&EPA
United States
Environmental Protection
Agency
                       Great Lakes
                       National Program Office
                       77 West Jackson Boulevard
                       Chicago, Illinois 60604
EPA 905-R94-022
October 1994
Assessment and
Remediation
Of Contaminated Sediments
(ARCS) Program
MINERAL PROCESSING PRETREATMENT OF
CONTAMINATED SEDIMENTS
                     •) United States Areas of Concern

                     P ARCS Priority Areas of Concern
                                 printed on recycled paper

-------
                    MINERAL PROCESSING PRETREATMENT
                        OF CONTAMINATED SEDIMENT

                         Ashtabula River, Buffalo River,
                       Indiana Harbor/Grand Calumet River,
                           and Saginaw River and Bay
                              Areas of Concern

                                     by

                                James P. Allen
                      United States Department of the Interior
                               Bureau of Mines
                         Salt Lake City Research Center
                              729 Arapeen Drive
                           Salt Lake City, Utah 84108
Prepared for the Assessment and Remediation of Contaminated Sediments (ARCS) Program
                       Great Lakes National Program Office
                 U. S. Environmental Protection Agency, Chicago, IL
                   U S  Environmental Protection Agency
                   Region 5, Library (PL-12J)
                   77 West Jackson Boulevarjd, 12th Floor
                   Chicago, IL  60604-3590

-------
                                         DISCLAIMER

The information in this document has been funded in whole or in part by the U. S. Environmental Protection
Agency (EPA) under Interagency Agreements No. DW14934541-1 and DW14934541-2 with the U. S.
Bureau of Mines.  It has been subjected to the Agency's peer and administrative review and it has been
approved for publication as an EPA document. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.

-------
                                            ABSTRACT

The U.S. Bureau of Mines (USBM), Department of the Interior, was requested by the Great Lakes National
Program Office, U.S. Environmental Protection Agency, to evaluate various mineral processing techniques for
their effectiveness in assisting with remediation of contaminated sediments in rivers and harbors around the
Great Lakes.  Samples from the Ashtabula River, Buffalo River, Indiana Harbor/Grand Calumet River, and
Saginaw River and Bay Areas of Concern were received by USBM and evaluated in this study. USBM used
grain size analysis, surface area, and electron microscopy techniques to characterize the sediments and their
contaminants. Grain size separation, magnetic separation, gravity separation, attrition scrubbing, and froth
flotation were evaluated for their effectiveness in concentrating contaminants from sediment samples
representing each of the four sites.

The concept found to show promise in assisting remediation has been called pretreatment and involves using
mineral processing technology such as size separation to separate a contaminant-laden portion from the bulk
of the sediment.  The result of this is that the size and cost of the final treatment or disposal effort can be
reduced. Other potential benefits include improved effectiveness of any treatment process to follow and
possible beneficial use of cleaner sediment fractions.  This report shows that grain size separation applied to
a coarse-grained sediment such as that from the Saginaw River has potential to concentrate metallic and
organic contaminants in approximately 20 percent of the sediment mass. Potential applications of magnetic
separation at Indiana Harbor, and froth flotation at Saginaw River, are also deemed to show limited
application.

-------
                                     TABLE OF CONTENTS

DISCLAIMER	 li

ABSTRACT  	iii

LIST OF FIGURES	vi

LIST OF TABLES	  vii

INTRODUCTION 	 1
       Assessment and Remediation of Contaminated Sediments	 1
       Sample Description	 2
       Evaluating Results of Physical Separations	 7
       Special Procedures in Evaluating Organic Contaminants  	 8
               Technical Approach  	 8
               Data Quality  	 8
               Froth Flotation Separations for Organic Contaminants	 9

LABORATORY PROCEDURES	   11
       Methods Selection  	  11
       Deoiling	  12
       Working Size Fractions	     12
       Sample Preparation  	  12
       Chemical Analysis  	  12
       Grain Size Analysis	  14
       Gravity Separation  	  15
       Magnetic Separations	  15
       Froth Flotation	  15
               Metallic Contaminants	  15
               Organic Contaminants 	  16
       Attrition Scrubbing	  17

SAMPLE CHARACTERIZATION 	  18
       Density Measurements	  18
       Specific Surface Area Determinations  	  18
       Mineralogical Characterization	  20

RESULTS AND DISCUSSION	  23
       Ashtabula River  	  23
               Grain Size Separation	  23
               Gravity Separation  	  24
               Froth Flotation-Organic Contaminants	  24
                      Amine ethoxylate surfactants  	  25
                      Ethoxylated alcohol surfactants	26
       Buffalo River	  26
               Grain Size Separation	  26
               Gravity (Density) Separations	  _.. .  27
               Magnetic Separation	  28
               Froth Flotation-Metallic Contaminants	  29
               Attrition Scrubbing	  29
               Froth Flotation-Organic Contaminants	  29
                      Amine ethoxylate surfactants  	  29

                                              iv

-------
                     Ethoxylated alcohol surfactants	31
       Indiana Harbor Ship Canal and Grand Calumet River       ....         	33
              Grain Size Separation	33
              Gravity (Density) Separations	       	34
              Magnetic Separations	  35
              Froth Flotation-Metallic Contaminants	  36
              Attrition Scrubbing	  36
              Froth Flotation-Organic Contaminants	  36
                     Ethoxylated alcohol surfactants	  37
                     Grand Calumet River-Amine ethoxylate surfactants  	  37
              Size Classification-Organic Contaminants	  38
       Saginaw River	  39
              Grain Size Separation	  39
              Gravity (Density) Separations-Metallic Contaminants	  40
              Magnetic Separations	  40
              Froth Flotation-Metallic Contaminants	  41
              Attrition Scrubbing	  41
              Froth Flotation-Organic Contaminants	41
                     Anionic surfactants 	  41
                     Amine ethoxylate surfactants  	42
                     Nonionic surfactants  .  .     	         . .           .    .     ....  43
              Grain-size Separation-Organic Contaminants	                . .    43
              Gravity Separation-Organic Contaminants 	  44

CONCLUSIONS                                   .                                  .        46
       General Comments ....            	      ....                   	46
              Gram Size Separation    	                  	46
              Potential for Recycling of Metallic Contaminants	        	  46
       Summaries 	  46
              Ashtabula  River	  46
              Buffalo River	47
              Indiana Harbor   	  47
              Saginaw River	  47
       Feasibility Matrix	  48

RECOMMENDATIONS  	  48

REFERENCES	       	          	49

GLOSSARY OF MINERAL PROCESSING TERMINOLOGY 	       	  50

APPENDIX	  A-1

MINERALOGY OF CONTAMINATED GREAT LAKES SEDIMENTS   .         	       	  B-1
       INTRODUCTION  	       B-2
       RESULTS AND DISCUSSION	           	        .   .B-2
              Ashtabula River  Sediment	          ...          .    .           B-2
              Buffalo River Sediment	          	B-3
              Indiana Harbor Canal Sediment          	                                   B-3
              Saginaw River Sediment                                  .        .          .   B-4
       CONCLUSIONS            	      	B-5

APPENDIX-QUALITY ASSURANCE PROJECT PLAN	C-1

-------
                                         LIST OF FIGURES
Figure No.                                                                                 Page No.
1.   Location of Ashtabula River Sediment Sample	  3
2.   Location of Buffalo River Sediment Sample  	  4
3   Locations of Indiana Harbor Canal and Grand Calumet River Sediment Samples	  5
4.   Locations of Saginaw River Sediment Samples 	  6
5.   Partitioning of mass and contaminants among multiple products in a separation process	7
6.   Schematic diagram of froth flotation laboratory setup	  16
7.   Specific Surface Area of Great Lakes Sediments by Particle Size  	  19
8.   Correlation of selected contaminant levels with surface area for a) Indiana Harbor Ship Canal, and
    b) Saginaw River sediment	  20
9.   Photomicrograph of Saginaw River sediment #1, showing wood chip with inclusions of iron- and zinc-
    bearing phases (300x)	  22
10  Photomicrograph of minus-400-mesh Saginaw River #1 sediment, showing micron-sized lead-bearing
    particles as bright specks (1000x)	   22
11.  Partitioning of Contaminants by Particle Size-Ashtabula River Sediment	23
12.  Levels of PCB Contamination in Size Fractions of Ashtabula River Sediment	       	24
13.  Partitioning of Contaminants by Particle Size-Buffalo River Sediment     .           	    27
14.  Effects  of reagent composition, aeration rate, and reagent concentration on flotation of oil and grease
    from Buffalo River sediment	               .        	31
15.  Effect of aeration rate and Igepal CO-530 concentration on effectiveness of oil and grease separation
    from Buffalo River sediment	  32
16.  Partitioning of Contaminants by Particle Size-Indiana Harbor Canal Sediment  	  34
17.  Effects  of reagent composition, reagent concentration, and solids content on flotation of oil and grease
    from Grand Calumet River sediment	38
18.  Partitioning of Contaminants by Grain Size Separation on Saginaw River Sediment #2	  39
                                                VI

-------
                                          LIST OF TABLES
Table No.                                                                                  Page No.
1.   Analyses of Contaminated Great Lakes Sediments	        .       .     .  .        2
2.   Characterization and Separation Methods Investigated     ..    .      .               	  11
3.   Objectives for Precision, Accuracy, and Completeness
    in Chemical Analysis of Sediment Samples	13
4.   Sizes of Standard Testing Sieves   	  14
5.   Surfactants for Oily Sediment Flotation	  17
6.   Densities of Great Lakes Sediments	  18
7.   Flotation of Oil and Grease from Ashtabula River Sediment Using Amine Ethoxylate Surfactants--
    Parameter Screening	  25
8.   Flotation of Oil and Grease from Ashtabula River Sediment Using Ethoxylated Alcohol Surfactants-
    Parameter Screening	  26
9.   Example of Gravity Separation Effect:  Buffalo River Sediment, +400 mesh 	  28
10. Flotation of oil and grease from Buffalo River sediment using Amine Ethoxylate surfactants	  30
11. Effects of Aeration Rate and  Reagent Concentration on Flotation of Buffalo River Sediment with
    Ethoxylated Alcohol Surfactant	  33
12. Example of Gravity Separation Effect:  Indiana Harbor Sediment, 100 x 400 mesh	  35
13. Flotation of oil and grease from Indiana Harbor Canal sediment using Surfactants	  36
14. Flotation of oil and grease from Grand Calumet River sediment using amine ethoxylate surfactants  ...  37
15. Distribution of PCBs  and oil and grease by particle size in Grand Calumet River sediment	  38
16. Example of Gravity Separation Effect:  Saginaw River Sediment #1,+200 mesh	  40
17. Flotation of Saginaw River Sediment #2 with Anionic Surfactants  	42
18. Flotation of oil and grease from Saginaw  River sediment #2 with amine ethoxylate surfactants 	  42
19. Flotation of Saginaw River Sediment #2 with Nonionic Surfactant    	43
20. Distribution of organic constituents by particle size in Saginaw River sediment #2             	44
21. Particle size separations at 425 urn and 12 urn in the presence of nonionic surfactant	  44
22. Elutriation of+200 mesh Saginaw  River Sediment #2 for PCB Recovery	        	45
23. Feasibility Matrix-Mineral Processing Pretreatment of Contaminated Sediment            	48
                                                 VII

-------
                                          INTRODUCTION

Assessment and Remediation of Contaminated Sediments

The USBM's effort in contaminated Great Lakes sediment characterization and treatment was performed
under an Interagency Agreement between USBM and the U.S. Environmental Protection Agency (USEPA).
Work commenced in April, 1990 by the Chemical Remediation research group at the Salt Lake City Research
Center (SLRC).

The project was conducted under the direction of the Engineering/Technology Work Group in the Assessment
and Remediation of Contaminated Sediments (ARCS) Program to investigate mineral processing
technologies as removal or remediation alternatives for contaminated sediments.  The ARCS Program is a
6-year effort authorized by the Water Quality Act of 1987. Under this program, the Great Lakes National
Program Office (GLNPO) of the USEPA carried out studies emphasizing the removal of toxic pollutants from
sediments in the Great Lakes system. The objectives of the ARCS Program were to assess the extent of
sediment pollution in designated Areas of Concern and to identify and demonstrate options for the removal
and/or treatment of the contaminated sediments.

The USBM's contribution to the ARCS Program has been to evaluate the application of mineral processing (or
physical separation) technologies to removal of low levels of contamination from large volumes of sediment.
Physical separation techniques are widely used in the mining industry to recover valuable minerals or metals
from ores. Methods such as size classification, magnetic separation, gravity  separation, or froth flotation,
collectively known as mineral processing, can be applied in some cases to separate contaminants from the
bulk of polluted sediment. The objective is to reduce the volumes of contaminated material requiring
treatment by concentrating the contaminants, in the same way an ore is beneficiated.  This concept is familiar
to the mining industry,  but it's application to full-scale contaminated sediment remediation projects in the U.S.
is relatively new. Since these methods are economically applied on a very large scale to ores of low value-to-
mass ratio, they are among the least expensive processes in modern industry. Although these processes are
applied relatively inexpensively in the mineral processing industry, the modifications and environmental
controls that will be required in a sediment  remediation application will cause these costs to increase.

As applied to an environmental remediation situation, such as contaminated sediments, a few important
points about mineral processing technology (i.e., physical separation) should be noted. First, mineral
processing makes particle-particle separations. No chemical bonds are  broken, and no phases are
destroyed.  This is in contrast to many other remediation technologies, where a process such as incineration
actually destroys the contaminant.  Second, mineral processing separations are based on differences in the
physical properties of particles, so that no separation can be achieved if all particles are physically similar.
Third, mineral processing is usually a high-capacity operation. Since most ores that are processed contain
low valuable metal levels, this technology must have a low cost per unit of mass processed.  For example, in
processing copper, five or six separate mineral processing operations will be performed, plus smelting and
refining, all on  an ore that contains less than $10 worth of copper per ton. (A typical Western-U.S. ore
contains 0.5 percent copper, or 500 mg/kg.) Finally, the  capacity and efficiency of most mineral processing
operations decreases with particle size.

Some contractors offer pretreatment or "soil-washing" technologies, usually consisting of one or more mineral
processing operations.  Best results will be obtained when the pretreatment system  is chosen based on a
detailed knowledge of the physical and chemical characteristics of the sediment.  Mineral processing unit
operations appropriate to the physical characteristics of the sediment can then be arranged into an integrated
system.  Bench-scale mineral processing testing to verify performance is inexpensive, and scale-up reliability
is well-documented (Weiss, 1985). Thorough guidance on applying these and other sediment remediation
technologies is published by USEPA in the ARCS Program's Remediation Guidance Document (USEPA,
1994).

-------
Sample Description

USBM  received a total of six samples of contaminated sediment, from the Ashtabula River, Buffalo River,
Grand Calumet River/Indiana Harbor Ship Canal, and Saginaw River Areas of Concern. Characterization and
testing  of these samples forms the basis for this report.  These samples, with their chemical analyses, are
reported in Table 1.  The locations where the samples were collected are shown on the maps in Figures 1 -4.
The Saginaw River #2 sample was taken from a  reputed "hot spot" in the river near the Bay City Wastewater
                      Table 1.—Analyses of Contaminated Great Lakes Sediments

                                          Analysis, mg/kg dry weight
Element
As
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Zn
Sb
Ba
Se
Ag
PCBs
Oil&
Grease
Ashtabula
River
5.4
3.8
168
121
24,600
44
0.24
80
192
NA
NA
NA
NA
1.2
15,600
Buffalo
River
10.3
2.0
60
70
29,700
100
0.47
37
167
2.4
140
0.37
0.27
2.1
5,910
Grand
Calumet
River
22.5
1.3
351
110
108,000
316
3.98
76
1000
NA
NA
NA
NA
152
33,000
Indiana
Harbor Ship
Canal
30.8
8.0
440
278
142,000
830
0.74
94
3,280
7.8
200
1.6
2.7
NA
49,800
Saginaw
River #2
5.6
4.0
160
78
14,000
41
0.22
61
273
2.0
59
0.19
0.7
18
2,530
Saginaw
River #1
8.5
1.2
95
63
22,900
36
0.14
34
206
NA
NA
NA
NA
NA
NA
  NA=not analyzed
Treatment Plant outfall.  The Saginaw River #1 sample was considered in some respects to be more
representative of the contaminated sediment in the river than the #2 sample.  It was a composite of sediment
from 3 sampling stations in the Federal navigation channel of the lower Saginaw River in the vicinity of Bay
City, Michigan. The Grand Calumet River sample was used entirely for separation testing on organic
contaminants, and was collected about 0.75 miles east of Bridge St. in Gary, Indiana, near the U. S. Steel
coke ovens.  The Indiana Harbor sample was taken from the  Indiana Harbor Ship Canal near Columbus
Drive. The Buffalo River sample was collected from the upstream end of the Federal navigation channel,
near Mobil Oil. The Ashtabula sediment was a composite of samples collected throughout the river. These

-------
                                                    Ashtabula River
                                 The sediment sample was a composite comprised of
                                 subsamples taken throughout the Ashtabula River system
Figure 1.—Location of Ashtabula River Sediment Sample

                                           3

-------
Figure 2.—Location of Buffalo River Sediment Sample

-------
                                          Indiana Harbor sediment sample
                                                    Grand Calumet River sediment sample

                                                                       u    u    t.e
                                                                       I	1	1 mliet
Figure 3.—Locations of Indiana Harbor Canal and Grand Calumet River Sediment Samples


                                             5

-------
      Sediment sample Saginaw #1 was a composite of
      three samples collected throughout the Saginaw River
                           Sediment Sample Saginaw #2
                                                   Uorcteri
                                                1         J
                                      0   75   i
.75   1
   nta
Figure 4.—Locations of Saginaw River Sediment Samples

                                             6

-------
sediment samples were collected by the ARCS Engineering/Technology Working Group for use in a series of
bench-scale treatment technology evaluations; USBM received representative aliquots of the samples for its
work.

Evaluating Results of Physical Separations

Evaluation of the success of a mineral-processing separation fora given constituent is made  using two
parameters: the mass of sediment found to report, after separation, to a given process stream, and the
concentration of the constituent of interest in that process stream.  Metallurgical accounting is used to
determine the removal efficiency, or distribution, of a constituent in a given process stream or streams. This
is illustrated in Figure 5, which shows a hypothetical process separating a feed material of  mass M0 and
contaminant concentration a0  into n product streams, each of mass M, and contaminant level  a,. The
                  Feed
                KEY
        a - contaminant concentration
        M = Mass
                                         Physical
                                        Separation
                                         Process
                                                                    Products
                                                                  •*» a,, M,
                                                                      a?M2
                                                                   *- a, M3
-#»- a,, M
Figure 5.—Partitioning of mass and contaminants among multiple products in a separation process.
distribution of a contaminant into one of n product streams is calculated as

                             aM     aM
                      D   -
                        0,1
                                                                                             (1)

-------
where
        Da, = distribution of constituent a into product stream i,
        a, = concentration of constituent a in product stream i,
        M, = mass (or mass flowrate) of product stream i,
        a0 = concentration of constituent a in the feed,
        M0 = mass (or mass flowrate) of the  feed to the process, and
        n = number of products.

It is customary to multiply by 100 and speak of product distributions in percentages.

This is somewhat different from a more typical remediation process, where a constituent is destroyed yielding
a single product stream nearly equal in mass to the initial feed mass.  In this case, M, for the single product is
equal to M0, and the removal efficiency is equal to one minus the ratio of tne final constituent concentration a,
to the initial constituent concentration a0.

Most physical separation processes employed here result in two products;  one enriched in the contaminants
of interest, and another depleted in the same contaminant(s). Usual mineral-processing practice is to speak
of the enriched product as the "concentrate," while the depleted product is called the "tailings" or simply
"tails." The most successful physical separations in environmental remediation are those that result in a
concentrate product of low mass, and a tailings product of low contaminant concentration.

Special Procedures in Evaluating Organic Contaminants

                                         Technical Approach

Subsequent to evaluation of separability of metallic contaminants by mineral processing methods, a program
was undertaken to evaluate certain methods for the applicability to removal of organic contaminants. This
work concentrated on particle size separation and froth flotation. The work was intended to be preliminary in
nature, identifying promising avenues for remediation process development, but not necessarily predicting
performance of specific treatment methods.

Owing to the expense of PCB analysis on each of the large number of samples generated by a mineral
processing test campaign, and the possible mix of organic contaminants found from one AOC to another, the
oil and grease analysis was used as an indicator of partitioning  of organic contaminants among the products
of separation tests. It was assumed that toxic constituents, such as PCBs, would distribute in separation
experiments in the same manner as  oil and grease.  A few PCB analyses were obtained to verify the outcome
of certain  promising experiments.  It should  be understood,  however, that a given toxic organic constituent
may reside in a number of phases, such as oil, grease, coal, or detritus, and that the oil and grease analysis
does not fully reflect this diversity.

                                            Data Quality

The Bureau of Mines was authorized by the ARCS Quality Assurance Officer to use the oil and grease
analysis of separation products only for identification of general trends.  It was not intended to use oil and
grease analyses obtained by the Bureau to define the character of sediment samples, or to predict sediment
cleanup effectiveness in more than a general way.

Through the course of the test campaign, difficulty in obtaining a satisfactory mass balance on the oil and
grease through a separation procedure was  observed  infrequently. No systematic errors were identified to
account for this, but the following observations are offered to describe the scope of the problem. The mass
balance closure (i.e. that fraction of the amount of a constituent initially found in the sample that can  be
accounted for after a separation process) is  used in process engineering to assess the quality of chemical
analyses.

                                                  8

-------
        •       Samples with low oil and grease levels (below 1,000 mg/kg) seemed to contribute to poor
                mass balance closure. This was especially true of the coarse-grained Saginaw River
                sample.

        •       Most mass balance deviations were in the direction of reporting an excess of oil and grease
                relative to the measured starting concentrations, so that researchers feel comfortable in
                reporting the results of oil-and-grease separations as having removed "at least" the stated
                amount of oil and grease. Again, the data presented are best evaluated by looking  for trends
                within a data set, rather than concentrating on the result of a single experiment.

        •       Mass balance closures were usually good for the Buffalo River, Indiana Harbor, and
                Ashtabula River samples, at 6,000, 25,000, and 18,000 mg/kg oil and grease, respectively.
                As mentioned above, the Saginaw River sample,  at about 1,000 mg/kg oil and grease, was
                more troublesome.  Also  difficult was the extremely polluted Grand Calumet River sample.
                Most analyses from the Bureau's contractor placed  the oil and grease  content at about
                30,000  ppm for this sample; soxhlet extractions performed by the Bureau  resulted in
                estimates of about 140,000 mg/kg of extractable material.

                          Froth Flotation Separations for Organic  Contaminants

Froth flotation is  used to  process  millions of tons of ore daily.  Copper, iron, phosphates, coal, and potash are
a few of the materials that can be economically concentrated by this  method. It is based on manipulating the
surface properties of minerals with reagents so that the mineral of  interest has a hydrophobic (water-hating)
surface, like wax. The minerals to be rejected have, or are made to have, a hydrophilic (water-loving)
surface. When air bubbles are introduced, the hydrophobic minerals attach to the bubbles and are carried to
the surface and skimmed away. Flotation has been successfully applied to particles as small as 10 microns.
Almost all flotation is conducted in stirred,  aerated tanks of up to 56 cubic meters (2,000 cubic feet), although
vertical columns and air-sparged  hydrocyclones are used occasionally.

When using flotation to remove oily contaminants from sediment, the role of a surfactant resembles that of a
detergent. Most organic contaminants are naturally very hydrophobic, and the objective in using a surfactant
is to  reduce the oil phase's hydrophobicity to the point where it will be wetted by the water phase and detach
itself from solid surfaces. Surfactants are  able to accomplish this because such molecules have a lipophillic
(fat-soluble) head, which is absorbed into the oil phase, and a hydrophillic tail, which extends into the water
phase. The result of this is that the overall hydrophobicity of the oil phase is decreased. The strength of a
surfactant's attachment to an oil phase is approximated by the hydrophile-lipophile balance (HLB) of the
surfactant.  Once freed of the solid surface, an oil droplet is assisted to the surface by air bubbles and
skimmed away.

Two  factors are required to evaluate  the effectiveness of a flotation separation:  distribution of mass toward
the clean tailings, and distribution of  oil and grease toward the oily froth concentrate. The coefficient of
separation can be used to show the best combination of these two results as a single parameter. In  this
case, the coefficient of separation is  calculated by subtracting the weight distribution to the concentrate from
the distribution of oil and grease in the concentrate:

           Coefficient  of separation  =  DOGconc - Dwtconc                                            (2)


where
        DOGconc= distribution  of oil and grease to the froth concentrate, and
        Dwt,conc=  distribution of weight to the concentrate,  Mconc/Mtota|.

Again,  it is customary to  calculate these values as percentages by multiplying by 100.

-------
The coefficient of separation is useful because a successful flotation separation requires both high
contaminant distribution and low weight distribution to the concentrate. Neither result alone can identify a
good separation. For example, the froth concentrate may contain 80 to 90 percent of the contamination, but if
it also represents 80 to 90 percent of the mass of material then no useful separation has been effected. The
ideal (though impossible) result would be for 100 percent of the contaminants to end up in 0 percent of the
weight, yielding a coefficient of separation of 100.
                                                  10

-------
                                    LABORATORY PROCEDURES
Methods Selection

The evaluation of these sediment samples roughly followed the course a typical metallurgical investigation on
a potential mineral ore: Based upon information about the physical properties of the material gleaned from
characterization studies and experience with similar materials, mineral-processing methods, reagents, and
test conditions were selected that had the potential to exploit the known physical characteristics of the
material. A matrix showing which separation methods were evaluated for each sediment is provided in Table
2. This is a stepwise approach; lessons learned from a given series of tests are used to design the next tests
performed, as opposed to a campaign where each sample is subjected to a predetermined battery of tests.
This approach is more likely to result in a useful separation procedure for a single sample, but allows for only
limited comparability among different samples,  since different procedures, or different conditions of the same
procedure, may have been applied.

	Table 2.—Characterization and Separation Methods Investigated	
                                                    Sediment Sample
Method
Grain-size
Separation-
metals
Grain-size
Separation-
organics
Density
Separation-
metals2
Density
Separation-
Ashtabula Grand Calumet Indiana Harbor
River Buffalo River River1 Ship Canal Saginaw River
X X X X X
X X X X X
X XX

                                          X


                                          X


                                          X
X


X


X
X

X


X


X
organics0

Mineralogy             X

Flotation-
Metals2

Flotation-
Organics               X

Magnetic
Separation-
metals2	X	X	X
1 Evaluation of this sample was limited to organic contaminants.
2 Methods that showed little promise in treating sediment of the three initial samples (Buffalo, Indiana Harbor, and Saginaw) were not
repeated on the Ashtabula River material.
3 Performed only as a "polishing" step following grain-size separation
As far as possible, generally-accepted mineral-processing laboratory procedures were followed in
characterizing the samples and conducting the laboratory-scale separation experiments. These procedures,
which allow for considerable modification by the user depending on the subject material, are summarized in
                                                  11

-------
the Society of Mining Engineers's Mineral Processing Handbook (Weiss, 1985.)

Deoilinq

Most of the bench-scale testing on these sediment samples was conducted using material that had been
freed of oil. Soxhlet extraction using 1,1,1 trichloroethane was used to prepare sediment samples for grain
size analysis and bench-scale magnetic, gravity, and grain-size separation tests. Samples so prepared are
identified as "deoiled" in this report.  The purpose of the deoiling procedure was  simply to improve the
consistency and physical-handling properties of the sediment.  It was assumed that removal of the oily portion
of the sediment would have no effect on the metals analyses of the sediment, or  of separation products
derived from deoiled sediment. The deoiling step was most important for dry sieving, dry magnetic
separation, and heavy-liquid separation, where clumping caused by the presence of oil or grease would skew
test results, foul the equipment, or both. The USBM's study was initially focused  only on metallic
contaminants-as evaluation of organic contamination was integrated into the study every effort was made to
accommodate sediment samples that had not been deoiled.

Working Size Fractions

Because  almost all separations are sensitive to particle size, each sediment sample was separated into
"working" size fractions by wet sieving.  In some cases, a further size split was made in the subsieve range
using a small glass hydrocyclone. This split was measured by the x-ray sedimentation instrument to be at
about 12  (Um. These fractions are not the same for each sample, but were chosen to obtain roughly equal
amounts  of material in each fraction. They were not intended as approximations of reasonable or effective
remedial  separations, but were made solely to increase the amount of information gained from separation and
characterization procedures in the laboratory. The tables reporting results of separation experiments will
provide data separately for each of the working size fractions.

Sample Preparation

To prepare the product of each separation test for chemical analysis, each sample was filtered if necessary,
dried at 100 °C, then pulverized to pass 100 mesh. If the material was to be analyzed for organic
contaminants, drying temperature was reduced to 60 °C.

Chemical Analysis

Evaluation of the effectiveness of separation processes is heavily dependent on chemical analyses of
separation products. The characterization and separation tests for metallic contaminants described  in this
document resulting in preparation and analysis of 802 samples, with approximately 7200 individual element
determinations. Chemical analysis for metals was performed by USBM according to a Quality Assurance
Project Plan (QAPP) approved by the ARCS Quality Assurance Officer and Project Manager. The QAPP can
be found  in Appendix C.  Amendments to the QAPP provided for analysis of certain organic constituents;
data quality issues regarding these analyses are discussed below.  This effort resulted in analysis of
approximately 580 additional samples. A data validation report was completed by the ARCS QA Officer and
is available from GLNPO. The analysis methods, with QA parameters, are summarized in Table 3.
                                                12

-------
                     Table 3.—Objectives for Precision, Accuracy, and Completeness
                               in Chemical Analysis of Sediment Samples
Digestion
Parameter Method
Antimony Hot HNO3 + H202
Arsenic Hot HNO3 + H2O2
Barium Hot HNO3 + H202
Barium
Cadmium Hot HNO3 + H2O2
Cadmium4
Chromium Hot HNO3 + H2O2
Chromium'1
Copper Hot HN03 + H2O2
Copper4
Iron Hot HNO3 + H2O2
Iron4
Lead Hot HN03 + H202
Lead4
Mercury Hot Aqua Regia
+ KMn04
Nickel Hot HN03 + H202
Nickel"
Selenium Hot HN03 + H202
Selenium Hot HN03 + K2S2Oa
Hot HCI reduction
Silver Hot HN03 + H202
Silver4
Zinc Hot HN03 + H202
Zinc4
Analysis'
Method
GFAAS1
GFAAS2
FAAS2
GFAAS'
FAAS3
GFAAS'
FAAS3
GFAAS'
FAAS3
GFAAS'
FAAS3
GFAAS'
FAAS3
GFAAS'
Cold Vapor
FAAS3
GFAAS'
GFAAS2
HGAAS3
FAAS2
GFAAS'
FAAS3
GFAAS'
Reference
SOW #788
SOW #788
SOW #788

SOW #788

SOW #788

SOW #788

SOW #788

SOW #788

SW 846 Method
7471 9/86
SOW #788

SOW #788
Used when SO42
is >2000 ppm
SOW #788

SOW #788


Rel Std.
Dev
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
1 Graphite Furnace Atomic Absorption Spectrometry. Palladium may be substituted for nickel
2 Flame Atomic Absorption Spectrometry
3 Hydride Generation Atomic Absorption Spectrometry
4 GFAAS is substituted for FAAS when necessary to achieve a lower detection limit.
Precision
Accuracy
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
matrix modifier.

Complete-
ness
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%


Mineraloqical Analysis

Sediment samples from each location were mounted in resin, sectioned, and polished. A Hitachi S530
scanning electron microscope (SEM) was used for visual examination of particles and qualitative elemental
analysis by energy-dispersive x-ray spectrography. Particle size, shape, degree of liberation, and texture
were determined visually. Elemental information was obtained by using an energy-dispersive spectrometer
(EDS) to analyze x-rays emitted from the sample.  Generation of x-ray energies due to the interaction of the
electron beam with the atoms in the sample creates peaks characteristic of each element. The system is
capable of performing simultaneous analysis of elements beginning with sodium in the periodic table. An
area of a few cubic microns can be analyzed.  A thin layer of gold applied to the surface of the sample is
required for this analysis and is the reason for the gold peaks seen on the EDS spectra. Calibration of the
peaks to their respective energies was performed regularly.
                                                 13

-------
Grain Size Analysis

Size analysis of sediment samples was performed using essentially the same procedure ordinarily employed
for analysis of ores.  Tyler mesh sieves of stainless steel mesh were used for all analyses. The Bureau's
procedure was to sieve the sample wet on 400 mesh and then dry the fine and coarse fractions by heating as
described above. The coarse (plus 400 mesh) material was placed on a nest of sieves in the usual square-
root-of-two series and shaken by machine for 15 minutes. The material remaining on each sieve was
weighed and reserved for analysis. The material passing the finest sieve (400 mesh) was combined  with the
minus 400 mesh material from wet sieving for analysis. The mesh size openings in microns are shown in
Table 4.


                              Table 4.—Sizes of Standard Testing Sieves
Tyler Mesh U.S
no.
10
14
20
28
35
48
65
100
150
200
270
400
500
. Standard Nominal
no. opening, urn
12
16
20
30
40
50
70
100
140
200
270
400
500
1700
1180
850
600
425
300
212
150
106
75
53
38
25
In the square-root-of-two sieve series, each sieve is chosen so that its nominal opening size differs from the
next sieve larger or smaller by /2. The reader can easily verify this in Table 4 by observing that each sieve
opening is approximately double that of the sieve two sizes smaller. Using sieve sets in which the openings
are in constant ratio is customary in mineral processing; in such cases, the weight w retained on a sieve is
related to the opening size x by the equation
                            w = ex
                                                                                               (3)
where Cand mare experimental constants (Gaudin and Hukki, 1946).
                                                 14

-------
Gravity Separation

The separability characterization work was performed on sediment which was deoiled and classified into
working size fractions. Unless otherwise indicated, each working size fraction was separated into three
products based on specific gravity (S.G.) a light product (S.G. < 1.9), a middling product (1.9 <; S.G. < 2.9),
and a heavy product(S.G.> 2.9) by suspending the sediment sequentially in two "heavy liquids":
bromochloromethane (BCM, S.G. = 1.9), and tetrabromoethane (TBE, S.G.  ~ 2.9). The liquids were chosen
to isolate lightweight solid organic  matter (wood bits, leaf fragments, coal, charcoal, etc.) in the light product,
common, naturally-occurring nonmetallic minerals in the middlings, and metallic phases in the heavy product.
Separatory funnels with specially-constructed outlets were used to perform these separations. The sediment
that floated in BCM was the light product, the sediment that sank in BCM but floated in TBE was the
middlings product, and the sediment that sank in  BCM and again in TBE was the  heavy product.

Magnetic Separations

As-received (i.e., not deoiled) sediment was  used in  the working size fractions. Before placing the material in
the separator, a hand-held permanent magnet was passed through each working size fraction, with the object
of collecting the most magnetic material and  preventing its plugging  the separator. On the Indiana Harbor
sample, a Davis Tube separator was employed in place of the hand magnet, owing to the expected large
amount of recoverable ferromagnetic iron ore. The Davis Tube is a low-intensity wet magnetic separator
commonly used in the iron ore industry to evaluate the recoverability of magnetite (Fe304). The working size
fractions coarser than 200 mesh were dried and passed through a Carpco-induced roll magnetic separator at
two settings. The fractions finer than 200 mesh were treated on the wet high-intensity magnetic separator
(WHIMS), also a Carpco instrument. The difference between these two separators is in the size of the
material they treat--the WHIMS is ineffective on coarse material, while the induced-roll separator suffers
dramatic losses in capacity on treating fine material. The procedure followed in these tests with both
instruments was to begin with a low magnet  current, then returning the uncollected (and therefore less
magnetic) material to the separator at a  higher setting, repeating to collect fractions of steadily decreasing
magnetic susceptibility. Owing to the difficulty of making magnetic flux measurements with these small
separators, the common metallurgical practice is  to report the electromagnet current in amperes; the
magnetic field strength is directly proportional to the  magnet current.

Froth Flotation

                                         Metallic Contaminants

In these tests, as-received sediment was used without drying or deoiling. A bench-top Denver flotation
machine with a 1.2-liter volume was used. Air was supplied by natural aspiration, and agitation was at about
1,700 rpm,  sufficient to maintain suspension  of the sediment.  In the normal practice of flotation, reagents
called collectors are added to render the surfaces of the desired mineral particles hydrophobic. These
mineral  particles then attach to air bubbles and rise to the surface, where they are skimmed off in a froth.
This is illustrated in the schematic  in Figure 6. Often a frothing agent (usually a long-chain alcohol) is used to
stabilize the mineral-laden froth. For the current tests on sediment, three reagent schemes were tested.  In
the first, the sediment was subjected to flotation without a collector.  In the second, oleic acid  (a fatty or
carboxylic acid collector) was added. Fatty acids are widely used anionic collectors for floating metal oxides
and salt-type minerals such as apatite, calcite, and fluorite.  In the third configuration, potassium amyl
xanthate was added with copper sulfate (CuS04). Xanthates are anionic collectors used for recovering metal
sulfides, and CuS04 is sometimes used  to enhance the floatability of certain metal sulfides such as sphalerite
(ZnS). Obviously, since copper is added to the system, the effectiveness of xanthate-collector flotation at
separating copper as a contaminant cannot be evaluated. Since collector adsorption is usually dependent on
pH (at least for ionic collectors), various  pH levels were tested as appropriate for the expected collector-
mineral  system.  It can be assumed that if heavy  metal contamination in the sediment is present as discrete
                                                 15

-------
                                AIR
                                                                MINERAL OR
                                                                CONTAMINANT SKIMMED
                                                                AND COLLECTED AS
                                                                FROTH
                                                      HYDROPHOBIC MINERAL OR
                                                      CONTAMINANT ATTACHES TO
                                                      AIR BUBBLES
          IMPELLER
          Figure 6.—Schematic diagram of froth flotation laboratory setup.

oxide or sulfide phases, it would report to the flotation concentrate in at least some of these tests.

                                        Organic Contaminants

The equipment and procedures used for evaluation of organic contaminant recovery by flotation are the same
as those described above in connection with metallic contaminants.  The only exception is that usually a
frothing agent was not added to the organic experiments, since the surfactants used typically have very
strong foaming properties, alleviating the need for enhancement of froth formation.

It is common practice to refer to these flotation reagents by their trade names, and that practice will be
followed here; exact chemical composition is usually available in publications such as "McCutcheon's
Emulsifiers and Detergents-North American Edition," published annually (McCutcheon, 1991), or in literature
from the manufacturer.  In this study, a variety of surfactants were initially selected, based on their detergent
properties, for investigation with contaminated sediments. These surfactants are listed in Table 5. These
surfactants were selected for their reputed efficacy in oil/water systems, based on consultation with the
manufacturers. The most promising appear to be ethylene oxide-based amine and alcohol surfactants. The
nonionic ethoxylated alcohol surfactants are among the most popular detergents owing to their effectiveness
in industrial and household cleaners and easy biodegradation.  Polyoxyethylene-derivative  surfactants range
from water-insoluble, oil-soluble reagents with three or  fewer moles of ethylene oxide per mole of reagent, to
the water-soluble, oil-insoluble forms with up to 30 moles of ethylene oxide.
                                                 16

-------

Charge
Anionic
Anionic
Anionic
Anionic
Nonionic
Nonionic/cationic
Table 5. — Surfactants
Chemistry
Phosphate Esters
Alkyl aryl sulfonates
Lauroyl sarcosinates
Fatty acid sulfates
Ethoxylated alcohols
Amine ethoxylates
for Oily Sediment Flotation
Trade Name(s)
Rhodafac PL
Nacconol
Hamposyl

Igepal CO
Triton X
Triton RW

Common Uses
Emulsifiers, hand
cleaners, dry cleaning
Detergents, oxide
mineral flotation
Mouthwash, shampoo,
rug cleaners
Soaps, oxide and
silicate mineral flotation
Detergents, wetting
agents, personal care
products
Emulsifiers, detergents,
secondary oil recovery
Attrition Scrubbing

Attrition scrubbing is used to remove superficially-bound contaminants such as adsorbed metal cations from
solid particles.  Bulk as-received samples of sediment were subjected to attrition scrubbing in a modified
Denver flotation machine. To evaluate the possible benefits of chemical additions, the pH of the attrition
solution was adjusted to 3.0 with hydrochloric, nitric, or sulfuric acid.  Three mineral acids were used because
of their differing oxidizing capability and the differing solubilities of metal salts in chloride, nitrate, or sulfate
systems.  For example, lead would be soluble in chloride media, but not in a sulfate solution. Nitric acid was
employed for its oxidizing capability.  Attrition was accomplished by agitating the sediment in the specified
solution for 10 minutes at 1,700 rpm. The sediment was filtered, washed, dried, and analyzed for metals.
The filtrate was collected and adjusted with NaOH to pH 9.5.  Adjusting the solution after attrition and filtration
verified that any metals solubilized in the attrition process could be recovered from solution.
                                                   17

-------
                                   SAMPLE CHARACTERIZATION
Density Measurements
The density of each dry, deoiled sample was measured using a helium pycnometer. In addition to measuring
the density of the entire sample, the portion of the sample passing a 400 mesh sieve was measured
separately. These data are reported in Table 6. In each case, the data reported are the mean and standard

                             Table 6.—Densities of Great Lakes Sediments
                                          g/cm3 dry solids
                Ashtabula River
Buffalo River
Indiana Harbor
Sagmaw River #1
                            -400                  -400                 -400                  -400
              Complete    mesh    Complete     mesh     Complete    mesh     Complete     mesh
Mean
Density
Standard
Deviation
Rel. Std.
Dev. (pet)

2.500

0.011

0.44

2.598

0.002

0.07

2.716

0.012

0.45

2.737

0.004

0.14

2.529

0.015

0.58

3.124

0.002

0.08

2.697

0.002

0.08

2.682

0.002

0.09
deviation of 5 analyses of a single sample.  Geologists and mineral processing engineers frequently observe
that the most common naturally-occurring nonmetallic minerals in the earth's crust have densities near
2.7 g/cm3. As an approximation, the samples measuring higher than 2.7 g/cm3 can be assumed to contain
significant amounts of metallic phases, while those with densities lower than 2.7 g/cm3 can be assumed to
contain significant amounts of low-density organic matter, perhaps detrital material such as wood bits, or solid
carbonaceous phases such as coal or coke. The separate analysis of the minus-400-mesh material was
made to determine if the finest portion of the sediment had a significantly different composition from the bulk
of the sediment.  It is interesting to look at the Indiana Harbor samples in this light, since the coarser material
has lower density, and the finer  material has higher density.  Visual inspection through a binocular
microscope detected carbonaceous material,  probably coke, in the coarser fraction;  the higher density of the
finer fraction is undoubtedly due to the presence of fine iron oxides. The Sagmaw River sample appears to
have the  least amount of organic material, as  its measured density is very close to 2.7.  Both the Ashtabula
River and Buffalo River samples appear to have significantly more organic material, especially in the coarser
fractions.  Data regarding density of the solids are useful in calculating dredging and processing parameters,
as well as evaluating the feasibility of density-based separations.

Specific Surface Area Determinations

Specific surface area of the sediment is another useful characterization parameter. Sized fractions of each
deoiled sediment were analyzed for specific surface area by nitrogen adsorption using the Brunauer, Emmett,
Teller (BET) method (Brunauer,  Emmett, and  Teller, 1938). These results are shown in Figure 7. The
significance of these data is that the finer particles (usually) have higher surface area per unit mass than
coarser particles.  This is especially true  of the Saginaw and Indiana Harbor samples. The reason for the
upward trend in surface area with  the coarse material from Ashtabula and Buffalo is the presence of organic
material (detritus) in the coarse  fractions, as mentioned in connection with the density measurements. It can
be expected that material with higher surface  area will have a higher capacity for adsorption of contaminants,
such as metal cations, from water. The calculated surface area of nonporous spheres of a given diameter is
shown for reference, indicating that these sediment samples have from ten to one thousand  times the surface
                                                18

-------
area of hard spherical particles of equivalent size.
          100
           10
          0.1
         001
       0001
       00001
                                                                        J	•
                              Saginaw River #1    Saginaw River #2
                          10
                                      30           100         300
                                     Mean Particle Diameter,
1,000
3,000
Figure 7.—Specific Surface Area of Great Lakes Sediments by Particle Size
If adsorption is the mode of occurrence of much of the metallic contamination, one could expect to observe a
correlation between specific surface area and metals concentration. Two attempts at such a correlation are
shown in Figure 8. These figures are generated by separating the material into discrete size ranges by
sieving, and then measuring and plotting the specific surface area and metal content of each of these size
fractions.  The figure on the left, for Indiana Harbor sediment, shows an apparent correlation between levels
of arsenic, cadmium, chromium, and lead with surface area.  This is circumstantial evidence that a significant
portion of each of these metals is present as adsorbed species on the surfaces of particles. The figure on the
right, from Saginaw River, is less conclusive.  Some correlation appears for zinc, and possibly for cadmium
and chromium, but arsenic does not appear to be  correlated with surface area. This suggests that a different
mode of occurrence should be sought for arsenic, and that other modes of occurrence cannot  be ignored  for
the other metals mentioned.  Conventional mineral processing methods may not be successful on
contamination of this kind, since the elements of interest are not confined to discrete, liberated grains.
Physical/chemical methods such as attrition scrubbing in acidic or surfactant solutions may be required to
reduce the metal content of such material.
                                                 19

-------
        1,000
         SCO
         600
         400
         200
                                          o-
                                            35
                                            30
                                            25
                                            20
                                            15
                                            10
                                                       1.000
                                                        800  -
                                                        600  -
                                                        400  -
                                                        200  -
                                                                                           35
                                                                                           30
                                                                                           25
                                                                                           20
                                                                                            15
                                                                                            10
             235      10     20  30    50
                   Specific Surface Area, m'/g

             a. Indiana Harbor Ship Canal
02   05  1   2    5   10  20   50
      Specific Surface Area, m'/g

     b. Saginaw River #1
Figure 8.—Correlation of selected contaminant levels with surface area for a) Indiana Harbor Ship Canal, and
b) Saginaw River sediment.

Mineralogical Characterization

Mineralogical characterization can also be useful in characterizing sediment. The purpose of mineralogy in
this context is to attempt to identify contaminant-bearing phases, providing valuable information for designing
systems to recover these phases from the bulk of the sediment.  Samples from each of the Areas of Concern
mentioned above were analyzed using optical electron microscopy.  To aid in characterization, the samples
were deoiled and separated into several discrete size fractions. The complete mineralogist's report is
included later in this document: some of the more important findings are summarized here.

The coarsest (plus 10 and 10X14 mesh) fractions of the Indiana Harbor sample contained principally non-
mineral slag phases. Two types of slag are present:  glassy balls containing high amounts of sulfur with
minor vanadium, copper, and zinc, and a dull, vuggy material also high in sulfur but containing minor silicon,
aluminum, potassium, chromium, iron, copper, and zinc.  Iron oxide was also a common phase in the coarse
material. Again, the amount of slag present appears to decrease with decreasing particle size.

Naturally-occurring quartz is the most common phase in the 65><100 mesh Indiana Harbor sediment. Iron
oxide is quite common in this size fraction and is mostly liberated. Calcite, feldspar, and clay are also
mentioned. Pyrite was seen as inclusions in slag particles.

Abundant iron oxide was observed in the minus 400 mesh Indiana Harbor sample.  The slag phases, while
not as plentiful as in the coarser material, were still common. Single very dense grains about 10 /urn were
seen of lead/iron and lead/tellurium composition (one grain of each).  It was not possible to determine if these
lead-bearing grains were lead sulfide, since the lead and sulfur spectral lines overlap. The abundant
presence of iron ore is not surprising given the proximity of what were formerly some of the world's largest
steel mills. Assuming hematite (Fe203) stoichiometry for the purposes of making a calculation, the sample is
                                                 20

-------
estimated to contain about 20 percent potential iron ore. This suggests that the mass of contaminated
material might be reduced up to 20 percent by recovering the iron oxides as iron blast furnace feed. This is
discussed in connection with magnetic separation later in this document.

The most common phases in the coarser fractions of the Buffalo River sediment are (1) a silicon- and iron-
bearing slag phase containing chromium, copper, and nickel, (2) quartz, (3) feldspar, and (4) an iron oxide.
Clay, mica, and calcite are found in minor amounts. In the coarse fractions, wood is very common.  Also
observed in the coarse fraction are a brass colored unidentified metal (uncommon) and pyrite and
chalcopyrite inclusions in the iron- and silicon-bearing slag grains.  As particle size decreases, less slag and
wood are found, and the amount of mineral increases.

In the minus 400 mesh Buffalo River sediment, the same four principal phases were observed, but a larger
number of minor phases were identified. Calcite, sphalerite, zircon, barite, pyrite (common), and some
chalcopyrite are mentioned in the mineralogist's notes. Some copper metal was seen, as well as a single
20 ^m grain of silver.

Slag phases were also found in the Saginaw River samples, although in these cases, the slag contained iron,
aluminum, silicon, calcium, titanium and other non-target metals. Also observed through the microscope in
the Saginaw samples were 5-15 urn iron and zinc sulfide grains adhering to wood chips. This is significant
because  low-density phases containing mostly wood and leaf fragments were found to be high in metallic
contaminants in gravity separation tests discussed below. Such a  particle is shown in Figure 9. When copper
was detected, it was in sulfide mineral phases known to occur in nature. No phases were detected containing
cadmium or mercury. Lead  was observed concentrated in the finer fractions, most frequently as micron-sized
metallic or sulfur-containing  phases.  The photomicrograph in Figure 10 shows these lead-bearing grains as
bright spots dispersed throughout the sediment.

The occurrence of slag phases  in all of these samples is interesting for two reasons. First, slag materials are
usually glasslike in  appearance and chemical properties, refractory to most leaching systems. While this
must be verified, it seems likely that the metals associated with these slag phases are tightly bound and
biologically unavailable.  Second, most slags have lower density than common naturally-occurring minerals,
indicating that a gravity separation  might be successful in recovering this material, if necessary.
                                                 21

-------
Figure 9.—Photomicrograph of Saginaw River sediment #1, showing wood chip with
inclusions of iron- and zinc-bearing phases (300x).	
Figure 10. — Photomicrograph of minus-400-mesh Saginaw River #1 sediment,
showing micron-sized lead-bearing particles as bright specks (1000x)
                                    22

-------
Ashtabula River
                                     RESULTS AND DISCUSSION
                                         Grain Size Separation
The complete particle size and metallic contaminant distributions for this material are provided in Table A1 of
the appendix. The size chemistry analysis can be used to predict the effects of a size separation operation
(such as hydrocyclone treatment) on distribution of chemical components in a sample. It requires separating
the sediment into discrete size fractions, and analyzing each for contaminants. Part a in Table A1 gives the
size distribution and the elemental analyses of the individual size fractions expressed in mg/kg. Part b
provides the fraction of each contaminant found in each size fraction, while part c presents this same
information in cumulative form, showing the amount of each pollutant that is found with particles finer than the
stated size.  Part d presents the element analysis that would be expected in the coarser fraction if a perfect
size separation were performed at the stated size.

The Ashtabula River sediment is a very fine-grained material, with 64 percent of the mass finer than 37 urn
(400 mesh.) It shows no distinct trends of contaminant distribution according to grain size. Cadmium is
moderately enriched  in the fines, and chromium, iron, mercury, nickel, and zinc show very small enrichment in
the fine-grained portion of the sediment. There is also some enrichment of metals in the coarsest size
fractions resulting from the presence of organic detritus mentioned above. There is nothing in this data to
warrant further investigation of grain-size separation as a pretreatment option for concentration of metallic
contaminants from Ashtabula River sediment. This is shown in Figure 10, which  plots the amount of a
contaminant associated with particles finer than a given size versus the amount of mass associated with the
same particles.  When contaminants are evenly distributed throughout all sizes of particles, this plot is always
a straight line from the origin with slope=1. If contaminants are preferentially associated with either coarse or
fine material, downward or upward deviation from the diagonal (indicated on the figure as a dashed line) will
be observed.
                   700
                                              40          60

                                            Mass Distribution, pet finer
                                                                    80
                                                                                100
              Figure 11.—Partitioning of Contaminants by Particle Size-Ashtabula River
              Sediment.
                                                  23

-------
Figure 12 is a simple bar graph showing PCB concentration and distribution according to grain size for this
material. While the data show a small enrichment of PCBs in the coarse fractions (undoubtedly associated
with detritus), the overall variation is not large, indicating that this sediment is not a suitable candidate for
concentration of PCBs by grain-size separation.
                                 +48       100x150     200x270       -400
                                      48X100     150x200     270x400
                                         Size Fraction, Tyler Mesh
                    Figure 12.—Levels of PCB Contamination in Size Fractions of
                    Ashtabula River Sediment.
                                           Gravity Separation

 The mineralogist's report identified no contaminant-bearing phase in the Ashtabula sample that significantly
 differed in density from naturally-occurring sediment minerals.  This observation, in conjunction with the fine-
 grained character of the sediment, indicates that density separation testing on this material, with the objective
 of recovering metallic contaminants, would be unsuccessful. The report indicated that the coarser fractions
 contain significant amounts of detritus, suggesting that organic contaminant levels might be slightly reduced
 by removal of this material, although this was not tested in the laboratory.


                                  Froth Flotation-Organic Contaminants

 As will be shown in treating some of the other sediment samples, the fine-grained character of the sediment
 poses special challenges in performing particle separations. The many variables of a flotation system
 (reagent selection, reagent concentration, aeration rate, agitation intensity, pulp density, temperature, etc.)
 can all dramatically influence the quality of a separation, and fine-grained materials are usually more sensitive
 to changes in process variables.  For this reason, organic contaminant flotation testing on the Ashtabula River
 sediment attempted to screen variables to identify those process parameters that significantly influence
 separation quality.

 Seven common flotation process  parameters were selected for use in the study: surfactant composition,
 surfactant dosage, pulp density, aeration  rate, pH, agitation intensity, and conditioning time. A Plackett and
 Burman experiment design (Plackett and  Burman, 1946) was employed to identify the main effects, if any, of
 each of the variables. The effects of variable interactions are not detected by these experimental designs.
                                                  24

-------
The experimental campaign was conducted twice: once with the cationic/nonionic amine ethoxylate
surfactants of the Triton RW series, and again with the nonionic ethoxylated alcohol surfactants of the Igepal
CO series.  Results of these test series are provided in Table 7 and Table 8.—.

  Table 7.—Flotation of Oil and Grease from Ashtabula River Sediment Using Amine Ethoxylate Surfactants-
                                         Parameter Screening
Surfactant

Triton
reagent1
RW-50
RW-50
RW-50
RW-50
RW-150
RW-150
RW-150
RW-150

Cone.,
pet
0.01
0.01
0.10
0.10
0.01
0.01
0.10
0.10
Slurry
density,
pet solids
10
20
10
20
10
20
10
20
Aeration
rate,
L/min
1.0
3.5
1.0
3.5
3.5
1.0
3.5
1.0


pH
4
7
7
4
7
4
4
7
Agitation
speed,
rpm
900
900
1,700
1,700
1,700
1,700
900
900
Condit-
ioning
time,
mm
5
20
20
5
5
20
20
5
Concen-
trate
weight
dist., pet
29.8
49.3
56.7
92.3
220
346
184
23.6
Oil & Grease
Analysis, mg/kg

Concen-
trate
47,300
19,900
23,100
10,600
39,600
20,100
40,100
40,000


Tails
6,820
1,690
3,850
3,660
7,240
6,310
10,700
3,150
Coeffic-
ient of
Separ-
ation,
pet
448
42.7
32.0
47
38.7
28.2
274
56.1
  1 The code numbers associated with the reagent designation indicate ten times the number of moles of ethylene oxide per mole of
  surfactant (e.g., Triton RW-150 contains 15 moles EtO/mole surfactant.)	
Amine ethoxylate surfactants Like the nonionic Igepal CO reagents discussed in the following paragraphs,
the Triton RW surfactants are derived from ethylene oxide, although the use of the alkyl amine group gives
the surfactant some cationic character. Triton RW will form a stable oil/water emulsion at pH greater than 10;
reducing the pH to 7 or below breaks the emulsion and releases the oil. If recovery of the surfactant is
desired, this can be accomplished by again raising the pH to 10 or greater. All tests were conducted at either
10 or 20  percent solids using tap water for dilution.

For these amine ethoxylate surfactants, slurry density, agitation intensity, and surfactant composition were
found to  be  the significant variables  (with greater than 95 percent confidence) affecting oil and grease levels
in the tailings over the range investigated.  This is an important point:  the data show, for example, that 20
minutes of conditioning had  no significant effect over only 5 minutes of conditioning, but we cannot conclude
that conditioning is unnecessary, or  that a time shorter than 5 minutes would also be satisfactory.  Analysis of
variance indicates a  probability of 0.2 percent that the responses observed are the results of random effects
rather than the variables indicated.  Slurry density, surfactant composition, and agitation  speed were the
significant variables affecting weight distribution to the concentrate (at 90 percent confidence). There is a
5 percent probability that the responses observed are the results of random effects rather than the variables
indicated.

For this surfactant, best results were observed  at the lower conditions of pH and slurry density (pH 4 and
10 percent solids) and at the higher  level of reagent composition (15 moles EtO/mole surfactant).  Agitation
intensity  was found to have a strong deleterious effect on the separation, with the higher level resulting in
much higher losses of sediment weight to the oily concentrate.  These tests were conducted to identify the
important parameters controlling froth flotation separation of organic contaminants from sediment; no attempt
was made to optimize the system.
                                                  25

-------
 Table 8.—Flotation of Oil and Grease from Ashtabula River Sediment Using Ethoxylated Alcohol Surfactants-
                                        -Parameter Screening
Surfactant

Igepal
Reagent
CO-210
CO-210
CO-210
CO-210
CO-850
CO-850
CO-850
CO-850

Cone ,
pet
0.01
0.01
010
0 10
001
001
0 10
0.10


density,
pet solids
10
20
10
20
10
20
10
20


rate,
L/min
1 0
35
1 0
35
35
1 0
35
1 0



PH
5
9
9
5
9
5
5
9


speed,
rpm
900
900
1,700
1,700
1,700
1,700
900
900
Condit-

time,
mm
5
20
20
5
5
20
20
5
Content-

weight
dist., pet
138
137
143
29.2
743
64.4
4.9
7.8
Oil & Grease Analysis,
mg/kg

Concen-
trate
104,000
99,900
120,000
59,800
32,100
45,100
77,100
95,900


Tails
14,300
37,900
20,800
25,600
4,700
15,200
22,200
19,300
Coeffic-
ient of
Separ-
ation,
pet
74.1
58.8
709
409
12.9
105
72.8
75.4

Ethoxvlated alcohol surfactants The results were similar, though not identical, when evaluating the
ethoxylated alcohol reagents of the Igepal CO series (manufactured by Rhone-Poulenc Inc., Surfactant and
Specialty Chemicals Division). This is a group of nonylphenoxypoly(ethyleneoxy) ethanol reagents; each
member of the series has a different ethylene oxide content, and hence differing solubility in either oil or
water.  The numbers less than CO-530 denote oil-soluble, water-insoluble members, while those greater than
CO-530 are oil-insoluble, water-soluble. The number following the CO- designation is approximately ten
times the ethylene-oxide content of the surfactant, in weight percent. The Igepal CO series is well-suited to
research and process development because of its well-defined progression in reagent composition and
hydrophile-lipophile balance (HLB). An equivalent series of reagents is offered by Union  Carbide under the
trade name Triton N.

Again,  reagent formulation and slurry density were found to be significant parameters affecting oil and grease
levels in the tailings, but, as might be expected of a nonionic surfactant, pH had no significant effect (with
greater than 85 percent confidence). Higher agitation intensity again had the effect of producing excessive
weight losses to the concentrate.  The system is also sensitive to conditioning time. There is a 11.6 percent
probability that the observed  responses are the results of random effects rather than the variables indicated.
The oil and grease analyses  in this test series clearly contain a systematic error, as mass balances for the
tests show about twice the oil and grease accounted for compared to the analysis of the feed.  The source of
the error is not known, and these  data should be used only for comparative purposes within this data set.

The variables affecting weight distribution to the concentrate are reagent composition and reagent
concentration (with greater than 98 percent confidence).  The statistical analysis also indicates agitation
speed  is significant although this effect may also be attributed to the interaction of the two variables already
mentioned. There is a 0.2 percent probability that the observed responses are the results of random effects
rather than the variables indicated.
Buffalo River
                                        Gram Size Separation
Table A2, part b, for the Buffalo River sample shows that most of the contaminants reside in the minus
400 mesh (less than 38 ^m) fraction. The coarsest fractions have high metal analyses (Table A2-a), but this
represents only a small amount of the contamination, since only a small fraction of the weight is in these
sizes.  The coarse fractions contain considerable amounts of wood chips and other organic matter, and it is
possible that contaminant ions have adsorbed on the organic fibers, accounting for the high analyses shown

                                                 26

-------
in Table A2-a.  Table A2-d shows the analysis that would be obtained in a coarse sand product if a "perfect"
size separation were performed. The table shows that the lead levels, for example, could be cut roughly in
half by such an operation making a size split at 400 mesh (38 ^m), but only at the expense of retaining
63.3 percent of the sediment weight with the contaminated fraction.

As with the Ashtabula River sample, Figure 13 shows the distribution of selected contaminants versus mass
distribution.  The figure shows only a slight deviation upward from the diagonal, indicating that these
                       100
                     £  60  -
                     I
                     .8
                     Q
                        20  -
                                    20
                                              40         60
                                          Mass Distribution, pet finer
                                                                           100
              Figure 13.—Partitioning of Contaminants by Particle Size-Buffalo River
              Sediment

contaminants cannot be effectively concentrated by particle size separation.

                                     Gravity (Density) Separations

The gravity separation results in Table A3 show that for each size fraction where a separation was possible,
both the light (S.G.<1.9) and heavy (S.G.>2.9) products contained  higher levels of contaminants than the
middling product (1.9
-------
the sediment's contaminants.  Removal of the light and heavy products leaves 53 percent of the overall
sediment mass but on the average only 24 percent of the contaminants.
Tabte 9.— Example of Gravity Separation Effect: Buffalo River
Before Gravity Separation

Element
Mass
As
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Zn
Analysis,
mg/kg
-
6.9
1.1
51
54
21,100
39
0.2
28
98
Distribution2,
pet
56.6
40.8
28.4
42.1
39.3
39.1
26.6
28.4
40.3
27.4
Sediment, +400 mesh
After Gravity Separation1
Analysis,
mg/kg
-
4.6
0.9
39
38
15,100
22
0.1
22
71
Distribution,
pet
53.3
26.0
23.1
30.8
26.2
26.3
14.3
19.6
30.5
18.7
Difference in
analysis, mg/kg
NA
-2.3
-0.2
-12.0
-16.0
-6,000.0
-17.0
-0.1
-6.0
-27.0
1 Gravity separation on +400 mesh fraction only. Gravity separations made at 1 9 and 2.9 g/cm3. the material reported in these
columns is that found between these two densities.
2 Distribution of constituent based on the complete sediment, i.e. before gravity separation, the +400 mesh portion of the sediment
amounted to 56 6 percent of the total sediment mass, and contained 40.8 percent of the total amount of arsenic in the sediment.
                                         Magnetic Separation

As-received (i.e., not deoiled) sediment was used and separated at 200 and 400 mesh by wet screening. A
further size split was made in the subsieve (less than 38 /^m) range using the glass hydrocyclone.  This split
was at about 12 ,um. The plus 200 mesh material was dried and passed through the Carpco-induced roll
magnetic separator at two settings. The 200*400 mesh, 400 mesh * 12 jum, and -12 /^m fractions were
treated on the wet high-intensity magnetic separator (WHIMS).  Low intensity or hand magnet fractions were
not taken on this sample because preliminary magnetic fractionations showed less than 0.5 percent of the
material is recovered at low intensity.

Results of the magnetic fractionation are shown in Table A4.  Chemical analyses of the various magnetic
fractions are shown in part a. of the table, and the corresponding element distributions, calculated against the
total sediment weight, are shown in part b. of the table.

By comparing the analysis of the non-magnetic product with that of the total, Table A4 shows that the
contamination levels of all but the finest fraction can be reduced by magnetic separation   For example, the
plus 200 mesh fraction shows significant reductions in the levels of arsenic, chromium, copper, lead, nickel,
and zinc. Performance on the 200X400 mesh fraction is similar, although from lower starting  levels. Some
progress can even be made  on the 400 mesh * 12  ^m fraction. Although some classification did occur on
WHIMS treatment of the finest fraction (minus 12 /^m), the non-magnetic product in this category still shows
high levels of many contaminants.
                                                 28

-------
                                 Froth Flotation-Metallic Contaminants

The results of flotation testing on Buffalo River sediment are shown in Table A5. Flotation of Buffalo River
sediment without collector or with xanthate collector produced concentrates of 4-15 percent of the sediment
mass, with some enrichment of metal contaminants. The material  reporting to the  concentrate was the
extremely fine-grained sediment and was likely entrained in the froth  rather than actually attached to bubbles.
It seems apparent, then, that the enrichment observed is the natural enrichment of the fine particles
previously noted, and not the result of a selective flotation separation based on mineralogy.

Flotation with oleic acid appears to be slightly selective, with the test  at pH 7 recovering 48-70 percent of the
contamination in about 32 percent of the sediment. This concentrate also contained mostly fine material, but
it appears from a comparison to the size chemistry data that some  (albeit small) degree of selective mineral
separation has occurred. This was  not observed,  however, in the test at pH 10. This test did not sustain a
stable froth and recovered only a small amount of material.

                                          Attrition Scrubbing

Attrition scrubbing with any of the acids did not significantly reduce the metals content of the sediment. The
analysis of the sludge  shows that some solubilization of arsenic, copper, iron, nickel, and zinc occurred, but
this is a very small amount of material. The final filtrate contained less than 1 mg/kg Fe and much lower
levels (below detection) of the other metals.  These results are provided in Table A6.  It seems likely that the
fine size distribution is the reason for the poor attrition performance-fine particles simply cannot attain
sufficient momentum to collide forcefully with other particles and produce the abrading effect required.
Equipment capable of higher energy input to the pulp may be more effective.  It is also possible that much of
the contamination has adsorbed on the surfaces of internal pores of the material, so that superficial abrasion
will not reach these bound contaminant ions.

                                 Froth Flotation-Organic Contaminants

When evaluating the Buffalo River sample, only the alcohol and amine ethoxylate (Igepal CO and Triton RW)
surfactants were used. The fine grain-size distribution of this material made selection of flotation operating
parameters a more critical concern than reagent variety, so that the number of surfactants used was limited to
allow investigation of other parameters.  Since these reagents are mostly nonionic in character, pH was
eliminated as a parameter in the Buffalo River investigation. When using the Igepal reagent, tests were
conducted at the natural pH (i.e., without pH adjustment) of the sediment;  when using the Triton RW
reagents, the test was conditioned at pH 10, then acidified to slightly  less than 7 for flotation. The Triton RW's
are derived from ethylene oxide, although the use of the alkyl amine group gives the surfactant some cationic
character. Triton RW will form a stable oil/water emulsion at pH  greater than 10; reducing the pH to 7 or
below breaks the emulsion and releases the oil. If recovery of the surfactant is desired, this  can be
accomplished by again raising the pH to 10 or greater. The code numbers associated with the reagent
designation indicate ten times the number of moles of ethylene oxide per mole of surfactant  (e.g., Triton RW-
150 contains 15 moles EtO/mole surfactant.)  Except as noted, all tests were conducted at about 10 percent
solids using tap water for dilution.

Amine ethoxvlate surfactants Flotation of Buffalo River sediment had limited success when  Triton RW
surfactants were used. The results are provided in Table 10. The  tailings were cleaned from an original head
assay of 5,910 mg/kg oil and grease to less than 3,000 mg/kg. One test lowered the assay of the tailings to
slightly less than 1,000 mg/kg oil and grease. In obtaining lower concentrations of oil and grease in the
tailings, as much as 76 percent of the material was floated. Coefficients of separation varied from 19.5 to
48.6 percent.
                                                 29

-------
   Table 10.—Flotation of oil and grease from Buffalo River sediment using Amine Ethoxylate  surfactants
                              Wt. Dist.     Oil and Grease            Oil & Grease
  Triton                          to        Analysis, mg/kg	Distribution, pet         coeff. of
   RW-     Triton      Air       Cone.,                                                    Separation
  Type     Cone.    (Umin)       pet       Cone.      Tails       Cone.        Tails           pet
50
50
50
50
100
100
100
150
150
150
150
0.01
0.01
0.05
0.05
0.03
0.03
0.03
0.01
0.01
0.05
0.05
1.00
3.50
1.00
3.50
2.25
2.25
2.25
1.00
3.50
1.00
3.50
60.4
63.9
76.0
74.7
69.7
70.7
66.6
12.4
27.4
58.9
69.4
9,600
11,300
9,240
9,380
8,960
8,970
9,450
31,900
17,900
11,500
8,260
1,830
1,600
1,390
904
1,580
1,370
1,180
2,890
2,760
1,580
1,530
88.9
92.6
95.5
96.8
92.9
94.1
94.1
61.0
71.0
91.3
92.5
11.1
7.4
4.5
3.2
7.1
6.0
5.9
39.0
29.0
8.7
7.5
28.5
28.7
19.5
22.1
23.2
23.4
27.5
48.6
43.6
32.4
23.1

The low coefficients of separation are the result of an inverse relationship between weight distribution to the
concentrate and oil and grease analysis in the tailings.  As the system is pushea to float more contaminants, it
carries more sediment with it. This is illustrated in Figure 14, which shows sediment mass in the concentrate
and oil and grease levels remaining in the tailings located against axes of the three parameters investigated.
As the Triton RW surfactant type was changed from 50 to 150 (effectively tripling the EtO content, vertical
axis), the weight distribution to the concentrate significantly decreased while the amount of oil and grease in
the tailings increased. A similar effect was  seen when the Triton RW concentration was reduced from .05
percent to .01 percent in solution (first horizontal axis).  Once again the weight percent in  the concentrate
decreased but the oil and grease in the tailings increased, a typical effect of amine surfactants.  The cube plot
is frequently used to illustrate variable effects in a three-parameter system.

Two results were discovered when the air rate was reduced from 3.5 to  1.0 liters per minute (second
horizontal axis). When Triton RW-150 was used (15 moles EtO), the weight of the concentrate decreased
while the oil and grease  in the tailings remained basically unchanged. Conversely, when Triton RW-50 was
used  (5 moles EtO), the  oil and grease quantity increased and the weight percent was unchanged.  It is
expected that this result  is due to the stronger foaming  properties of the RW-150 surfactant, resulting in
decreased selectivity in this application.

This test series is a 23 factorial design with  replicated centerpoint (Box, Hunter, and Hunter, 1978).
Regression analysis showed Triton concentration and composition to be the significant variables affecting oil
and grease levels in the  tailings (with greater than 85 percent confidence), and identified a first-order
interaction between these terms. The effect of aeration rate on oil and grease levels in the tailings was
insignificant.  Analysis of variance gives a probability of 0.5 percent that the measured responses are the
results of random effects rather than the variables indicated.  Regression analysis for the weight distribution
to the concentrate identified all three variables as significant (with greater than 90 percent confidence), along
with a first order interaction between reagent type and reagent concentration.  Analysis of variance showed
that there is a probability of 0.03 percent that the observed responses are the results of random effects rather
than the variables indicated.
                                                  30

-------
                                         27.4
                       69.4
                                        '2,760
            15 moles
            Reagent's
             ethylene
              oxide
             content
                          72.4
            5 moles
                        2,890
                       1,530
         58.9
     \   1,580

     69.0
     1,380
i  63.9
'1,600
                         -60.4
        76.0
       KEY

      Weight
   Distribution to
    Concentrate
       (pet)

       Oil&
     Grease in
      tailings
      (ppm)
35L/m
                                                                  Aeration
                                                                    rate
                 1 OL/m
                        1,830
                   .01 pet          Reagent
                               concentration
       1,390
            05 pet
          Figure 14.—Effects of reagent composition, aeration rate, and reagent concentration on
          flotation of oil and grease from Buffalo River sediment.
Determining which combination of parameters is best depends on the objective of the pretreatment process.
If the remaining sediment needs to be as clean as possible regardless of the amount of volume reduction,
then the content of ethylene oxide in the surfactant should be minimized while the concentration of the
surfactant and the air rate should be maximized within the range studied. If a higher concentration of oil and
grease is acceptable, but maximum volume reduction is required, then the content of ethylene oxide  in the
surfactant should be maximized and the surfactant concentration and air rate should be minimized, again
within the range studied.

A likely approach is to operate somewhere near the center of the cube.  In our laboratory testing, the midpoint
of each of the three parameters would lower the oil and grease in the tailings to about 1,200 mg/kg while
leaving the concentrate with 67 percent of the original amount of material to be further treated

Ethoxvlated alcohol surfactants  In using  Igepal surfactants, selection of appropriate operating parameters
was again deemed to be more important  than testing a  broad range of reagent formulations  Consequently,
the oil flotation work on Buffalo River sediment using Igepal CO surfactants investigated aeration rate and
reagent concentration, but considered only a single surfactant formulation, Igepal  CO-530 at 54  percent
ethylene  oxide  CO-530 is a middle-of-the-road choice among surfactants in this class, as it is on  the
borderline between oil and water solubility. (The numbers following CO- in the surfactant designation are
approximately ten times the ethylene oxide content of the surfactant  in weight percent.) Solids content for
these experiments was held constant at about 10 percent solids. Higher solids loading of fine-grained
sediments with these surfactants produces a voluminous, stiff froth which entrains almost all of the fine
particles  regardless of hydrophobicity.  Results of experiments with the nonionic Igepal  surfactant  are given in
Table 11.
                                                  31

-------
            100
             80
             60
 Coefficient
     of
 Separation,
     pet
             40
             20
                                                                                    LEGEND

                                                                                     Air Rate

                                                                                    3.5L/min
                                                                                     1.0 L/min
                            0.02          0.04          0.06          0.08
                                        Igepal CO-530 Concentration, pet
0.1
            0.1
            2
Figure 15.—Effect of aeration rate and Igepal CO-530 concentration on effectiveness of oil and grease
separation from Buffalo River sediment.


The tests with Igepal CO-530 show that as the surfactant concentration is reduced from 0.1 to 0.03 percent
the oil and grease distribution to the concentrate remains relatively unchanged.  This is illustrated in Figure
15. Decreasing the concentration below 0.03 percent will affect the oil and grease distribution, however.
Reduction of surfactant concentration from 0.1 to 0.03 percent also causes a significant decrease in the
weight distribution to the concentrate. Lowering the weight in the concentrate while maintaining the same
percent of oil and grease yields an increased coefficient of separation as the surfactant concentration
approaches 0.03 percent.

As the air rate was changed to 3.5 or 1.0 L/min similar trends were observed. The weight distribution to the
concentrate was again decreased and the coefficient of separation accordingly improved. Table 11  uses the
coefficient of separation (defined above) to show that the combination of air rate and concentration of Igepal
CO-530 that will  produce a maximum coefficient of separation 1.0 L/min and 0.02 percent. The laboratory
results in the table  indicate that this combination produced an oil and grease level in the tailings of 2,200
mg/kg and limited the weight of the concentrate to 13.5 percent.
                                                 32

-------
Table 1 1 .—Effects of Aeration Rate and Reagent Concentration on Flotation of Buffalo River Sediment with
Ethoxylated Alcohol Surfactant
Aeration
Rate
(L/min)
3.50
3.50
3.50
3.50
3.50
1.00
1.00
1.00
1.00
0.50
0.50
0.50
0.50
Reagent
Cone.
wt. pet
0.10
0.05
0.03
0.02
0.01
0.10
0.05
0.03
0.02
0.10
0.05
0.03
0.02
Wt. dist.
to cone.,
pet
75.4
59.1
45.3
16.2
11.3
45.2
35.8
18.3
13.5
35.8
28.2
18.5
9.6
Oil & Grease
Analysis, mg/kg
Cone.
8,760
10,100
15,100
29,800
37,300
22,500
18,300
29,600
44,100
35,500
20,800
25,500
41,700
Tails
541
729
518
2,180
2,660
2,870
1,930
1,950
2,200
3,160
1,940
2,100
2,520
Oil & Grease Coeff. of
Distribution, pet Separation,
Cone.
98.0
95.2
96.0
72.6
64.2
86.6
84.1
77.2
75.8
86.3
80.9
73.4
63.8
Tails
2.0
4.8
4.0
27.4
35.8
13.4
15.9
22.8
24.2
13.7
19.1
26.6
36.2
pet
22.7
36.1
50.7
56.4
52.9
41.4
48.3
59.0
62.3
50.4
52.7
54.9
54.2

Lower levels of oil and grease removal from the tailings or lower weight recovery in the concentrate can be
achieved, but not without compromises.  The oil and grease in the tailings can be lowered to near 500 mg/kg
by increasing the air rate to 3.5 L/min. This will result in an increase in the weight lost to the concentrate to
45.3 percent. Alternatively, the weight distribution of the concentrate can be reduced to less than 10 percent
by reducing the air rate to 0.5 L/min.  This would cause the oil and grease in the tailings to increase to about
2,500 mg/kg.


Indiana Harbor Ship Canal and Grand Calumet River

                                        Grain Size Separation

The size chemistry data with the Indiana Harbor Ship  Canal sample, shown in Table A7, are similar to the
Buffalo River sample, except that the coarse fractions are not as heavily polluted as the Buffalo River
sediment.  The coarsest fractions, which represent only a small amount of the total sediment weight, are
relatively low in pollutants. Microscopic examination of these coarse fractions showed them to be composed
of mostly glasslike slag particles and droplets.  This slag  material is probably quite inert to the extraction
procedures used in determining element compositions for the size chemistry.  The sediment becomes more
concentrated in all pollutants as the particle size decreases.  As shown in Table A7, part c, about 60 percent
of the sediment weight and 73 to 86 percent of the metals analyzed are finer than 400 mesh (38 /^m). Figure
16 indicates a minor distribution of contaminants toward the fine fraction, so that about 80 percent of the
                                                 33

-------
                   100
                                            40          60
                                       Mass Distribution, pet finer
80
100
             Figure 16.—Partitioning of Contaminants by Particle Size-Indiana Harbor Canal
             Sediment


contaminants plotted in the figure are found in the finest 60 percent of the sediment. This is probably not
sufficient partitioning to justify particle size separation pretreatment for separating metallic contaminants from
the sediment.

                                    Gravity (Density) Separations

The results shown in Table A8 for the Indiana Harbor sediment show that each working size fraction
responded differently to gravity separation.  The plus 100 mesh (150 urn) material contained a large portion of
light non-polluted material which appeared to be slag and other man-made materials. This size fraction also
contained a small amount of heavy material which contained high levels of each contaminant. The heavy
fraction from the 100x400 mesh material contained the most heavy metal contaminants and the light fraction
was the least polluted. Table 12 shows the expected analyses of the 100x400 mesh fraction both before and
after removal of the heavy and light material (the light fraction was removed because it would likely contain
high levels of organic contamination). Again considering the minus 400 mesh material to be unseparable,  the
unseparated 100x400 mesh  material contains 26 percent of the sediment mass and on the average 17
percent of the sediment's contaminants.  Removal of the light and heavy products leaves 15 percent of the
sediment mass and 8 percent of the contaminants. The removal of these products also lowered the
contaminant analyses by an  average  of *\2 percent.
                                                34

-------
        Table 12.—Example of Gravity Separation Effect:  Indiana Harbor Sediment, 100 x 400 mesh.
Before Gravity Separation
Element
Mass
As
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Zn
Analysis, Distribution2,
mg/kg pet
-
19
3.2
265
187
108,000
425
1.2
71
1470
26.4
18.0
11.0
15.2
16.7
23.1
14.5
17.4
22.4
11.7
After Gravity Separation1
Analysis, Distribution2, Difference in
mg/kg pet analysis, mg/kg
-
17
3.0
240
156
61,600
380
1.3
52
1480
15.5
9.7
6.1
8.1
8.1
7.7
7.6
11.1
9.7
6.9
NA
-2.0
-0.2
-25.0
-31.0
-46,400.0
-45.0
0.1
-19.0
10.0
1 Gravity separation on 100x400 mesh fraction only Gravity separations made at 1 9 and 2 9 g/cm3, the material reported in these
columns is that found between these two densities
2 Distribution of constituent based on the complete sediment; i e before gravity separation, the 100x400 mesh portion of the sediment
amounted to 26 4 percent of the total sediment mass, and contained 18.0 percent of the total amount of arsenic in the sediment
                                         Magnetic Separations

As-received (i.e., not deoiled) Indiana Harbor sediment was used and separated at 100 and 400 mesh by wet
screening. A further size split was made in the subsieve range (i.e. finer than the smallest sieve used, 400
mesh or 37 urn) using the glass hydrocyclone.  This split was at about 12 /^m. Because of the high levels of
magnetic ferrous material in the sample, all four size fractions (plus 100 mesh, 100X400 mesh, 400 mesh *
12 /^m, and -12 ,um) were first run through the Davis tube separator. Following recovery of a strongly
magnetic product in the Davis tube, the plus 100 mesh material was passed through the dry high-intensity
induced roll separator at 0.25 and 2.5 A magnet current settings. (In magnetic separation, magnetic field
intensity is directly proportional to magnet current.)  The 100X400 mesh, 400 mesh x 12 urn, and -12 ^m
fractions were treated by wet high-intensity magnetic separation following the Davis tube separation. Magnet
current settings were 0.5 A and then 5 A, with a magnetic product recovered  each time. In  all size fractions,
material that passed through the Davis tube and the high-intensity magnetic separators without being
collected in a magnetic product was designated "non-magnetic."

Results of the magnetic fractionation are shown in Table A9.  Chemical analyses of the various magnetic
fractions are shown in part a. of the table, and the corresponding element distributions, calculated against the
total sediment weight, are shown in part b. of the table.

Although patterns of classification of pollutant metals can be observed in magnetic separation of the Indiana
Harbor size fractions, very little progress is made with respect to the pollutional classification of the sediment.
The Davis tube product was considerably concentrated in iron, ranging from 28 percent iron in the  plus
100 mesh material up to 43.6 percent iron in the -12 /^m material.  It was expected that the  Davis tube
magnetic product would consist mostly of iron oxides free of heavy metal  contamination.  While in some size
fractions the Davis tube product is less heavily  contaminated than the next fraction, this is not uniformly the
case, and the levels of pollution in the  Davis tube product are still quite high.  In addition, while the levels of
iron in the Davis tube product are quite high, they are not near the levels that would be expected of a
relatively pure iron oxide that might be used as a component of steel mill blast furnace feed.
                                                  35

-------
                                 Froth Flotation-Metallic Contaminants

The flotation tests on Indiana Harbor sediment, shown in Table A10, were difficult to conduct and control
owing to the remarkable frothing capabilities of this sediment, especially at an acidic pH. It is likely that this is
the effect of an unidentified nonionic surfactant thought to be present in the Indiana Harbor sediment. Most of
the tests produced voluminous, persistent froth with high mass loading, making selective separation almost
impossible. All of the tests floated large amounts of the sediment, and selectivity (as judged by the
differences between distribution of weight to the concentrate and distributions of contaminants to the
concentrate) was very low.

                                          Attrition Scrubbing

The attrition scrubbing results on Indiana Harbor sediment are similar to those for Buffalo River. Attrition
scrubbing with any of the acids did not appreciably reduce the metals content of the sediment. Besides iron,
small amounts of chromium, copper, nickel, and zinc were found in the sludge formed by neutralization of the
attrition scrubbing liquid, but the effect on the sediment was minor. Again, the final filtrates contained
extremely low levels of heavy metals.  These results are presented in Table A11.

                                 Froth Flotation-Organic Contaminants

Two sediment samples were used in organic contaminant flotation studies from the Indiana Harbor
Canal/Grand Calumet River area of concern. The first, mentioned here as the Indiana Harbor sample, is the
same as the sample identified earlier taken from the Indiana Harbor Ship Canal near Columbus Drive. A
second, more contaminated sample, was obtained and used later in the  work. This sample was taken from
the upstream end of the East Branch of the Grand Calumet River. This sample was analyzed by the Bureau's
contractor at about 33,000 mg/kg oil and grease, although  losses from soxhlet extraction suggest that this
number may be as high as 140,000 mg/kg. Total PCB  content was about 150 mg/kg.

Indiana Harbor Canal-Amine ethoxvlate surfactants  Table 13 provides the results of testing Triton RW-
surfactants on the Indiana Harbor Canal sample. As with the Buffalo River sediment, the fineness of this
sediment contributed to excessive foaming (especially with the amine ethoxylate Triton RWs). The result of
this excessive foaming, as noted above, is entrainment of most of the solids in the froth.  This is manifest in
the high weight distributions to the concentrate found in Table 13. What can be concluded from this table
Table
Triton
Reagent
Type
RW-150
RW-150
RW-150
RW-50
RW-50
RW-50
X-114
X-114
X-114
13.— Flotation of
Triton
Cone., pet "
0.10
0.05
0.01
0.10
0.05
0.01
0.10
0.05
0.01
oil and grease
Wt. dist.
to cone.,
pet
71.1
51.3
42.8
59.3
59.3
50.5
61.7
73.1
49.0
from Indiana Harbor Canal sediment using Surfactants
Oil and Grease
Analysis, mg/kg
Cone. Tails
81,900 7,300
110,000 7,200
105,000 22,900
104,000 13,500
87,400 10,700
99,700 16,800
102,000 14,900
90,400 7,750
98,500 18,400
Oil & Grease
Distribution, pet
Cone. Tails
99.6 3.9
94.2 5.8
77.4 22.6
91.8 82
92.3 7.7
85.8 14.2
91.7 8.3
96.9 3.1
83.7 16.3
Coeff. of
Separation
pet
28.5
42.9
346
32.5
33.0
35.3
30.0
23.8
34.7
                                                 36

-------
is that an increase in reagent concentration will increase weight loss to the concentrate.  It also seems that
the RW-150 surfactant is more effective at cleaning sediment than the others tested.  For these tests, Triton
RW-150 produced the best results at 0.05 percent; the lower concentration did not clean as effectively, and
the higher concentration only served to divert more of the sediment mass to the frothy concentrate.

Ethoxvlated alcohol surfactants  It was observed early in the test program that the Igepal surfactants
produced a solids-laden, stiff froth that defied control.  These reagents were deemed to be too foamy for the
Indiana Harbor and  Grand Calumet River material, and results from such tests are not reported here. Results
from tests on the Indiana Harbor Canal sediment with the similar Triton X-114 surfactant  are presented with
the Triton RW- data in Table 13.

Grand  Calumet River-Amine ethoxylate surfactants  Investigation of the Grand Calumet  River material
followed a pattern similar to that used on the Buffalo River sample with Triton surfactants; a 23 factorial
experiment design was employed to evaluate the effects of reagent formulation, reagent  concentration, and
solids content. The results of these tests are presented in Table 14. In addition,  a cube  plot is provided  in
Figure  17.  Regression analysis shows that the oil and grease content of the tailings is influenced by reagent
composition and slurry density (percent solids) with greater than 98 percent confidence.  There is a
0.2 percent probability that the observed responses are the results of random effects. For weight distribution,
these same two variables and their first-order interaction are significant with greater than 99 percent
confidence. There is a 0.01 percent probability that the observed responses are the results of random
effects.

      Table 14.—Flotation of oil and grease from Grand Calumet River sediment using amine ethoxylate
                                             surfactants
Triton
RW-
code
50
50
50
50
100
100
150
150
150
150
reagent
cone.,
pet
0.1
0.1
0.3
0.3
0.2
0.2
0.1
0.1
0.3
0.3
Solids
content,
pet
10
30
10
30
20
20
10
30
10
30
wt. dist.
to cone.,
pet
58.4
75.3
58.5
74.0
71.8
75.2
67.1
76.1
69.2
76.3
Oil and grease
mg/kg
Cone.
39,200
45,000
47,000
53,100
35,800
40,600
44,300
60,300
45,200
43,000
analysis,
Tails
13,000
5,670
17,300
9,550
3,260
1,750
6,270
2,950
6,270
1,580
Oil and
dist.
Cone.
80.9
96.0
79.3
94.1
95.6
98.6
93.5
98.5
94.2
98.9
grease
, pet
Tails
19.1
4.0
20.7
5.9
3.5
1.4
6.5
1.5
5.8
1.1
Coeff.
- ofsep'n,
pet
22.5
20.7
20.8
20.1
23.8
23.4
26.4
22.4
25.0
22.6

Examination of the cube plot shows the interaction that exists between reagent formulation and percent
solids. A higher ethylene oxide formulation results in a higher weight loss to the concentrate at 10 percent
solids, but has no effect at 30 percent solids.  Higher solids content results in cleaner tailings at all conditions
tested. Overall, these results are discouraging in that the plot shows no area where an oil and grease content
of lower than 3,000 mg/kg can be obtained while achieving a weight distribution to the concentrate of less
than 70 percent.  It seems likely that, owing to the very high oil and grease content of this sediment, it is
impossible to add sufficient surfactant to liberate all contaminant without creating a flotation situation so
foamy that almost all of the solids report to the frothy concentrate.  A more effective approach might be to
float the sediment in two or more stages at lower surfactant concentrations, or employ a treatment that

                                                 37

-------
minimizes bubbling.
                                          76.1
                                                                76.3
            15 moles
            Reagent's
             ethylene
               oxide
             content
             5 moles
       KEY

      Weight
   Distribution to
    Concentrate
       (pet)

       Oil&
     Grease in
      tailings
       (ppm)
30 pet
                                                                   Solids
                                                                   content
                          13,000                  17,300
                    10 pet         Reagent          .30 pet
                               concentration
                                                         10 pet
          Figure 17.—Effects of reagent composition, reagent concentration, and solids content on
          flotation of oil and grease from Grand Calumet River sediment.
                               Size Classification-Organic Contaminants

As with the Saginaw River samples to be discussed later, the Grand Calumet River material was investigated
for distribution of organic contaminants throughout the range of particle sizes. These data are provided in
Table 15.  Unlike the Saginaw material discussed below, the Grand Calumet sample showed no clear
Table
15. — Distribution of PCBs and oil and grease by particle size in Grand Calumet River sediment
Oil and grease
Size
fraction,
urn
+425
425x75
75x20
20x10
-10
Total
1 Calculated
Wt.
dist.,
pet
11.4
37.6
34.3
7.8
8.9
100.0
Cum.
wt dist.,
pet1
886
51 0
16.7
8.9
0
NA
Analysis,
mg/kg
60,200
55,500
22,500
19,900
17,500
33,000
mass or analysis of this fraction combined with all
Dist.,
pet
17.8
54.1
20.0
4.0
4.1
100.0
Cum.
analysis,
mg/kg1
60,200
56,600
42,600
40,600
38,600

Total PCBs
Analysis,
mg/kg
143
149
150
200
169
152
Dist.,
pet
10.6
36.4
33.4
10.1
9.8
100.3
Cum.
analysis,
mg/kg1
143
148
149
153
154

coarser fractions
                                                 38

-------
correlation between grain size and PCB contamination. Moreover, the oil and grease content of this sample
showed an opposite trend to what is commonly expected--the coarser material has significantly higher
levelsof oil and grease.
Saqinaw River
                                        Grain Size Separation
Table A12 describes the size chemistry of the Saginaw River #2, Series TRP-6 sample.  This Saginaw
sample is different from the Indiana and Buffalo samples because it has a much coarser size distribution, with
only 12.4 percent of the weight finer than 400 mesh (38 /urn). Like the Buffalo River sample, the plus 35 mesh
size fraction contains wood chips and metal analyses are high, but this is only 0.2 percent of the weight and
makes an insignificant contribution to the net metal content of the sample.  Part b of the table shows that 48
to 68 percent of the metals (except for arsenic and iron) are concentrated into the minus 400 mesh (37 ^m)
fraction, which represents only 12 percent of the sediment weight.  Size separation to produce a clean sand
product in this case becomes more plausible. Table A12-c shows that if a perfect separation were made at a
size coarser than 400 mesh (37 urn), such as 150 rnesh (106 ^m),  69.3 percent of the cadmium, 64.4 percent
of the chromium, and 79.5 percent of the lead would be removed in 23.7 percent of the sediment weight.
Such an operation would reduce metal concentrations from 4.3 to 1.7 mg/kg cadmium, from 155 to 72 mg/kg
chromium, and from 48 to 13 mg/kg lead (see Table A12-d.)  Arsenic appears to be anomalous in its
distribution, with about 52 percent of the arsenic occurring in the 35 to 100  mesh size range.  About
67 percent of the sample weight occurs in this size range. Figure 18  shows the distribution of selected
contaminants  by particle size. The chart shows that cadmium, chromium, and lead are strongly distributed
towards the fines, so that 60-80 percent of these contaminants can be removed with the finest 20 percent of
the material. As will be discussed later, PCB contamination also follows this pattern of association with the
                   I
                   .s
                      40 -
                      20 -
                                    20
             ~T
   40         60
Mass Distribution, pet finer
                                                                    80
              Figure 18.—Partitioning of Contaminants by Grain Size Separation on Saginaw
              River Sediment #2.
                                                39

-------
finest particles in the sediment.

The Saginaw River #1 sample has a slightly finer size distribution, but the element distribution trends are
essentially the same as in the sample just described.  The size chemistry data for Saginaw River #1 are
presented in Table A13.

                           Gravity (Density) Separations-Metallic Contaminants

The results shown in Table A14 for the Saginaw River #1 sediment show that for each working size fraction
the light and heavy products contained higher contaminant levels than the middling product, but the mass
associated with these products was small when compared to the overall mass of the sediment.  The
combined results for the plus 65 and 65x200 mesh products are shown in Table 16. For these size fractions
removal of the light and heavy products removes less than 1 percent of the overall sediment weight and it


         Table 16.—Example of Gravity Separation Effect: Saginaw River Sediment #1, +200 mesh.
Before Separation
Element
Mass
As
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Zn
Analysis, Distribution2,
mg/kg pet
-
3.3
0.6
49
16
6200
16
0.2
10.7
58
68.1
45.9
40.6
59.9
33.0
34.9
29.0
70.7
37.0
16.3
After Separation1
Analysis, Distribution2, Difference in
mg/kg pet analysis, mg/kg
-
3.1
0.6
49
13
5900
11
0.2
98
51
67.3
42.4
37.1
58.3
25.1
32.7
19.0
68.2
33.2
14.0
NA
-0.2
0.0
0.0
-3.0
-300.0
-5.0
0.0
-0.9
-7.0
 1 Gravity separation on +200 mesh fraction only.  Gravity separations made at 1.9 and 2.9 g/cm3; the material reported in these columns
 is that found between these two densities.
 2 Distribution of constituent based on the complete sediment; i.e. before gravity separation, the +200 mesh portion of the sediment
 amounted to 68.1 percent of the total sediment mass, and contained 45.9 percent of the total amount of arsenic in the sediment.	

 removes roughly 4 percent of the metallic contaminants. Removal of these products lowered the contaminant
 levels in the remainder by an average of 11  percent.

 Additional gravity separation tests on the Saginaw River sediments showed that the light material, along with
 its associated contaminants, can be removed from the coarse sediment by water elutriation of closely sized
 material. In water elutriation,  closely-sized material is allowed to settle against a rising current of water.
 Denser particles, with a  terminal settling velocity greater than the rising current, will settle to the bottom of the
 device, while lighter particles will be carried  by the current over the top.  As shown in Table A15, this process
 removes 1.5 percent of the  overall sediment as light material which contains roughly 5 percent of the
 sediment's contaminants.

                                         Magnetic Separations

 As-received (i.e., not deoiled) sediment was used and separated at 65, 200, and 500 mesh by wet screening.
 The plus 65 mesh and 65><200 mesh material were each dried and run through the Carpco induced roil
 magnetic separator at two magnet current settings, 0.5 A and 2 A. The 200X500 mesh and minus 500 mesh

                                                 40

-------
fractions were treated on the wet high-intensity magnetic separator (WHIMS), also at 2 magnet current
settings, 1 A and 5 A. Low intensity or hand magnet fractions were not taken on this sample because
preliminary magnetic fractionations showed less than 0.5 percent of the material is recovered at low intensity.

Results of the magnetic fractionation are shown in Table A16.  Chemical analyses of the various magnetic
fractions are shown  in part a. of the table, and the corresponding element distributions, calculated against the
total sediment weight, are shown in part b. of the table.

Again, very little progress can be made by application of magnetic separation to sized Saginaw River #1
sediment. As seen with the Buffalo River sediment, the most magnetic products are enriched in
contaminants, while the least magnetic products are slightly reduced  in contamination. While this suggests
that a small degree of separation occurred, the magnetic fractions collected are only small portions of the total
amount of sediment, so that most of the metallic contaminants remain with the  nonmagnetic fraction at a
slightly reduced analysis. The degree of separation decreases in the finer size fractions.

                                 Froth  Flotation-Metallic Contaminants

The results of flotation testing on Saginaw River sediment #1 are shown in Table A17. The pattern is similar
to that seen with the Buffalo River  sediment; flotation without collector and with xanthate produced small
concentrates with some enrichment of metal contaminants.  As before, the material reporting to the
concentrate was the extremely fine-grained sediment and may have been entrained in the froth rather than
actually attached to  bubbles.  The  lower mass distribution to the  concentrate, compared to Buffalo River
sediment, is probably the result of  the lower concentration of fine particles  in the Saginaw sample.

The greatest enrichment of contaminants was observed at pH 7  using oleic acid. In this case, 31-73 percent
of the metallic contaminants can be concentrated in 18 percent of the original sediment mass. This result is
similar to that expected from the hypothetical 150 mesh size separation discussed above.

                                          Attrition Scrubbing

The Saginaw River #1 sediment contains low initial levels of heavy metal contaminants, and very little
reduction in these levels was achieved through attrition  scrubbing with acids. Small amounts of arsenic,
nickel, and zinc were observed in the sludge obtained by neutralizing the scrubbing liquids, and the final
filtrate contained very low levels of all metals. These results are shown in  Table A18.  The poor results from
scrubbing of  Saginaw River sediment are unexpected in light of the surface area-contaminant correlations
discussed earlier.

                                 Froth Flotation-Organic Contaminants

Three series of tests were conducted on the Saginaw River #2 sediment.  In the first, anionic surfactants were
evaluated at  fixed concentration under acidic, natural, and basic conditions.  Second, the cationin/nonionic
Triton RW surfactants were evaluated.  In the third series, the nonionic surfactant series Igepal CO was
evaluated.

Anionic surfactants  The results of the anionic flotation series are presented in Table 17. Starting at about
1,000 mg/kg  oil and grease, the data show that oil and grease levels were consistently removed to less than
300 mg/kg.  There is variation in the amount of sediment mass associated with the oily concentrate: with the
sarcosinate,  phosphate ester, and lauryl sulfate reagents, an increase in pH resulted in a higher weight loss
to the oily concentrate. This is probably the result of more voluminous and persistent frothing observed at
alkaline pH.
                                                 41

-------
Table 17.— Flotation of Saginaw River Sediment #2 with Anionic Surfactants
Surfactant
Na alkylaryl
sulfonate
Na alkylaryl
sulfonate
Na Alkylaryt
sulfonate
Na lauroyl
sarcosinate
Na lauroyl
sarcosinate
Na lauroyl
sarcosinate
Phosphate ester
Phosphate ester
Phosphate ester
Na lauryl sulfate
Na lauryl sulfate
Na lauryl sulfate
1 Natural pH (nat ) is the
between 6 4 and 7 2
PH
4
Nat.
10
4
Nat.
10
4
Nat.
10
4
Nat.
10
Wt. Dist'n
to Cone.,
pet
10.7
19.3
14.5
14.6
23.5
24.4
17.6
38.4
39.6
15.5
20.4
21.1
pH of the sediment diluted with
Oil & Grease
Analysis, mg/kg
Cone.
5,130
3,150
5,350
6,210
3,610
2,310
3,950
1,640
2,080
4,120
3,680
3,090
tap water, without
Tails
185
100
203
281
136
220
96
<3
94
346
121
12
adjustment
Oil & Grease dist.,
pet
Cone.
76.9
88.3
81.7
79.1
89.1
77.3
89.7
99.9
93.5
68.7
88.6
98.6
For these tests,
Tails
23.1
11.7
18.3
20.9
10.9
22.7
10.3
0.1
6.5
31.3
11.4
1.4
natural pH
Coefficient
of separation,
pet
66.2
69.0
67.2
64.5
656
52.9
72.1
61.6
53.9
53.1
68.2
77.5
was observed
Amine ethoxylate surfactants A second reagent group investigated in connection with the Saginaw sediment
was the nonionic/cationic Triton RW surfactant series. The most significant aspect of the Triton RW flotation
data for Saginaw River is the apparent loss of oil and grease at the highest pH level, corresponding to the
area of stable oil/water emulsion.  In other words, the amount of oil and grease found in both flotation
products was less than that known to exist in the untreated sample, indicating a contaminant loss,  probably
due to solubilization.  Certainly some form of water treatment would be required to capture the mobilized
contaminants. These results are provided in Table 18.
Table 18. — Flotation of oil
Triton RW-
Reagent
150
150
150
50
50
50
PH
4
7
10
4
7
10
and grease from Saginaw River sediment #2 with amine ethoxylate surfactants
Wt. dist to
cone , pet
11.1
19.4
17.7
23.8
25.9
14.6
Analysis, mg/kg
Cone. Tails
27,600 144
15,900 106
1,560 25
1,820 178
12,000 125
1,590 19
Distribution
Cone.
96.0
97.3
93.1
76.2
97.1
93.6
, pet
Tails
4.0
2.7
6.9
23.8
2.9
6.4
Coefficient
of sep'n, pet
84.9
779
75.4
52.4
71.2
790
                                                 42

-------
Nonionic surfactants  Table 19 shows the results of flotation tests with the nonionic Igepal CO and other
surfactants.  Almost all conditions tested removed oil and grease to below detectable levels.  Curiously, only
tests with the end members of the Igepal series produced measurable levels of oil and grease in the tailings.
As might be  expected from a nonionic reagent, there is no pH effect shown in the Igepal data. The Saginaw
River sediment seems to release oily contamination quite readily under all conditions tested.  Indeed, tests
using only methylisobutylcarbinol (MIBC) also showed extensive cleaning of this sediment. This is significant
because MIBC is typically used only as a froth-building agent, and is not expected to have
detergentproperties.
Table 19. — Flotation of Saginaw River Sediment #2 with Nonionic Surfactant
Reagent
EtO
content,
wt pet
PH
Wt. Dlst'n
to Cone.,
pet
Oil and grease
Analysis, mg/kg Distribution, pet
Cone.
Tails
Cone.
Tails
Coefficient
of
Separation,
pet
Igepal CO series ethoxyated alcohol surfactants
CO-210
CO-210
CO-430
CO-530
CO-530
CO-630
CO-730
CO-730
CO-850
CO-850
23
23
44
54
54
65
75
75
80
80
5
Nat.1
Nat.1
5
Nat.1
Nat.1
5
Nat.1
5
Nat.'
6.2
14.8
17.3
16.9
11.2
15.2
16.1
17.7
19.2
17.1
12,400
8,680
7,580
6,310
7,030
4,510
6,570
5,280
5,800
5,460
560
<3
<3
<3
<3
<3
<3
<3
<3
100
59.6
99.8
99.8
998
99.7
99.6
99.8
99.7
99.8
91.8
40.4
0.2
0.2
0.2
03
0.4
0.2
0.3
0.2
8.2
53.4
85.0
82.5
829
88.5
84.4
83.7
82.0
80.6
74.7
Other nonionic surfactants
MIBC
MIBC
MIBC
Triton X-1 14
Triton X-1 14
Triton X-1 14
1 Natural pH (nat.)
between 6.4 and 7
NA
NA
NA
58
58
58
4
Nat.1
Nat.1
4
Nat.1
10
10.7
7.5
4.9
19.7
19.4
14.9
is the pH of the sediment diluted with tap water,
.2.
26,700
33,200
25,600
16,400
16,400
1,760
242
273
813
137
183
12
without adjustment.
93.0
90.8
61.9
96.7
96.6
96.3
For these tests,
7.0
9.2
38.1
3.3
3.4
3.7
natural pH
82.3
83.3
57.0
77.0
77.2
81.4
was observed
                              Grain-size Separation-Organic Contaminants

The predicted effect of grain size separation on distribution of organic constituents was also investigated for
the Saginaw River #2 sediment sample. The distributions of oil and grease, PCBs, and total organic carbon
through selected size fractions are shown in Table 20.

Several laboratory-scale hydrocyclone separations were performed in the presence of nonionic Igepal CO
surfactants to determine if organic contamination could be further removed from coarse particles by adding
surfactant to the sediment. These data are shown in Table 21. This separation sequence involves sieving on
a 35-mesh screen, followed by pumping the minus-35-mesh  material through the small glass hydrocyclone,

                                                43

-------
making a separation at approximately 12 urn. The underflow is the material that discharges from the cyclone
apex and is the coarser product; the fines are designated as the overflow. These data indicate that
contaminants from Saginaw River can be concentrated in the very fine portion of the sediment, amounting to
about 9 percent of the initial mass. It is apparent, however, that the amount and type of surfactant profoundly
affects the quality of the particle size separation performed by the device, since the amount of material
reporting as fine particles is much greater when the water-insoluble Igepal CO-210 is used.

	Table 20.—Distribution of organic constituents by particle size in Saginaw  River sediment #2	
                          wt rikt'n         To*a' Organic Carbon                 Total PCBs
Size fraction
plus 48 mesh
48x65 mesh
65x100 mesh
100x200 mesh
200x400 mesh
400 mesh x 12 urn
minus 12 urn
Total (calculated)
pet Analysis, mg/kg
17.8
38.3
12.6
12.6
7.6
2.0
9.1
100.00
4,750
4,390
3,850
7,210
13,700
14,400
58,200
10,536
Dist'n, pet Analysis, mg/kg
8.0
16.0
4.6
8.6
9.9
2.7
50.2
100.00
11.1
0.8
2.2
5.6
14.7
22.0
142.3
17.8
Dist'n, pet
11.1
1.8
1.5
4.0
6.3
2.5
72.8
100.00
       Table 21.—Particle size separations at 425 ,um and 12 pm in the presence of nonionic surfactant
                                                Oil and grease                 Oil and grease
               Weight Distribution, pet            analysis, mg/kg                 distribution, pet
Igepal +35 Under- Over- +35 Under- Over- +35 Under- Over-
Reagent mesh1 flow2 flow3 mesh1 flow2 flow3 mesh1 flow2 flow3
CO-210 3.5 66.7 29.9
CO-530 3.2 89.5 10.5
CO-850 2.6 91.3 8.7
<3 503 4031 0.0 21.8 78.2
<3 <3 11,024 0.0 02 99.8
<3 <3 8880 0.0 0.4 99.6
1 >425 Mm
2 <425 Mm, >12 |jm
3 <12 pm
                               Gravity Separation-Organic Contaminants

 To demonstrate the effectiveness of combining unit operations on a single sediment, Saginaw River #2
 sediment previously treated by grain-size separation at 200 mesh was further cleaned of PCBs by water
 elutriation. In this two-step process, the sediment was sieved at 200 mesh, and the sand was then separated
 into narrower size fractions, also by sieving.  Each of these fractions was treated separately in the water
 elutriation column, a density separation device.  The objective of this was to remove the lightweight detrital
 material known to absorb organic contaminants.

 This type of process should be viewed as a "polishing" operation, since 95.2 percent of the PCBs are already
 removed in the minus 200 mesh material.  The results from this test, shown in Table 22,  indicate that 68
 percent of the PCBs remaining in the plus 200 mesh sediment can be removed in roughly 1 percent of the

                                                 44

-------
material which was elutriated. A relatively simple density separation method such as water elutriation could
be used in conjunction with hydrocycloning to isolate PCBs from the sediment.
Table 22.— Elutriation
Elutriation
Product
+35 Lights
+35 Heavies
35X48 Lights
35X48 Heavies
48X65 Lights
48X65 Heavies
65X100 Lights
65X100 Heavies
100X1 50 Lights
100X1 50 Heavies
150X200 Lights
150X200 Heavies
-200
Total Lights
Total Heavies
of +200 mesh Saginaw
Wt. Distribution1
pet
0.2
3.9
0.1
26.2
0.1
3.3
0.2
21.2
0.3
11.0
0.5
6.1
26.8
1.5
71.7
1 Based on the complete sediment; for example, -200 mesh
and contains 64 0 mg/kg PCBs.
River Sediment #2
PCB analysis,
mg/kg
183
0.05
36.1
0.19
32.0
0.13
22.0
0.24
11.4
0.30
10.2
1.78
64.0
41.1
0.35
material comprises 26.8
for PCB Recovery
PCB distribution,
pet
2.3
0.0
0.2
0.3
0.2
0.0
0.2
0.3
0.2
0.2
0.3
0.6
95.2
3.4
1.4
percent of the sediment
                                                  45

-------
                                           CONCLUSIONS

General Comments

                                        Grain Size Separation

Because of its relatively low cost and straightforward concept, grain size separation should be the first mineral
processing pretreatment operation evaluated on any sediment. In this study, grain size separation made
significant progress on the Saginaw River sediment, and might be applicable on the other samples as well.
Grain size separation should always be carefully evaluated when the contaminated sediment is sandy-
perhaps consisting of more than 40 percent plus 200 mesh particles.  As has been shown, detrital material,
which is usually coarser than mineral matter in sediment, also contains much contamination, and can be
simply removed by either size classification or density separation. Although grain-size separation on
contaminated sediment is usually made to separate sand from silt (i.e., at about 75 urn or 200 mesh),
remediation planners should note that significant progress can be made in removing contaminants by
separating in the subsieve size  range-perhaps as fine as 12 urn.

                             Potential for Recycling of Metallic Contaminants

It has been observed that many of the metallic contaminants found in sediment are valuable metals when
refined, and it is appropriate to ask if any of the metals found in sediments can be recovered and put to some
use. Except for the case of  iron ore contained in Indiana Harbor sediment and discussed below, USBM found
no situations where a metal  or metals  could be isolated and concentrated to the point where they might be
recycled to the mining or manufacturing industry. The reason for this is that the majority of the metallic
contamination appears to be bound to nonmetallic, "natural" sediment particles, rather than concentrated in
discrete phases, as in an ore deposit.  Recovering metals from such material would likely require dissolving or
melting the entire sediment matrix.  Such energy- and chemical-intensive processes would be unable to
produce usable metallic products at a  cost even close to these metals relatively-low value in current or likely
future markets.  In addition,  these processes might produce a waste product of greater environmental
concern than the original contaminated sediment.  The environmental harm from generating the energy
required for separating and refining metals from contaminated sediment also likely offsets any environmental
gain from recycling the metals in the sediment.

Summaries

                                           Ashtabula River

The Ashtabula River material was a fine-grained material of relatively high metallic and organic
contamination.  No discrete  contaminant-bearing phases were identified to account for the metallic
contamination measured, and the USBM does not believe that processing techniques that rely strictly on
phase separations, such as  gravity separation, magnetic separation, or froth flotation, will be successful in
isolating metallic contamination from this material. Metallic and organic contamination are evenly distributed
through all particle  sizes, and the USBM's analysis does not predict that contamination can be isolated by
grain-size separation.

Some degree of removal of  organic contamination is possible using froth flotation with ethylene oxide-based
surfactants, although the process is far from optimized. The investigation showed that reagent formulation,
pulp density, and agitation intensity were found to  be the most important operating parameters, and careful
manipulation of these variables will certainly produce better results than those presented herein.
                                                 46

-------
                                            Buffalo River

The Buffalo River sediment was the most difficult of the samples tested to successfully apply mineral
processing pretreatment.  The very fine size distribution is the primary reason for this, aggravated by the lack
of any clearly identifiable contaminant phases or associations. Grain size separation has the potential to
make a small impact on contaminant distributions, but probably not sufficient to justify the additional cost of
this treatment. There is contamination associated with the coarsest material that can be quickly removed with
a size separation device such as a stationary screen, but this is a small amount of the total sediment.

Froth-flotation treatment of the Buffalo River sediment was very similar in effect to treatment of the Ashtabula
River material: significant, but hardly thorough, separation of oily contamination can be expected from a well-
optimized flotation circuit based on the parameters and  surfactants investigated in this study.

                                  Indiana Harbor/Grand Calumet River

Although the bench-scale efforts did not produce furnace-grade iron ore, the data demonstrate that a
significant amount of iron  can be recovered by magnetic separation, carrying with it considerable
contamination. It seems likely, given the choice between feeding this material to the steelmaking process, or
paying fora separate remediation process to recover contaminants, that a case can be made for the recycling
option. A benefit is realized because the amount of material remaining after recovery of the iron ore is
reduced, which may translate into a lower remediation cost.  Some of the elements  that were found to follow
iron in the magnetic separation process are known to be harmful to steelmaking furnaces or deleterious to
steel quality.  It would be necessary to blend sediment-derived iron feedstock with fresh iron ore to assure
that these elements did not reach harmful levels in the furnaces. Organic pollutants following the iron into the
steelmaking process and  not destroyed by the extreme  temperatures could  be captured in existing pollution
control equipment.

The Indiana Harbor Canal sample was extremely difficult to treat by froth flotation owing to its extreme
foaming action, effectively precluding selective recovery of metallic or organic contaminants. Although some
progress was made using flotation on organic contaminants in the Grand Calumet River sample, the single-
stage, single-surfactant systems investigated  here did not produce cleaned  sediment of sufficiently low
contaminant levels.  It should be noted, however, that almost all froth-flotation circuits used to recover
valuable minerals in the mining industry rely upon multiple stages of flotation to achieve economic levels of
recovery and/or grade.  In addition, most successful detergent formulations  are combinations of surfactants,
rather than a formula based on a single surfactant.

                                            Saginaw  River

In all of the mineral processing techniques evaluated in this study, the Saginaw River samples were
significantly easier to treat than any of the other sediments.  The reason for this is undoubtedly that the
Saginaw sediment is much coarser than any of the other samples. In addition, the concentration levels of
most contaminants were lower than in the other samples.

Clearly grain size separation on Saginaw River sediment stands out as the  most promising application of
mineral  processing technology on these samples. This approach has the potential to reduce by up to
80 percent the amount of material requiring remediation. Significant reductions in all metallic contaminants
were observed in the coarse fraction relative to the bulk sediment.  Additionally, PCS contamination can be
significantly reduced by removing the finer portion of the sediment.  This was demonstrated in  Saginaw Bay in
a pilot-scale demonstration project sponsored by the ARCS Program (USACE Detroit District,  1994).
                                                 47

-------
Feasibility Matrix

The purpose of the USBM's study was to evaluate the effectiveness of selected mineral processing
technology to separation of contaminants from river sediment.  The approach is similar to the practice in the
mining industry of using these techniques to concentrate (or beneficiate) very small concentrations of valuable
metals to the point where other, more intensive, processes economically recover and refine the metals.  A
well-designed and executed mineral processing circuit will leave behind a tailings product almost barren of
the metal sought.  This is analogous to the objective of the pretreatment approach to sediment remediation.
While not intended as a definitive statement of the effectiveness of any mineral processing  approach to all
sediments, Table 23 is presented as a guide to the present study of these processing methods to the samples
indicated.  As always, results could be made more conclusive with further research, and the most promising
of these possibilities are noted on the table.

	Table 23.—Feasibility Matrix-Mineral Processing Pretreatment of Contaminated Sediment
Operation

Grain Size
Separation
Froth Flotation
Attrition Scrubbing
Density
Separations
Magnetic
Separations
F=Feasible
L=Limited Applications
U=Unlikely
M=More study recommended
Ashtabula
River

U
L,M
U

U

U




Buffalo
River

L
L.M
U

L

U




Grand Calumet River-
Indiana Harbor Canal

L
L,M
U

L

L,M





Saginaw River

F
L
L,M

F

U




                                        RECOMMENDATIONS

The current research accents the need for thorough, systematic characterization procedures for
contamination in sediment, specifically aimed at identifying those phases in sediment that are the hosts of
contaminants. In the mining industry, no operator would presume to design a mineral processing circuit
without a precise and thorough knowledge of the mineral phases present, and the specific identity of the
phase that contains the metal he wishes to recover.  Further work is needed to adapt the techniques used in
that industry for identification of contaminant-bearing phases. The specific surface area and scanning
electron microscope studies mentioned herein are perhaps the most promising avenues for gaining further
characterization  data. Undoubtedly, given further time and advanced equipment and techniques, the SEM
could become a  valuable tool in identifying those phases in the sediment that bind contaminants. Variable
pressure ("environmental") SEM's and microprobe techniques are two possibilities for gaining more detailed
characterization  information.  Detailed characterization knowledge makes selection of a processing method
more scientific, and certainly more likely to be successful.
                                                 48

-------
                                     ACKNOWLEDGEMENTS

Mr. William B. Schmidt, Chief, Environmental Technology, Washington, DC, served as project manager for
USBM's study. Supervision at the Salt Lake City Research Center was provided at various times by David A.
Rice, Donna D. Harbuck, Kenneth S. Gritton, Charles F. Davidson, and Thomas A. Phillips.  Robert R. Hall
and Israel G. Torres, chemical engineers, performed much of the experimental work and contributed to
preparation of the tables and graphs in this document. Laboratory assistance was provided by Michael W.
Willis, Scott R. Clifford, Mark Lemmon, Michael V. Hoffman, and Cynthia M. Sanchez. Dwight D.
Hammargren developed the QAPP and, with David L. Neylan, supervised the chemical analysis of samples.
Analysts were Wesley R. Roach, Robert A. Davidson, Jolene G. Jacobsen, and John A. Bertagnolli.
Mineralogical analysis was by Audie L. King and Denise Chirban.

The discussion of organic contaminant flotation from Buffalo River sediment is adapted from a paper
prepared by Scott R. Clifford, Physical Science Technician, U. S. Bureau of Mines for the Society for Mining,
Metallurgy, and Exploration 1992 Student Paper Competition. Mr. Clifford won first place in the
Undergraduate division of that contest.  The paper was published in the November 1993 issue of MINERALS
AND METALLURGICAL PROCESSING (Clifford, 1993).


                                         REFERENCES

Box, G. E. P., W. G. Hunter, and J. S. Hunter, 1978. Statistics for Experimenters, John Wiley and Sons, New
York.

Clifford, S. R., 1993.  Flotation of Organic Contaminants from Buffalo River Sediment, Minerals and
Metallurgical Processing, Nov. 1993.

Gaudin, A. M., and R. T. Hukki, 1946. Principles of Comminution—Size and  Surface Distribution,
Transactions of the American Institute of Mining, Metallurgical, and Petroleum Engineers, New York, Vol. 169,
pp. 67-87.

Leja, J., 1982. Surface Chemistry of Froth  Flotation, Plenum Press, New York, pp. 313-16.

Plackett, R. L. and J. P. Burnham, 1946. The Design of Optimum Multifactorial Experiments, Biometrika, Vol.
33, p. 305.

USAGE Detroit District, 1994. Pilot Scale Demonstration of Sediment Washing for the Treatment of
Contaminated Sediments from the Saginaw River. The USEPA Great Lakes National Program Office,
Chicago, IL.

USEPA, 1994. ARCS Remediation Guidance Document, EPA-905-B94-003, USEPA Great Lakes National
Program Office, Chicago, IL.
                                               49

-------
                       GLOSSARY OF MINERAL PROCESSING TERMINOLOGY
Concentrate     The product of a separation process that is enriched in the constituent(s) of interest.

Conditioning     In flotation, the time alloted for mixing of water, solids, and flotation reagents prior to
                introducing air bubbles and beginning flotation of particles.  The amount of conditioning time
                required is governed by the kinetics of adsorption of reagents at the particle surfaces.

Cut size         The size at which a separation is to be made.

Distribution      Fraction of the initial amount of a constituent that reports to a given product in a separation
                process; partitioning.

Filtrate          Liquid passed tthrough a filtration device.

Head assay     The chemical analysis of the raw material, prior to any treatment.

Mesh           A designation of the opening size in a woven-wire screen, equivalent to the number of
                square apertures per inch. A larger mesh number, therefore, indicates a smaller opening
                size. Particles coarser than a  given size "x" are often referrred to as "plus-x mesh," while
                those finer are spoken of as "minus-x  mesh."

Pulp            A mixture of solids and liquid, a slurry.

Subsieve        The range of particle sizes smaller than the finest sieves commonly employed;  in this study,
                the range finer than 400 mesh, or 37 urn.


Tailings         The product of a separation  process that is depleted in the constituent(s) of interest

Vuggy           Having numerous small cavities  or hollows.
                                                 50

-------
APPENDIX A
  A-1

-------
Table A1 --Size and Element Distributions in Ashtabula River Sediment

Size
Fraction,
Tyler Mesh
+20
20x48
48x65
65x100
100x150
150x200
200x270
270x400
-400

Bottom
Size,
urn
425
300
212
150
106
75
53
38
-

Weight
Dist'n ,
pet
1 0
08
1 1
1 2
28
38
72
180
64 1
a
Cumulative
Wt Dist'n ,
pet finer
990
982
97 1
959
93 1
893
82 1
64 1
00
- Analysis of Individual Size Fractions
Element Analysis of Size Fraction, mg/kg
As
80
18
15
13
12
99
95
71
T:
Cd
1.5
39
42
53
40
32
2.7
21
8.5
Cr
272
188
159
159
102
96
49
3
84
Cu Fe
486 27.100
144 20.200
133 21.600
121 28.400
90 26,100
77 25.700
62 20,800
47 20.800
84 33,300
Pb
165
206
223
240
194
152
128
95
138
Hg
048
1 28
200
1 72
1 33
1 11
091
086
1 66
Ni
32
65
57
62
32
30
21
20
35
Zn
353
236
515
695
598
444
351
278
670
b - Mass and Element Distribution
Size
Fraction,
Tyler Mesh
*20
20x48
48x65
65x100
100x150
150x200
200x270
270x400
-400
Bottom
Size,
um
425
300
212
150
106
75
53
38
-
Weight
Dist'n,
pet
1 0
08
1 1
1 2
28
38
7.2
180
64.1
Cumulative
Wt Dist'n.
pet finer
990
982
97 1
959
931
893
82.1
64 1
00
Element Distribution to Size
As
08
1 4
1 6
1.6
3.2
3.7
68
127
68 1
c - Cumulative
Cd
0.2
0.5
07
1.0
1.7
1.9
3.0
59
850
Mass and
Cr
38
21
24
2.6
4.0
5.0
49
07
745
Element
Cu Fe
60 09
14 06
18 08
18 12
31 2.5
36 33
5.5 5.1
104 12.8
664 728
Distributions
Fraction,
Pb
1 2
1 2
1 8
21
40
4.3
68
127
657

pet
Hg
03
07
1 5
1 5
26
30
4 6
109
74 9


Ni
1 0
1 6
20
23
28
36
48
11 3
/06


Zn
06
0.3
10
1.5
30
30
45
90
770

Cumulative Llomont Distribution
Sizo
Fraction,
1 yler Mesh
<20
20x48
48x65
65x100
100x150
150x200
200x270
270x400
•400
Bottom
Size
um
425
300
212
150
106
75
53
38
.
Weight
Disl'n
pet
1 0
08
1 1
1 2
28
3.8
72
18.0
64 1
Cumulalivo
Wl Dist'n
pet liner
990
982
97 1
959
93.1
89.3
821
64.1
00
pel (mar than Bottom Si/o
As
992
978
961
94.5
91.3
87.6
808
68.1
00
Cd
998
993
98.6
97.6
95.8
939
90.9
85.0
00
Cr
962
94 2
91 7
89.1
851
80.1
75.2
745
00
Cu r-o
94 0 99 1
92 6 98 5
90.8 97 7
89 0 96 6
85.9 94,1
82.3 90 7
76 8 85 6
66 4 72 8
0.0 0.0
Pb
988
97 6
957
936
89.6
853
784
65.7
00
Hg
997
989
974
959
93.3
904
857
749
00
Ni
990
97 4
954
930
902
866
81 9
706
00
Zn
994
990
98.0
965
935
905
860
77.0
00
                          A-2

-------
Table A1 -Size and Element Distributions in Ashtabula River Sediment
d - Cumulative Analysis of Coarse Fraction
Size
Fraction,
Tyler Mesh
+20
20x48
48x65
65x100
100x150
150x200
200x270
270x400
-400
Bottom
Size,
Mm
425
300
212
150
106
75
53
38
-
Weight
Dist'n,
pet
1.0
08
1 1
1.2
28
38
72
180
64 1
Cumulative
Wt Dist'n, -
pet finer
99.0
98.2
97.1
95.9
93 1
893
821
64 1
00
Cumulative Analysis o( All Fractions
Coarser than Bottom Size, mg/kg
As
8.00
124
135
134
12.7
11 7
108
8.9
10.1
Cd
1.5
2.6
3.2
3.8
3.9
3.6
3.3
27
64
Cr
272
235
206
192
156
134
100
51
72
Cu
486
334
258
218
166
134
105
76
81
Fe
27,100
24.000
23,100
24.700
25.200
25,400
23,600
22.200
29,300
Pb
165
183
198
211
204
185
162
129
135
Hg
048
084
1 28
1 41
1 38
1 28
1 13
1.00
1 42
Ni
32
47
51
54
45
40
32
26
32
Zn
3530
301.0
3822
473 7
5242
495 7
437 5
3570
557 8
                         A-3

-------
Table A2.--Size and
Element Distributions in Buftalo River Sediment
a. - Analysis
Size
Fraction
T"tor Merh
+ 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400

Size
Fraction
Tv/lor Mtjch
4 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
•400
Bottom
Size
425
300
212
150
106
75
53
38
-

Bottom
Size
425
300
212
150
106
75
53
38

Weight
Dist'n
P0.4
0.5
1.1
3.1
7.4
9.6
9.1
5.5
63.3

Weight
Dist'n
0.4
0.5
1 1
3.1
7 4
9.6
9.1
5.5
63.3
Cumulative
Wt. Dist'n
99.6
99.1
98.0
94.9
87.5
77.9
68.8
63.3
0.0
b.
Cumulative
Wt. Dist'n
pet fm«r
99.6
99.1
98.0
94 9
87.5
77.9
68.8
63.3
0.0

As
207
30.3
14.6
9.40
6.77
5.69
6.70
6.04
12.8
- Mass

As
7.2
1.3
1.4
2.5
4.3
4.7
5.3
2.9
70.3
of Individual Size Fractions

Cd
6
3
2
1
1
1
1
1
2
Element
Cr
590
330
150
70
40
30
20
20
60
Analysis of Size
Cu
544
237
119
108
72
54
42
54
102
Fe
109,00
45,100
25,400
19,100
15,000
13,200
15,000
22,600
41,700
Fraction, mg/kg
Pb
390
150
80
70
60
40
50
50
130
Hg
0.84
0 49
0.66
0.29
0.26
0.24
0.17
0.17
0.71
Ni Zn
170 367
83 144
46 90.8
34 101
27 86.7
19 81.5
20 85.9
24 97.6
50 305
and Element Distribution

Cd
1.4
0.9
1.3
1.9
4.4
5.7
5.4
3.3
75.6
c - Cumulative Mass and
Size
Fraction
T"lor Mflch
+ 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400
Bottom
Size
425
300
212
150
106
75
53
38
.
Weight
Dist'n
0.4
0.5
1.1
3.1
7.4
9.6
9.1
5.5
63.3
Cumulative
Wt. Dist'n
99.6
99.1
98.0
94.9
87.5
77.9
68.8
63.3
0.0
Element
Cr
4.3
3.0
3.0
4 0
5.4
5.3
3.3
2.0
69.6
Distribution to Size Fraction,
Cu
2.4
1.3
1 5
3 7
5.9
5.8
4.3
3.3
71.8
Fe
1.3
0 7
0.8
1 8
3.4
3.8
4.1
3.8
80.2
Pb
1.5
0.7
0.9
2.1
4.3
3 7
4.4
2.7
79.7
pet
Hg
0.6
0.5
1 3
1 7
3.6
4.3
2.9
1.7
83.4

Ni Zn
1.6 0.6
1.0 0.3
1.2 0.4
26 14
4.8 2.8
4.4 3.5
4.4 3.5
3.2 2.4
76.7 85.1
Element Distributions
Cumulative
As
92.8
91.5
90.1
87.6
83.2
78.5
73.2
70.3
0.0
Cd
98.6
97.7
96.4
94.5
90.1
84.3
78.9
75.6
0.0
Cr
95.7
92.7
89.6
85.7
80.2
74.9
71.6
69.6
0.0
Cu
97.6
96.3
94.8
91.1
85.2
79.4
75.1
71.8
0.0
Element
Fe
98.7
98.0
97.1
95.3
92.0
88.1
84.0
80.2
0.0
Distribution
Pb
98.5
97.8
96.9
94.8
90.5
86.8
82.4
79.7
0.0

Hg
99.4
98.9
97.6
95.9
92.3
88.1
85.2
83.4
0.0

Ni Zn
98.4 99.4
97.3 99.0
96.1 98.6
93.6 97.2
88.7 94.4
84.3 90.9
79.9 87.5
76.7 85.1
0.0 0.0
A-4

-------


Table
A2.--Size and
Element
d. - Cumulative
Size
Fraction
Tyler Mesh
+ 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400
Bottom
Size
fjm
425
300
212
150
106
75
53
38
-
Weight
Dist'n
pet
0.4
0.5
1.1
3.1
7.4
9.6
9.1
5.5
63.3
Cumulative
Wt. Dist'n
pet finer
99.6
99.1
98.0
94.9
87.5
77.9
68.8
63.3
0.0
Distributions in Buffalo River Sediment
Analysis
of Coarse Fraction
Cumulative
Analysis of
Coarser than Bottom
As
207
109
57.0
28.1
15.5
1 1.2
9.90
9.32
11.5
Cd
6
4
3
1
1
1
1
1
1
Cr
590
446
283
154
86
62
50
45
55
Cu
544
373
234
157
107
84
72
69
90
Fe
109,00
73,500
47,000
30,100
2 1 , 1 00
17,700
16,900
17,800
32,900
All Fractions
Size, mg/kg
Pb
390
257
160
105
78
62
58
57
103
Hg
0.84
0 65
0.65
0.43
0 33
0.29
0.26
0.24
0.54
Ni
170
122
80
52
37
29
27
26
41
Zn
367
243
159
124
102
93.0
90.9
91.9
227
A-5

-------
Table A3.--Density Separation of Buffalo River Sediment
a. Contaminant Analyses
+ 200 mesh material
Density
Fraction
S.G.<1.9
1.92.9
Total
Wt Dist
pet
0.73
23.74
1.13
25.59

As
36
4.9
58
8.2

Cd
6.2
1.0
2.2
1.20

Cr
155
61
324
75
Analysis, mg/kg
Cu Fe
242 26,600
42 15,200
388 155,000
63 21,700

Pb
230
10
386
33

Hg
1.8
0.14
1.0
0.22

Ni
81
24
160
32

Zn
375
59
618
93
200x400 mesh material
Density
Fraction
S.G.<1.9
1.92.9
Total
400 mesh x 1 2 /vm
Density
Fraction
S.G.<2.9
S.G.>2.9
Total
Wt Dist
pet
0.25
29.57
1.15
30.96
material
Wt Dist
pet
15.50
0.64
16.14

As
28
4 4
36
5.8


As
7.4
42
8.7

Cd
7.6
0.9
2.4
1.01


Cd
1.7
2.4
1.7

Cr
127
22
233
31


Cr
36
466
53
Analysis, mg/kg
Cu Fe
220 24,900
35 15,000
301 166,000
46 20,700

Analysis, mg/kg
Cu Fe
53 27,000
349 380,000
65 41,000

Pb
220
32
311
44


Pb
54
235
61

Hg
3 7
0.20
0.69
0.25


Hg
0.33
0.75
0.35

Ni
64
21
110
25


Ni
31
200
38

Zn
339
81
628
103


Zn
172
488
184
- 1 2 //m material
No Separation
Possible
Total
Wt Dist
pet
27.30

As
15

Cd
4.7

Cr
113
b. Contaminant
Analysis, mg/kg
Cu Fe
134 44,000
Distributions

Pb
186


Hg
1.0


Ni
63


Zn
431

+ 200 mesh material
Density
Fraction
S.G.<1.9
1.92.9
Total
Wt Dist
pet
0.73
23.74
1.13
25.59

As
2.77
12.25
6.94
21 .96

Cd
2.06
10.88
1.14
14.08

Cr
1.65
21.24
5.36
28.25
Distribution, pet
Cu Fe
2.27 0.63
12.87 11.80
5.65 5.72
20.78 18.15

Pb
2.02
2.87
5.27
10.16

Hg
2.73
7.05
2 42
12.20

Ni
1.51
14.61
4.63
20.74

Zn
1.34
6.89
3.43
1 1.66
200x400 mesh material
Density
Fraction
S.G.<1.9
1.92.9
Total
400 mesh x 12 /i/m
Density
Fraction
S.G.<2.9
S.G.>2.9
Total
Wt Dist
pet
0.25
29.57
1.15
30.96
material
Wt Dist
pet
15.50
0.64
16.14

As
0.75
13.74
4.38
18.87


As
12.07
2.79
14.86

Cd
0.87
12.19
1.26
14.33


Cd
12.08
0.70
12.78

Cr
0.47
9.54
3.92
13.93


Cr
8.19
4.37
12.55
Distribution, pel
Cu Fe
0.71 0.20
13.35 14.50
4.45 6.22
18.52 20.93

Distribution, pet
Cu Fe
10.61 13.69
2.88 7.94
13.48 21.63

Pb
0.67
1 1.45
4.31
16.43


Pb
10.13
1.82
1 1.95

Hy
1.99
12.55
1 68
16.22


Hg
10.86
1.02
1 1.88

Ni
0.41
15.92
3.23
19.56


Ni
12.32
3.27
15.60

Zn
0.42
1 1 79
3.54
15.75


Zn
13.13
1.53
14.66
-1 2 /j
-------
Table A4.-Magnetic Separation of Buffalo River Sediment
a. - Element Analyses
+ 200 mesh material - dry high-intensity induced roll separator
Product
0.75 amp
2amp
non-mag
Total
200x400 mesh
Product
1 amp
Samp
non-mag
Total
400 mesh x 1 2
Product
1 amp
Samp
non-mag
Total
Wt Dist
pet
4.7
1.0
14.2
19.9
material - wet
Wt Dist
pet
1.7
2.0
23.2
26.9
//m material -
Wt Dist
pet
1.0
1.1
11.6
13.7

As
29.5
39.0
5.5
12.8
high-intensity

As
27.9
12.7
3.8
6.0

Cd
1.5
1.6
0.5
0.8
separator

Cd
2.0
1.2
0.8
0.9

Cr
249
98
93
130


Cr
210
84
16
33
Element Analysis
Cu Fe
407 63,200
326 34,200
82 29,200
171 37,500

Element Analysis,
Cu Fe
252 100,000
140 56,700
27 11,200
50 20,200
, mg/kg
Pb
188
107
20
64

mg/kg
Pb
286
96
30
51

Hg
0.77
0.35
0.16
0.31


Hg
0.83
0.32
0.69
0.67

Ni
104
76
36
54


Ni
96
66
15
24

Zn
255
178
56
109


Zn
434
268
63
102
wet high-intensity separator

As
24.6
13.7
6.5
8.5

Cd
2.4
1.9
1.6
1.7

Cr
380
106
20
55
Element Analysis
Cu Fe
314 276,000
184 103,000
51 15,600
82 42,600
, mg/kg
Pb
226
97
46
64

Hg
1.6
1.1
1.6
1.6

Ni
66
94
23
32

Zn
507
348
117
166
Minus 1 2 /jm material - wet high-intensity separator
Product
1 amp
Samp
non-mag
Total

Wt Dist
pet
2.0
2.2
35.3
39.4


As
36.6
18.8
16 5
17.6


Cd
6.1
4.2
4.5
4.6


Cr
283
137
82
95
b. -
Element Analysis
Cu Fe
404 172,000
307 101,000
120 34,100
144 44,700
Distribution
, mg/kg
Pb
352
186
165
175


Hg
1.1
0.47
0.44
0.47


Ni
98
107
32
39


Zn
775
409
373
395

+ 200 mesh material - dry high-intensity induced roll separator
Product
0.75 amp
2amp
non-mag
Total
200x400 mesh
Product
1 amp
5 amp
non-mag
Total
400 mesh x 12
Product
1 amp
Samp
non-mag
Total
Wt Dist
pet
4.7
1.0
14.2
19.9
material - wet
Wt Dist
pet
1.7
2.0
23.2
26.9
//m material -
Wt Dist
pet
1.0
1.1
1 1 6
13.7
Element Distribution, pet
As
1 1.3
3.2
6.4
20.8
high-intensity
Cd
2.9
0.7
2.9
6.5
separator
Cr
14.7
1.2
16.6
32.5

Cu Fe
16.5 8.2
2.8 0.9
10.1 11.4
29.5 20.5

Pb
8.5
1.0
2.7
12.2

Hg
5.6
0.5
3.5
9.6

Ni
13.2
2.0
13.8
29.0

Zn
5.3
0.8
3.5
9.6

Element Distribution, pet
As
3.8
2.1
7.2
13.1
Cd
1.4
1.0
7.6
10.0
Cr
4.4
2.1
4.7
11.2
Cu Fe
3.7 4.6
2.5 3.2
5.4 7.1
11.5 14.9
Pb
4.6
1.9
6.7
13.1
Hg
2.1
1.0
24.7
27.8
Ni
4.3
3.6
9.4
17.3
Zn
3.2
2.4
6.4
12.0
wet high-intensity separator
Element Distribution, pet
As
2.1
1.2
6.1
9.5
Cd
1.0
0.9
7.6
9.5
Cr
5.0
1.5
2.9
9.4
Cu Fe
2.8 8.0
1.8 3.2
5 1 5.0
9.7 16.1
Pb
2.3
1 0
5 1
8.4
Hg
2.6
2 0
29.2
33.8
Ni
1.9
2 8
7 2
11.8
Zn
2.3
1 7
5 9
10.0
                 A-8

-------
Table A4.--Magnetic Separation of Buffalo River Sediment
Minus 1 2 fjm material - wet high-intensity separator

Product
1 amp
Samp
non-mag
Total
Wt Dist
pet
2.0
2.2
35.3
39.4

As
5.8
3.4
47.4
56.6

Cd
4.9
3.8
65.3
74.0

Cr
6.9
3.8
36.2
46.9
Element
Cu
6.8
5.8
36.6
49.3
Distribution
Fe
9.2
6.1
33.1
48.4
, pet
Pb
6.6
3.9
55.7
66.2

Hg
3.2
1.6
23.9
28.7

Ni
5.2
6.3
30.4
41.8

Zn
6.7
3.9
57.8
68.4
                 A-9

-------
Table A5.-Froth Flotation of Buffalo River Sediment

Flotation
Conditions
pH 4
No Reagents








Oleic Acid
pH 7








Oleic Acid
pH 10








Copper Sulfate
and Potassium
Amyl Xanthate
pH 7
(Note, Cu
analyses are omitted
due to
the addition of
copper sulfate)

Copper Sulfate
and Potassium
Amyl Xanthate
pH 10
(Note, Cu
analyses are omitted
due to the addition
of
copper sulfate)



Element
Mass
As
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Zn
Mass
As
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Zn
Mass
As
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Zn
Mass
As
Cd
Cr
Fe
Pb
Hg
Ni
Zn
Mass
As
Cd
Cr

Fe
Pb
Hg
Ni
Zn
Initial
Analysis,
mg/kg

10.3
2
60
70
29,700
100
0.47
37
166.8

10.3
2
60
70
29,700
100
0.47
37
166.8

10.3
2
60
70
29,700
100
0.47
37
166.8

10.3
2
60
29,700
100
0.47
37
166.8

10.3
2
60

29,700
100
0.47
37
166.8
Concentrate
Analysis,
mg/kg
(49.5 g)
27.5
6
140
190
51,100
280

68
430
(118.9 g)
24.4
5
110
160
52,200
220

65
330
(21.1 g)
44.1
5
190
300
61,700
370

94
430
(58.1 g)
24.5
5
110
48,800
240

64
370
(16.0 g)
59.3
8
180

66,300
470

96
520
Distribution,
pet
13.35
34.73
34.79
28.98
36.02
24.05
39.58

25.14
35.32
31.77
64.70
69.96
48.23
66.82
48.51
70.44

48.61
58.28
5.75
19.66
13.24
14.94
23.38
9.42
19.54

12.02
11.60
15.39
31.42
31.27
23.54
22.10
35.60

23.46
81.78
4.37
20.66
15 45
11.71

7.90
19.98

10.60
12.93
Tails
Analysis,
mg/kg
1321.0 g)
9.19
2
61
60
28,700
76

36
140
(255.3 g)
6.2
1
55
37
25,800
43

32
110
(345.1 g)
1 1.0
2
66
60
36,200
93

42
200
(319.1 g)
9.73
2
65
31,300
79

38
15
(349.4 g)
10.4
2
62

35,300
86

37
160
Distribution,
pet
86.65
65.27
65.21
71.02
63.98
75.95
60.42

74.86
64.68
68.23
35.30
30.04
51.77
33.18
51.49
29.56

51.39
41.72
94.25
80.34
86.76
85.06
76.62
90.58
80.46

87.98
88.40
84.61
68.58
68.73
76.46
77.90
64.40

76.54
18.22
95.63
79.34
84.55
88.29

92.-10
80.02

89.40
87.07
                     A-10

-------
Table A6. -Acidic Attrition Scrubbing of Buffalo River Sediment
a. -Hydrochloric Acid
1 .-Contaminant Analyses
Attrition
Product
Solids
Precipitate
Filtrate
Total

Attrition
Product
Solids
Precipitate
Filtrate
Total

Wt Dist
pet
98.96
1.04
100.00

Wt Dist
pet
98.96
1.04
100.0

As
14.3
99.2
<0.006
15.2

As
93.23
6.77
100.0

Cd
Cr
1.80 117
0.90 114
<0.003 <0.03
1.79 117
2
Cd
99.48
0.52
100.0

.-Contaminant
Cr
98.99
1.01
100.0
b. -Nitric
Analysis, mg/kg
Cu Fe
108 44.400
39 428,000
0.03 0.84
107 48,375
Pb
105
18
<0.16
104
Hg
0 49
0.17
0.02
0 48
Nil
53
392
<0.03
57
Zn
168
1310
<0.005
180
Distributions
Distribution, wt pet
Cu Fe Pb
99.62
0.38
100.0
Acid
90.83
9.17
100.0

99.82
0.18
100.0

Hg
99.64
0.36
100.0

Ni
92.81
7.19
100.0

Zn
92.45
7.55
100.0

1 .-Contaminant Analyses
Attrition
Product
Solids
Precipitate
Filtrate
Total

Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
99.15
0.85
100.00

Wt Dist
pet
99.15
0.85
100.0
As
12.7
78.0
<0.006
13.3

As
95.02
4.98
100.0
Cd
1.80
1.00
<0.003
1.79
2.
Cd
99.53
0.47
100.0
Cr
80
20
<0.03
79
-Contaminant
Cr
99.79
0.21
100.0
Analysis, mg/kg
Cu Fe
79 31,000
30 376,000
<.03 0.40
79 33,922
Pb
99.6
<70
<0.16
98.8
Hg
0.49
0.16
<.01
0.49
Ni
43
390
<0.03
46
Zn
169
1120
<0.005
177
Distributions
Distribution, wt pet
Cu Fe Pb
99.68
0.32
100.0
90.61
9.39
100.0
100.00
0.00
100 0
Hy
99.72
0.28
100.0
Ni
92 81
7.19
100 0
Zn
94.64
5.36
100.0
c.-Sulfunc Acid
1 .-Contaminant Analyses
Attrition
Product
Solids
Precipitate
Filtrate
Total

Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
99.09
0.91
100.00

Wt Dist
pet
99.09
0.91
100.0
As
12.2
75.5
<0.006
12.8

As
94.60
5.40
100.0
Cd
1.60
2.00
<0.003
1.60
2.
Cd
98.86
1.14
100.0
Cr
76
30
<0.03
76
-Contaminant
Cr
99.64
0.36
100.0
Analysis, mg/kg
Cu Fe
76 28,800
34 447,000
<.03 0.81
76 32,621
Pb
94
<30
<0.16
93
Hg
0.49
0.57
<.01
0.49
Ni
38
384

-------
Table A7.~Size and Element Distributions in Indiana Harbor Sediment

Size
Fraction
Tyler Mesh
TTo
10x14
14x20
20x28
28x35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400

Bottom
Size
fjm
1680
1 190
841
595
425
300
212
150
106
75
53
38
-

Weight
Dist'n
pet
4~T~
2.6
2.2
1.8
1.6
1.8
2.1
4.7
5.0
4.0
5.2
5.3
59.6
a. •
Cumulative
Wt. Dist'n
pet finer
95.9
93.3
91.1
89.3
87.7
85.9
83.8
79.1
74.1
70.1
64.9
59.6
0
• Analysis of Individual Size
Fractions




Element Analysis of Size Fraction, mg/kg
As
3.6
4.2
6.0
8.2
8.6
8.6
8.4
8.0
14.8
23.7
30.2
27.1
37.6
Cd
0.3
0.5
0.4
0.7
4.0
1.0
2.0
2.0
3.0
5.0
6.0
6.0
11.0
Cr
30
30
49
40
57
70
98
79
140
248
307
338
676
Cu Fe
23 14,700
32 18,400
53 23,100
24 22,300
58 26,500
79 31,600
134 50,100
129 46,000
127 42,400
233 78,200
343 155,000
380 202,000
400 115,000
Pb
20
30
49
60
86
159
118
118
239
455
604
656
1 170
Hg
< 12
<.1 1
0.30
0.84
0.44
0.44
0.20
0.25
0 44
0.69
0.99
1.10
1.84
Ni
1 14
69
48
46
52
49
57
53
52
82
1 18
1 19
80
Zn
557
1 13
470
180
262
508
1024
942
1220
2240
3036
3760
2330
b. - Mass and Element Distribution
Size
Fraction
Tyler Mesh
7T6
10x14
14x20
20x28
28x35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
•400
Bottom
Size
/jrrt
1680
1 190
841
595
425
300
212
150
106
75
53
38

Weight
Dist'n
pet
4.1
2.6
2.2
1.8
1.6
1.8
2.1
4.7
5.0
4.0
5.2
5.3
59.6
Cumulative
Wt. Dist'n .
pet finer
95.9
93.3
91.1
89.3
87.7
85.9
83.8
79.1
74.1
70.1
64.9
59.6
0
Element Distribution to
As
0.5
0.4
0.5
0.5
0.5
0.5
0.6
1.3
2.6
3.3
5.5
5.0
78.7
c - Cumulative
Size
Fraction
Tyler Mesh
+ 10
10x14
14x20
20x28
28x35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400
Bottom
Size
fjm
1680
1190
841
595
425
300
212
150
106
75
53
38

Weight
Dist'n
pet
4.1
2.6
2.2
1.8
1.6
1.8
2.1
4.7
5.0
4.0
5.2
5.3
59.6
Cumulative
Wt. Dist'n
pet finer
95.9
93.3
91.1
89.3
87.7
85.9
83.8
79.1
74.1
70.1
64.9
59.6
0.0
Cd
0.2
0.2
0.1
0.2
0.8
0.2
0.
1.2
1.9
2.6
4.0
4.1
84.0
Mass and
Cr
0.3
0.2
0.2
0.2
0.2
0.3
0.4
0.8
1.5
2.1
3.4
3.8
86.6
Element
Cu Fe
0.3 0.6
0.3 0.5
0.4 0.5
0.1 0.4
0.3 0.4
0.5 0.6
0.9 1.1
2.0 2.2
2.1 2.1
3.0 3.2
5.8 8.2
6.6 10.8
77.7 69.4
Distributions
Size Fraction, pet
Pb
0.1
0.1
0.1
0.1
0.2
0.4
0.3
0.7
1.5
2.2
3.9
4.3
86.2

Hg
0.0
0.0
0.5
1.2
0.5
0.6
0.3
0.9
1.7
2.1
3.9
4.5
83.8

Ni
5.9
2.2
1.3
1.0
1.0
1.1
1.5
3.1
3.3
4.1
7.7
7.9
59.8

Zn
1.1
0.1
0.5
0.2
0.2
0.5
1.1
2 2
3.0
4.4
7.8
9.9
68.9

Cumulative Element Distribution,
pet finer than Bottom Size
As
99.5
99.1
98.6
98.1
97.6
97.1
96.5
95.2
92.6
89.2
83.7
78.7
0.0
Cd
99.8
99.7
99.6
99.4
98.6
98.4
97.8
96.6
94.7
92.1
88.1
84.0
0.0
Cr
99.7
99.6
99.3
99.2
99.0
98.7
98.3
97.5
96.0
93.8
90.4
86.6
0.0
Cu Fe
99.7 99.4
99.4 98.9
99.0 98.4
98.9 98.0
98.6 97.6
98.1 97.0
97.2 95.9
95.2 93.7
93.2 91.6
90.1 88.4
84.3 80.2
77.7 69.4
0.0 0.0
Pb
99.9
99.8
99.7
99.5
99.4
99.0
98.7
98.0
96.6
94.3
90.4
86.2
0.0
Hg
100.0
100.0
99.5
98.3
97.8
97.2
96.9
96.0
94.3
9.2.2
88.3
83.8
0.0
Ni
94.1
91.9
90.6
89.5
88.5
87.4
85.9
82.8
79.5
75.4
67.7
59.8
0.0
Zn
98.9
98.7
98.2
98.0
97.8
97.4
96.3
94.1
91.1
86.6
78.8
68.9
0.0
                       A-12

-------
Table A7.--Size and Element Distributions
m Indiana Harbor Sediment
(continued)
d. - Cumulative Analysis of Coarse Fraction
Size
Tyler Mesh
+ 10
10x14
14x20
20x28
28x35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400
Bottom
Size
//m
1680
1190
841
595
425
300
212
150
106
75
53
38

Weight
Dist'n
pet
4.1
2.6
2.2
1.8
1.6
1.8
2.1
4.7
5.0
4.0
5.2
5.3
59.6
Cumulative
Wt. Dist'n
pet finer
95.9
93.3
91.1
89.3
87.7
85.9
83.8
79.1
74.1
70.1
64.9
59.6
0.0
Cumulative Analysis of All Fractions
Coarser than Bottom Size, mg/kg
As
3.6
3.8
4.4
5.0
5.5
5.9
6.2
6.6
8.2
10.3
13.2
15.0
28.5
Cd
0.3
0.4
0.4
0.4
0.9
0.9
1.1
1.3
1.6
2.1
2.6
3.1
7.8
Cr
30
30
35
36
38
42
50
56
72
96
127
155
465
Cu
23
26
33
32
35
41
53
70
81
101
137
169
307
Fe
14,700
16,100
17,900
18,600
19,600
21,200
24,900
29,700
32,100
38,300
55,600
74,800
98,800
Pb
20
24
30
35
42
57
65
77
108
154
221
278
812
Hg
0.00
0.00
0.07
0.20
0.23
0.26
0.25
0.25
0.29
0.34
0.44
0.52
1.31
Ni
1 14
97
85
78
75
71
70
66
63
66
73
79
80
Zn
557
385
406
368
354
374
458
567
693
900
1220
1550
2010
A-13

-------
Table AS.-Density Separation of Indiana Harbor Sediment
a. - Contaminant Analyses

Density
Fraction
S.G.<1.9
1.92.9
Total
100x400 mesh
Density
Fraction
S.G.<1.9
1.92.9
Total
400 mesh x 12
Density
Fraction
S.G.<1.9
1.92.9
Total

Wt Dist
pet
13.69
9 6.86
0.48
21.03
material
Wt Dist
pet
5.73
9 15.45
5.21
26.40
//m material
Wt Dist
pet
0.32
.9 9.75
5.93
16.01


As
5.1
5.8
48
6.3


As
18
17
25
19


As
37.40
28.60
22.40
26.48


Cd
0.0
1.0
4.0
0.4


Cd
3.0
3.0
4.0
3.2


Cd
13.0
9.0
5.0
7.6
+ 100 mesh

Cr
10
80
710
49


Cr
190
240
420
265


Cr
640
530
430
495
material
Analysis, mg/kg
Cu Fe Pb
1 2 4400 20
30 13,200 50
385 357,000 570
26 15,300 42

Analysis, mg/kg
Cu Fe Pb
109 23,800 370
156 61,600 380
366 337,000 620
187 108,000 425

Analysis, mg/kg
Cu Fe Pb
388 91,000 1140
317 130,000 880
256 302,000 630
296 193,000 793


Hg
0.19
0.23
0.63
0.21


Hg
0.68
1.33
1.51
1.22


Hg
2.66
1.82
1.32
1.65


Ni
55
15
206
45


Ni
49
52
149
71


Ni
88
82
124
98


Zn
84
270
1210
170


Zn
1080
1480
1880
1472


Zn
4310
3540
2350
3115
-1 2 fjm material
Density
Fractions
S.G.<1.9
1.92.9
Total

Wt Dist
pet
2.41
.9 33.13
1.03
36.57


As
59
47
39
47


Cd
19.0
15.0
8.0
15.1
b -

Cr
970
820
620
824
Contaminant
Analysis, mg/kg
Cu Fe Pb
606 103,000 1730
530 166,000 1440
378 298,000 940
531 166,000 1445
Distributions

Hg
3.42
3.38
1.90
3.34


Ni
104
107
148
108


Zn
6500
6630
4010
6548

+ 100 mesh material
Density
Fraction
Wt Dist
pet
S.G.<1.9 13.69
1.92.9 0.48
Total 21.03
100x400 mesh
Density
Fraction
S.G.<1.9
1.92.9
Total
400 mesh x 1 2
Density
Fraction
S.G.<1.9
1.92.9
Total
material
Wt Dist
pet
5.73
.9 15.45
5.21
26.40
fjrrt material
Wt Dist
pet
0.32
.9 9.75
5.93
16.01

As
2.50
1.43
0.84
4.77


As
3.62
9.67
4.68
17.97


As
0.43
10.03
4.78
15.24

Cd
0.00
0.90
0.25
1.15


Cd
2.24
6.05
2.72
11.02


Cd
0.55
11.46
3.87
15.88

Cr
0.30
1.19
0.74
2.23


Cr
2.36
8.05
4.75
15.16


Cr
0.45
11.22
5.53
17.20
Distribution, pet
Cu Fe Pb
0.55 0.49 0.35
0.69 0.74 0.44
0.63 1.40 0.35
1.87 2.62 1.15

Distribution, pet
Cu Fe Pb
2.11 1.11 2.73
8.13 7.73 7.56
6.44 14.27 4.16
16.68 23.11 14.46

Distribution, pet
Cu Fe Pb
0.42 0.24 0.47
10.43 10.30 11.05
5.12 14.55 4.81
15.98 25.09 16.34

H9
1.40
0.85
0.16
2.42


Hg
2.10
1 1.09
4.25
17 43


Hg
0.46
9.57
4.22
14.26

Ni
9.04
1.24
1.19
1 1.47


Ni
3.37
9.65
9.33
22 35


Ni
0.34
9.60
8.83
18.78

Zn
0.35
0.56
0.18
1.08


Zn
1.86
6.89
2.95
11.71


Zn
0.42
10.41
4.20
15.03
                    A-14

-------
Table A8.-Density Separation of Indiana Harbor Sediment
-12 fjm material
Density
Fraction
S.G.<1.9
1.92.9
Total
Wt Dist
pet
2.41
33.13
1.03
36.57
Distribution, pet
As
5.10
55.49
1.43
62.02
Cd
5.97
64.90
1.08
71.95
Cr
5.07
58.96
1.39
65.41
Cu
4.92
59.24
1.31
65.47
Fe
2.01
44.68
2.49
49.18
Pb
5.36
61.45
1.25
68.06
Hg
4.44
60.40
1.06
65.89
Ni
3.01
42.57
1.83
47.41
Zn
4.72
66.22
1.25
72.18
                   A-15

-------
Table A9 --Magnetic Separation of Indiana Harbor Sediment
a. - Element Analyses
+ 100 mesh material - dry high-intensity induced roll separator
Product
Davis
0.25amp
2. Samp
non-mag
Total
100x400 mesh
Product
Davis
0.5 amp
5 amp
non-mag
Total
400 mesh x 12
Product
Davis
0.5amp
5amp
non-mag
Total
Wt Dist
pet
0.5
5.4
6.2
9.5
21.6
material - wet
Wt Dist
pet
2 4
3 5
5.2
12 1
23.2
fjm material -
Wt Dist
pet
4.1
4.7
6 8
3.3
18.9

As
26
18
5.9
5.1
9.0
high-intensity

As
25
32
29
17
23

Cd
1.0
2.5
0.9
0.6
1.2
separator

Cd
0.5
8.2
7.9
3.0
4.6

Cr
604
139
52
23
73


Cr
427
610
422
175
322
Element Analysis,
Cu Fe
350 280,000
740 27,000
269 10,000
59 10,000
296 20,104

Element Analysis,
Cu Fe
165 393,000
405 233,000
406 107,000
149 23,600
247 112,000
mg/kg
Pb
260
306
95
42
128

mg/kg
Pb
263
901
804
360
531

Hg
0.23
0.21
0.12
0.12
0.14


Hg
0.42
1.36
1.48
0 68
0 94

Ni
210
93
48
41
60


Ni
203
1 13
97
46
84

Zn
962
655
301
128
328


Zn
1520
3840
2570
992
1828
wet high-intensity separator

As
20
35
28
24
27

Cd
0 9
9.7
8.6
7.9
7.1

Cr
298
619
521
375
472
Element Analysis,
Cu Fe
126 320,000
436 183,000
376 141,000
255 70,600
316 178,000
mg/kg
Pb
236
966
901
767
751

Hg
0 16
1 4
1.6
1.3
1.2

Ni
1 10
110
1 10
97
1 10

Zn
1800
4880
3640
2480
3350
Minus 1 2 fjm material - wet high-intensity separator
Product
Davis
O.Samp
5amp
non-mag
Total

Wt Dist
pet
0.7
7.6
13.4
14.7
36.4


As
29
49
45
45
46


Cd
3.6
14
16
17
16


Cr
468
803
817
732
773
b. -
Element Analysis,
Cu Fe
223 436,000
518 227,000
547 159,000
451 100,000
496 155,000
Distribution
mg/kg
Pb
447
1270
1380
1310
1310


Hg
0.30
1.5
2.0
2.0
1.9


Ni
189
134
1 16
95
1 13


Zn
4820
6530
6410
5220
5925

+ 100 mesh material • dry high-intensity induced roll separator
Product
Davis
0.25amp
2 Samp
non-mag
Total
100x400 mesh
Product
Davis
O.Samp
Samp
non-mag
Total
Wt Dist
pet
0.5
5.4
6.2
9.5
21.6
material - wet
Wt Dist
pet
2.4
3.5
5.2
12.1
23.2
Element Distribution, pet
As
0 4
3.3
1.3
1.7
6.7
high-intensity
Cd
0.1
1.6
0.7
0.7
3.0
separator
Cr
0.6
1.6
0.7
0.5
3.4

Cu Fe
05 1.1
11.0 1.2
4.6 0.5
1.5 0.8
17.7 36

Pb
0.2
2.1
0.8
0.5
3.6

Hg
0 1
1.0
0.7
1.0
2.7

Ni
1 0
5.4
3.2
4.2
13.7

Zn
0.1
1.1
0.6
0,4
2.2

Element Distribution, pet
As
2.0
3.8
5.3
7.3"
18.3
Cd
0.1
3.4
4.9
4.3
12.7
Cr
2.2
4.6
4.8
4.6
16.2
Cu Fe
1.1 7.7
3.9 6.7
5.9 4.7
5.0 2.4
15.8 21.5
Pb
0.8
4.1
5.4
5.7
16.0
Hg
0.9
4.1
6.7
7.2
18.9
Ni
5.1
4.2
5.4
5.9
20.7
Zn
1.1
4.1
4.1
3.7
12.9
                 A-16

-------
Table A9.-Magnetic Separation of Indiana Harbor Sediment
400 mesh x 1 2 //m material

Product
Davis
O.Samp
5amp
non-mag
Total
Minus 1 2 fjm

Product
Davis
0.5amp
5amp
non-mag
Total
Wt Dist
pet
4.1
4.7
6.8
3.3
18.9
material - wet
Wt Dist
pet
0.7
7.6
13.4
14.7
36.4
- wet high-intensity separator

As
2.8
5.8
6.4
2.7
17.8
high-intensity

As
0.7
12.7
20.9
23.0
57.2

Cd
0.4
5.5
6.9
3.1
16.0
separator

Cd
0.3
12.9
25.8
29.3
68.3

Cr
2.6
6.4
7.7
2.7
19.4


Cr
0.7
13.2
23.8
23,3
61.0
Element
Cu
1.4
5.7
7.1
2.3
16.6

Element
Cu
0.4
10.8
20.4
18.3
49.9
Distribution,
Fe
10.9
7.2
8.0
2.0
28.1

Distribution,
Fe
2.4
14.3
17.8
12.3
46.8
pet
Pb
1 3
6.0
8.0
3.3
18.5

pet
Pb
0.4
12.5
24.1
25.0
62.0

Hg
0.6
5.7
9.3
3.7
19.3


Hg
0.2
10.2
23.0
25.8
59.1

Ni
4.8
5.6
8.0
3.4
21.9


Ni
1.3
10.8
16.7
14.9
43.7

Zn
2.2
7 i
7.5
2.5
19.3


Zn
1.0
15.0
26.2
23.4
65.6
                 A-17

-------
Table AID.--Froth Flotation of Indiana Harbor Sediment

Flotation
Conditions
Oleic Acid
pH7








Oleic Acid
pH 10








Copper Sulfate
and Potassium
Amyl Xanthate
pH 4
(Note: Cu analyses
omitted due to
theaddition of
copper sulfate)


Copper Sulfate
and Potassium
Amyl Xanthate
pH 7
(Note- Cu analyses
omitted due to
theaddition of
copper sulfate)


Copper Sulfate
and Potassium
Amyl Xanthate
pH 10
(Note: Cu analyses
omitted due to
theaddition of
copper sulfate)




Element
Mass
As
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Zn
Mass
As
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Zn
Mass
As
Cd
Cr
Fe
Pb
Hg
Ni
Zn
Mass
As
Cd
Cr
Fe
Pb
Hg
Ni
Zn
Mass
As
Cd

Cr
Fe
Pb
Hg
Ni
Zn
Initial
Analysis
mg/kg
-
30.8
8
440
278
142,000
830
0.74
94
3,280

30.8
8
440
278
142,000
830
0.74
94
3,280
-
30.8
8
440
142,000
830
0.74
94
3,280

30.8
8
440
142,000
830
0.74
94
3,280

30.8
8

440
142,000
830
0.74
94
3,280
Concentrate
Analysis,
mg/kg
(286.0g)
35
10
570
340
140,000
900

100
3,870
I192.7g)
32.7
10
550
330
117,000
920

100
3,570
(299.0g)
39.4
10
540
1 44.000
920
1.05
86
3,680
(211.7g)
37.8
10
530
142,000
900
1.04
88
3,610
(130.4g)
39.3
9

480
93,300
890
1.87
93
3,320
Distribution,
pet
73.41
85.49
87.34
85.82
86.22
79.44
87.04

79.09
92.29
50.47
57.73
59.28
56.59
58.35
46.90
63.45

55.41
58.79
77.51
91.63
91.99
89.43
87.01
91.88
94.02
85.56
93.58
56.03
64.85
67.99
63.99
59.92
66.80
44.39
59.29
66.47
34.45
40.61
37.15

34.92
24.88
39.71
41.95
38.22
34.52
Tails
Analysis,
mg/kg
(103. 6g)
16.4
4
260
150
100,000
370

73
893
(189. Ig)
24.4
7
430
240
135,000
540

82
2,550
(86. 8g)
12.4
3
220
74,100
280
0.23
50
870
(166. 2g>
26.1
6
380
121,000
570
1.66
77
2,320
(349.4g)
30.2
8

470
148,000
710
1.36
79
3,280
Distribution,
pet
26.59
14.51
12.66
14.18
13.78
20.56
12.96

20.91
7 71
49.53
42.27
40.72
43.41
41.65
53.10
36.55

44.59
41.21
22.49
8.37
8.01
10.57
12.99
8.12
5.98
14.44
6.42
43.97
35.15
32.01
36.01
40.08
33.20
55.61
40.71
33.53
65.55
59.39
62.85

65.08
75.12
60.29
58.05
61.78
65.48
                      A-18

-------
Table Al 1 .-Acidic Attrition Scrubbing of Indiana Harbor
Sediment
a. - Hydrochloric Acid
1 . - Contaminant Analyses
Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
96.60
3.40
100.00
As
34.8
22.6
<0.0006
34.4
Cd
8.40
41.10
<0.003
9.51
Analysis, mg/kg
Cr Cu Fe
514 297 124,000
247 1120 454,000
<0.03 <0.03 0.50
505 325 135,226
Pb
851
255
0.00
831
Hg
1.09
1.28
<0.012
1.10
Ni
89
341
<0.03
98
Zn
3100
5430
< 0.005
3179
2. - Contaminant Distributions
Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
96.60
3.40
100.0
As
97.8
2.2
100.0
Cd
85.3
14.7
100.0
Distribution, wt pet
Cr Cu Fe
98.3 88.3 88.6
1.7 11.7 11.4
100.0 100.0 100.0
Pb
99.0
1.0
100.0
Hg
96.0
4.0
100.0
Ni
88.1
1 1.9
100.0
Zn
94.2
5.8
100.0
b. - Nitric Acid
1 - Contaminant Analyses
Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
96.3
3.7
100.0
As
36.0
25.0
<0.0006
35.6
Cd
8.80
7.80
<0.003
8.76
Analysis, mg/kg
Cr Cu Fe
Pb
538 323 137,000 871
366 34 434,000 100
<0.03 <0.03 0.10 <0.16
532 312 148,000 843
Hg
0.96
0.39
<0.012
0.93
Ni
98
291
<0.03
105
Zn
3220
4100
<0.005
3250
2. - Contaminant Distributions
Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
96.3
3.7
100.0
As
97.4
2.6
100.0
Cd
96.7
3.3
100.0
Distribution, wt pet
Cr Cu Fe
97.5 99.6 89.2
2.5 0.4 10.8
100.0 100.0 100.0
Pb
99.6
0.4
100.0
Hg
98.5
1.5
100.0
Ni
89.8
10.2
100.0
Zn
95.4
4.6
100.0
c. - Sulfunc Acid
1 . - Contaminant Analyses
Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
93.6
6.4
100.0
As
34.2
20.1
<0.0006
33.3
Cd
8.60
4.70
<0.003
8.35
Analysis, mg/kg
Cr Cu Fe
Pb
448 308 126,000 859
1260 100 354,000 195
<0.03 <0.03 O.35 <0.16
500 295 140,500 817
Hg
1.07
0.30

-------
Table A12.--Size and Element Distributions in Saginaw River Sediment #2

Size
Fraction
Tyler Mesh
+ 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400

Bottom
Size
/jm
425
300
212
150
106
75
53
38


Weight
Dist'n
pet
0.2
16.9
29.2
21.6
8.4
5.6
3.9
1.8
12.4
a.
Cumulative
Wt. Dist'n -
pet finer
99.8
82.9
53.7
32.1
23.7
18.1
14.2
12.4
0.0
- Analysis of Individual Size Fractions
Element Analysis of Size Fraction, mg/kg
As
16.6
2.8
2.6
4.1
3.3
3.3
3.4
5.0
9.7
Cd
7
2
1
2
3
3
3
6
21
Cr Cu
180 226
40 78
80 50
90 71
60 84
60 53
70 87
90 140
740 484
Fe
13,600
4,180
18,700
23,700
6,200
5,240
6,780
10,300
29,700
Pb
133
9
9
10
39
58
98
146
230
Hg
0.78
0.25
0.15
0.18
0.24
0 15
0.20
0.30
1.26
Nli
97
18
31
42
42
54
67
94
319
Zn
610
29
75
60
89
146
170
265
1 160
b. - Mass and Element Distribution
Size
Fraction
Tyler Mesh
+ 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400
Bottom
Size
fjm
425
300
212
150
106
75
53
38
-
Weight
Dist'n
pet
0.2
16 9
29.2
21.6
8.4
5.6
3.9
1.8
12.4
Cumulative
Wt. Dist'n -
pet finer
99.8
82.9
53.7
32.1
23.7
18.1
14.2
12.4
0.0
Element Distribution to Size Fraction, pet
As
0.8
11.7
18.8
21.9
6.9
4.6
3.3
2.2
29.8
c. - Cumulative
Size
Fraction
Tyler Mesh
+ 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400

Size
Fraction
Tyler Mesh
+ 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400
Bottom
Size
pm
425
300
212
150
106
75
53
38


Bottom
Size
//m
425
300
212
150
106
75
53
38

Weight
Dist'n
pet
0.2
16.9
29.2
21.6
8.4
5.6
3.9
1.8
12.4

Weight
Dist'n
pet
0.2
16.9
29.2
21.6
8.4
5.6
3.9
1.8
12.4
Cumulative
Wt. Dist'n
pet liner
99.8
82.9
53 7
32.1
23.7
18.1
14.2
12.4
0.0
d. -
Cumulative
Wt. Dist'n
pet finer
99.8
82.9
53.7
32.1
23.7
18.1
14.2
12.4
0.0
Cd
0.3
7.8
6.8
10.0
5.8
3.9
2.7
2.5
60.2
Mass and
Cr Cu
0.2 0.4
4.4 11.0
15.1 12.2
12.6 12.8
3.3 5.9
2.2 2.5
1.8 2.8
1.0 2.1
59.4 50.2
Fe
0.2
4.3
33.6
31.5
3.2
1.8
1.6
1.1
22.6
Pb
0.6
3.2
5.5
4.5
6.8
6.8
8.0
5.5
59.3
Hg
0.5
13.0
13.5
12.0
6.2
2.6
2.4
1.7
48.1
Ni
0.3
4.2
12.6
12.6
4.9
4.2
3.6
2.4
55.1
Zn
0.6
2.3
10.3
6.1
3.5
3.9
3.1
2.3
67.9
Element Distributions
Cumulative Element Distribution,
pet finer than Bottom Size
As
99.2
87.5
68.7
46.7
39.9
35.3
32.0
29.8
0.0
Cd
99.7
91.9
85.1
75.1
69.3
65.4
62.7
60.2
0.0
Cr Cu
99.8 99.6
95 4 88.6
80.3 76.4
67.7 63.5
64.4 57.6
62.2 55.2
60.5 52.3
59.4 50.2
0.0 0.0
Fe
99.8
95.5
61.9
30.4
27.2
25.4
23.8
22.6
0.0
Pb
99 4
96.3
90.8
86.3
79.5
72.7
64.8
59.3
0.0
Hg
99 5
86.5
73.0
61 .0
54.8
52.2
49.8
48.1
0.0
Ni
99 7
95 5
82.9
70.2
65.3
61.1
57.5
55.1
0.0
Zn
99 4
97.1
86.8
80.7
77.1
73.3
70.1
67.9
0.0
Cumulative Analysis of Coarse Fraction
Cumulative Analysis of All Fractions
Coarser than Bottom Size, mg/kg
As
16.6
3.0
2.7
3.2
3.2
3.2
3.2
3.2
4.0
Cd
7.0
2.1
1.4
1.6
1.7
1.8
1.9
2.0
4.3
Cr Cu
180 226
42 80
66 61
74 64
72 66
71 65
71 66
72 68
1 54 1 20
Fe
13,600
4,300
13,400
16,700
15,500
14,800
14,400
14,400
16,300
Pb
133
10
10
10
13
16
20
22
48
Hg
0.78
0.26
0.19
0.19
0.19
0.19
0.19
0.19
0 32
Ni
97
19
27
31
33
34
36
37
72
Zn
610
36
61
60
64
69
74
78
212
                          A-20

-------
Table A13. - Size and Element Distributions in Sagmaw River Sediment #1

Size
Tyler Mesh
+ 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400

Bottom
fjm
425
300
212
150
106
75
53
38


Weight
pet
3.3
14.1
32.1
10.1
5.6
2.7
2.6
3.5
26.0
a
Cumulative
pet finer
96.7
82.6
50.5
40.4
34.8
32.1
29.5
26.0
0.0
. - Analysis of Individual Size Fractions
Element Analysis of Size Fraction, mg/kg
As
10.1
3.0
5.6
6.2
4.9
2.7
5.6
9.2
9.8
Cd
0.5
0.3
0.3
0.4
0.5
0.8
0.8
1.1
2.3
Cr Cu
89 70
51 36
92 67
95 71
81 64
28 21
37 38
42 49
77 74
Fe
23,500
13,100
25,000
25,000
20,100
4,600
6,200
8,300
25,200
Pb
23
17
16
16
23
47
63
50
90
Hg
0.09
0.12
0.11
0.12
0.1 1
0.12
0.16
0.17
0.36
Ni
33
20
31
32
29
12
19
27
42
Zn
67.0
13.8
16.1
32.5
48.7
89.2
174
214
695
b. - Mass and Element Distribution
Size
Fraction
Tyler Mesh
+ 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400
Bottom
Size
//m
425
300
212
150
106
75
53
38

Weight
Dist'n
pet
3.3
14.1
32.1
10.1
5.6
2.7
2.6
3.5
26.0
Cumulative
Wt. Dist'n
pet finer
96.7
82.6
50.5
40.4
34.8
32.1
29.5
26.0
0.0
Element Distribution

As
5.1
6.5
27.6
9.6
4.2
1.1
2.2
4.9
38.9
c. - Cumulative
Size
Fraction
Tyler Mesh
+ 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400

Size
Fraction
Tyler Mesh
+ 35
35x48
48x65
65x100
100x150
150x200
200x270
270x400
-400
Bottom
Size
/vm
425
300
212
150
106
75
53
38


Bottom
Size
//m
425
300
212
150
106
75
53
38

Weight
Dist'n
pet
3.3
14.1
32.1
10.1
5.6
2.7
2.6
3.5
26.0

Weight
Dist'n
pet
3.3
14.1
32.1
10.1
5.6
2.7
2.6
3.5
26.0
Cumulative
Wt. Dist'n
pet finer
96.7
82.6
50.5
40.4
34.8
32.1
29.5
26.0
0.0
d. -
Cumulative
Wt. Dist'n
pet finer
96.7
82.6
50.5
40.4
34.8
32.1
29.5
26.0
0.0

Cd
1.8
4.6
10.6
4.4
3.1
2.4
2.3
4.2
66.7
Mass and

Cr Cu
3.8 3.7
9.3 8.2
38.4 34.6
12.5 11.5
5.9 5.8
1.0 0.9
1.2 1.6
1.9 2.8
26.0 31.0

Fe
3.6
8.6
37.5
1 1 8
5.3
0.6
0.8
1.4
30.6
to Size Fraction, pet

Pb
1 9
6 1
13 1
4 1
3.3
3.2
4.2
4.5
59.6

Hg
1 6
9 4
19.6
6.7
3.4
1.8
2.3
3.3
51.9

Ni
3.5
9 0
31.7
10.3
5 2
1.0
1.6
3.0
34.8

Zn
1 1
0 9
2 5
1.6
1.3
1.1
2.1
3.6
85.9
Element Distributions
Cumulative Element Distribution,
pet finer than
As
94.9
88.4
60.9
51.3
47.1
46.0
43.8
38.9
0.0
Cd
98.2
93.6
83.0
78.6
75.5
73.1
70.9
66.7
0.0
Cr Cu
96.2 96.3
86.8 88.1
48.5 53.5
36.0 42.0
30.1 36.2
29.2 35.3
27.9 33.7
26.0 31.0
0.0 0.0
Fe
96.4
87.8
50.3
38.5
33.3
32.7
31.9
30.6
0.0
Bottom Size
Pb
98.1
92.0
78.9
74.8
71.5
68.2
64.1
59.6
0.0

Hg
98.4
89.0
69 4
62.7
59.3
57.5
55.2
51.9
0.0

Ni
96.5
87.6
55.9
45.6
40.4
39.4
37.8
34.8
0.0

Zn
98.9
98.0
95 6
94.0
92.7
91.6
89 4
85.9
0.0
Cumulative Analysis of Coarse Fraction


Cumulative
Analysis
of All Fractions
Coarser than Bottom Size, mg/kg
As
10.1
4.34
5.16
5.34
5.29
5.19
5.21
5.39
6.53
Cd
0.5
0.3
0.3
0.3
0.3
0.4
0.4
0.4
0.9
Cr Cu
89 70
58 42
80 58
83 61
83 61
80 59
79 58
77 58
77 62
Fe
23,500
15,100
21,500
22,100
21,900
21,200
20,700
20,100
21,400
Pb
23
18
17
17
17
18
20
21
39
Hg
0 09
0.11
0.11
0.11
0.11
0.1 1
0.11
0.12
0.18
Ni
33
22
28
29
29
28
28
28
31
Zn
67.0
23.9
18.8
21.2
23.5
26.1
31.6
40.2
210
                         A-21

-------
Table A 14. -Density Separation
of Saginaw
River Sediment 01
Contaminant Analyses
+ 65 mesh material
Density
Fraction
S.G.<1.9
1.92.9
Total
Wt Dist
pet
0.27
50.39
0.22
50.88
Analysis, mg/kg
As
29.8
2.90
8.24
3.07
Cd
3.0
0.4
2.0
0.4
Cr
60
40
90
40
Cu
120
8.0
430
10
Fe
19.700
4,410
42,600
4,660
Pb
150
4.0
670
8.0
Hg

0.40
0.13
0.12
0.13
Ni
98
7.0
92
8.0
Zn
450
21
560
26
65x200 mesh material
Density
Fraction
S.G.<1.9
1.92.9
Total
Wt Dist
pet
0.20
16.90
0.14
17.24
Analysis, mg/kg
As
33.7
3.60
6.94
3.97
Cd
4.0
1.0
10
1.1
Cr
130
74
220
76
Cu
320
26
500
33
Fe
21,200
10,400
51,300
10,900
Pb
180
30
1050
40
Hg

0.60
0.53
2.2
0.54
Ni
90
18
64
19
Zn
1200
140
540
160
200x500 mesh material
Density
Fraction
S.G.<1.9
1.92.9
Total
Wt Dist
pet
0.07
3.24
0.02
3.33
Analysis, mg/kg
As
34.2
2.04
NSR
2.69
Cd
4.0
0.5
9.0
0.6
Cr
170
24
420
29
Cu
190
19
540
25
Fe
23,300
5,930
91,000
6.690
Pb
200
31
1620
42
Hg

NSR
0.11
NSR
0.1 1
Ni
96
13
210
16
Zn
1300
100
620
130
-500 mesh material
Density
Fractions
S.G.<2.9
S.G.>2.9
Total
Wt Dist
pet
28.51
0.05
28.55
As
8.95
NSR
8.94
Cd
2.0
8.0
2.0
Cr
75
260
75
Contaminant
Analysis, mg/kg
Cu Fe
75
560
76
26,800
150,000
27,000
Pb
87
670
88
Hg

0.22
NSR
0.22
Ni
41
570
42
Zn
700
1200
700
Distributions
+ 65 mesh material
Density
Fraction
S.G.0.9
1.92.9
Total
Wt Dist
pet
0.27
50.39
0.22
50.88
Distribution, pet
As
1.64
29.91
0.38
31.93
Cd
0.81
20.17
0.45
21.42
Cr
0.29
35.96
0.36
36.60
Cu
0.96
12.03
2.86
15.85
Fe
0.44
18.25
0.78
19.47
Pb
1.08
5.40
4.00
10.48
Hg
0
28
0
29

.48
.95
.12
.54
Ni
1.33
17.84
1.04
20.21
Zn
0.50
4.33
0.51
5.34
65x200 mesh material
Density
Fraction
S.G.<1.9
1.92.9
Total
Wt Dist
pet
0.20
16.90
0.14
17.24
As
1.35
12.45
0.20
14.01
Cd
0.78
16.91
1.43
19.13
Cr
0.45
22.31
0.56
23.33
Distribution, pet
Cu Fe
1.87
13.11
2.13
17.11
Pb
0.34 0.94
14.44 13.58
0.60 4.02
15.38 18.54
Hg
0.
39.
1.
41.

52
29
36
16
Ni
0.89
15.38
0.46
16 74
Zn
0.96
9.69
0.32
10.97
200x500 mesh material
Density
Fraction
S.G.<1.9
1.92.9
Total
Wt Dist
pet
0.07
3.24
0.02
3.33
Distribution, pet
As
0.48
1.35
0.00
1.84
Cd
0.27
1.62
0.14
2.04
Cr
0.21
1.39
0.12
1.71
Cu
0.39
1.84
0.25
2.48
Fe
0.13
1.58
0.12
1.83
Pb
0.37
2.69
0.68
3.74
Hg

0.00
1.58
0.00
1.58
Ni
0.33
2.13
0.17
2.63
Zn
0.37
1.33
0.04
1.73
A-22

-------



Table A14
.-Density Separation of Saginaw
River Sediment #1
-500 mesh material
Density
Fractions
S.G.<2
S.G.>2
Total
.9
.9
Wt Dist
pet
28.51
0.05
28.55
As
52.23
0.00
52.23
Cd
57.05
0.37
57.41
Cr
38.14
0.21
38.35
Distribution, pet
Cu Fe Pb
63.79
0.76
64.55
62.75
0.56
63.32
66.41
0.82
67.23
Hg
27.72
0.00
27.72
Ni
59.10
1.32
60.42
Zn
81.73
0.22
81.95
A-23

-------
Table A15.--Water Elutnation of Saginaw River Sediment
a. - Contaminant Analyses
Elutnation
Product
+ 35 Lights
+ 35Heavies
35x48 Lights
35x48Heavies
48x65Lights
48x65Heavies
65x100Lights
65x100Heavies
100x1 SOLights
100x1 SOHeavies
150x200Lights
1 50x200Heavies
200x270Lights
200x270Heavies
270x400Lights
270x400Heavies
-400
Total
TotalLights
TotalHeavies
Wt Dist
pet
0.18
2.98
0.17
18.61
0.1 1
27.03
0.31
10.75
0.41
4.77
0.32
2.41
0.53
2.06
0.50
1.24
27.64
100.00
2.52
69.84
Analysis, mg/kg
As
74
5.3
39
3.8
42
3.0
20
4.0
18
1 7
31
2.9
18
2.6
15
2.9
8.1
5.2
26
3.3
Cd
2.0
0.8
4.0
0.3
4.7
0.3
2.8
0.4
2.2
4.3
2.4
0.6
2.1
0.6
1.8
1.9
2.8
1.3
2.4
0.66
Cr
86
44
120
57
160
40
97
42
79
31
94
51
83
44
67
59
61
51
88
45
Cu
140
17
180
53
481
39
201
46
208
31
340
45
346
94
296
1 18
630
213
276
45.6
Fe
23,500
7,300
25,700
19,800
25,900
14,400
14,600
14,800
14,600
4,850
21,000
6,440
18,300
8,000
16,200
9,370
12,900
14,100
18,400
14,400
Pb
260
21
68
20
270
20
95
50
77
23
130
47
100
35
78
35
59
38
1 10
27
Ni
110
14
130
24
170
17
740
18
62
9
78
12
61
15
49
19
33
26
160
18
Zn
430
50
490
12
760
11
390
40
430
580
600
74
620
94
470
270
510
200
510
65
b - Contaminant Distribution
Elutnation
Product
+ 35 Lights
+ 35Heavies
35x48Lights
35x48Heavies
48x65Lights
48x65Heavies
65x100Lights
65x100Heavies
100x150Lights
100x1 SOHeavies
150x200Lights
150x200Heavies
200x270Lights
200x270Heavies
270x400Lights
270x400Heavies
-400
Total
TotalLights
TotalHeavies
Wt Dist
pet
0.18
2.98
0.17
18.61
0.11
27.03
0.31
10.75
0.41
4.77
0.32
2.41
0.53
2.06
0.50
1.24
27.64
100.00
2.52
69.84

As
2.5
3.0
1.2
13.5
0.9
15.3
1.2
8.1
1 .4
1 5
1.9
1.3
1.8
1.0
1.4
0.7
43.1
100.0
12.4
44.5

Cd
0.3
1.8
0.5
4.3
0.4
6.3
0.7
3.3
0.7
15.8
0.6
1.1
0.9
1.0
0.7
1.8
59.8
100.0
4.7
35.5

Cr
0.3
2.6
0.4
20.9
0.4
21.3
0.6
8.9
0.6
2.9
0.6
2.4
0.9
1.8
0.7
1.4
33.3
100.0
4.4
62.3
Distribution
Cu
0.1
0.2
0.1
4.6
0.3
5.0
0.3
2.3
0.4
0.7
0.5
0.5
0.9
0.9
0.7
0.7
81.8
100.0
3.3
14.9
, pet
Fe
0.3
1.5
0.3
26.2
0.2
27.6
0.3
11.3
0 4
1.6
0.5
1 .1
0.7
1.2
0.6
0.8
25.3
100.0
3.3
71.4

Pb
1.2
1.7
0.3
9.9
0.8
14.4
0.8
14.3
0.8
2.9
1 1
3.0
1.4
1.9
1.0
1.1
43.3
100.0
7.5
49.2

Ni
0.8
1.6
0.8
17.2
0.7
18.2
9.0
7.6
1.0
1.6
1 0
1.2
1.3
1.2
1.0
0.9
35.1
100.0
15.4
49.5

Zn
0.4
0.7
0.4
1.1
0.4
1.5
0.6
2.2
0 9
13 9
1 0
0.9
1.6
1.0
1.2
1.7
70.5
100.0
6.5
23.0
                    A-24

-------
Table Al 6.--Magnetic Separation of Saginaw River Sediment tn
a. - Element Analyses
+ 65 mesh material - dry high-intensity induced roll separator
Product
0.5 amp
2 amp
non-mag
Total
Wt Dist
pet
0.6
0.8
48.5
49.9

As
19
14
2.9
3.3

Cd


2.0
1.0
0.3
0.3
65x200 mesh material - dry high-intensity induced
Product
0.5 amp
2amp
non-mag
Total
200x500 mesh
Product
1 amp
5amp
non-mag
Total
Wt Dist
pet
0.9
0.5
16.6
18.0
material - wet
Wt Dist
pet
0.3
0.2
3.7
4.1
-500 mesh material - wet high
Product
1 amp
5amp
non-mag
Total

Wt Dist
pet
0.5
0.8
26.8
28.0


As
7.9
11
2.8
3.3
high-intensity

As
1 1
9.0
3.8
4.5

Cd
roll


1.0
1.0
0.3
0.4

Cr
220
140
59
62
separator

Cr
130
130
67
72
Element
Cu
1110
330
59
76

Element
Cu
470
340
76
103
Analysis,
Fe
46,000
37,300
17,600
18,300

Analysis,
Fe
33,100
18,400
20,200
20,800
mg/kg
Pb
680
470
53
67

mg/kg
Pb
260
260
57
73

Hg
0.68
0.20
0.12
0.13


Hg
0.27
0.21
0.12
0.13

Ni
160
130
29
32


Ni
100
67
31
35

Zn
290
240
10
17


Zn
530
390
140
167
separator

Cd


2.0
3.0
1.0
1.1

Cr
130
84
21
30
Element
Cu
170
220
26
43
Analysis,
Fe
50,900
25,500
4,050
7,900
mg/kg
Pb
960
840
300
363

Hg
0.51
0 39
0.14
0.17

Ni
120
64
14
23

Zn
1080
860
720
749
-intensity separator

As
15.1
13,1
10.1
10.3


Cd


3.0
2.0
2.0
2.0



Cr
230
120
72
76
b. -
Element
Cu
410
290
85
96
Distribution
Analysis,
Fe
92,400
71,500
25,400
27,800

mg/ky
Pb
420
220
1 10
1 18


Hg
0.84
0.54
0.28
0.30


Ni
180
100
45
49


Zn
1080
860
720
730

-f 65 mesh material - dry high-intensity induced roll separator
Product
0.5 amp
2 amp
non-mag
Total
Wt Dist
pet
0.6
0.8
48.5
49.9

As
2.1
2.0
26.6
30.8

Cd


1.4
0.9
17.3
19.6
65x200 mesh material - dry high-intensity induced
Product
0.5 amp
2amp
non-mag
Total
200x500 mesh
Product
1 amp
5amp
non-mag
Total
Wt Dist
pet
0.9
0.5
16.6
18.0
material - wet
Wt Dist
pet
0.3
0.2
3.7
4.1

As
1.3
1.1
8.8
11.2
high-intensity

As
0.6
0.3
2.7
3.5

Cd
1
0
5
7
roll


.1
.6
.9
.6

Cr
2.0
1.7
43.0
46.6
separator

Cr
1.8
1.0
16.7
19.5
Element
Cu
7.8
3.0
33.6
44.5

Element
Cu
5.0
2.1
14.8
21.9
Distribution, pet
Fe
1.3
1.4
40.7
43.4

Pb
4.3
3.9
27.1
35.3

Hg
2.3
0.9
32.8
36.0

Ni
2.6
2.8
38.0
43.3

Zn
0.6
0.7
1.8
3 1

Distribution, pet
Fe
1.4
0.5
16.0
17.9
Pb
2.5
1 4
10.0
13.9
Hg
1.4
0.6
1 1.2
13.2
Ni
2.4
0.9
13.9
17.3
Zn
1.7
0 7
8.5
1 1.0
separator

Cd
0
0
4
5


.6
.6
.4
.6

Cr
0.5
0.2
1.2
1.9
Element
Cu
0.5
0.4
1.1
2.1
Distribution, pet
Fe
0.6
0.2
0.7
1.6
Pb
2.7
1.4
11.7
15.9
Hg
0.8
0.3
2.9
4.0
Ni
0.9
0.3
1.4
2.5
Zn
1.1
0.5
9.8
11.3
                   A-25

-------
Table A16.-Magnetic Separation of Saginaw River Sediment #1
-500 mesh material - wet high-intensity separator

Product
1 amp
Samp
non-mag
Total
Wt Dist
pet
0.5
0.8
26.8
28.0

As
1.3
1.9
51.2
54.5

Cd
1.7
1.8
63.6
67.1

Cr
1.6
1.4
29.0
32.0
Element
Cu
2.3
2.6
26.7
31.6
Distribution
Fe
2.1
2.6
32.4
37.1
. pet
Pb
2.1
1.8
31.1
34.9

Hg
2.2
2.3
42.2
46.8

Ni
2.3
2.1
32.5
36.9

Zn
1.8
2.4
70.4
74.6
                   A-26

-------
Table A17.-Froth Flotation of Saginaw River Sediment

Flotation
Conditions
pH 4
No Reagents








Oleic Acid
1.3 Ib/ton
pH 7







Oleic Acid
1 .5 Ib/ton
pH 10







Copper Sulfate
and Potassium
Amyl Xanthate
0.25 Ib/ton
pH 10
(Note: Cu analyses
omitted due to
theaddition of
copper sulfate)

Copper Sulfate
and Potassium
Amyl Xanthate
0.25 Ib/ton
pH7
(Note: Cu analyses
omitted due to
theaddition of
copper sulfate)



Element
Weight
As
Cd
Cr
Cu
Fe
Pb
HO
Ni
Zn
Weight
As
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Zn
Weight
As
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Zn
Weight
As
Cd
Cr

Fe
Pb
Hg
Ni
Zn
Weight
As
Cd
Cr
Fe
Pb
Hg
Ni
Zn
Initial
Analysis,
mg/kg
NA
8.5
1.2
95
63
22900
36
0.14
34
206

10.3
2.0
60
70
29700
100
0.47
37
206

10.3
2.0
60
70
29700
100
0.47
37
206

10.3
2.0
60

29700
100
0.47
37
206

10.3
2.0
60
29700
100
0.47
37
206
Concentrate
Analysis,
mg/kg
(22.60 g)
8.5
1.2
95
63
22900
36
0.14
34
206
(71.84g)
15.3
3.0
120
100
17700
110
0.25
54
833
(15. 4 g)
14.7
3.0
110
100
24800
120
0.58
53
804
(19.9 g)
26.1
5.0
200

35700
200
0.57
86
1460
(26.84 g)
21.2
4.0
150
33800
140
0.51
66
1190
Distribution,
pet
6.12
24.76
20.68
20.68
26.10
24.22
34.93
26.69
25.14
35.32
18.37
56.41
40.30
41.54
60.00
41.91
69.23
31.91
48.31
73.09
4.58
15.31
12.58
10.09
15.58
10.81
20.74
5.96
12.38
17.49
5.58
37.88
22.82
23.26

22.36
45.79
16.64
24.11
43.52
7.10
30.23
23.42
26.38
24.18
36.03
26.17
25.17
39.05
Tails
Analysis,
mg/kg
(346.7 g)
3.7
1.0
35
24
7200
17
0.13
15
148
(319. 3 g)
2.7
1.0
38
15
5520
1 1
0.12
13
69
(321. Og)
3.9
1.0
47
26
9820
22
0.44
18
182
(336. 6 g)
2.5
1.0
39

7330
14
0.17
16
112
(351.1 g)
3.7
1.0
32
8100
19
0.11
15
142
Distribution,
pet
93.88
75.22
79.32
79.32
73.90
75.78
65.07
73.31
78.24
67.60
81.63
43 59
59.70
58.46
40.00
58.09
30.77
68.09
51.69
26.91
95.42
84.69
87.42
89.91
84.42
89.19
79.26
94.04
87.62
82.51
94.42
62.12
77.18
76.74

77.64
54.21
83.36
75.89
56.48
92.90
69.77
76.58
73.62
75.82
63.97
73.83
74.83
60.95
                       A-27

-------
Table A18.-Acidic Attrition Scrubbing of Sagmaw River Sediment #1
a. - Hydrochloric Acid
1 . - Contaminant Analyses
Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
99.5
0.5
100.00
As
6.95
69.70
<0.0006
7.29
Cd
0.60
<1.00
< 0.003
0.60
Analysis, mg/kg
Cr Cu Fe
73 60 18,600
46 20 291,000
<0.03 <0.03 0.10
73 60 20,100
Pb
28
<60
<0.16
28
Hg
0.15
0.23
<0.012
0.15
Ni
30
476
<0.03
32
Zn
192
2600
<0.005
205
2. - Contaminant Distributions
Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
99.5
0.5
100.0
As
94.8
5.2
100.0
Cd
100.0
0.0
100.0
Distribution, wt pet
Cr Cu Fe
99.7 99.8 92.1
0.3 0.2 7.9
100.0 100.0 100.0
Pb
100.0
0.0
100.0
Hg
98.9
1.0
100.0
Ni
92.0
8.0
100.0
Zn
93.1
6.9
100.0
b. - Nitric Acid
1 . - Contaminant Analyses
Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
99.5
0.5
100.0
As
6.94
126
<0.0006
7.57
Cd
0.50
2.00
<0.003
0.51
Analysis, mg/kg
Cr Cu Fe
74 59 18,600
30 30 270,000
<0.03 <0.03 0.21
74 59 19,900
Pb
30.0
<70
<0.16
29.8
Hg
0.17
0.41
<0.012
0.17
Ni
30
484
<0.03
32
Zn
202
2770
<0.005
216
2. - Contaminant Distributions
Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
99.5
0.5
100.0
As
91.2
8.8
100.0
Cd
97.9
2.1
100.0
Distribution, wt pet
Cr Cu Fe
99.8 99.7 92.8
0.2 0.3 7.2
100.0 100.0 100.0
Pb
100.0
0.0
100.0
Hg
98.7
1.3
100.0
Ni
92.1
7.9
100.0
Zn
93.2
6.8
100.0
c. - Sulfuric Acid
1 . - Contaminant Analyses
Attrition
Product
Sohds
Precipitate
Filtrate
Total
Wt Dist
pet
99.6
0.4
100.0
As
6.91
155
0.001
7.47
Cd
0.50
<1.0
<0.003
0.50
Analysis, mg/kg
Cr Cu Fe
55 44 14,000
98 20 322,000
<0.03 <0.03 0.22
55 44 15,200
Pb
25
<70
<0.16
25
Hg
0.13
1.06
<0.012
0.13
Ni
23
614
<0.03
25
Zn
190
2300
<0.005
198
2. - Contaminant Distributions
Attrition
Product
Solids
Precipitate
Filtrate
Total
Wt Dist
pet
99.6
0.4
100.0
As
92.1
7.9
100.0
Cd
100.0
0.0
100.0
Distribution, wt pet
Cr Cu Fe
99.3 99.8 91.9
0.7 0.2 8.1
100.0 100.0 100.0
Pb
100.0
0.0
100.0
Hg
97.0
3.0
100.0
Ni
90.8
9.2
100.0
Zn
95.6
4.4
100.0
                        A-28

-------
MINERALOGY OF CONTAMINATED GREAT LAKES SEDIMENTS

               Ashtabula River, Buffalo River,
             Indiana Harbor/Grand Calumet River,
                 and Saginaw River and Bay
                    Areas of Concern

                           by

                      Denise Chirban

           United States Department of the Interior
                     Bureau of Mines
               Salt Lake City Research Center
                    729 Arapeen Drive
                 Salt Lake City, Utah 84108
                           B-1

-------
                                          INTRODUCTION

The objective of mineralogical analysis of contaminated sediment samples is to identify those phases that
contain the metallic contaminants of interest. In this study, host phases for arsenic,cadmium, chromium,
copper, iron, lead, mercury, nickel, and zinc were sought using microscopy techniques commonly used for
evaluation of rocks and minerals.

Both naturally-occurring minerals and anthropogenic substances were found to contain the toxic elements
mentioned above. An important distinction is made between naturally-occurring phases and those which
appear to be of human origin. Both of these will contribute to the measured metal concentration of the
sediment, and it is often impossible for the mineralogist to determine how much of a  measured metal content
is the result of man's activities rather than natural background levels. The naturally-occurring metal-bearing
phases found in river sediment are usually very insoluble, and are  probably not available for uptake by living
organisms. In many cases the contaminants were found at concentrations of a few hundred parts-per-million
(mg/kg) or less, so the search for their contaminant-bearing phases was an laborious (and often fruitless)
task.  Since metal ions have the capacity to adsorb individually on  the surfaces of many materials found in
sediment, a sample may have elevated metal levels, and yet reveal no discrete metal-bearing phases.  It is
also possible  that much of the apparent metal contamination is from naturally-occurring minerals.  A glossary
of mineral  names, classes, and formulas in provided in table B-1.

As with much of the experimental work, sediment samples for mineralogical analysis were prepared by
separating the sediment into discrete size fractions by sieving.  The purpose for this  is to identify the changing
consist of the sediment with changes in grain size. A small, dry aliquot of each size  fraction was placed in a
mold and set  into epoxy resin; this resin mount was then polished on one surface to reveal cross-sections of
the sediment  particles. These epoxy mounts werei nspected directly under the optical microscope to
recognize gross features, and were also used for the scanning-electron microscope  (SEM) inspections   To
assist the SEM in identifying textural detail, an thin gold coating was applied to each polished surface by
sputtering.  Estimates regarding individual particle size, shape, degree of liberation,  average atomic number
(density) and  texture were made by visual inspection of the SEM image. Backscattered electron (BSE)
imaging was used to discriminate between low  and high atomic number phases  (Goldstein, Newbury, Echlin,
Joy, Fiori,  and Lifshin, 1981)  Simultaneous analysis of elements with atomic number greater than sodium
was available by EDS to provide an estimate of the element's abundance in the sample.  This EDS analysis
is used as an aid to the mineralogist in identifying phases, and is not meant to quantitatively determine
contaminant levels.

                                     RESULTS AND DISCUSSION

Ashtabula River Sediment

The naturally-occurring minerals found in Ashtabula River sediment, in decreasing order of abundance, were
quartz, feldspar, calcite, dolomite,  clay, iron oxide, mica/sericite, siderite, pyrite (both framboidal and
nonframboidal), pyroxene, ilmenite, zircon, chalcopyrite, ferriferous sphalerite,  galena, barite, apatite, iron
phosphate, titanium dioxide, arsenopyrite, monazite and xenotime. Materials other than  naturally-occurring
minerals found in the sediment include wood bits and other organic matter, shells, fibers, paint chips, metal
(including  brass  steel, aluminum,  and jewelry),  slag fragments, silver, calcium and lead chlorides  (perhaps
these are naturally-occurring minerals), titanium chloride, and transparent rods composed of calcium,
aluminum, and silicon

The 48x65 mesh size fraction contains more organic material than minerals, while the reverse is true in the
270x400 and minus 400 mesh size fractions. The 100x150 and 150x200 mesh size fractions contain
approximately equal amounts of minerals and organic  material. Only a small amount of  contaminant-bearing
material was  found in each size fraction.
                                                 B-2

-------
Iron occurs in pyrite, chalcopyrite, arsenopyrite, and other iron sulfide minerals, iron oxides, mica/sericite,
pyroxene, ilmenite, siderite, sphalerite and iron phosphate. There is a sulfur-bearing organic-like material that
also contains iron and was found in traces in all size fractions.  Figure B-1  represents this low-atomic-number
material which contains calcium, iron, sulfur, titanium, chromium, copper, and zinc. Traces of iron-bearing
slag were found in all size fractions.  Usually this slag contains other metals as well, as in Figure B-2, which
shows a glassy slag particle containing iron, copper, silicon, and zinc. Occasionally iron was found as
metallic fragments, and with zinc in sphalerite.

Copper occurs in chalcopyrite, slag, brass, a sulfide of iron, copper and zinc, and in the low-atomic-number
material that resembles organic matter

Zinc was found in naturally-occurring sphalerite, slag fragments, organic material,  brass, zinc metal or oxide,
and a sulfide of zinc, iron, and copper.

Slag fragments, organic material, and metallic particles were found to be chromium-bearing.

Lead occurs in galena, lead oxide and/or metallic lead, lead-bearing paint chips, a chloride of titanium and
lead, and a phosphate of lead and  calcium. Figure B-3 is a photomicrograph of a fine-grained lead-bearing
calcium phosphate found in the 270x400 mesh fraction.  Free grains of lead-bearing phases were found in
both the 270x400 and the minus 400 mesh material.

Only two nickel-bearing phases were found  In one phase nickel occurs with iron and chromium in stainless
steel. Nickel, copper,  and gold are components of the other phase, possibly from jewelry

Arsenopyrite, a natural mineral, was the only arsenic-bearing phase found in this sediment.

Buffalo River Sediment

Quartz, plagioclase, potassium-feldspar,  calcite, iron oxide, mica, pyrite, clay, pyroxene,  sphalerite,
chalcopyrite, ilmenite,  barite, zircon, and  monazite  are the naturally-occurring  minerals found (in decreasing
order of abundance) in the sediment.  Wood and other organic material, fabric, shells, slag, glass, and paper
are also constituents of the  sediment.

Particles  of vuggy metallic and smooth vitreous slags are the major unnatural  contributors to the iron
concentration of this sediment.  The slag particles are more abundant in the coarser size fractions.  Iron,
chromium, copper, nickel, zinc, aluminum, sulfur, potassium, calcium, and titanium are constituents of the
slag. Often the slag contains inclusions of silicates, iron  oxides, and sulfides of copper and/or iron.  Iron
oxide, mica, pyrite, pyroxene, ilmenite, and chalcopyrite are the minerals that contribute to the iron
concentration in the sediment.

Copper was found as metallic grains, chalcopyrite, and in trace amounts in the slag fragments

The zinc-containing phases found in this  sample were1 1) a dense phase composed of copper, iron  and zinc,
possibly a metal, 2) and an  occasional trace of zinc found in some of the slag  particles

Chromium and nickel were found only in the slag particles mentioned above

No lead, arsenic, cadmium, or mercury phases were detected in this sample.  Again, the cadmium could be a
component of the mineral sphalerite.

Indiana Harbor Canal Sediment

The majority of this sediment sample is composed  of slag fragments and naturally-occurring minerals. Both

                                                 B-3

-------
vesicular irregularly-shaped slag fragments and smooth, rounded slag particles occur. The vesicular slag has
fairly low density and is composed mostly of sulfur with traces of aluminum, silicon, calcium, potassium,
titanium, iron, copper, chromium, and zinc Inclusions of iron oxide, mica, potassium-feldspar,  pyrite, and
quartz were found in the slag fragments. The smooth, rounded  slag particles are more abundant than the
vesicular type and contain sulfur with lesser amounts of vanadium and occasionally copper and zinc.

Iron oxide, quartz, feldspar,  calcite, apatite, pyrite (both framboidal  and nonframboidal), pyroxene, amphibole,
mica, iron phosphate, and clay are the naturally-occurring  minerals found in the sediment. Fragments of
bone, wood, seeds, steel, bronze, wire, shells, aluminum foil, paint, and plastics were found in  addition to the
slag  particles and naturally-occurring minerals.

Iron concentration in the Indiana Harbor sediment is the highest of  all the sediment locations studied and is
the result of the presence of slag particles, iron oxides, pyrite, pyroxene, amphibole, iron phosphate, mica,
and steel fragments.  Due to the proximity of this site to iron ore shipping lanes and steel  manufacturers, a
high  concentration of iron can be expected.

Copper, zinc, and chromium were  identified only in  the anthropogenic slag particles and bronze shavings No
naturally-occurring minerals containing these elements were found.

Lead is concentrated in the  finer fractions.  Lead phases were only found in the -400  mesh size fraction
where the concentration of lead is  the highest  Figure B-4 represents a lead and iron-bearing phase. A lead
and tellurium-bearing phase was also found in the -400 mesh sediment.

Barium was found in  naturally-occurring barite  No other barium-containing phases were found.

No discrete phases containing arsenic, nickel, cadmium, mercury,  silver, antimony, or selenium were found  in
the sediment sample examined from Indiana Harbor Ship  Canal.

Saginaw River Sediment

Two samples were examined from the Saginaw River area, SR #1  and SR #2. The naturally-occurring
minerals found in both sites, in decreasing order of abundance, include quartz, potassium-feldspar,
plagioclase, calcite, iron oxides, iron phosphates, pyrite, ilmenite, titanomagnetite, chalcopynte, apatite, clays,
mica, galena, sphalerite, gypsum,  barite, zircon, monazite, pentlandlte, tetrahedrite/tennantlte, scheellte, and
rutile.  Shells, wood,  aluminum foil, paint chips, glass, slag, bone, plastic, steel,  bronze, and wire were also
found

In SR #1, a considerable number of  iron phases were found  The  most abundant  of these were slag
fragments, iron oxides, pyrite, and metallic particles The iron oxide occurs as discrete grains, as coatings on
inorganic minerals, and as mineral inclusions in slag Occasionally pyrite and sphalerite  were found as
inclusions in wood (Figure 9, mam text) and quartz  Two types of vesicular nonmagnetic slag  occur, a silica-
rich  glassy phase similar to Figure B-2, and a iron-rich metallic  phase  Varying concentrations of aluminum,
sulfur, silicon, calcium, titanium and  iron were found in the slag particles. In addition  to iron oxides and pyrite,
pyroxene, ilmenite, and to a lesser degree, mica are the major  iron-bearing minerals.

Copper was found in three  phases  (1) in its  native state or as  an  oxide; (2) as a magnesium,  sodium,
aluminum, silicon, iron, nickel, copper, zinc chloride; and  (3) as bronze metal.

Zinc occurs as metallic grains with iron, lead, copper or tin and in the sulfide mineral  sphalerite.

Lead was found to be concentrated  in the finer size fractions as micron-sized metallic or sulfur-bearing
 phases. Some lead  was found as metallic grains with zinc, copper and/or iron.  As depicted in Figure 10 in
the preceding text, the lead-bearing  particles are abundant and on the order of 1-2 microns in  diameter

                                                 B-4

-------
Chromium was found in iron-rich slag and/or metal fragments containing iron, copper, and/or zinc

No arsenic, cadmium,  mercury, or nickel phases were found, although the cadmium may occur naturally in
sphalerite by substitution of cadmium for zinc

In SR #2, iron compounds were again the most abundant contaminant-bearing phase found. Metallic iron
phases containing chromium, nickel, and/or vanadium were found in the sediment (Figure B-5). Slag
fragments containing iron, zinc, titanium, aluminum, potassium, calcium, silicon,  and/or sulfur in varying
concentrations were also identified.  Pyrite inclusions in wood, free pyrite grains (both framboidal and
nonframboidal), iron phosphate, chalcopyrite, iron silicates  and iron oxide minerals also contribute to the iron
concentration in the sediment sample.  Often the iron phosphate was found to contain small amounts of
chromium, copper, nickel, and zinc

A dense zinc-bearing phase corresponding to zinc metal  or zinc oxide was found as inclusions in calcite, or
cemented by calcite to an iron silicate, and as free particles. Zinc was also a constituent of the slag particles
mentioned above.  The naturally-occurring zinc minerals  sphalerite and tetrahedrite were also found.

Nickel and chromium are constituents of steel, slag, and  a transition metal-bearing phosphate (Figure B-6).
These metals are also present in the natural  mineral pentlandite, a sulfide of nickel and iron.

Copper was found only in naturally-occurring chalcopyrite and tetrahednte/tennantite  The
tetrahedrite/tennantite  solid solution series is a natural sulfosalt containing arsenic and/or antimony, iron, zinc,
copper, and sulfur. Atomic substitutions into this mineral cause variations in the  chemical formula.  It is in the
tetrahedrite/tennantite  that some or all of the arsenic and antimony is found

Lead is concentrated in both the coarsest and finest size fractions. Inclusions of galena that were found in
the larger size  fractions became liberated in the  finer fractions (Figure B-7)

No cadmium or mercury phases were detected,  although cadmium and mercury  can  substitute into some of
the minerals found in this sample.  Mercury can  substitute into tetrahedrite/tennantite, and cadmium can
replace zinc in zinc-bearing phases.  Perhaps cadmium and mercury are constituents of these  phases, but
are below the detection limits of the EDS system used.

                                          CONCLUSIONS

It is apparent from the  examination of these samples that some material has an affinity for adsorption of the
metallic contaminants, since the phases identified do not account for all of the metallic content of these
samples. The  possibility of adsorbtion as a mode of occurence for metallic contaminants is discussed  in the
main text of this document in connection with specific surface area measurements Some (apparently)
precipitated phases were also identified, although it is impossible to determine if they are natural or
anthropogenic.  For example, phosphates containing  lead (as in Figure B-2) or chromium, iron, copper, zinc
and nickel (as in  Figure B-6) were common.  The identity of this phosphate phase(s)  is not known, but is in
some cases similar in composition to the shells of aquatic mollusks Perhaps metal ions  have  become
adsorbed or coprecipitated in the fairly porous matrix of animal's shell during growth  Organic materials
containing iron, chromium, zinc and/or copper as depicted in Figure B-1 were also frequently observed
Pyrite, iron oxide, titanium dioxide, lead-bearing  phases and chalcopyrite particles were often found trapped
within wood and other organic material.

Many of the metals are bound  in naturally-occurring minerals and will contribute to high metals concentration
levels, although not necessarily present as the result of man's activities.

Framboidal  pyrite is found in sedimentary rocks, and probably came to be in the sediment associated with
coal.  This type of pyrite is notorious for its high surface area and reactivity. The presence of framboidal

                                                 B-5

-------
pyrite (an example of which is shown in Figure B-8) is significant because it indicates either a) fairly recent
deposition of the pyrite-bearing material into the river, or b) anaerobic conditions preventing oxidation of the
very small grains.

                                            REFERENCE

Goldstein, J. I., D. E. Newbury, P. Echlin, D. C. Joy, C. Fiori, and E. Lifshm, 1981.  Scanning Electron
Microscopy and X-Ray Analysis, Plenum Press, New York, pp. 75-87.
                                                 B-6

-------
Figure B-1.—Calcium, iron, chromium, copper, zinc, titanium and sulfur bearing
organic(?) in the 48x65 mesh Ashtabula River sediment. 1000X, BSE.
Figure B-2.—Iron, zinc, copper and silicon bearing slag fragment from Ashtabula
River sediment, 1500X, BSE.
                                    B-7

-------
Figure B-3.—Photomicrograph of a lead-bearing calcium phosphate in the 270X400
mesh Ashtabula River sediment. 2000X, BSE
Figure B-4 —Photomicrograph of an iron- and lead-bearing particle in the -400 mesh
Indiana Harbor sediment. 4850X, BSE.

                                    B-8

-------
Figure B-5.—Photomicrograph of a steel particle (white) with clay, pyrite and calcite.
Saginaw River sediment #2, 400X, BSE.
Figure B-6.—Photomicrograph of a chromium, calcium, iron, copper, nickel and
zinc-bearing phosphate in -400 mesh Saginaw River sediment #2  3000X, BSE.
                                     B-9

-------
Figure B-7.—Photomicrograph of lead-bearing grains (indicated with arrows) in
-400 mesh Saginaw River sediment #2. 1125X, BSE.
Figure B-8.—Framboidal pyrite in the Ashtabula River sediment.  3000X, BSE.

                                    B-10

-------
                     Table B-1.—Glossary of Minerals Found in Sediment Samples
Mineral name
Class
Formula
amphibole
apatite
arsenopyrite
barite
calcite
chalcopyrite
chlorite
clay
 montmorillonite
 vermiculite
 illite
 kaolin
dolomite
feldspar
 albite
 plagioclase
ferriferous sphalerite
galena
gypsum
ilmenite
iron oxide
 goethite
 hematite
 magnetite
iron phosphate
 graftonite
 ludlamite
 strengite
 triphylite
 vivianite
mica
 muscovite
 sericite
pyrite
silicate
phosphate
sulfide
sulfate
carbonate
sulfide

silicate
carbonate
silicate
sulfide
sulfide
sulfate
oxide
oxide
phosphate
silicate
sulfide
a complex group with general formula
A0.1B2Y6Z8022(OH,F,CI)2
 A=Ca,Na,K
 B=Ca,Fe2M_i,Mg,Mn2t,Na
 Y=AI,Cr,Fe2t,Fe3t,Mg,Mn2*,Ti
 Z=AI,Si,Ti
Ca5(P04)3(F,OH,CI)3
FeAsS
BaS04
CaC03
CuFeS2
(Mg,Fe2*)2AI4Si2010(OH)4

(Na1Ca)033(Al1Mg)2Si4010(OH)2-nH20
(Mg,Fe,AI)3(AI,Si)4O10(OH)2-4H20
AI2Si205(OH)4
CaMg(C03)2
aluminosilicates of Na, K, Ca, or Ba
NaAISi3O6
(Na,Ca)AI(AI,Si)Si208
(Zn,Fe)S
PbS
CaSCV2H20
FeTi03

FeO(OH)
Fe203
Fe304

(Fe,Mn,Ca)3(PO4)2
Fe3(P04)2-4H20
FeP04-2H20
LiFeP04
Fe3(P04)2-8H20

KAI2(Si3AI)010(OH,F)2
 KAI3Si3010(OH)2-nH20
FeS2
                                              B-11

-------
                     Table B-1 —Glossary of Minerals Found in Sediment Samples
Mineral name
Class
Formula
pyroxene

quartz
rutile
scheelite
sidente
sphalerite
tennantite
tetrahedrite
titanium dioxide
titanomagnetite
xenotime
zircon
silicate         Mineral group of general formula: ABSi206
                A=Na,Ca,Mg,Fe2'
                B=Mg,Fe2',AI,Fe3'1Cr3',Mn2t,Sc
silicate         Si02
oxide          Ti02
oxide          CaW04
carbonate      FeC03
sulfide         ZnS
sulfide         (Cu,Fe)12As4S13
sulfide         (Cu,Fe)12Sb4S,3
oxide          Ti02
spinel          (Fe,Ti)3 04
phosphate      YP04
silicate	ZrSiQ4	
                                               B-12

-------
APPENDIX-QUALITY ASSURANCE PROJECT PLAN
                 C-1

-------
                         QUALITY ASSURANCE  PROJECT  PLAN

     CONTAMINATED GREAT LAKES SEDIMENT-METALS CHARACTERIZATION AND TREATMENT


                              U.S. BUREAU OF MINES

                         SALT LAKE CITY RESEARCH CENTER

                                 August 7,  1990
 Approved:
 J . y^P .  Allen, Principal Investigator
 S^lt Lake City Research Center


               t   /r
 D. D. ^famroargren ,  Laboratory (^ Officer
 Salt Lake City Research Center
 D. A. Rice, Project Manager
 Salt Lake City Research Center
 S. Yaksich, AR(|S E/T Working Group Chair

  />
                         9 — .
  .  Schumacher  ARCS QA Officer
 D.  G. Easterly, EMSL-LV |PA QA Officer
/R.  6hristensen, EPA Project Officer
 D. Cowgill, ARCS Frografft Manager

-------
                               TABLE OF CONTENTS









 INTRODUCTION  	   1




 PROJECT DESCRIPTION  	   1




 PROJECT ORGANIZATION AND  RESPONSIBILITY  	   2




 QUALITY ASSURANCE OBJECTIVES	   2




 SAMPLING PROCEDURES  	   4




 SAMPLE CUSTODY  	   5




 CALIBRATION PROCEDURES AND FREQUENCY  	   5




 ANALYTICAL PROCEDURES  	   6




 DATA REDUCTION, VALIDATION, AND REPORTING  	   7




 INTERNAL QUALITY CONTROL  CHECKS 	   9




 PERFORMANCE AND SYSTEM AUDITS 	 11




 PREVENTIVE MAINTENANCE 	 11




ROUTINE DATA ASSESSMENT PROCEDURES 	 12




CORRECTIVE ACTION 	 13




QUALITY ASSURANCE REPORTS 	 14

-------
                                 INTRODUCTION

The Contaminated Great Lakes Sediments-Metals Characterization and Treatment
project is part of the Minerals Separations research group at the Salt Lake
City Research Center (SLRC).  The Minerals Separations group at the SLRC has
long-standing expertise in mineral processing technology.  The group is
working under the Engineering/Technology Work Group in the Assessment and
Remediation of Contaminated Sediments (ARCS) program to investigate common
mineral processing technologies as removal or remediation alternatives for
contaminated sediments.

The ARCS program is a 5-yr effort authorized by the Water Quality Act of 1987.
Under this program, the Great Lakes National Program Office (GLNPO) of the
Environmental Protection Agency (EPA) will carry out studies emphasizing the
removal of toxic pollutants from sediments in the Great Lakes' system.  The
objectives of the ARCS program are to assess the extent of sediment pollution
in designated Areas of Concern and to identify and demonstrate options for the
removal and/or treatment of the contaminated sediments.  The ARCS program is
to be completed in the years 1988-1992.

The Bureau of Mines' participation in the ARCS program will be divided into
research work and analytical work.   This quality assurance project plan (QAPP)
will separately address the analytical and research aspects of the project.
The analytical aspects of this QAPP are taken from the SLRC Analytical
Chemistry group's (ACG) quality assurance plan already in effect for
environmentally sensitive samples.

                              PROJECT DESCRIPTION

The EPA's ARCS program is interested in assessing a number of technologies
which may be applicable to the remediation of contamination associated with
the sediments in the five Areas of Concern (AOC).   Since metals contamination
has been identified as one of the problems associated with these sediments,
the Bureau has been asked to characterize sediments from three AOC (Buffalo
River,  Indiana Harbor,  and Saginaw) and assess options for their treatment.

Chemical,  physical, and mineralogical characterization work will be conducted
on sized and unsized samples of sediment supplied by the EPA.   Based on this
characterization information,  treatment options such as flotation, tabling,
and wet high-intensity magnetic separation will be assessed.

Samples received from the EPA, and additional samples produced by sizing and
separation testing on these samples,  will be analyzed for the EPA's nine
inorganic priority pollutants (antimony, arsenic,  barium, cadmium, chromium,
lead,  mercury,  selenium,  and silver)  plus other elements determined by the
laboratory to be of interest.   A maximum of 5,000 element determinations will
be made.   This corresponds to about 550 sample analyses if all nine priority
pollutants are assayed in each sample.

-------
                    PROJECT ORGANIZATION AND RESPONSIBILITY

The  SLRC  is  one  of nine  research centers  operated by the Bureau.   It  is headed
by a research  director,  with  an assistant director and  two research
supervisors.   The individual  research  groups report through group  supervisors
to the research  supervisors,  while  the chemical and mineralogical  analysis
groups report  through  group supervisors to  the assistant director.  The SLRC
organization is  illustrated in Figure  1.

The  Contaminated Great Lakes  Sediments-Metals Characterization and Treatment
project is under the direction of Dr.  David A. Rice, Group Supervisor,
Minerals  Separations.  Day-to-day technical direction is by Mr. James P.
Allen, Principal Investigator.  Project staff includes  a chemical  engineer and
physical  science technician.  Mineral  separation tests  and physical
characterizations, such  as size distribution and zeta potential, are conducted
by the project staff,  while chemical and  mineralogical  determinations are made
by members of  the SLRC analytical group.  The flow of communication for
management of  the Contaminated Great Lakes  Sediments-Metals Characterization
and  Treatment  project  is shown in Figure  2.

                         QUALITY ASSURANCE  OBJECTIVES

The  purpose  of this quality assurance  project plan is to define the procedures
to be taken  to assure  that the information  produced reliably represents, to
the  extent possible from the  condition of the samples when they are received,
the  potential  of selected mineral processing technologies to remove heavy
metal contaminants from Great Lakes sediment.  This plan also seeks to assure
that the physical properties  of the sediment samples are accurately
characterized.

Representativeness - Representative samples from the separation tests will be
dried and submitted for elemental analysis.  Representativeness for all
samples will be  assured through use of dry  riffle splitters (five passes).

Completeness - As shown in Table 1,  the objective for completeness in all
elemental determinations is 90 pet.   Researchers will endeavor to submit
sufficient samples to  the analytical section to assure that all nine elemental
determinations can be made.  It is likely, however,  that some separations
procedures will not produce amounts of material sufficient for complete
analysis.

Comparability  - Data comparability will be assured by reporting of element
determinations in weight percent for flame and parts-per-million (ppm or pg/g)
for furnace.   Results of separations tests  (e.g.  flotation,  magnetic
separation, sizing)  will be reported in percent distribution.   Particle sizes
will be reported both  in Tyler mesh sizes (where appropriate)  and in microns
(/m).

Precision and Accuracy - Specific objectives for precision and accuracy of
elemental  determinations are given in table 1.   Known standards,  certified
reference  materials,  duplicates,  blanks, and spiked  samples  will be used to
assess  precision, accuracy, and detection limits  of  elemental  analyses.  The

-------
                       TABLE 1.  -  Data quality objectives for precision,  accuracy,  and completeness
Measurement
Parameter
Antimony
Arsenic
Barium
Cadium
Chromium
Lead
Mercury
Selenium
Silver
Digestion
Method
Hot HN03
Hot HN03
Hot HN03
Hot HN03
Hot HN03
Hot HN03
Hot Aqua
+ KMn04
Hot HN03
Hot HN03
Hot HC1
Hot HN03
+ H202
+ H202
+ H202
+ H202
+ H202
+ H202
Regia
+ H202
+ K2S208
reduction
+ H202
Analysis1
Method
GFAAS2
GFAAS2
FAAS3
GFAAS2
FAAS3
GFAAS2
FAAS3
GFAAS2
FAAS3
GFAAS2
Cold vapor
GFAAS2
HGAAS*
FAAS3
GFAAS2
Precision
Reference Rel. Std. Dev.
SOW #788
SOW #788
SOW #788
SOW #788
SOW #788
SOW #788
SW 846 method
7471 Sep. 1986
SOW #788
Used when S042"
is high (>2000 ppm)
SOW #788
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
+20%
Accuracy
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
±20%
+20%
Completeness
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
90%
2 Graphite Furnace Atomic Absorption Spectrometry.
3 Flame Atomic Absorption Spectrometry.
* Hydride Generation Atomic Absorption Spectrometry.

-------
 objective for precision of separation tests will be ±10 pet of the mass
 balance.   Routine duplication of separations tests to assess accuracy is not
 planned,  although significant results may be verified by duplication of a
 selected test to assure accuracy.

 Detection Limits - The ARCS detection limits of 2 ppm for analysis of
 antimony, arsenic, barium, chromium,  lead,  and selenium and 0.1 ppm for
 analysis  of cadmium,  mercury,  and silver will be the objectives.   These
 detection limits incorporate the analytical limit and will be used to test for
 sample contamination  via blanks as well as  for validity of low analyte
 concentrations.   Detection limit determination for the ARCS program is done by
 the multiple analyses (between 15  and 25 replicates)  of a standard with a
 concentration of 5 and 10 times the instrument detection limit (preferably
 near or at the detection limit).   The detection limit is then calculated as
 three times the  standard deviation of the replicate low-level standard
 measurements.

                              SAMPLING PROCEDURES

 Research  Group -  Initial procurement  of the sediment  samples  is  the
 responsibility of GLNPO.   The  SLRC will be  provided with representative splits
 of  these  samples;  splitting will be under the  direction of ERL-Duluth.
 Further splitting when the samples are  received at the SLRC will be
 accomplished by mixing the received material until visual  homogeneity is
 observed,  then splitting on a wet  rotary splitter.  These  and all  wet sample
 splits  will be stored in the dark  at  U"  ±1°  C.   These  wet-sample splits will
 be  consumed through the course  of  the project,  so  holding  times may be  as  long
 as  1  yr.   No special  storage procedures  are  required  for samples that have
 been  dried.  Holding  times for  dry solid samples are  indefinite.
 Representative samples  of products of separation tests  will be obtained by  the
 use of  suitable wet and dry rotary and  riffle  splitters.

 Sampling  and EPA-recommended preservation procedures are the  responsibility of
 the research group.   Solid samples  submitted by  the research  group  for
 analysis  will  be  dried  at 105"  C to constant weight.  Only dry samples  will be
 submitted for  analysis.   Oven-dried samples will be disaggregated,  rolled, and
 riffle-split (5 passes)  to assure  homogeneity  and  representativeness.   Drying
 facilities  used will be  connected  to  ventilation to prevent exposure  to any
 volatile  components.

 Analytical  Chemistry Group  - Since sampling procedures are beyond the control
 of the ACG,  effort  is directed  at  supplying reliable analysis once  the  samples
 are received.  The  samples  are used "as  is" after being received by the ACG.
 Once received, the  laboratory staff will do what they can to preserve the
 sample such  as refrigeration if specified by the submitter.  The EPA
 recommended holding times begin when  the ACG receives the samples.   Those
 samples will be given priority in order to meet specified holding times.  Good
 communication between the research group and the laboratory is essential to
meet holding times.

-------
                                SAMPLE CUSTODY

Research Group - Upon receipt of the samples at the SLRC,  they will be split
into quantities suitable for characterization and testing,  and each split will
be numbered and labeled.  When a researcher takes one or more of these splits
for characterization or processing studies, its number, date, name of
researcher, and purpose will be entered in a log.  As further samples are
generated, they will be assigned numbers consistent with the original split.
When these samples are prepared for chemical analysis,  they will be assigned
an analytical code of up to six characters, and this will be referenced with
the research number in the record book.  The principal  investigator will serve
as sample custodian for the purposes of receiving samples and verifying sample
custody records.

When a sample is assigned an analytical number, it is written on both the
sample container and a log sheet; the samples and sheets are placed in the
analytical sample receiving area.  Custody of the sample is transferred to the
analytical group when the sample sheet is posted to the research group
clipboard in the analytical section.

Analytical Chemistry Group - The samples will remain in the custody of the ACG
until all analyses are completed.  When QC review is complete, the samples are
returned to the research group.

Research groups assign samples a two-letter project code and a sequential
four-digit number.  This six-character sample number is placed on the sample
and listed on a log sheet along with analyte requests.   The sample number ties
the sample results to the log sheet as it moves between analysis stations.  To
identify and segregate those environmental samples requiring analysis as
specified in this QAPP, all such requests are on a green analytical request
form, as shown in figure 3, or labeled in the computer  data base,  so sample
data can be easily identified, safeguarded, and reviewed.

                     CALIBRATION PROCEDURES AND FREQUENCY

Research Group - Research equipment, capable of calibration, will be
calibrated once during each bimonthly period in which the device is used, with
the exception of pH meters which will be calibrated daily against two
standards when in use.  Calibration of laboratory equipment will be recorded
in the laboratory data books.

Calibration procedures for particle size analysis equipment will be limited to
checking the devices by testing a sample of known size  distribution.  Sieves
will be visually examined for defects prior to each use.  Faulty sieves will
be repaired or replaced, while adjustments to the Cyclosizer and X-ray
sedimentometer will be made in consultation with the manufacturer.
Calibration of the helium pycnometer will be checked by measuring the, volume
of a steel ball of known size, and required adjustments made in consultation
with the manufacturer.

Analytical Chemistry Group - Instrument calibration takes place prior to the
analysis of an analytical batch or any set of samples.   Interruptions,  for

-------
 whatever reason,  that  cause  a delay  will  result  in calibration  of  the
 instrument and repeat  of the samples.  An analyst produces a hard  copy
 calibration curve for  review and comparison with previous curves which  are
 kept on file.   The coefficient of determination  (R2) of the curve  is recorded
 to  gauge the closeness of fit of the standards to the generated curve.
 Calibration is acceptable when a coefficient of  determination greater than
 0.97 is obtained.   For the ARCS program,  a calibration standard in the  mid-
 calibration range is analyzed at the beginning,  middle, and end of an
 analytical batch.   Calibration continues  to be acceptable as long  as the
 analysis of this  mid-range standard  falls  within 10 pet of its  "true" value.
 All calibration curves consist of a  minimum of two points plus  a blank  or zero
 point.   Preferably, four points are  used  in a second-order regression.
 Readings of samples that are above the most concentrated standard  are diluted
 and rerun until the range of the standard  curve  is obtained.

 For mineralogical  analysis,  calibration of X-ray diffraction equipment  will be
 checked weekly using a quartz standard, and a calibration failure  warrants
 service by the manufacturer.   An energy calibration will be made weekly on the
 scanning electron microscope with failure  of the calibration warranting
 outside service.

                             ANALYTICAL PROCEDURES

 Research Group - The research group  will perform minerals characterization and
 separation procedures  on the  samples.  In  most cases, the samples will  have
 been filtered,  deoiled,  dried,  and redispersed in water.  For measurement of
 physical  properties and separability by flotation,  gravity,  or magnetics,
 samples will be prepared according to manufacturer's instructions.   Care will
 be  taken  in preparation of all  slurry samples (i.e.  samples that have been
 dried,  then redispersed in water) to assure that solid particles are
 thoroughly  dispersed.   Methodology will be systematic and based on sound
 metallurgical  laboratory practice using applicable suggestions in references
 such as  the Mineral Processing Handbook of AIME.

 Froth flotation is a physical separation procedure based on the hydrophobicity
 of  a particle  surface.    The separation is usually carried out in an agitated
 tank into which air bubbles are  introduced.  The hydrophobic minerals attach
 to  the air bubbles and  float  to  the  surface,  where  they are  skimmed off.
 Reagents are usually added to flotation operations  to render one or more of
 the  mineral phases present hydrophobic.  Usually frothing agents are added to
 insure a stable foam.   Laboratory flotation is  carried out in specially
 designed flotation machines or appropriately constructed glass vessels with
 porous bottoms for producing air bubbles.

 In  the laboratory, gravity separation will be accomplished by suspension of
 the  sediment in a fluid having a density of 2.9  g/cm3.  The particles with
density greater than 2.9 will sink, and those with  lower density will float.
The separation can be effected by drawing off the "sink"  fraction using a
separatory funnel.  A shaking table may be used  for  gravity  separation of
large samples.

-------
Laboratory magnetic separations are performed by passing a stream of the
material to be separated through a magnetic field that can be varied in
intensity.  The magnetic particles will be attracted toward one of the magnet
poles and pass over a splitter into a collection bin, while the non-magnetic
paarticles pass through without disruption.  By increasing the intensity of
the field and returning the unattracted portion to the separator, fractions of
strongly magnetic and then weakly magnetic material can be collected.
Magnetic separation is usually performed on dry material for coarse particles,
while for fine particles, wet magnetic separation is more successful.

Initially, a sample of each sediment will be separated into 10 to 15 size
fractions by sieving and cycloning.  These size fractions will be analyzed for
element concentrations as described below.  Laboratory separation and
characterization procedures such as flotation will be performed on as-received
and on sized material.  The sized samples for this work will be produced by
sieving or cycloning at sizes determined to be appropriate after elemental
analysis of the complete size distribution.  In this case, sieving and
cycloning are not analyses, but sample preparation procedures.  Separation
procedures may be performed in series, i.e. the products of a magnetic
separation test may be used as the feed for a gravity separation test.

Analytical Chemistry Group  The ACG will perform elemental concentration
determinations on dry samples submitted by the research group as products of
separation or characterization work.  Determinations of elemental
concentrations in solid samples shall be made using the methods specified in
table 1. The instruments to be used will include the Perkin-Elmer Models 2100
and 5100 Flame and Zeeman corrected graphite furnace Atomic Absorption (AA)
Spectrometers.

                  DATA REDUCTION,  VALIDATION, AND REPORTING

Research Group - During both characterization and separation treatment of
sediments, the samples will be separated based on size, density,  magnetic
properties, flotation properties,  or other physical properties; therefore, as
is customary, data reduction will  focus on the amount of material in each
separation product and its contaminant contents.  A materials balance will be
used to validate data integrity for each separation.

Analytical Chemistry Group - The Perkin-Elmer 5110AA & 2100AA interface to
computers which perform the data reduction.  Information such as sample
weight, dilution factors, and size of the volumetric flask are entered, and a
final report is generated on a printer.  Other instruments read out in peak
height, peak area or absorbance,  or directly in concentration.  These kinds of
data are calculated by hand or with the aid of computers which establish
concentration curves and calculate concentrations.   Standard curves
established by the computer are usually the result of linear or second-order
regression.  Equations used to convert sample readings in concentration units
to final answers are:

for liquid samples

                  ppm(L)  = ppm(instrument) x Dilution factor

                                       7

-------
for solid samples

                             ppm(L) x mL(Volumetric)
                    ppm(S) -     g(sampie weight)

                             USL of analvte   --_   ppm(S)
                   percent -     7     ,   X 1UU - 1r>r»r»n
                   r         Hg of sample          10000

For analysis of triplicate samples, Relative Standard Deviation (RSD) is
defined as the standard deviation of the elemental analyses divided by the
mean of the elemental analyses, expressed as a percent.

Spike recovery (SR) is calculated as follows:


                            SR - ^ x 100

                         SS - Spiked sample result
                         S  - Sample result
                         SP - Spike amount

Data are valid if the Group Supervisor-Analytical Chemistry Group (GS-ACG)
reviews and approves the sample results and the associated QC data, and if at
the discretion of the GS-ACG repeats of the calculations by another analyst
show no discrepancies or if they do they are corrected.  Analyses are
performed and data listed in reports or notebooks in the following sequential
order for review and validation: a blank, series of standards, 10 samples, a
blank, a standard, 10 samples, a blank, a standard, a duplicate, triplicate,
and a spike of one of the previous samples, standard reference material, a QC
sample, a blank, and a standard.  This liquid QC sample is used to verify
calibration of the instrument.  At the discretion of the GS-ACG, a reference
sample or an unknown prepared by the GS-ACG may be included.  The GS-ACG may
require that water spikes be performed to check the accuracy of the spiking
technique.  Spike volumes will be less than 1 pet of the sample's original
volume.  If a sample concentration exceeds the highest standard, it is
unacceptable.

Analytical results are reported using a spreadsheet program.  Sample numbers,
volumetric size, and sample weights obtained electronically from the
analytical balance are stored in a computer file and are subsequently read
into the spreadsheet to avoid transcription errors.  Instrument readings and
dilution factors are typed into the spreadsheet which does the calculations
and prints the report.  Once the analyst completes an analyte for a set of 20
samples or less, the GS-ACG reviews the report containing sample data and
supporting QC information.   The report is returned to the anlyst after the GS-
ACG initials the report and makes any written comments.  Reports returned with
no initials require additional work.   The analyst submits the approved report
to the sample room and shows analyte completion by highlighting, initialing,
and dating the analyte column on the green analytical request form (fig. 3).

Mineralogical analysis and identification is essentially a subjective
procedure relying on the experience and knowledge of the geologist.   X-ray

                                       8

-------
diffraction and energy-dispersive scanning electron microscope data will be
used to complement optical determinations made by the geologist.

Data will be kept on file for 5 yr.

                        INTERNAL QUALITY CONTROL CHECKS

Research Group - At the discretion of the researcher, duplicate tests will be
conducted to ensure the accuracy and reproducibility of the research.
Typically, these replicate tests will be conducted on tests with exceptional
or unexpected results.  Routine duplication of separation and characterization
work is not planned.  These repeat analyses will be a double check on
analytical work to help ensure that consistent results are obtained throughout
the entire research program.

Analytical Chemistry Group - Internal quality control procedures include the
use of triplicate analyses, matrix spikes,  blanks, certified reference
materials, on-going calibration standards,  QC samples, and control charts.
These procedures are summarized in table 2.

                TABLE 2. - Summary of internal quality control
                      procedures  for  elemental  analyses
Procedure
Triplicate analyses
Matrix spike
Blanks
Reference material
Calibration standard
QC sample
Frequency
1 per set
1 per set
3 per set
1 per set
3 per set
1 per set
Requirement
±20 pet RSD
±15 pet of spike

-------
value when  doing  a  total  analysis.  Since the methods listed  in  table  1  are  a
digestion using an  extraction  technique and not a true, complete digestion,
some elements may be  lower  due  to NIST 2704 and MP-1 being certified using
complete digestions or  techniques.

Precision is checked  by analyzing one sample from each sample set  in
triplicate.  Acceptable precision is ±20 pet RSD.

Blanks will be run  at the rate  of three per sample set to check detection
limits.  Acceptable readings on blanks are below the detection limits.

One matrix  spike  of a sample from the sample set will be run with  each set.
The spike target  amount is  .5  to 2 times the concentration of the  sample;
acceptable  spike  recovery is 85 to 115 pet.

A middle-range calibration  standard will be run three times with each  sample
set to check instrument calibration.  Acceptable readings for the  calibration
standard are 90 to  110  pet  of  the known value.

QA control  charts will  be kept  on file for blanks, reference material, mid-
range calibration standards, and the QC (independent calibration standard)
sample.

A QC sample is run  each time a batch is analyzed.  The purpose of  the  QC
sample is to provide  a  check on calibration from an independent standard
source.  QC samples are produced on the automatic diluter from a separate
NIST-traceable standard at  a dilution specified by the computer between 10 and
100 pet of  the highest  standard and unknown to the analyst.   A batch is a set
of samples  run with the same reagents, under the same conditions on a  single
instrument, usually on  the  same day.  Whenever an analyst changes  to a new
analysis, a QC sample must  be run.   QC samples of more than one concentration
may be required to  control  the samples at different concentrations.  To remove
bias from the analyst, QC samples may be diluted by using a well-defined range
of random dilution  factors.   The analyst makes the dilution,  not knowing the
dilution factor and presents the diluted QC sample for analysis.   After
analyzing the diluted QC  sample, the computer-recorded or concealed dilution
factor is used to normalize the value to that used on the control chart.   The
analyst enters the QC sample result into the computer data base and observes
the QC chart which  shows  if the analysis is in control.   Analyzing QC  samples
is a process of self monitoring conducted by the analysts as part of their
daily routine.

A mean of repeat analyses of a QC sample may be different than the true value.
The GS-ACG may at any time- present reference samples to analysts  to monitor
the accuracy of the work.   A reference sample is an EPA QC sample or a
purchased sample (certified reference material) which has statistics on
analyses from numerous laboratories.   All analytes present may or may not be
analyzed.

All standard curves will include an analyzed blank,  or the instrument will be
initially zeroed on blanks.   Rezeroing in the course of the  run is not
permitted.   Analysts usually prepare standards from stock solutions purchased

                                      10

-------
 from reputable  companies.   Each  analyst  is  responsible  for preparing
 calibration curves.   One  standard  and  its instrument response  is monitored by
 the  analyst,  and the  record is made available  to  the GS-ACG.   Occasionally the
 GS-ACG will submit a  calibration check which consists of a standard in  the
 same concentration as that  used  to measure  the  instrument response but
 prepared  from stock solutions which are  different from  those used by  the
 analyst.

 whenever  a  new  reagent bottle is used, it will  be checked for  contamination by
 the  analyte of  interest and not  used if  that analysis exceeds  the detection
 limit.

 Occasionally the GS-ACG will see that  reference or unknown samples are  given
 to researchers  to be  disguised and submitted as regular samples.  Other
 samples can include spiked  or unspiked samples  analyzed previously.
 Comparison  of the reference  sample results  to true values are made available
 to the analysts  and researchers.   Comparison of the results to previously
 analyzed  samples, and the percent  spike  recoveries are also made available.

                         PERFORMANCE AND  SYSTEM AUDITS

 The  following audits  will be performed:

 Internal  Systems  Audit - Performed by  the project manager at the initiation of
 the  project,  covering sample splitting procedures, work plans,  selection of
 methods,  and  data recording.

 External  Systems  Audit - Performed by Lockheed-ESC and GLNPO personnel.

 Internal  Performance Audits  - Initiated quarterly by the principal
 investigator using EPA solid reference sediment standards to assess
 performance of pollutant concentration measurement by the analytical section.

 External  Performance Audit - Initiated by the principal investigator at
 midpoint  of the project using samples  obtained from outside the Bureau to
 assess performance of pollutant analyses.

                            PREVENTIVE MAINTENANCE

 Research Group - No specific preventive maintenance is  planned on any of the
 research equipment used in this study.

Analytical Chemistry Group - All the  required maintenance is performed in-
house by the analyst according to the  procedure described in the instrument
manual except for that which can only  be  performed by a representative of the
 instrumentation company.

 Service contracts and maintenance agreements are not carried because of  the
additional high-operating costs involved. Vendor-specific  routine maintenance
 is generally done in conjunction with  a needed service  call.
                                      11

-------
                      ROUTINE DATA ASSESSMENT PROCEDURES

Research Group - Metallurgical balances will be used to validate data from all
separations.  Conservation of mass will be used to validate mass measurement
data.  The percent of mass conserved, A, in a separation process with n
product streams can be expressed as
                                A - £L x 100
                                    Mo

where M£ is the mass of separation product i and M0  is  the  mass  of the
starting material.  If A is less than 95 pet, the mass measurement data will
be flagged for examination, but the data will be validated.  If A is  less than
90 pet, the mass measurement data will be invalidated, and corrective action
will be required.

Similarly, elemental mass balances will be used to validate analytical data
for each desired element.  The percent of an element conserved B can be
expressed as
                                B - i=— x 100
where at is the analysis of separation product i,  M.L  is  the mass  of separation
product i, a0 is the analysis of the starting material,  and M0  is the  mass  of
the starting material.  If B is less than 90 pet or greater than 110 pet, the
elemental assay data will be invalidated, and corrective action will be
required.

For each set of sized separation tests, one size fraction will be selected at
random, split in two parts, and the separation will be duplicated on that
fraction.  The weights of corresponding products will be compared, then
combined further separation tests will be planned on that material.  For
example, with a set of four size fractions going to magnetic  separation, all
products going on to gravity separation, one size fraction will be split, and
duplicate magnetic separations will be conducted.   Weights will be compared
and if acceptable, recombined.  One product at random will be taken from
magnetic separation and split to run duplicate gravity separations with the
weights compared and recombined for analysis.  Acceptable precision is ±20 pet
relative difference (PRO) .

Particle size analyses will be duplicated.  Acceptable precision is ±20 pet
relative difference .

Analytical Chemistry Group - For QC samples, the acceptance limits are ±3
standard deviations from the mean.   Plus or minus 2.0 standard deviations from
the mean establishes a warning limit.   A QC chart containing  lines for the
acceptance and warning limits along with the mean will serve  as a background
upon which the values are plotted.   Charts will be kept for calibration and QC
samples, blanks, and reference materials.  The background will be updated

                                      12

-------
after the addition of each new point with the use of statistics and graphics
software.

Triplicate samples have acceptable precision if the RSD is ±20 pet.

Samples in a group of 20 or less is called a set.  Data from a set is
acceptable when spike recoveries fall in the range of 85 to 115 pet.  An
analytical batch is a group of samples run with the same reagents on the same
instrument under the same conditions.  The analytical batch has acceptable
accuracy when the QC chart shows the analysis to be in control, and the
analyses of the reference material is within 20 pet of the known value.

                               CORRECTIVE ACTION

Research Group - If mass measurement data are invalid, the separation will be
repeated.  If mass measurement data are still not valid, an investigation into
a plausible cause will be made, and data validity will be reassessed.  If
elemental analysis data are not valid, suspect analyses will be reanalyzed.
If elemental analysis data are still not valid, the separation will be
repeated.

Analytical Chemistry Group - QC sample data, which fall within the warning
limits,  are considered acceptable.   No action is required on the sample data.
Data falling outside of the acceptance limit invalidates all of the data
associated with that QC sample.  The entire analytical batch is rerun after
identifying and correcting the problem.   Those falling between the warning and
acceptance limits require less stringent action.   The analyst is alerted that
there might be a potential problem.  Instead of all the samples being rerun,
only a set of samples is rerun and a QC sample included and monitored closely.
Rapid feedback by analysts is maintained to spot and correct problems quickly.

If replicates are not within ±20 pet RSD,  the replicates are repeated.  If
still outside the range, the set is rerun.   If still no resolution to the
problem, data may be reported along with the RSD's to indicate a problem with
data precision.

If spike recovery does not fall within the range of 85 to 115 pet, the spike
is repeated along with a water spike of the same concentration used on the
sample.   If the spike still does not fall within the acceptable range, the
samples are put aside to investigate matrix interference.   A comparison is
made of spikes in water and in a simulated sample matrix.   These results may
indicate that the standards must be matched to the sample matrix.   Samples and
spikes are repeated with the new standards.   Time spent researching the cause
of the matrix problem and implementing a solution is proportional to the
number of samples that have a similar matrix.  If no resolution of the problem
is obtained,  a written explanation of the problem and its impact on the data
is reported along with the "as is"  determined data and spike recoveries.  If
the sample set is small, the set is analyzed using the method of additions
(MOA) and the letters MOA appear on the  report.

If the analysis of a reference sample does not fall within ±20 pet of the true
value, fresh standards are prepared,  and the samples are rerun.  This applies

                                      13

-------
only to standard reference materials which are certified using the digestion
method required by the samples.

                           QUALITY ASSURANCE REPORTS

The performance of the measurement systems and observations on data quality
will be reported quarterly as a part of the quarterly progress memo reports
called for in the Interagency Agreement.  The  report will address data
accuracy and precision, audit results, and identification and resolution of
significant QA problems.

The quarterly progress memo report will be prepared by the principal
investigator and approved by the project manager.

The results from the quarterly research records review will be reported with
the quarterly progress memo reports called for in the Interagency Agreement.
                                     14

-------
                             FIGURE 1
                      ORGANIZATION CHART
              SALT LAKE CITY RESEARCH CENTER
                                   .' »* * • •»*
    RESEARCH
     SUPPORT
    CHEMICAL
    ANALYSIS
  Group Supervisor
Dwlflht 0. Kanvnaro/en
   PHYSICAL
   ANALYSIS
  Group
  David L Nsylan
 OPERATIONS &
  FACIUTES
      I Nlsssn
                        RESEARCH DIRECTOR
                            Sltphen 0. HIR

                           ASSISTANT
                        RESEARCH DIRECTOR
                            Mont* 6. ShJrtS
   MINERALS
  TECHNOLOGY
  ASSESSMENT
 Rciearcri Supervisor
 Chirtei f. Davidson
   PARTICLE
 TECHNOLOGY
 Group Supervisor
James E. Gebriardt
   MINERAL
 ASSESSMENT
 Group Supervisor
James K MsysfDes



MINERAL
SEPARATIONS
Group Supervtaor
David A. Rice
   CHEMICAL 4
 METALLURGICAL
   PROCESSES
 Research Supervisor
 Richard Q. Sandberg
BIOTECHNOLOGY

 Group Supervisor
P«uJette B, Attrtnger
 REFRACTORY
  MATERIALS
 Group Supervtaor
 Glenn R. Palmer
                         ADVANCED
                         MATERIALS
                        Group Supervisor
                        Allan E. Petersen
 PERSONNEL
KenneWi L Kfflpack
                         ALTERNATE
                        TECHNOLOGY
                        Group Supervisor
                        Thomas K. Jeffers

-------
                         FIGURE 2
                    ORGANIZATION CHART
             CONTAMINATED GREAT LAKES SEDIMENT
           METALS CHARACTERIZATION AND TREATMENT
 PROJECT MANAGER
      D. A. RICE
        SLRC
I  RESEARCH GROUP

      J. P. ALLEN
 PRINCIPAL INVESTIGATOR
                     PROGRAM OFFICER
                        W. B. SCHMIDT
                       WASHINGTON, DC
   QA OFFICER
 D. D. HAMMARGREN
      SLRC
ANALYTICAL GROUP

 D. D. HAMMARGREN
      SLRC

-------
     Figure )—Analytical Request Form
UNITED STATES DEPARTMENT OF THE INTERIOR
            BUREAU OF MINES
No..
Report t<
Daterecc
S«fal
No.









0
1
2
3
U
5
6
















»Mr
iv«
-------