
71. 0
45. 5
55.0
61. 0
54. 5
37.2.,
47. 7
Hg (mg/kg,
dry weight)
92
37.6
12.8
22.6
1.76
80. 0
37.6
25. 3
56. 1
12. 5
15.8
51.4
18.7
19.8
56. 1
50.0
15. 5
14.7
13.5 1.56
21.0 0.62
42. 0
23.6
36.6
31.9
27.8
47. 0
52. 5
37.5
56. 5
57. 5
53.7
45. 5
32.4
24. 3
8. 35
28. 1
1 7.9
6.04
29.8
16.3
11.9
21.8
23. 1
35. 2
25.7
5. 18
Location
Reservoir No. 2,
Upper Section
Sudbur y River
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
145
-------
TABLE F-2 (Continued)
Sample Number
C-ll (0-2 inches)
C-ll (2-4 inches)
C-ll (4-6 inches)
C-12 (0-2 inches)
C-12 (2-4 inches)
C-12 (4-6 inches)
C-12 (6-8 inches)
C-13 (0-2 inches)
C-13 (2-4 inches)
C- 13 (4-6 inches)
C-13 (6-8 inches)
C-14 (0-2 inches)
C-14 (2-4 inches)
C-14 (4-6 inches)
C-15 (0-2 inches)
C-17 (0-2 inches)
C-17 (2-4 inches)
C- 17 (4-6 inches)
C-17 (6-8 inches)
C-19 (0-2 inches)
C- 19 (2-4 inches)
C-19 (4-6 inches)
C-19 (6-8 inches)
C- 19 (8-10 inches)
C-21 (0-2 inches)
C-21 (2-4 inches)
C-21 (4-6 inches)
C-21 (6-8 inches)
C-21 (8-10 inches)
C-21 (10-12 inches)
C-21 (12-14inches)
C-21 (14-16 inches)
Hg (mg/kg,
wet weight)
17. 3
17. 3
6.65
2.42
3.2
0. 278
1. 51
16. 5
19.7
13. 0
4.48
24. 6
7. 5
5.4
13.2
8.45
1. 34
4. 35
4.67
5.4
9.65
14. 60
7. 01
9.81
27. 2
14. 1
20.2
8.4
10.9
11. 35
9. 5
18.7
6.95
8. 39
2.47
1.63
3.89
% Moisture
72. 5
50.0
36.8
38.9
35.6
32. 0
29.0
76.0
53.0
46.4
36.2
64. 5
32.9
34. 0
53.6
37.6
24.6
40.7
32.8
46. 3
86.2
48.8
46. 1
69. 0
49.6
42.4
48.6
40. 3
45.3
89. 1
52. 1
52.7
43.2
61. 0
37.4
64.0
69.0
Hg (mg/kg,
dry weight)
63.0
34.6
10.5
3.96
4.95
0.41
2. 13
68.9
42.8
24.3
7.01
69.4
11.2
8. 17
28.5
13.5
1.78
7. 5
6.95
10. 10
68.9
28.6
13.0
31.6
54.5
24.5
39.3
14.0
19.9
100.4
19.9
39.5
12.4
21. 5
3.94
4.52
12. 5
Location
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
Reservoir No. 2,
Upper Section
146
-------
TABLE F-2 (continued)
Sample Number
C-22 (0-2 inches)
C-22 (2-4 inches)
C-22 (4-6 inches)
C-22 (6-8 inches)
C-22 (8-10 inches)
C-22 (10-12 inches)
C-22 (12-14 inches)
C-22 (14-16 inches)
C-22 (16-18 inches)
C-22 (18-20 inches)
Hg (mg/kg,
wet weight)
9.25
22. 5
15.6
23.4
24.8
10. 8
3.42
4.78
10.2
0.29
% Moisture
89. 5
47.7
48. 3
49.4
41.8
30.9
33.9
67.4
55. 0
24. 5
Hg (mg/kg,
dry weight)
88.0
43. 0
30. 2
46. 1
42.6
15.6
5. 17
14.6
22.6
0. 38
Location
Reservoir No. 2,
Upper Section
147
-------
TABLE F-3
Water-Sample Analyses--Ashland, Massachusetts Area
oo
Sample
Number
W-6
W-7
W-8
1 W-9
1 w-io
I W-ll
1 W-12
1 W - 1 3
1 W-15
1 W-20
1 W - 1 6
|\V-17
IW - 2 1
jW-18
]
Date
8/27/71
8/27/71
8/27/71
10/10/71
10/10/71
10/10/71
10/10/71
10/10/71
10/27/71
10/27/71
10/27/71
10/27/71
10/27/71
10/27/71
DissolvedHg (0.45 filter), /ug/l (ppb)
Reflux
27. 0
1.0
1. 25
5.6
3.8
3. 2
Nonreflux
1.9
1.8
0. 65
0.9
0.5
<0.2
<0. 2
-------
Table F-3 (continued)
Sample
Number
W-25
W-26
W-31
W-32
W-34
W-51
W-42
W-43
W-44
W-45
W-46
W-47
W-48
W-49
W-50
Date
11/12/71
11/12/71
12/28/71
12/28/71
12/28/71
12/28/71
12/28/71
12/28/71
12/28/71
12/28/71
12/28/71
12/28/71
12/28/71
12/28/71
12/28/71
Dissolved Hg (0. 4?
Reflux
2. 2
3.4
filter), //g/1 (ppb)
Nonreflux
0. 3
0. 2
0.5
< 0. 2
Total Hg
Reflux
t • —
3.4
, fig /I (ppb)
Nonreflux
16.4
16.8
10. 2
3.6
3. 5
< 0.2
0. 2
10.2
5.4
2. 5
1.3
Location
Res. No. 2--upper sectn.
Res. No. 2--upper sectn.
Brook near Nyanza
(combined flow)
Brook near RR culvert
Cherry St. --brook
Brook midway between
Cherry St. and town
Brook, police sta. culvert
Res. No. 2--upper sectn.
Res. No. 2--upper sectn.
Concord St. --brook
Confluence of 2 brooks
Brook confluence with
Sudbury River
Brook confluence with
Sudbury River
Sudbury R. after conflu.
Sudbury R. after conflu.
-------
Table F-3 (continued)
?
Sample
Number
W-68
W-60
W-65
W-66
W - 6 3
W-67
W - 6 8
W-69
W-7Q
W-75
W-76
W-77
W-78
W-81
W-82
W-87
Date
2/10/72
2/10/72
3/31/72
3/31/72
3/31/72
3/31/72
3/31/72
3/31/72
3/31/72
3/31/72
3/31/72
3/31/72
3/31/72
|4/12/72
14/12/72
4/12/72
Dissolved Hg (0. 45
Reflux
15. 0
a. 6
4. 0
1.8
filter) fjg/l (ppb)
Nonreflux
2. 7
1. 4
0. 2
Total Hg
Reflux
2.0
3.8
22. 0
18.0
5.6
8.6
4.0
2.6
21. 0
fig /I (ppb)
Nonreflux
3.9
4.4
3.7
0. 2
3. 0
Location
Res. No. 2~-upper sectn.
Res. No. 2--upper sectn.
Brook near Nyanza
(right branch)
Brook near Nyanza
(right branch)
Brook near Nyanza
(left branch)
Brook near RR culvert
_
Brook near RR culvert
Cherry St. --brook
Cherry St. --brook
Sudbury R. after conflu.
Sudbury R. after conflu.
Sudbury R. before Res. #2
Sudbury R. before Res. #2
Res. No. 2--upper sectn.
Res. No. 2--upper sectn.
Confluence of 2 brooks
-------
Table F-3 (continued)
Sample
Number
W-88
W-89
W-90
W-91
W-92
W-93
W-94
W-95
W-96
W-97
W-98
W-99
W-100
W- 101
W-102
Date
4/12/72
4/12/72
4/12/72
4/12/72
4/12/72
4/12/72
4/12/72
4/12/72
4/12/72
4/12/72
4/12/72
4/21/72
4/21/72
4/21/72
4/21/72
Dissolved Hg (0. 45 filter), //g/l(ppb)
Reflux
5.6
9.0
2.4
Nonreflux
0. 2
1. 1
3. 5
3. 4
0. 2
0.6
Total He
Reflux
16.6
2. 4
fig/I (ppb)
Nonreflux
3.0
5. 1
6.6
6.0
0.27
4.6
Location
Confluence of 2 brooks
Confluence of 2 brooks
Confluence of 2 brooks
Brook near RR culvert
Brook near RR culvert
Brook near Nyanza
(right branch)
Brook near Nyanza
(right branch)
Brook near Nyanza
(left branch)
Brook near Nyanza
(left branch)
Brook near Magunco Rd.
Brook near Magunco Rd.
Sudbury R. after conflu.
Sudbury R. after conflu.
Confluence of 2 brooks
Confluence of 2 brooks
-------
Table F-3 (continued)
ui
to
1
Sample
Number
W-103
W- 104
W-105
W-106
W-107
W-108
W-109
W-110
W-lll
W-112
Date
4/21/72
4/21/72
4/21/72
4/21/72
4/21/72
4/21/72
4/21/72
4/21/72
4/21/72
4/21/72
Dissolved Hg(0. 45 filter), fig /I (ppb)
Reflux
6.8
41. 0
-------
Table F-4
Fish-Sample Analyses--Ashland, Massachusetts Area
Specie
Largemouth Bass
Large mouth Bass
Largemouth Bass
Largemouth Bass
Large mouth Bass
Largemouth Bass
Largemouth Bass
Largemouth Bass
Smallmouth Bass
Smallmouth Bass
Smallmouth Bass
Smallmouth Bass
Bluegill
Bluegill
Bluegill
Bluegill
Bluegill
Yellow Perch
Crappie
Crappie
Weight (Ibs)
3.3
2.9
2.7
1.8
0.9
0.7
0.2
0. 1
0.96
0.61
0. 55
0. 13
0. 36
0.31
0.26
0. 19
0. 18
0. 13
0.36
0.32
Length (in. )
17. 0
17. 0
16. 0
14. 5
12. 0
10.6
7.8
6.9
12.0
10.0
9.0
5.0
6.0
6.0
5.0
4. 5
4.5
4. 5
6.5
6.0
Hg(mg/kg, ppm)
7.6
6. 0
6. 0
6.9 .
4. 2
1.6
0. 64
3. 5
3. 25
2.4
1. 3
1.9
1. 5
1. 5
2.4
1.6
2.4
2.7
1.45
1.70
NOTE: Analyses made by the Massachusetts Department of
Public Health, Lawrence, Mass., Laboratory.
153
-------
Table F-5
Water Quality Parameters- -Framingham Reservoir Watershed
Location
Sudbury
River,
Concord
Street,
Ashland- -
Before
brook frrn
Nyanza
joins rvr.
Cherry St.
Ashland--
Below
Nyanza on
brook
Framing-
ham Res.
§2, Foun-
tain St.
Date
3/23/70
4/21/70
5/18/70
6/15/70
11/30/70
2/1/71
4/26/71
9/27/71
10/26/71
3/23/70
4/21/70
5/18/70
6/15/70
11/30/70
2/1/71
4/26/71
9/27/71
10/26/71
3/23/70
4/21/70
5/18/70
6/15/70
11/30/70
2/1/71
4/26/71
9/27/71
10/26/71
Turbidity
1. 5
4.0
1. 5
3.4
2.8
2.6
1. 4
2.8
2.2
15.0
0.5
0. 3
0.2
1.2
7. 2
4. 5
0.2
3.8
1.0
10. 0
1. 1
2.7
4.7
3.7
1. 0
2.4
2. 7
Color
65
80
90
110
130
55
52
90
75
1800-mhgny
2750
3500-mhgny
5000
2000
250
900
3000
300
75
55
80 (pink)
125
88
55
55
36
52
Chloride
34
72
47
60
60
57
88
78
86
725
2650
2600
6900
850
500
300
1850
620
55
74
70
105
80
63
Alkalinity
7.0
7. 5
9.0
12
10
9.0
7.0
14
9.0
44
96
430
1000
115
450
1000
180
140
8. 0
7. 5
9.0
16
11
6.0
12
Hardness
28
33
26
16
44
38
43
75
72
565
1000
930
1680
200
285
200
750
320
34
38
35
60
55
62
pH
6. 6
6. 8
6. 5
6.8
6.8
6. 2
6.6
6.8
6.8
6. 7
8. 5
11.4
11.9
8.4
11.6
12. 0
9.0
10. 3
6.7
6.7
6.9
7.6
6. 3
6.7
7. 1
°F
34
58
60
60
38
32
48
60
54
44
44
62
70
40
34
56
60
54
36
50
62
64
32
42
69
58
Tryptone, G. E. Agar
20°, 48 hr
120
320
1000
2200
900
800
700
1300
2600
1600
3500
0
2
1800
1
9
220,000
1300
60,000
2600
30,000
6000
20,000
2000
800
7600.
300
35°, 24hr
180
130
400
1100
600
130
160
900
1100
900
3300
30
1
1200
6
160
150, 000
3000
1200
420
18,000
600
2700
160
90
5000
180
Coliform
/"* r\1 f\r^ » A a
\-s OJL Oli 16 B
100ml, MF
1200
25
2
0
0
60
25
0
14
200
1000
0
0
0
0
0
10
0
800
2700
800
50
160
500
900
4
0
-------
TABLE F-6
Average Flow Volumes, Framingham Reservoir No. 2 (1968)
Month
January
February
March
April
May
June
July
August
September
October
November
December
Flow Volume
(million gallons per day)
44.9
62. 5
266.7
83. 5
51.3
77.6
39.9
10.9 (Some water through gates)
7.76 (All water through gates)
4. 2 (No flow in 14 days)
28. 5 (No flow in 10 days)
52.6
155
-------
compound •which we believe may be amercurated anthraquinone
derivative. The particular complex was not being digested by the
cold-digestion procedures normally employed on water samples.
Many of the water samples were then refluxed with nitric acid and
sulfuric acid, resulting in mercury levels up to 10 times higher than
those obtained with the cold procedure. The reflux procedures are
described in Appendix E. Results of both reflux and nonreflux
analyses are reported in Table F-3. The soluble mercury fraction
was determined by analyzing samples which had been passed through
a 0.45 micron filter.
Beginning with sample number W-63 in Table F-3, the odd-number
samples were total-mercury samples and the even-number samples
•were used for dissolved-mercury determination. Successive odd-
even number samples are from the same location.
156
«U.S. GOVERNMENT PRINTING OFF!CE:1972 514-149/73 1-3
-------
SELECTED WATER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
w
rtD,
Control of Mercury Contamination in Freshwater
Sediments
Yeaple, D. S. ; Johanson, E. E. ; Feick, G.
JBF Scientific Corporation
16080 GWU
68-01-0060
Environmental Protection Agency report
number EPA-R2-T2-OT7, October 1972.
Methods for controlling the release of mercury from sediments have been
developed, and the effects of dredging on the redistribution of mercury have
been evaluated. A program of laboratory studies was conducted concurrently
with a field survey where the extent of mercury contamination at a typical site
was evaluated.
Laboratory studies consisted of both partitioning and aquarium experiments
using artificially contaminated sediments as well as sediments from the polluted
field site. Inorganic sulfides and long-chain alkyl thiols with suitable modifica-
tions were found to be the most effective binding agents. A number of factors
were identified which affect the decision to decontaminate a polluted sediment
or to remove the material by dredging. If the material is to be dredged, pre-
cautions must be taken when land disposal methods are used. The field survey
consisted of determining both the horizontal and vertical extent of the mercury
contamination as well as pertinent hydraulic parameters.
From results of the laboratory and field work, a pilot field project is described
whereby techniques for controlling mercury contamination can be evaluated at a
site where the field conditions have been fully established.
*Water Quality Control, *Metals, *Mercury, Water Pollution Treatment,
Sediments
Sludge Treatment, Ultimate Disposal, Sediment Treatment, Field Surveys,
Pilot Project
05F
nty
of
Send To:
WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON, D C 2O24O
Donald S. Yeaple
JBF Scientific Corporation
------- |