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

-------